USF Libraries
USF Digital Collections

Associations between individual, social, and service factors, recovery expectations and recovery strategies for individu...

MISSING IMAGE

Material Information

Title:
Associations between individual, social, and service factors, recovery expectations and recovery strategies for individuals with mental illness
Physical Description:
Book
Language:
English
Creator:
Walby, Gary W
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
Severe mental illness
Stigma
Social support
Empowerment
Anxiety
Dissertations, Academic -- Public Health -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: Clinical and consumer recovery efforts continue to advance more successful recovery agendas. Limited research into the effect of the expectation to recover and what strategies are most effective in enhancing recovery impedes coherent recovery programming. What factors are significantly associated with recovery expectancy and recovery strategies is still largely unknown. Thus, this study addresses three key gaps in the field.^ ^First, which factors associate with the expectation of recovery and is expectation related to strategy choice? Second, what are common strategies for recovery from mental illness and what factors are associated with each strategy? Third, does recovery expectancy or severity of mental illness mediate or moderate the relationship between clinical, social and service factors and recovery strategies? This study surveyed a sample of 350 randomly chosen participants from a community mental health organization with varying degrees of mental illness in a cross-sectional study utilizing instruments that measured clinical/historical, social, and service factors and recovery. The results were examined in multivariate analysis targeted to address the gaps noted above. The six recovery strategies included: (1) effective illness management, (2) positive future orientation, (3) meaningfulness, personal control, and hope, (4) recognizing support, (5) help seeking, and (6) symptom eradication.^ ^Recovery expectancy was not significantly predicted by any of the clinical, social, or service factors. Although 25% of the variance was explained by the full model, factors associated with expectancy differ from recovery strategies and require further investigation. Except for symptom eradication, recovery expectancy was significantly negatively associated with each recovery strategy.Main effects models were significant for all six recovery strategies. The social factor constructs (social support, empowerment, stigma) were most consistently and robustly associated with all recovery strategies. Variance explained in full models ranged from 71% for positive future orientation to 19% for symptom eradication. However, no mediating or moderating effects were detected for recovery expectancy or illness severity. The results of this study further the understanding of recovery and provide information for development of recovery programs.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2006.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Gary W. Walby.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 693 pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001949947
oclc - 243474410
usfldc doi - E14-SFE0002203
usfldc handle - e14.2203
System ID:
SFS0026521:00001


This item is only available as the following downloads:


Full Text

PAGE 1

Associations Between Individua l, Social, and Service Factor s, Recovery Expectations and Recovery Strategies for I ndividuals With Mental Illness By Gary W. Walby A dissertation submitted in partial fulfillment Of the requirements for the degree of Doctor of Philosophy Department of Community and Family Health College of Public Health University of South Florida Co-Major Professor: Marion Becker, Ph.D. Co-Major Professor: Me linda S. Forthofer, Ph.D. Jeannine Coreil, Ph.D. Sondra Fogel, Ph.D. Date of Approval: September 28, 2006 Keywords: severe mental illness, stig ma, social support, empowerment, anxiety Copyright 2007, Gary W. Walby

PAGE 2

Dedication The effort and focus needed to comple te a dissertation cannot be sustained on ones own. There are, of course, many indivi duals that support, take on extra tasks, forgive your absence, long for your absence wh en you arrive with surliness in tow, and applaud every small milestone. My wife, Ma dalaine, did all that and more. Thank you for sustaining me, supporting me, and encour aging me. Thank you for listening even when you had heard it all before. Thank you for running interference with all the family and friends that I was forced to ignore Thank you for marrying me even though you knew that all this was in front of us. Other family requires special mention as we ll. To my mother, Lorraine, that gave birth to me and was around for a few other important moments over all these years as well. The memory of my father that sustained me when I wondered if I had enough brain power left and realized that he had bequeathed that much a nd more. My sister Linda, her wonderful husband Keith, my brill iant nieces Debbie and Deni se, and brother Mike. My daughter Heather who kept life interesting, but always surprised, Im proud of you. My beautiful granddaughter Kaidence that never failed to make me smile and who embodies unconditional love. To Tevlin, my gra ndson, and future defensive lineman or professional wrestler. To Criss, my mother-inlaw that was always cheerful even when in pain or dealing with a pain (meaning me). To special friends Ed, Arlene, Sue, Ri ck, Greg, Teresa, and are newly reunited family members, Donna and Claude. Im looking forward to making up for lost time with all of you.

PAGE 3

Acknowledgements My committee blended like the ingredients of a rich and addictive desert. If knowledge is power, then I was surrounded by enough to light a small city. To Dr. Melinda Forthofer a mentor and wise sage that was (somewhat de pressingly) a decade my junior. You challenged, pr aised, and provided in the righ t measure at all times and I cannot begin to list all that Ive learned from you. Perhap s the most important is the confidence to use new talents and knowledge to better the lives of ot hers. To Dr. Marion Becker you provided words of wisdom and experience as a healer and teacher to encourage, coach, and smooth the road. It woul d sometimes take me days to realize that you taught me a profound life lesson that applie d to much more then academia. To Dr. Jeannine Coreil I learned more about theory and practice in your c ourses than I thought possible and you taught me to appreciate subtleties, depths, patterns and the big picture all at once. How you did that without my head exploding, Ill never know. Thank you for your time and kindness. To Dr. Sondra F ogel you are a kindred spirit in your love of the clinical realm and your passionate pursu it of social justice and the betterment of all. You inspired, provided solace, and al ways uplifted with your genuine joy in our conversations. I always left with more energy than when I arrived. Special thanks to other professors that brought me knowledge and joy, including Dr. Karen Liller, Dr. Suzanne Perry-Casler, and Dr. Kay Perrin. Finally, thanks to the students that I was honored to te ach and who taught me much in return. Special thanks to those students that volunteered to assist with data collection for this project: Amy BraddLee, Jessica Burns and Christina Rickus I couldnt have done it without you.

PAGE 4

i Table of Contents List of Tables...............................................................................................................vi ii List of Figures...............................................................................................................xiv Abstract....................................................................................................................... ...xv Chapter 1...................................................................................................................... ....1 Introduction.........................................................................................................1 Purpose of Study and Significance......................................................................5 Study Rationale...................................................................................................9 Limitations of Existing Knowledge Base..........................................................13 Research Questions...........................................................................................15 Overview of Design...........................................................................................16 Implications for Public Health...........................................................................17 Delimitations.....................................................................................................18 Limitations.........................................................................................................19 Select Definitions..............................................................................................20 Chapter 2: Comprehensiv e Literature Review..............................................................23 Introduction/Overview......................................................................................23 Mental Illness....................................................................................................26 Recovery............................................................................................................28 Domains of Factors Associ ated with Recovery.................................................37 Domain 1: Individual, Histor ical, and Clinical Factors.......................38 Diagnosis..................................................................................38 Schizophrenia...........................................................................39 Major Depression.....................................................................42 Bipolar Disorder.......................................................................44 Current Symptom Levels..........................................................45 Age of Onset.............................................................................47 Trauma and Abuse....................................................................48 Hospitalization History.............................................................50 Employment History and Current Employment.......................51 Current Use of Substances and Substance Use History...........52 Familial History of Mental Illness............................................53 Synthesis and Conceptualiza tion of the Importance of Domain 1 factors.................................................................55

PAGE 5

ii Domain 2: Social Factors that Influence Recovery.............................58 Social Support..........................................................................59 Social Integration or Co mmunity Connectedness....................64 Empowerment...........................................................................67 Empowerment and Recovery.......................................73 Stigma and Discrimination.......................................................76 Synthesis and Conceptualiza tion of the Importance of Domain 2 Factors.....................................................................85 Domain 3: Service Factors that Influence Recovery...........................86 Synthesis and Conceptualiza tion of the Importance of Domain 3 Factors................................................................95 Chapter 3: Methodology...............................................................................................97 Study Design.....................................................................................................97 Population/Sample.............................................................................................99 Inclusion/Exclusion Criteria..............................................................................99 Sample Description and Procedures................................................................100 Sample 1: Individuals with Severe Mental Illness.............................100 Sample 2: Individuals with Mild to Borderline Mental Illness Attending Outpatient Services.......................................................102 Power and Sample Size Estimation.................................................................103 Research Questions and Hypotheses...............................................................104 Study Setting...................................................................................................115 Data Collection Procedures.............................................................................116 Pre-Data Collection Procedures..........................................................116 Institutional Review Bo ard Authorization..............................116 Training of Research Assistants.............................................116 Data Collection Procedures.................................................................117 Participant Contact.................................................................117 Confidentiality and Data Storage...........................................118 Informed Consent and Verification of Inclusion/Exclusion Criteria...............................................................................118 Incentives................................................................................119 Data Collection.......................................................................119 Dependent Variables: Recovery Expectancies and Strategies.......................121 Recovery Expectation Checklist.........................................................121 Recovery Assessment Scale................................................................122 Personal Vision of Recovery Questionnaire.......................................126 Measurement of Potential Correlates (Independent Variables) Across Multiple Domains.......................................................................................128 Domain 1: Individual, Histor ical and Clinical Factors......................128 Background Characteristics, Related Aspects of History and Demographics.............................................................129 Background Short Form.........................................................129 Symptom Checklist 90-Revised.............................................129

PAGE 6

iii Chart Abstraction Tool...........................................................133 Domain 2: Social Factors that Influence Recovery...........................133 Support and Community Connection Survey.........................133 Empowerment Scale...............................................................138 Internalized Stigma of Mental Illness Scale...........................141 Domain 3: Service Factors that Influence Recovery.........................144 Service Satisfaction Questionnaire.........................................144 Moderating and Mediating Variables.................................................146 Data Analysis Procedures................................................................................147 Data Entry, Cleaning and Verification................................................147 Factor Analysis...................................................................................153 Univariate Analyses............................................................................159 Bivariate Analyses..............................................................................159 Research Question 1: To What Degree Does Illness Severity Influence Beliefs in Recovery for the Individual...........................................................................160 Multivariate Analyses.........................................................................161 Research Question 2: Are I ndividual, Social, or Service Factors Associated w ith Recovery Expectancy?...............163 Research Question 3: Are I ndividual, Social, or Service Factors Associated with Recovery Strategies....................165 Research Question 4a: Does the Expectation of Recovery Mediate the Relationship between Individual, Social or Service Factors and Recovery Strategies?.........................167 Research Question 4b: Does the Expectation of Recovery Moderate the Relationship between Individual, Social and Service Factors and Recovery Strategies?..................169 Research Question 5: Does Severity of Mental Illness Moderate the Relationship between Individual, Social and Service Factors and Recovery Strategies....................170 Chapter 4: Results.......................................................................................................183 Study Participants............................................................................................184 Demographic Characteristics..............................................................184 Univariate Assessment of Clini cal and Historical Factors by Sample: Furthering the Understanding of the Respondents.........187 Univariate Assessment of Social Factors by Sample: Furthering the Understanding of the Respondents..........................................193 Univariate Assessment of Service Factors by Sample: Furthering Understanding of the Respondents................................................196 Results of Comprehensiv e Bivariate Analysis................................................197 Associations between Age, Ge nder, Education, Income and Recovery........................................................................................198 Sample Differences in Recovery Expectancy.....................................201

PAGE 7

iv Differences in Recovery Stra tegies between Samples and Recovery Expectations...................................................................204 Bivariate Analysis for Research Ques tion 2: Are Indivi dual, Social or Service Factors Associated with Recovery Expectancy?................................206 Research question 2.1: Are Indivi dual Factors Associated with Recovery Expectancy?...................................................................206 Research question 2.2: Are Social Factors Associated with Recovery Expectancy?...................................................................212 Research question 2.3: Are Serv ice Factors Associated with Recovery Expectancy?...................................................................214 Research Question 3: Are Individual, Social or Serv ice Factors Associated with Recovery Strategies?..........................................................................218 Research question 3.1: Are Indivi dual Factors Associated with Recovery Strategies?......................................................................218 Research question 3.2: Are Social Factors Associated with Recovery Strategies?......................................................................231 Research question 3.3: Are Serv ice Factors Associated with Recovery Strategies?......................................................................235 Multivariate Analysis......................................................................................242 Modifications to the Analysis Plan.....................................................242 Overview.............................................................................................243 Focus on Recovery Expectancy..........................................................244 Research question 2.1: Are Indi vidual Factors Associated with Recovery Expectancy?..............................................245 Research question 2.2: Are Social Factors Associated with Recovery Expectancy?..............................................247 Research question 2.3: Are Service Factors Associated with Recovery Expectancy?..............................................247 Research question 2: Are Individual, Social, or Service Factors Associated with Recovery Expectancy? Main Effects Model....................................................................249 Summary of Research Question 2, Investigating Recovery Expectancy.....................................................................................257 Focus on Recovery Strategies.............................................................259 Research question 3: Are Individual, Social or Service Factors Associated with Recovery Strategy 1: Effective Illness Management......................................................260 Research question 3.1: Ar e Individual Factors Associated with Eff ective Illness Management?..260 Research question 3.2: Are Social Factors Associated with Eff ective Illness Management?..263 Research question 3.3: Are Service Factors Associated with Eff ective Illness Management?..265

PAGE 8

v Research question 3: Are Individual, Social or Service Factors Associated with Effective Illness Management? Main Effects Model......................266 Research question 3: Are Individual, Social or Service Factors Associated with Recovery Strategy 1: Positive Future Orientation.............................................................272 Research question 3.1: Ar e Individual Factors Associated with Po sitive Future Orientation?.......275 Research question 3.2: Are Social Factors Associated with Po sitive Future Orientation?.......275 Research question 3.3: Are Service Factors Associated with Po sitive Future Orientation?.......278 Research question 3: Are Individual, Social or Service Factors Associated with Positive Future Orientation? Main Effects Model........................279 Research question 3: Are Individual, Social or Service Factors Associated w ith Recovery Strategy 1: Meaningfulness, Personal Control, and Hope...................287 Research question 3.1: Ar e Individual Factors Associated with Meaningfulness, Personal Control, and Hope?...............................................288 Research question 3.2: Are Social Factors Associated with Meaningfulness, Personal Control, and Hope?...............................................290 Research question 3.3: Are Service Factors Associated with Meaningfulness, Personal Control, and Hope?...............................................292 Research question 3: Are Individual, Social or Service Factors Associated with Meaningfulness, Personal Control, and Hope? Main Effects Model....................................................................293 Research question 3: Are Individual, Social or Service Factors Associated w ith Recovery Strategy 1: Recognizing Support.........................................................298 Research question 3.1: Ar e Individual Factors Associated with Recognizing Support?................301 Research question 3.2: Are Social Factors Associated with Recognizing Support?................303 Research question 3.3: Are Service Factors Associated with Recognizing Support?................305 Research question 3: Are Individual, Social or Service Factors Associated with Recognizing Support? Main Effects Model..............................305

PAGE 9

vi Research question 3: Are Individual, Social or Service Factors Associated with Recovery Strategy 1: Help Seeking..............................................................................313 Research question 3.1: Ar e Individual Factors Associated with Help Seeking?.............................314 Research question 3.2: Are Social Factors Associated with Help Seeking?.............................316 Research question 3.3: Are Service Factors Associated with Help Seeking?.............................317 Research question 3: Are Individual, Social or Service Factors Associated with Help Seeking? Main Effects Model..............................................318 Research question 3: Are Individual, Social or Service Factors Associated w ith Recovery Strategy 1: Symptom Eradication........................................................323 Research question 3.1: Ar e Individual Factors Associated with Symptom Eradication?...............326 Research question 3.2: Are Social Factors Associated with Symptom Eradication?...............328 Research question 3.3: Are Service Factors Associated with Symptom Eradication?...............330 Research question 3: Are Individual, Social or Service Factors Associated with Symptom Eradication? Main Effects Model........................330 Summary of Research Question 3, Investigation of Recovery Strategies........................................................................................338 Social Support........................................................................347 Empowerment.........................................................................348 Stigma.....................................................................................349 Focus on Mediation.............................................................................351 Research Question 4a: What is the Mediating Effect of Recovery Expectancy on Domains 1-3 for Each of the Recovery Strategies?...................................................353 Focus on Moderation..........................................................................357 Research question 4b: Evaluation of Recovery Expectancy Moderating Effects............................................................359 Research question 5: Evaluation of Illness Severity Moderating Effects............................................................363 Summary of Associations with Recovery Strategies..........................368 Strategy 1: Effective illness management.............................369 Strategy 2: Positive future orientation...................................373 Strategy 3: Meaningfulness, personal control, and hope.......374 Strategy 4: Recognizing support...........................................376 Strategy 5: Help seeking........................................................377 Strategy 6: Sympto m eradication..........................................379

PAGE 10

vii Chapter 5...................................................................................................................... 381 Synthesis of Research Findings.......................................................................381 Recovery Expectancy......................................................................................382 Recovery Expectancy and Seve rity of Mental Illness........................383 Neurocognitive Deficit as Impedime nt to Recovery Expectancy.......386 Recovery Strategies.........................................................................................387 Explanatory contribution of i ndependent variable domains............................392 Domain 1: Clinical/h istorical factors.................................................392 Domain 2: Social factors....................................................................396 Domain 3: Service factors..................................................................402 Implications for Underlying Theories Used to Generate and Assess Study Results........................................................................................................404 Social Support Theory........................................................................404 Empowerment Theory........................................................................408 Stigma Theory.....................................................................................411 Additional Factors or Constructs that Might Influence Recovery Expectancy and Recovery Strategies..........................................................415 Study Limitations............................................................................................418 Limitations to Measurement of Recovery Expectancy.......................419 Limitations to Instrument Choice and Availability............................419 Limitations Due to Respondent Characteristics..................................422 Limitations to Generalizability...........................................................423 Study Strengths................................................................................................424 Implications for Public Health.........................................................................427 Suggestions for Future Research.....................................................................432 References.......................................................................................................435 Appendices...................................................................................................................471 Appendix A: Study Instruments.....................................................................472 Appendix B: Development of Social Support and Community Connection Survey.....................................................................................498 Appendix C: Other Forms and Co rrespondence Used in the Study...............544 Appendix D: Training Mate rial for Research Assistants Involved in the Study.....................................................................................................555 Appendix E: Partial Models Resulti ng from OLS Regression Testing the Impact of Recovery Expectancy as a Mediator and a Moderator and Testing Severity of Me ntal Illness as a Moderator.....................................588 Appendix F: Comprehensive Tables fo r All Bivariate Analyses between Dependent and Independent Variables No t Presented in the Main Text of Chapter 4....................................................................................................625 About the Author................................................................................................End Page

PAGE 11

viii List of Tables Table 1. Gender/Race/Ethnic ity of Individuals with a Severe Mental Illness Served at Pa rtnering Community Mental Health Center..............................101 Table 2. Summary of Contact Refusals and Ineligibility by Sample..........................103 Table 3. Summary of Independent Capacity Assessment Results..............................119 Table 4. Factor structure for Recovery Assessment Scale..........................................124 Table 5. Factor Structure for Personal Vision of Recovery Questionnaire................127 Table 6. Reliability Estimates for the SCL-90-R Original Reliability and Subsequent Estimates....................................................................................132 Table 7. Factor Structure for Social Support and Community Connectedness Survey...........................................................................................................135 Table 8. Empowerment Scale Factor Loadings..........................................................139 Table 9. Internalized Stigma of Mental Illness Fa ctor Loadings................................142 Table 10. Variable Descriptions.................................................................................148 Table 11. Factor Loadings and Items for Dependent Variables.................................155 Table 12. A Priori Multivariate Models.....................................................................172 Table 13. Mean and Standard Deviati on of Demographic Variable Age...................185 Table 14. Gender, Education, and Income for SMI, OP and Total Sample................186 Table 15. Proportion of Domain 1 Diagnoses by Sample..........................................188 Table 16. Univariate Statistics (Means and Standard Deviations) for Domain 1 Symptom Scales by Sample........................................................................189 Table 17. Univariate Means and Standard De viations for Continuous Domain 1 Independent Vari ables (Age of Onset, Hospitalizations, Years Employed, Familial Mental Illness) by Sample..........................................191 Table 18. Proportion of Categorical Domain 1 Independent Variables (Hospitalization, Employment, Substance Use, Medication, and Child Abuse) by Sample.......................................................................................192 Table 19. Univariate Domain 2 Means a nd Standard Deviations for Stigma, Social Support, and Empowerment by Sample...........................................194 Table 20. Univariate Means and Standard Deviations for Domain 3 Independent Variables: Service Factors by Sample.......................................................197 Table 21. Pearson Correlations fo r Age by Recovery Strategies................................199 Table 22. Independent Sample T-tests and One-Way ANOVA Results of the Relationship Between Gender, Income, Education and Recovery Strategies.................................................................................................... .200 Table 23. Chi-square Results for Recove ry Expectation Variables by Sample..........202 Table 24. Summary of Support for Hypotheses 1.1 to 1.3.........................................203 Table 25. Independent Sample T-test Resu lts, Means, and Standard Deviations for Recovery Strate gies (Continuous Dependent Variables) by Sample....204

PAGE 12

ix Table 26. Independent Sample T-test Results for Recovery Expectations by Recovery Strategies....................................................................................205 Table 27. Chi-square Results of Recovery Expectancy by Diagnosis........................207 Table 28. Independent Sample T-test Results of Recovery Expectancy by Symptom Scales..........................................................................................209 Table 29. Independent Sample T-test Results for Recovery Expectancy and Domain 1 Variables: Age of Onset, Hospitaliz ations, Employment, and Familial Mental Illness...............................................................................210 Table 30. Chi-square Results for Recove ry Expectancy and Hospitalization in Last Year, Current Employment, Substance Abuse, Medication, Abuse and Assault History.....................................................................................211 Table 31. Independent Sample T-test Results, Means and Standard Deviations, for Domain 2 Va riables (Stigma, Social Support, & Empowerment) and Recovery Expectancy...........................................................................213 Table 32. Independent Sample T-test Results for Recovery Expectancy and Domain 3 Service Variables.......................................................................215 Table 33. Summary of Bivariate Results for Research Question 2 Hypotheses.........216 Table 34. Independent Sample T-test Results for Recovery Strategies and Domain 1 Diagnosis Variables...................................................................219 Table 35. Pearson Correlation Results for Recovery Strategies and Symptom Scales..........................................................................................................222 Table 36. Pearson Correlation Results for R ecovery Strategies and Age of Onset, Lifetime Hospitaliza tion, Years Employed, and Familial Mental Illness...224 Table 37. Independent Sample T-test Results for Recovery Strategies and Hospitalization in Last Year, Employment, Substance Abuse, Medication and Abuse/Assault Variables...................................................226 Table 38. Pearson Correlation Results for Recovery Strategies and Domain 2 Variables (S tigma, Social Support, and Empowerment)............................232 Table 39. Pearson Correlation Results for Recovery Strategies and Domain 3 Service Variables........................................................................................235 Table 40. Summary of Bivariate Results for Research Question 3 Hypotheses.........237 Table 41. Control Variable L ogistic Regression Results............................................245 Table 42. Domain 1 (Clinical and Historical Factors) Logistic Regression Results..246 Table 43. Domain 2 (Social Factor s) Logistic Regression Results............................428 Table 44. Domain 3 (Service Factor s) Logistic Regression Results..........................249 Table 45. Main Effects: Logistic Regression Results...............................................250 Table 46. Summary Table: Logistic Regre ssion Results for Research Question 2....252 Table 47. Summary of Bivariate Results for Research Question 2 Hypotheses.........255 Table 48. Results of OLS Regression Testing the Associat ion of Control Variables (Domai n 1) on Recovery Strate gy 1 (Effective Illness Management)..............................................................................................261 Table 49. Results of OLS Regression Testing the Associat ion of Clinical Historical Variables (Domain 1) on Recovery Strategy 1 (Effective Illness Management)...................................................................................262

PAGE 13

x Table 50. Results of OLS Regression Testing the Associat ion of Social Variables (Domai n 2) on Recovery Strate gy 1 (Effective Illness Management)..............................................................................................264 Table 51. Results of OLS Regression Tes ting the Impact of Service Variables (Domain 3) on Rec overy Strategy 1 (Effective Illness Management)........266 Table 52. Main Effects Results of OLS Re gression Testing the Impact of Domain 1-3 Variables on R ecovery Strategy 1 (Effective Illness Management).....267 Table 53. Reduced Model Results of OLS Regression Testing the Impact of Domain 1-3 Vari ables on Recovery Strate gy 1 (Effective Illness Management)..............................................................................................270 Table 54. Summary Table: Results of OL S Regression Testing the Impact of Domain 1-3 Vari ables on Recovery Strate gy 1 (Effective Illness Management)..............................................................................................273 Table 55. Results of OLS Regression Tes ting the Impact of C ontrol Variables on Recovery Strategy 2 (Positive Future Orientation).....................................274 Table 56. Results of OLS Regression Tes ting the Impact of C linical/Historical (Domain 1) on Recovery Strategy 2 (Positive Future Orientation)............276 Table 57. Results of OLS Regression Testing the Associat ion of Social Variables (Domain 2) on Recovery Strategy 2 (Positive Future Orientation)............277 Table 58. Results of OLS Regression Tes ting the Impact of Service Variables (Domain 3) on Recovery Strategy 2 (Positive Future Orientation)............279 Table 59. Main Effects Results of OL S Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 2 (Positive Future Orientation).................................................................................................280 Table 60. Reduced Model Results of OLS Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 2 (Positive Future Orientation).................................................................................................283 Table 61. Summary Table: Results of OL S Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 2 (Positive Future Orientation).................................................................................................286 Table 62. Results of OLS Regression Testing the Impact of Control Variables on Recovery Strategy 3 (Meaningfulness, Personal Control, and Hope)........288 Table 63. Results of OLS Regression Tes ting the Impact of C linical/Historical (Domain 1) on Recove ry Strategy 3 (Meaningfulne ss, Personal Control, and Hope)....................................................................................................289 Table 64. Results of OLS Regression Tes ting the Impact of Social Variables (Domain 2) on Recove ry Strategy 3 (Meaningfulne ss, Personal Control, and Hope)....................................................................................................291 Table 65. Results of OLS Regression Tes ting the Impact of Service Variables (Domain 3) on Recove ry Strategy 3 (Meaningfulne ss, Personal Control, and Hope)....................................................................................................292 Table 66. Main Effects Results of OLS Re gression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 3 (Meaningfulness, Personal Control, and Hope)......................................................................................294

PAGE 14

xi Table 67. Reduced Model Results of OLS Regression Testing the Impact of Domain 1-3 Va riables on Recovery Stra tegy 3 (Meaningfulness, Personal Control, and Hope).......................................................................297 Table 68. Summary Table: Results of OL S Regression Testing the Impact of Domain 1-3 Va riables on Recovery Stra tegy 3 (Meaningfulness, Personal Control, & Hope).........................................................................299 Table 69. Results of OLS Regression Tes ting the Impact of C ontrol Variables on Recovery Strategy 4 (Recognizing Support)..............................................301 Table 70. Results of OLS Regression Tes ting the Impact of C linical/Historical (Domain 1) on Recovery Strategy 4 (Recognizing Support)......................302 Table 71. Results of OLS Regression Tes ting the Impact of Social Variables (Domain 2) on Recovery Strategy 4 (Recognizing Support)......................304 Table 72. Results of OLS Regression Tes ting the Impact of Service Variables (Domain 3) on Recovery Strategy 4 (Recognizing Support)......................305 Table 73. Main Effects Results of OLS Re gression Testing the Impact of Domain 1-3 Variable s on Recovery Strategy 4 (Recognizing Support)...................306 Table 74. Reduced Model Results of OLS Regression Testing the Impact of Domain 1-3 Vari ables on Recovery Strate gy 4 (Recognizing Support).....309 Table 75. Summary Table: Results of OL S Regression Testing the Impact of Domain 1-3 Vari ables on Recovery Strate gy 4 (Recognizing Support).....311 Table 76. Results of OLS Regression Tes ting the Impact of C ontrol Variables on Recovery Strategy 5 (Help Seeking)...........................................................314 Table 77. Results of OLS Regression Tes ting the Impact of Control Variables on Recovery Strategy 5 (Help Seeking)...........................................................315 Table 78. Results of OLS Regression Tes ting the Impact of Social Variables (Domain 2) on Recovery Strategy 5 (Help Seeking)..................................316 Table 79. Results of OLS Regression Te sting the Impact of Service Variables (Domain 3) on Recovery Strategy 5 (Help Seeking)..................................318 Table 80. Main Effects Results of OLS Re gression Testing the Impact of Domain 1-3 Variab les on Recovery Strategy 5 (Help Seeking)...............................319 Table 81. Reduced Model Results of OLS Regression Testing the Impact of Domain 1-3 Va riables on Recovery Strategy 5 (Help Seeking).................321 Table 82. Summary Table: Results of OL S Regression Testing the Impact of Domain 1-3 Va riables on Recovery Strategy 5 (Help Seeking).................324 Table 83. Results of OLS Regression Tes ting the Impact of C ontrol Variables on Recovery Strategy 6 (Symptom Eradication).............................................326 Table 84. Results of OLS Regression Tes ting the Impact of C linical/Historical (Domain 1) on Recovery Strategy 6 (Symptom Eradication).....................327 Table 85. Results of OLS Regression Tes ting the Impact of Social Variables (Domain 2) on Recovery Strategy 6 (Symptom Eradication).....................329 Table 86. Results of OLS Regression Tes ting the Impact of Service Variables (Domain 3) on Recovery Strategy 6 (Symptom Eradication).....................331 Table 87. Main Effects Results of OLS Re gression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 6 (Symptom Eradication)..................332

PAGE 15

xii Table 88. Main Effects Results of OLS Re gression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 6 (Symptom Eradication)..................334 Table 89. Summary Table: Results of OL S Regression Testing the Impact of Domain 1-3 Variab les on Recovery Strategy 6 (Symptom Eradication)....336 Table 90. Summary of Multivariate Results for Research Question 3 Hypotheses....340 Table 91. Mediating Effect of Recovery Expectancy for the Association Between Domain 1-3 Vari ables and Recovery Strate gy 1 (Effective Illness Management)..............................................................................................354 Table 92. Mediating Effect of Recovery Expectancy in the Association Between Domain 1-3 Vari ables and Recovery Strategy 2 (Positive Future Orientation).................................................................................................355 Table 93. Mediating Effect of Recovery Expectancy in the Association Between Domain 1-3 Variable s and Recovery Strategy 3 (Meaningfulness, Personal Control, and Hope)......................................................................................355 Table 94. Mediating Effect of Recovery Expectancy in the Association Between Domain 1-3 Vari ables and Recovery Strate gy 4 (Recognizing Support)...356 Table 95. Mediating Effect of Recovery Expectancy in the Association Between Domain 1-3 Vari ables and Recovery Strategy 5 (Help Seeking)...............356 Table 96. Mediating Effect of Recovery Expectancy in the Association Between Domain 1-3 Variable s and Recovery Strategy 6 (Symptom Eradication)..357 Table 97. Moderating Effects of Recovery Expectancy in th e Association of Domain 1-3 Vari ables on Recovery Strate gy 1 (Effective Illness Management)..............................................................................................360 Table 98. Moderating Effects of Recovery Expectancy in th e Association of Domain 1-3 Variables on Recovery Strategy 2 (Positive Future Orientation).................................................................................................360 Table 99. Moderating Effects of Recovery Expectancy in th e Association of Domain 1-3 Va riables on Recovery Stra tegy 3 (Meaningfulness, Personal Control, and Hope).......................................................................361 Table 100. Moderating Effects of Recovery Expectancy in th e Association of Domain 1-3 Variables on Recovery Strategy 4 (Support Recognition)...361 Table 101. Moderating Effects of Recovery Expectancy in th e Association of Domain 1-3 Variables on Recovery Strategy 5 (Help Seeking)...............362 Table 102. Moderating Effects of Recovery Expectancy in th e Association of Domain 1-3 Variables on Recovery Strategy 6 (Symptom Eradication)..362 Table 103. Moderating Effects of Illness Se verity in the Association of Domain 1-3 Variables on Recovery Strategy 1 (Effectiv e Illness Management)...365 Table 104. Moderating Effects of Illness Se verity in the Association of Domain 1-3 Variables on Recovery Strategy 2 (Positive Future Orientation).......365 Table 105. Moderating Effects of Illness Se verity in the Association of Domain 1-3 Variable s on Recovery Strategy 3 (Meaningfulness, Personal Control, and Hope)....................................................................................366 Table 106. Moderating Effects of Illness Se verity in the Association of Domain 1-3 Variab les on Recovery Strategy 4 (Recognizing Support).................366

PAGE 16

xiii Table 107. Moderating Effects of Illness Se verity in the Association of Domain 1-3 Vari ables on Recovery Strate gy 5 (Help Seeking).............................367 Table 108. Moderating Effects of Illness Se verity in the Association of Domain 1-3 Variable s on Recovery Strategy 6 (Symptom Eradication)................367 Table 109. Summary of Full Model Significant Associati ons Between Recovery Strategies and Independent Variab les (Domains 1-3)..............................371

PAGE 17

List of Figures Figure 1. Recovery Model...........................................................................................6 Figure 2. Model of Conceptua lized Recovery Domains............................................12 Figure 3. Amount of Variance Explai ned by Each Domain of Recovery Expectancy................................................................................................254 Figure 4. Amount of Variance Explained by Each Domain for Effective Illness Management..............................................................................................272 Figure 5. Amount of Variance Explained by Each Domain for Positive Future Orientation................................................................................................285 Figure 6. Amount of Variance Explained by Each Domain for Meaningfulness, Personal Control, and Hope......................................................................298 Figure 7. Amount of Variance Explai ned by Each Domain for Recognizing Support......................................................................................................310 Figure 8. Amount of Variance Explained by Each Domain for Help Seeking........322 Figure 9. Amount of Variance Explai ned by Each Domain for Symptom Eradication................................................................................................335 xiv

PAGE 18

xv Associations Between In dividual, Social and Service Fact ors, Recovery Expectations and Recovery Strategies for Indi viduals with Mental Illness Gary W. Walby ABSTRACT Clinical and consumer recovery effort s continue to advance more successful recovery agendas. Limited research into the e ffect of the expectation to recover and what strategies are most effec tive in enhancing recovery impedes coherent recovery programming. What factors ar e significantly associated w ith recovery expectancy and recovery strategies is still largely unknown. Thus, this study addresses three key gaps in the field. First, which factors associate with the expectation of recovery and is expectation related to strate gy choice? Second, what are common strategies for recovery from mental illness and what factors are associated with each strategy? Third, does recovery expectancy or severity of mental illness mediate or moderate the relationship between clinical, social and service factors and recovery strategies? This study surveyed a sample of 350 randomly chosen participants from a community mental health organization with va rying degrees of ment al illness in a crosssectional study utilizing instruments that meas ured clinical/historical social, and service factors and recovery. The results were exam ined in multivariate analysis targeted to address the gaps noted above. The six recovery strategies included: (1) effective illness management, (2) positive future orientation, (3) meaningfulness, personal control, and hope, (4) recognizing support, (5) help s eeking, and (6) symptom eradication.

PAGE 19

xvi Recovery expectancy was not significantly predicted by any of the clinical, social, or service factors. Although 25% of the vari ance was explained by the full model, factors associated with expectancy differ from recovery strategies and require further investigation. Except for symptom eradica tion, recovery expectan cy was significantly negatively associated with each recovery strategy. Main effects models were significant for all six recovery strategies. The social factor constructs (social s upport, empowerment, stigma) were most consistently and robustly associated with all r ecovery strategies. Variance explained in full models ranged from 71% for positive future orientation to 19% for symptom eradication. However, no mediating or moderating effects were dete cted for recovery expectancy or illness severity. The results of this study furthe r the understanding of recovery and provide information for development of recovery programs.

PAGE 20

1 Chapter One Introduction Mental illness is a leading cause of disability and morb idity in the United States and internationally (Murray & Lopez, 1996). With regard to years lived with a disability, depression is the leading cause worldwide (U stun, Ayuso-Mateos, Ch atterji, Mathers, & Murray, 2004). Mental illness contributes to a reduced life span, especially for individuals diagnosed with schizophrenia, bu t is not itself usually a cause of mortality, with the obvious exception of in creased risk for suicidal beha vior. Nonetheless, in terms of disability, mental illness surpasses all fo rms of cancer and HIV/AI DS and is nearly as disabling as cardiovascular and re spiratory diseases (Ustun, 1999). The financial impact of mental illness is noteworthy. In 1988, the financial burden of mental illness, cons idering direct and indirect costs, was $129.3 billion for the United States, with the largest expenditure for schizophrenia (Wacylenki, 1994). Direct and indirect costs increased to appr oximately $328.4 billion in 2002 (Levy, 2005). Annual costs combined with the prevalence of mental illness makes mental illness more costly to society than other health problems For example, a comparison of expenditures for individuals with HIV/AIDS and serious mental illness finds that those with HIV/AIDS (a disease with relatively low prevalence) spend on average $7,400 annually on illness-related care, whereas those with a severe mental illness (a high prevalence category) spend approximately $5,800 per year. Individuals with ne ither category of illness spend approximately $1,800 annually (Rothbard, Metraux, & Blank, 2003).

PAGE 21

2 Addressing prevalence more specifically, the Centers for Disease Control and Prevention (CDC) estimates the occurrence of HIV infected individuals in the U.S. as approximately 850,000-950,000 with an additional estimated 180,000-280,000 undiagnosed (CDC, 2005). Contrast this with individuals with se vere and persistent mental illness (4.8 million or 2.6% of the popula tion), those with serious mental illness (10 million or 5.4% of the population) or thos e with any mental disorder (44.2 million or 23.9% of the population) (Kessler & Zhao, 1999), and the burden becomes obvious. With the prevalence of severe mental illnesse s alone at twelve times that of HIV/AIDS, it can be easily seen that the cost of mental illness is extraordinary. Additionally, the untreated cost of mental illness may add an a ddition $70 billion or more of uncounted expenditures annually (Kramer, 2000). The cost estimates of mental illness pertai n to all individuals with mental illness, from mild to severely symptomatic. Costs, symptoms and diagnoses all point to mental illness existing on a continuum from transient symptoms related to stressful situations (e.g., divorce) to severe and persistent ment al illness (Kessler, Walters, & Forthofer, 1998; Kessler & Zhao, 1999). However, a ps ychiatric diagnosis alone does not by itself constitute mental illness. A disruption in social, occupational or academic function is required for an individual to be considered ment ally ill. For those with the label severe mental illness (SMI), psychiatric disorders will cluster with one or more primary and severe mental disorders and one or more less severe disorders (Kessler, Davis, & Kendler, 1997; Kessler & Zhao, 1999). The diagnoses with the highest probability of leading to the SMI label are schizophrenia, sc hizoaffective disorder, bipolar disorder, and major depression. A potential problem with di agnosis when comparing across studies is

PAGE 22

3 the consistency of the diagnostic system. For this research, the Diagnostic and Statistical Manual for Mental Disorders, Fourth Edition criteria were used to identify mental disorders by the partnering mental health cente r. Further, all diagnoses were performed by a licensed psychiatrist or by a licensed nurse practitioner and then validated by a psychiatrist. Extreme cases of normally less severe dia gnoses can qualify an individual for the SMI label as well, (e.g., complex post-trauma tic stress disorder or severe obsessivecompulsive disorder). In most cases, indi viduals with mild to moderate mental illness are typically diagnosed with one or more di sorders of less severity and experience less compromise in functioning (e.g., simple phobia or dysthymic disorder). Nevertheless, mild disorders should not be ignored as ther e is evidence that mild disorders place a person at considerable risk for a major ment al disorder (Kessler, Merikangas, Berglund, Eaton, Koretz, & Walters, 2003). Schizophrenia, bipolar disorder and other severe diagnoses often have an earlier onset than less severe disorders and have a period of deteriorati ng functioning (known as the prodromal phase) before meeting criteria for a major mental illness (Lencz, Smith, Auther, Correll, & Cornblatt, 2004; Miller et al., 2002; Yung, et al., 2003). Those with mild to moderate illnesses will experience a shorter or no apparent prodromal phase and their illness is more likely to be (but not exclusively) a response to environmental occurrences (Eaton, Badawi, & Melton, 1995). I ndividuals with limited and contextually stimulated and maintained disorder may remit spontaneously or with limited assistance, while those with severe disord ers typically require more intensive clinical assistance and take longer to remit, if remission ever occurs (DellOsso, Pini, Casano, Mastrocinque,

PAGE 23

4 Seckinger, Saettoni, et al., 2002; Gitli n, Nuechterlein, Subotnik, Ventura, Mintz, Fogelson, et al., 2001). For individuals with severe mental illness, recovery has been defined as living a satisfying life within the constraints of the mental illness (Anthony, 1993; Deegan, 2001). This is a common definition in first person accounts of recovery. Processes of recovery that have been examined include hope, indepe ndence, stigma resistance, treatment, help seeking, self-management and empowerm ent (Borowitz-Ensfield, 1998; Corrigan, Giffort, Rashid, Leary, & Okeke, 1999; Corri gan, Salzer, Ralph, & Okeke, 2005). For individuals with less severe mental illness, recovery has rarely been addressed as a relevant concept even though there can be periods of intense symptoms and compromised functioning (Hayden & Klein, 2001; Sheppard & Teasdale, 2004). Whether recovery requires treatment, can occur spontaneously, or is a combination of both is still a matter of debate. However, part of the appeal of the increasing focus on recovery is that it has reintroduced the idea of hope and meaningfulness, considered essential components of the recovery construct (C orrigan & Ralph, 2005). Among mental health professionals and me ntal health servic es researchers, conceptualizations of recovery have been focused primarily on indi vidual and clinical factors such as symptom control and psyc hiatric self-management skills (Repper & Perkins, 2003; Smith, Bellack, & Liberman, 1996 ; Smith et al., 1996). Previous research has paid limited attention to the role of social context and social in teractions as factors that may explain recovery, reintegration experiences and clinical outcomes (Chadwick, 1997). Consequently, recovery has generally be en equated with treatment factors such as medication compliance. Even with state of the art medication, most consumers with

PAGE 24

5 severe mental illnesses feel unaccepted and lack a basic trust in their ability to care for themselves or to be cared for by others (Walby, 2003a). Furthermore, many individuals with a mental illness have become isolated from the greater society and maintain a supportive network comprised on ly of mental health provi ders, caretakers, and other mentally ill persons (Cressw ell, Kuipers, & Power, 1992; Dickinson, Green Hayes, Gilheany, & Whittaker, 2002). Limiting the concep t of recovery to reduction in clinical symptoms ignores potentially powerful contextual factors that may be a barrier to, or an asset for, an enhanced state of recove ry (Nelson, Lord, & Ochocka, 2001; Repper & Perkins, 2003). Purpose of Study and Significance The purpose of this study was to inve stigate factors that influence the process of recovery for individuals on a continuum of mental illness. In this study recovery was defined as the individuals belief that they will reach life goals and contentment either through elimination of mental illness or within the restrictions of mental illness. Different processes or strategies for recovery have been identified, but what factors support these conduits has had little systematic inves tigation (Borowitz-Ensfield, 1998; Corrigan, Giffort et al., 1999; Corriga n, et al., 2005). For instance, hope, personal challenge, professional assistance, and cont rol of symptoms are all potential individual strategies important for recovery. For this research, the term strategies is used for the different processes of recovery. To describe aspects of recovery as strategies is deliberate to help differentiate between the process of recovery and recovery as an outcome. Further, strategies imply choice and action, supporting the consumer ge nerated recovery literature that suggests such concepts are crucial fo r recovery (Chadwick, 1997). Whether these

PAGE 25

strategies operate independently, and whether some or all are necessary for the outcome of recovery has not been estab lished. Figure 1 illustrates the recovery model tested in this study. Specifically, the domains of indi vidual (referencing clin ical and historical factors), social, and service factors were tested for direct associations with expectation of recovery and recovery strategies. Further, recovery expectancy c ould either mediate or moderate the relationship between the individua l, social, or service domains and recovery strategies. Both of these poten tial relationships we re tested. Finally, severity of mental illness was also examined for potential moderating affects on the relationship between Figure 1 Recovery Model 6 Individual Factors Domain Social Factors Domain Service Factors Domain Recovery Strategies Recovery Expectancy Severity of Mental Illness

PAGE 26

7 individual, social, and service dom ains and recovery strategies. Recovery has been defined in the context of life satisfaction within the constraints of some level of psychological impairment (Anthony, 1993; Deegan, 2001). This is based on the assumption that recovery is only relevant to individuals with severe mental illness. Another assumption is that individuals with mental illness be lieve that they will always be mentally ill and that all indivi duals pursuing recovery accepts some level of impairment as inevitable. These beliefs and expectations have not been tested empirically. However, pilot studies for this project have demonstrated that allowing for both elimination and restriction of mental illness as recovery outcomes is important because some individuals with differing symptom severity and illness trajectories have endorsed the ability to eradicate mental illness from their lives while others believe that mental illness is something they will cope with forever (Walby, 2003a, 2003b). This pattern of findings runs counter to the preva iling yet largely untested assumption that all individuals with severe mental illness believ e they will always have mental illness and that less severe cases always believe that they will eventually eliminate or be cured of mental illness. One unique contribution of this study was to investigate expectations of recovery from mental illness. Whether a person has an inherent belief in recovery may impact the degree an individual chooses to endorse differe nt recovery strategies This objective was important objective because other studies ha ve assumed that recovery is a process embraced by all and one that each individual wi ll define and accomplish in his or her own way (Chadwick, 1997; Deegan, 1998, 2001). Belief that a goal is attainable or that a

PAGE 27

8 process is within the individua ls behavioral capacity will a lter the degree of effort and perseverance committed to accomplishing the tasks necessary to complete the goal. Another contribution of this study was to broaden understanding of the recovery concept by applying recovery to individuals wi th different severities of mental illness. Recovery may or may not be relevant to in dividuals who are clini cally compromised but not severely mentally ill. If recovery is relevant to nonSMI individuals, do they expect to recover completely and what strategies do they endorse? How is this different from those with a severe mental illness? Engagi ng individuals with less severe symptoms and diagnoses into the process of recovery ma y normalize the prevalence of mental illness decreasing stigma or, at least, lower the bur den of isolation carried by the majority of those with severe illnesses. Following investigation into recovery expectancy and the pote ntial relevance of recovery to the less severely impaired, the next task was to begin to understand what factors affect choice of recovery path a nd expectation of recovery outcome. These choices can be sorted into individual/clinical-historical, so cial, and service domains. Previous recovery research had been focuse d on individual and clini cal factors such as symptom control and psychiatric self-managem ent skills (Smith et al., 1996). Similar factors (e.g., age of onset, tr eatment compliance, diagnosis and other variables) were tested for association with reco very strategies or pathways identified by other researchers. Thus, considerable previous research has targ eted clinical aspects of recovery and, while not completely at odds with the consumer vision of recovery, the research has not included social and service f actors to any great degree.

PAGE 28

9 The roles of social factors in recove ry strategies and outcomes are poorly understood. Levels of support, satisfaction w ith relationships, expe riences of belonging to ones community and internaliz ation of stigma are potential factors that could affect recovery beliefs and choices. A thorough exam ination of these factors is presented in Chapter 2. Finally, the degree that services affect recovery strategies and the recovery expectancy was investigated. This serves an important purpose because service intensity and satisfaction have not been investigated for an association with different recovery strategies. In fact, there is controversy about whether being involved in services supports or inhibits the individual path to recovery (Chadwick, 1997). The results of this research will inform the development of linkages between mental health agencies and community-based initiatives to facilitate recovery and reintegration of persons with severe mental illness into communities. Understanding recovery strategies and what influences them will assist in the development of targeted recovery programs that will lessen the bur den of mental illness on individuals and society. Additionally, understanding recovery will assist in both primary and secondary prevention programs targeted to mental illness. Finally, the construc t of recovery can be expanded by exploring the generalizability of the concept of recovery to different severities of mental illness. Study Rationale The recovery paradigm first introduced in the mid-1980s continues to gain momentum. The paradigm, generated by consum ers of mental health services, is based on hope, empowerment and choice in a clinical atmosphere often believed by consumers

PAGE 29

10 to be targeted exclusively to symptom re lief and not personal growth (Fisher, 2003; Mueser et al., 2002; Resnick, Rosenheck, & Lehman, 2004). Rec overy from mental illness, however, has been difficult to define and in most cases has simply incorporated recovery concepts from physical disability and addictions research without considering the unique aspects of mental illness. To be gin to close this gap, this study addressed various aspects of mental illn ess recovery that, up to now have limited exposure in the literature. One aspect of recovery conspicuously abse nt from the literatur e is the degree that individuals believe they are going to recover from mental i llness. Expectancy beliefs may inhibit or promote whether recovery st rategies are utilized or the strategy for recovery chosen. In turn, various factors may influence expectation beliefs. The exploration of recovery expectan cy in this study took the following form: first, recovery expectations of the sample su rveyed was quantified, next fact ors significantly associated with expectancy were identified, and lastl y, recovery expectanci es relationship with recovery strategies was explored. The research targeting recovery strategies is limited as well. Further, factors that may be associated with such preferences (e .g., level of psychiatric impairment) have not been studied. Thus, this study identifies preferred strategies for recovery for the combined sample. These strategies reflect aspects of personal empowerment, support, professional assistance, and othe r relevant pathways identified in the literature (BorowitzEnsfield, 1998; Corrigan, Giffort et al., 1999; Corrigan et al., 2005). To more fully understand recovery, it is not enough to understand if individuals believe they will recover and what their curren t practices to operationalize their recovery

PAGE 30

11 journey entail. It is necessa ry to understand what factors are associated with each strategy chosen and how much of the varian ce of strategies they explain. The factors influencing recovery can be grouped into three domains. These domains are not arbitrary, but they are also not consistent with other studies in the literature. An exhaustive examination of the current recovery literature informed the choice to utilize three domains. To investigate what infl uences recovery, these three domains are hypothesized to interact to influence oppor tunities and preferen ces for recovery trajectories (Figure 2). The first domain is the individual factors, historical and clinical factors domain that includes psychiatric symp toms, age of onset, hospitalization history, familial history of mental illness, and employ ment factors. These factors are both clinical and historical because they assess current symptom levels and diagnosis as well as recollection of key aspects of mental health history asso ciated with disorder, e.g., hospitalization history and fa mily members with a diagnos ed mental illness. The second domain consists of the social c ontext/individual social fa ctors that each individual interacts with and is influenced by. The c onstructs in question in clude social support, stigma, and perceived cohesiveness between se lf and society. The social context domain had received the least attention in previous recovery research but may be vital for generation of successful interventions. The third domain (clinical and su pport services domain) encompasses the services in place to provide clinical intervention (therapy, medication), and assistance to promote commun ity living and to prevent relapse and rehospitalization (i.e., case management, supported living, supported employment, and residential treatment). However, this st udy did not examine the individual impact of different clinical and service support choices in this analysis but instead examined the

PAGE 31

Figure 2 Model of Conceptualized Domain s Associated with Recovery Individual, Historical and Clinical Factors Domain Social support Social connectedness Trust in others Empowerment Stigma internalization Total number of services Frequency of contact Satisfaction with services Relevancy of recovery Stability of housing Diagnosis Current symptom levels Medication Age of onset Abuse/trauma experiences Hospitalization history Employment Substance use/abuse Familial history of mental illness Clinical and Support Services Domain Social Factors Domain 12

PAGE 32

13 amount of contact and satisfac tion with services for potentia l association with recovery expectancies and strategies. An innovation of this study was the simultaneous attention to all three domains and how th ey contribute to recovery al ong the separate courses. Several surveys, each with multiple sub-scal es relevant to the domains outlined above were utilized to operationalize and quantify the constructs. Limitations of Existing Knowledge Base A lack of consensus on what constitute s recovery for individuals with mental illness continues to hinder research and program development. There are four main directions that recovery resear ch has taken thus far. Together, these have evolved into what is called the recovery movement in this study. The recovery movement can be theorized as a fifth direction and is described below. The first direction is a focus on individual narratives of recover y, fostering a belief that recove ry is a private and singular process that is very personal and individual with limited commentary on factors that may be common across those recovering (Chadw ick, 1997; Davidson, 2003; Deegan, 2001). The second direction begins the search for commonalities via qua litative research that has sought to identify themes within samples id entified as recovering from mental illness (Davidson, Sells, Sangster, & OConnell, 2005; Sells, Stayner, & Davidson, 2004; Young & Ensing, 1999). The next direction equates recovery with treat ment compliance and symptom reduction, reducing the concept of rec overy to measurable clinical concepts while ignoring aspects of re habilitation, services, employ ment, education, and other social factors (Holzinger, Loffler, Mulle r, Priebe, & Angermeyer, 2002; Svedberg, Backenroth-Ohsako, & Lutzen, 2003; Ziguras, Klimidis, Lambert, & Jackson, 2001). In other words, recovery is tantamount to clini cal success. Finally, recovery is viewed as

PAGE 33

14 synonymous with employment. A recovere d person, noting the language that views recovery as outcome versus process, is a pe rson able to obtain and maintain employment (Marwaha & Johnson, 2004; Provencher, Gregg, Mead, & Mueser, 2002). The recovery movement represents th e current philosophy, service provision, and policy of recovery. The recovery movement is an ongoing reconceptualization of mental health services that has beco me a guiding vision for mental health in the United States (Presidents New Freedom Commission on Mental Health, 2003). Initially begun in protest of and opposition to the perceived reductionistic practi ces of the clinical treatment establishment there has been recent efforts at rapprochement between clinical services and the consumer empowerment movement in recovery (Corrigan & Garman, 1997; Nelson, Lord, & Ochocka, 2001). The shared goal of recovery is reflected in the emphasis on empowerment and choice as key aspects of the Presidents New Freedom Initiatives vision, the primary policy doc ument driving the recovery movement. Recovery in the New Freedom Initiative prom otes the expectation that consumers make informed choices from available clinical serv ices that are, hopefully, evidence-based as well as supportive (e.g., supporte d housing) with the overa rching goals of defeating stigma, reducing unfair treatment limitations, an d making the fragmented service delivery system more cohesive. Though consistent with some of the key aspects of the recovery movement there are still obstacles in ope rationalizing the vision of the recovery movement. Lack of access to recovery base d services and continued mistrust of the consumers capacity for making informed j udgments remain key obs tacles (Corrigan & Ralph, 2005).

PAGE 34

15 The lack of consensus for a defined construc t of recovery that satisfies the clinical professionals, social service advocates, consum er run organizations, national, state, and local mental health lobbying organizati ons, and policy makers further restrict understanding of what promotes or limits recovery. Various factors culled from first person reports, qualitative studi es, empirical studies targeting treatment compliance and the growing body of recovery based literature demonstrates a wide array of potential covariates in the recovery process. There ha s been little effort as yet to systematically investigate these factors for direct, medi ating, or moderating effects. A further contributor to the uncertainty stems from whether recovery is a process or an outcome (Andresen, Oades, & Caputi, 2003; Corriga n, Calabrese et al., 2002). What would comprise a recovered individual? Is a person with mental illness continually on the path of recovery with some individuals further along or further behind (i.e. the addictions recovery model)? First person accounts a nd qualitative research are more likely to promote the process of recovery and agree w ith the concept of a process with no end than quantitative/empirical researchers who would prefer a measurable endpoint. This study addresses the issue of rec overy as a process, reflecti ng the ongoing use of strategies with the assumption that as an individual becomes non-symptomatic the option exists to discontinue active implementation of recove ry strategies until once again needed. Various strategies and covariat es of recovery were measured as indicators of an active recovery process and captured in five research questions that were targeted in this effort. Research Questions RQ1: To what degree does illness severity influence beliefs in recovery for the individual?

PAGE 35

16 RQ2: Are individual, social or service factors associated with recovery expectancy? RQ2.1: Are individual factors associated with recovery expectancy? RQ2.2: Are social factors associated with recovery expectancy? RQ2.3: Are service factors associat ed with recovery expectancy? RQ3 : Are individual, social or service fact ors associated with recovery strategies? RQ3.1: Are individual factors associated with recovery strategies? RQ3.2: Are social factors associated with recovery expectancy? RQ3.3: Are service factors associat ed with recovery expectancy? RQ4a: Does the expectation of recovery mediate the relationship between individual, social and service factors and recovery strategies? RQ4b: Does the expectation of recovery mode rate the relationship between individual, social and service factors and recovery strategies? RQ5: Does severity of mental illness moderate the relationship between individual, social, and service factors and recovery strategies? Overview of Design This study is a cross-sectional/eco logical, nonexperimental study with retrospective elements that utilizes primary data. Individual fact ors (e.g., symptoms) and contextual factors (e.g., social support), representing microand meso-level factors were investigated simultaneously. The population investigated for this research was individuals with mental illness. Further, this research addressed two clinical subgroups of this population, individuals with severe me ntal illness and indivi duals with mild to moderate mental illness. Random sampling of both samples was completed with assistance from the partnering agency. The samp le with severe mental illness is referred

PAGE 36

17 to throughout this document as the SMI samp le. The mild to moderate sample engaged in outpatient treatment is referred to as the OP sample. A full description of the population, sampling, inclusion and exclusion criteria are provided in Chapter 3. This study utilized a three-part design. First, the re spondents were assessed for their recovery beliefs and endorsement of diffe rent recovery strategies. This included comparing the SMI and OP samples on whether recovery is expected and if there is a significant relationship. Second, the study investigated recove ry strategies (dependent variables) and various factors believed to be associated with recovery (independent variables) in bivariate and multivariate analys es. Third, the mediating effect of recovery expectancy and the moderating effect of expectancy and severity of illness were explored. Implications for Public Health Implications for this study include both research and practic al benefits. In general, this project challenge s the Public Health core conc epts of equity and social justice through focusing on understanding the ne eds of a highly disenfranchised group of citizens. There exists an inherent bias towards and stigmatization of this population. Indeed, structural stigma in legislative bodies that control policy and funding for mental health services has been i nvestigated (Corrigan, et al ., 2005). Further, increased understanding of the mentally ill will help professionals in pub lic health to better understand the morbidity attached to mental illness in severe forms as well as less severe, and far more common, forms. This is essentia l as public health con tinues to move away from its traditional roots embedded in the study of mortality to encompass quality of life, functioning and wellness issues.

PAGE 37

18 The concept of recovery from mental illness has, before this study, been exclusively applied to severe forms of mental illness, psychological responses to physical disabilities or injuries, and addictions. By expanding the concept of recovery and testing its applicability to less severe forms of mental illness, this study tested the flexibility and applicability of the recovery concept and broadened the c oncept of population at risk. Delimitations The following are delimitati ons of this study enforced by the researcher. 1. The sample of individuals in the severely mentally ill sample was diagnostically limited to schizophrenia (any type), sc hizophrenia spectrum disorders (e.g. paranoid disorder, delusional disorder, schi zoaffective disorder), bipolar disorder (Bipolar I or Bipolar II), and major depression. Though other individuals with psychiatric diagnoses, (e.g., severe obse ssive-compulsive disorder, severe posttraumatic stress disorder) can meet the ge neral criteria for SMI, there is less consensus for their inclusion and they constitute only a small proportion of the SMI population (Kessler, 2000; Kessler et al., 1996; Kessler & Zhao, 1999). However, for individuals identified as SMI and not diagnosed with schizophrenia, schizoaffective disorder, bipolar, or major depressi on, a conference with their primary clinician was completed before de ciding whether to include or exclude the individual. 2. The outpatient sample contained any out patient therapy consumer who did not meet the criteria for SMI used in this study and also excluded consumers who were receiving only substance abuse treatment.

PAGE 38

19 3. Individuals randomly selected who are inca rcerated or hospitalized for psychiatric or medical reasons were ineligible due to access and stabilit y considerations. Similarly, consumers randomly selected who were living in the community but judged too unstable to complete data co llection were not eligible for the study. Limitations 1. Because of the study delimitations noted a bove and the fact that all participants will be selected from one mental health ag ency (albeit a large agency with 33 sites spanning three counties), this study does not generalize to all mentally ill individuals. 2. As this study focuses on individuals with a SMI or less severe type of mental illness who are in treatment, the results do not generalize to individuals who are not involved in mental health services. 3. This study is restricted to the instruments chosen or created for the purpose of the study and did not investigate in-depth the contextual milieu of the participants. 4. The instruments chosen to measure the cr iterion variable (rec overy) have only had limited use in studies and, though psychometri cally sound, lack a tr ack record that further establishes their validity. However, they were designed with the full ongoing participation of individuals with mental illness and the factors identified were congruent with the cu rrent recovery literature. 5. Because the design is ecological and not longitudinal there ar e risks of ecologic fallacies and the limitation of articulati ng associations versus causal pathways.

PAGE 39

20 Select Definitions Bipolar Disorder : Bipolar disorder is a cyclical disorder constituting both major depressive disorder and the excessive en ergy and poor judgment of manic disorder. Potential symptoms when depressed include sadness, helplessness and hopelessness, loss of energy and interest, negativ e thought processes and dist urbed sleep and appetite. When manic, possible symptoms include excessive energy, reduced need for sleep, grandiose assessment of abiliti es that can become delusion al, racing thoughts and flights of ideas, poor judgment and impulsive behavior For this study Bi polar I and Bipolar II disorders were eligible for inclusion. Bipol ar I diagnostic codes included 296.0 (single episode), 296.40 (most recent episode hypomanic), 296.4x (most recent episode manic), 296.6x (most recent episode mixed), 296.5x (most recent episode depressed), and 296.7 (most recent episode unspecified). The x provides additional information as follows: if x = 1 the disorder is mild, 2 = moderate, 3 = severe without psychotic features, 4 = severe with psychotic features, 5 = in partial remission, 6 = full remission, and 0 = unspecified (American Psychological Association, 1994) Bipolar I include manic and mixed episodes that are more severe then the hypoman ic episodes associated with Bipolar II. Bipolar disorder was previously called manic-depression. Ecologic Fallacy : The fallacy sometimes present when drawing inferences at the individual level (that is, rega rding relations between indivi dual level variables) based on group level data. In other words, assuming that the differences noted at the aggregated or group level is representative of a specific individual. The ecol ogical fallacy arises because associations between two variables at the group level (or ecological level) may differ from associations between analogous va riables measured at the individual level.

PAGE 40

21 Empowerment: Empowerment is a multi-level cons truct that includes an individuals feeling of personal power and control, an organizations ability to help its workers and customers or consumers feel that they have choices and a voice, and a belief within a community that its citizens can accomplish th eir goals and promote their own health. Thus, empowerment is the experience and be nefits of choice, independence, voice and control ecologically embedded at multiple le vels of experience by and within society. Major Depression : A major depressive episode is a period of time of at least two weeks where the individual experi ences extreme sadness, loss of hope, and a feeling of helplessness as well as other somatic and psychological symptoms. Though the major symptom picture of the disorder is only re quired to be present for two weeks for a diagnosis to be made, the majority of episode s last considerably longer and can be quite debilitating. Major depression disorder is when more than one episode of intense depression is reported. Multiple episodes over ti me indicate a chronic course with higher likelihood of being resistant to treatment. Individuals with a diagnostic code of 296.2x (single episode) or 296.3x (recurrent) were in cluded in the study. The use of the x parallels bipolar disorder. Recovery : Recovery is an elusive concept still poorly define d. Consumers in one study succinctly defined recovery as psychological recovery from the consequences of illness (Andresen et al., 2003). Other definitions have touched on recovery as a process of 1) overcoming "stuckness," 2) discovering and fo stering self-empowerment, 3) learning and self-redefinition, 4) returni ng to basic functioning, and 5) improving quality of life (Young & Ensing, 1999).

PAGE 41

22 Schizoaffective Disorder : This disorder combines schizophrenia with major depression or bipolar disorder. Diagnostic code is 295.7 for all cases. Schizophrenia : This disorder has, in general, two different symptom clusters. The first is the negative cluster where the individual does not experience what would be normally expected in a situation, (i.e., lack of emotion, lack of interest, disruption in experiencing pleasure from that which used to be pleasurab le). The second cluster is the positive cluster in which the individua l has experiences they are not supposed to have, (i.e., delusions, hallucinations, loose associations, di sruption in thought process). The disorder has also been called a psychotic or thought diso rder. The major disrup tion is in cognitive processes though depression and other mood re lated symptoms are common. Diagnostic codes identified in this study included: 295.30 (schizophrenia, paranoid type), 295.90 (undifferentiated), 295.60 (residual type), and 297.1 (delusional disorder). Severe Mental Illness : The combination of a major mental disorder, one or more relapses and disturbance in occupational, academic, or educational functioning defines SMI. Social Support : The level of assistance from others that an individual can count on in various domains, (i.e., emotional support, instrumental or practical support, knowledge support). Stigma : Stigma is defined as the co-occurren ce of its components: labeling, stereotyping, separation, status loss, and di scrimination. Stigma is both an occurrence outside and a process within an individu al (Link & Phelan, 2001).

PAGE 42

23 Chapter 2 Comprehensive Literature Review Introduction/Overview According to the Department of Health and Human Services (DHHS), individuals with severe mental illness (SMI) constitute one of the largest disenfranchised and stigmatized populations in the United States (DHHS, 1999). Indeed, mental illness is a factor in marginalization regardless of leve l of severity. Though ubiquitous with twentyfive percent of the population meeting criteria for a psychiatric diagnosis at any given time (Kessler & Walters, 2002; Kessler & Zhao, 1999), mental illness remains for many a mystifying and troubling event. There is no single cause of mental illness and multiple risk factors appear to be related to onset of most disorders. It also holds true that many risk factors are causally relevant for the onset of more then one disorder. Etiological uncertainty combined with a ttributions of dangerousness, disruption of identity cohesion, and manifestation of unpredictable behavior and emotions, limits responses to the mentally ill to primarily anger, pity or, fear (Corrigan, 2000; Corrigan, Markowitz, Watson, Rowan, & Kubiak, 2003). Despite causal uncertainty, there are effective treatments for many mental disorders. These evidence-based practices va ry in the degree that they incorporate recovery principles of choice and empow erment (Azrin & Goldman, 2005; Resnick, Fontana, Lehman, Rosenheck, 2005). Efforts to increase understanding, access, and

PAGE 43

24 evaluation of evidence-based practices continue though the majority of mental health services remain traditional, meaning not developed as evidence-based practices but are endorsed via longevity (Shem & Evans, 2005). However, evidence-based practices are more often developed for mild to moderate levels of symptoms and impairment and symptom resolution and the omnipresent threat of relapse are the best that conventional treatment can offer for severe mental illness. Further, the availability of efficacious treatment is limited and parity in coverage via insurance is unrealized (Abrahamson, Steel, & Abrahamson, 2003). Consumers have responded to limited and rigid services by leading the development of the concept of recove ry (Chadwick, 1997; Corrigan & Ralph, 2005; Deegan, 2001). Recovery, initially viewed with skepticism by the professional psychiatric community, is becoming more accep ted and mainstream. The President New Freedom Commission on Mental H ealth, (2003) was directly responsible for launching the concept of recovery into ma instream psychiatry. Detractors of this effort have argued that this policy documents main purpose was to disown clinical responsibility onto the afflicted and to consumer organizations that are often poorly funded and lacking skills in program development and evaluation. There are further accusations that instead of organizing and streamlining service provision while bringing evidence-baaed services to more individuals in need, the result is less funding and responsibility for the government while privatizing to the lowest bidder w ith limited accountability. There is limited evidence for or against these allegations and there are many others that believe that the Presidents Initiative is an accurate and stra ightforward analysis with reasonable policy

PAGE 44

25 suggestions. Regardless, the Commissions repo rt has served to raise recognition of the recovery concept. Research in recovery is attempting to de fine the construct in relation to similar concepts (e.g., reintegration and social inclusion), formalize the relationship between recovery and clinical interv ention, and to understand the factors that promote and inhibit recovery. At the same time, the concept of recovery is being defined by multiple stakeholders based on philosophical differences that are at times conflicted and others cooperative. The suspicion noted above colors some of the attempts in making recovery a beneficial empirically based c oncept but has, for the most part, been limited to rhetoric. Of more concern is a reactive integration of recovery based services into mainstream clinical services without prope r consideration of contextual factors, impact on already established clinical, but non-recovery based, services a nd without incorp orating proper evaluation design to meas ure service effects. To better understand recovery in its complexity, this chapter is organized into several sub-sections that provide an overview and synthesis of releva nt literature as well as providing the structure for this research effort. There is some variability in the degree of coverage for each section as the breadth and depth of reporting depends on the availability of empirical information. First, those with severe mental illness are described and then contrasted with indivi duals who have less severe form s of mental illness. Next, recovery strategies and their many dimensions will be elaborated. This will be followed by descriptions of three domains of associated factors that appear to impact recovery from mental illness. The first domain is comprised of individual factors, events, and familial influences that may be related to onset, maintenance, and recovery from mental

PAGE 45

26 illness. The second domain is consists of soci al factors that may a ffect recovery, (e.g., stigma, social support, and empowerment). The third domain is comprised of service factors that may increase or decrease the process and outcome of recovery. The risk of covering broad areas of inform ation deemed relevant to recovery is a fragmented discourse, reflecting, unfortunately, the current state of the field. Every effort has been made to integrate the information in relation to this research effort. The literature reviewed for each section will focus on the relevant points of the subject and its relation to recovery. Thus, the common or fo cusing theme that ties each of the following sections together is recovery. Mental Illness Mental illness exists along a continuum from severe to mildly symptomatic (van Os et al., 1999; Verdoux & van Os, 2002). Many individuals successfully function in their daily lives with mild to moderate symptoms (Kennedy, Abbot, & Paykel, 2004; Solomon, Haaga, & Arnow, 2001). Others e xperience symptoms that cause varying levels of disturbance in their ability to f unction in social, occupational, and educational settings (Klein, Schwartz, Rose, & Leader 2000; Ormel, Oldehinkel, & Brilman, 2001). The more severe, frequent and robust th e symptom pattern, the more likely that disturbance in functioning will result. Those labeled as severely mentally ill (SMI) are individuals who manifest re peated episodes of depressi on, psychosis, mania or other symptom clusters. Multiple episodes over a lifetime can result in substantially impaired functioning in key settings and in completing da ily tasks of living. Disorders that carry the label of severe mental illness are cons idered chronic because of their tendency to cycle from no discernible symptoms to highly symptomatic ov er the life course.

PAGE 46

27 Typical diagnoses of the SMI include major depression, schizophrenia, bipolar disorder and schizoaffective disorder. Approxi mately five to seven percent of the U.S. population is believed to suffer from a severe mental illness (Kessler & Walters, 2002). According to the American Psychiatric Asso ciation (APA), and others, individuals are qualified for this label via diagnosis with a major psychiatric diagnosis, a chronic course, and impairment in social, occupational a nd educational functioning (APA, 1994; Kessler, & Walters, 2002). Typically, these consum ers demonstrate a pattern of short hospitalization and discharge followed by repeated admissions. The most important characteristic of these patients is that they have rarely been treated successfully (Morin & Seidman, 1986), although the picture is confused further by several long term studies that indicates that partial or total remission has been observed for indi viduals with severe mental illness as they age (Calabrese & Corri gan, 2005). Investigation into the burning out of mental disorders over time appears to support that the effect does exist when considering the population as a whole, but with little knowledge of how the phenomenon occurs. Treatment for individuals labeled SM I is usually a combination of medication to reduce primary symptoms and psychosocial interventions to increase quality of life and successful adaptation to community livi ng (Barton, 1999; Zygmunt, Olfson, Boyer, & Mechanic, 2002). Individuals with mild to moderate ment al illness tend to present with a much wider array of potential symp toms across the population while the SMI population is more individually seriously symptomatic. Different manifestations of anxiety, depression, dissociation, or somatic symptoms are common in mild to moderate cases. At times, behavioral manifestations such as eating disorders, impulse control problems,

PAGE 47

28 and addictions are the presenting problem. What separates this group from individuals with a SMI is the level of symptom severi ty, less consistent a nd pervasive functional impairment, and greater potenti al for successful treatment. Categories of mental illness diagnoses are detailed in the Diagnostic and Statistical Manual of Mental Disorders, 4 th edition (APA, 1994). The vast majority of diagnoses relevant to the SM I are contained in the mood and schizophrenia and other psychotic disorder categories. While there is crossover in categories of diagnoses for SMI and non-SMI labeled individuals, this is in most cases differences in primary diagnosis and severity of the disorder. Individuals with le ss severe impairment often receive a primary diagnosis from many othe r diagnostic categories, e.g., anxiety, sleep, adjustment, and sexual disorders. Individuals with a SMI often meet the criteria for multiple diagnoses with one or more majo r diagnosis (e.g., schizophrenia or bipolar disorder) and often one or more less seve re diagnosis like the ones listed above. Important to the understanding of mental illn ess, especially severe mental illness, is that diagnoses are manifested in differe nt ways by different individuals and that a diagnosis alone does not a label of SMI make. Yet, a diagnostic label is a powerful carrier of perceived etiologies, behavioral expectations and stigma (Link, Phelan, Bresnahan, Stueve, & Pescosolido, 1999). Thus, to describe a mental illness will require at the least an understanding of the diagnos is, symptom expression, social, occupational or educational impairment, attributions of blame, stigma, and support factors. Recovery Recovery has been defined as living a sa tisfying life within the constraints of mental illness (Anthony, 1993; Deegan, 2001). Deegan (1998), a health professional

PAGE 48

29 successfully adapting to severe me ntal illness, stated recovery refers to the lived or real life experience of people as they accept and overcome the challenge of disability they experience themselves as recovering a new sense of self and of purpose within and beyond the limits of the disability (p 12). Recovery challenges tr aditional notions of psychiatry. Kraepelin (1913) or iginally postula ted that severe mental illness (especially schizophrenia) had an inevitable course leading to permanent impairment and dysfunction. However, longitudinal studies indicate that from one-half to two-thirds of individuals with schizophrenia, or other mental disorders severe enough to require ongoing hospitalization over the co urse of years, no longer needed to be hospitalized starting in their midto late-40s, were rela tively symptom free, ma ny were able to work in some capacity, and most were living with family or friends (Calabrese & Corrigan, 2005). Personal accounts and empirical evidence have led to different beliefs regarding recovery from the psychiatric and consumer perspectives, increas ing tension between these groups (Fisher, 2003; McGruder, 2001). However, recovery has succeeded in catching the attention of the Federal government that, desp ite the problems still facing acceptance and understanding of recovery, have made it the cornerstone of the report from the Presidents New Free dom Commission (Hogan, 2003). Recoveries conflict with clinical science is arguably based more on the limitations of the science differing with the desires of the consumers. Hi storically there is evidence that some treatments may exacerbate sympto ms. Further, diagnoses may be biased toward the view that certain mental il lnesses, especially schizophrenia, are insurmountable. For instance, treatments with conventional neurolep tics cause some of the negative symptoms (e.g., anhedonia, apathy) and slowed cognitive processes that can

PAGE 49

30 be misinterpreted as symptoms of disorder. This has been referred to as the neurolepticinduced deficit syndrome (NIDS) (Gerlach & Larsen, 1999; Velligan & Miller, 1999). Diagnostically, schizophre nia spectrum disorders (e .g. delusional disorder, schizoaffective, schizophreniform), vary in the degree of expe cted recovery. For instance, schizophreniform disorder is diagnosed in place of schizophrenia when the episode (prodromal, active, and residual) la sts between oneand 6-months, absence of blunted or flat affect, and requires g ood premorbid functioning (APA, 1994). This disorder is less severe and has a higher expectation of recovery than schizophrenia itself (Kruger, 2000b; McGorry, 1992). Thus, clinical ly and diagnosticall y, schizophrenia, and to a lesser extent major depression and bipolar disorder, is biased away from recovery and toward the expectation of permanent impairment. Before continuing on with a discussion of recovery, it is important to briefly mention other terms that have been used in the same context as r ecovery, specifically reintegration and social inclusion. Reintegrat ion appears to be a te rm that signifies the outcome of recovery. Individuals who are reintegrated live successfully and independently in the community (Hartman, 1996) Reintegration has also been closely aligned with research in stigma management and a reintegrated person is one who is either stigma resistant or is no longer stigma tized because they are fully reintegrated and accepted in the community (Herman, 1993). Rein tegration is also a term used with treatment compliance, especially medicati on compliance, and acquiring of appropriate life skills for independent living (Littrell & Littrell, 1998; Smith, et al., 1996). Thus, more narrowly, reintegration can also be viewed as the successful clinical intervention of individuals with mental illness who are back in the community.

PAGE 50

31 Social inclusion has a broader focus, akin to recovery, that takes into account the subjective experience of belongi ng to the community and the va rious processes that bring one from marginalization to inclusion (Reppe r & Perkins, 2003). So cial inclusion is a term used mostly in the lite rature from the United Kingdom and means that an individual with mental illness is accepted to the same extent as a person without mental illness while fully realizing potential functional limitations and influence of activ e symptoms. Social inclusion is bi-directional in that the indi vidual feels accepted by the community and in turn accepts the community. Social inclusion also implies that stigma and discrimination have limited impact on an individual. Further, the holistic and comprehensive characteristics of social inclusion pla ces emphasis on rehabilitation as a broadly encompassing construct that includes employme nt and independent living skills as well as social, relationship, health maintaining and other skill sets necessary for seamless immersion in the community. Though reflectiv e of positive values, realistic efforts to maintain SMI individuals in the community have focused on psychosocial intervention and clinical treatment with lim ited resources placed into re habilitation. Inclusion back into the community, and maintenance of an in tegrated community-based status, has been consistently secondary to c ontrol of psychiatric symptoms and the skills to support medication treatment. Much like recover y, there is limited understanding of the mechanisms that either promote or support social inclusion (Repper & Brooker, 1998; Repper & Perkins, 2003). Recovery has been investigated from clin ical and social perspectives. Clinically, psychiatry would likely define recovery as the cessation of symptoms and the restoration of functioning sufficient for self-care. This is consistently referred to as illness

PAGE 51

32 management (Mueser et al., 2002). There is generally one common route to clinical recovery. This includes compliance with medication and treatment regimens, limiting chronic and acute stress, engagement in supportive services as needed, and skill building (interpersonal and employment) through psychi atric rehabilitation. Thus, the focus is on objective measures of symptoms and functioning. This is an improvement over the traditional psychiatric view of severe mental illness where it is considered a chronic and permanent condition in which cure is impossi ble so a life of adaptation to illness, or succumbing to the illness, is guaranteed (Kr aepelin, 1919). Conspicuously absent from the older clinical views are concepts su ch as personal growth, empowerment, goal attainment or choice. Recovery as a social phenomenon parallel s growth of the consumer movement. Consumers have banded together to fill perceived gaps in services and have also sought more compassion and hope for their lives w ithin and without the traditional treatment establishment (Bassman, 2001; Dickerson, 1998). There are various pathways to recovery and individuals have choices in what they define as recove ry and the process to achieve it. Indeed, recovery is considered a very personal process w ith no clearly defined outcome. For some consumers, the concept of recovery equates to clinical recovery defined above, for others there is acceptance of the clinical pathway as part of the process but regard it as insufficient on its own, while for others clinical rec overy is a false door that leads to dependence and a ddiction to prescribed medicati ons that at first take the place of quality of life, and then commandeers it. Regardless, the clinical pathway is considered one of many possible avenues to recovery that an individual consumer may or may not endorse.

PAGE 52

33 On the other hand, the language of the c onsumer and related recovery literature suggests that recovery is inev itable or at least a normativ e process that all individuals with SMI are in some way part of (Bledsoe, 2001; Chadwick, 1997; Deegan, 1998, 2001). However, there is little evidence that r ecovery is either embraced by or relevant to all individuals with a SMI. Fu rther, there is little to s upport recovery as an important concept for individuals with less severe forms of mental illness. The concept of cure might be more relevant to individuals w ho assume that reduction in symptoms is equivalent to elimination of disorder, t hough whether this is accurate is unknown. In other words, cure might be relevant for thos e with mild to modera te disorders that are limited in duration and intensity and do not follow a typical chronic course. However, there is evidence that many disorders, for example obsessive-compulsive disorder, posttraumatic stress disorder, dysthymic disorder etc., are also recurrent disorders with periodic impact on functioning that, though less th an SMI, is still appreciable (Bystrirsky, Liberman, Hwang, & Wallace, 2001; Frank et al., 1991; Mintz, Mint z, Arruda, & Hwang, 1992). Therefore, recovery may have broader applications than cu rrently demonstrated. Within the population of individuals with an SMI, rec overy from schizophrenia has received the most attention from recove ry researchers (Andresen et al., 2003; Bender, 1995; Frese, 1998, Sells et al., 2004). Assessing differences in recovery processes for different diagnoses associated with SMI has only just begun. Some researchers stress that diagnosis is less important than other factors such as self-efficacy and empowerment (Bender, 1995; Yanos, Rosenfield, & Horwitz 2001). However, there are substantial differences in how different disorders mani fest (i.e., schizophre nia versus bipolar disorder), the expected cour se of the disorder, efficacy of treatment, stigma and

PAGE 53

34 acceptance, and impact on functioning. What may be a recovery related preference for individuals with schizophrenia (thought disorder) might be diffe rent than individuals with bipolar disorder or major depression (mood di sorders). The focus on schizophrenia has fostered growth in the development of recove ry stage models for schizophrenia that are, at times and perhaps inappropria tely, generalized to all indivi duals with a severe mental illness (Andresen et al., 2003). Both qualitative and quantita tive methods have been used to explore stages of recovery in schizophrenia. Andresen et al. (2003) developed a conceptual model of schizophrenia via review of the literature th at addressed schizophren ia and other mental disorders, consumer contributions and qualitati ve research. Stage one in their model is moratorium in which the individual is overw helmed by their symptoms, the weight of the diagnosis, and the changes that have come over their lives. The second stage is awareness, where recovery begins and th e process of delineating affected from nonaffected parts of the self is begun. Hope is kindled in this stage. The third stage, preparation, is a more deliberate resoluti on to recover followed by taking stock of abilities and finding resources th at help in recovery. Rebuild ing (stage four) is where the majority of recovery work takes place and involves goal setting and assessment of values. Risks and setbacks are common in this stage. The final stage is growth, where the individual is either symptom free or knows how to handle symp toms as they arise and has hope and faith in their future and is generally positive. A similar approach was used to summarize both qualitative a nd quantitative studies on r ecovery to identify common dimensions or factors relevant to recovery (Ralph, 2000). Internal factors (impact of illness, insight), self-managed care (idiosyncra tic methods of self-care), external factors

PAGE 54

35 (support, treatment), and empowerment (internal strength and caring) were identified as common across studies. These fact ors are likely to be most relevant for the stages of preparation, rebuilding and growth noted above (Andresen, et al., 2003). Thus, for schizophrenia at least, there is some understanding of the stag es relevant to recovery and the factors within th e stages that help promote recovery. Though there are several other models of recovery for schizophrenia, the results are markedly similar to the two models deta iled (Baxter & Diehl, 1998; Pettie & Triolo, 1999; Resnick et al., 2004; Spaniol, Wewiorski, Gagne, & Anthony, 2002; Young & Ensing, 1999). The two examples detailed suffice to point out that different research efforts have resulted in di fferent stages of recovery for schizophrenia that are thematically similar. Differences can be attributed in part to diverse values that researchers bring to their work, choice of research methodologies employed, and dissimilar methods of operationalizing recover y. Recovery as an outcome requires a set definition that continues to elude consensus in the recovery literature. Liberman & Kopelowicz (2005) have proposed criteria to operationalize recovery from schizophrenia. These include Remission of both positive and negative psychotic symptoms and signs, working or studying in normative employment or educational settings, independent living without supervision of money, self-care skills, and medicati on, social activities with peers, cordial family relations and contacts, recreational activity in normative settings (i.e., not in psychosocial clubhouses or day treatment programs), resilience and capacity for problem solvi ng when faced with new stressors or challenges in everyday life, subjective satisfaction with life, self-esteem and

PAGE 55

36 stable self-identity, participation as a citizen in voting, and self-advocacy, neighborliness, and other ci vic arenas (p. 108). This is a lofty list of standards that when applied might suggest no residual effects of mental illness at all. In fact, one might imagine that meeting these requirements would be a difficult task for anyone with no apprec iable challenge to his or her mental health status. However, it is a list developed from both empi rical research and consumer narratives and does provide a place to begin when designing recovery research. Projected as a set of potentially measurab le endpoints, the op erationalization of recovery noted above also points toward proce sses important to recovery. Consequently, it should be apparent that r ecovery could be viewed as both process and outcome. Whether recovery as an outcome (e.g., an endpoi nt identifiable as a recovered individual) is attainable is still much debated. Recovery for this research effort did not directly use the concepts outcome and process. Instead, the concepts of recovery expectancies and recovery strategies helped guide and opera tionalize the research. The more general concepts of outcome and process are not ab andoned, however, as recovery expectancies are related to recovery outcomes though not directly measuring an outcome attained but measuring an outcome projected. In turn, re covery strategies are related to the process of recovery. Referring back to Figure 1, introduced in Chapter 1 (p age 8), expectations for recovery include elimination of mental di sorder or ongoing restriction of functioning from mental disorder that nevertheless allows fulfillment, quality of life, and happiness. The expectation of recovery serves as a s ubjective measure of recovery as an outcome, for instance, if a respondent states, I have recovered or I am ha lfway to the point of

PAGE 56

37 being recovered or I will never recover, he or she is stating their expectation of recovery as a present or future outcome. Utilization of recovery strategies was defined as the degree that each respondent endorsed various strategies for recovery as im portant to their personal recovery. Chapter 3 discusses how research strate gies were detected and operati onalized for this research. Historically, these strategies have included how much help -seeking, reliance and support from others is necessary for recovery as well has the need to find and maintain hope (Lysaker, Buck, Hammoud, Taylor, & Roe, 2006; Resnick, Fontana, Lehman, & Rosenheck, 2005). In addition, symptom cont rol and viewing recovery as a personal challenge are identified strategies (Ochocka, Nelson, & Janzen, 2005). Domains of Factors Associated with Recovery The number of factors that might be involved in the recovery process is substantial. This section will review key fact ors that have either empirical or theoretical support for their influence on recovery. These factors are grouped into three domains. Figure 2 provided a graphical representation of the conceptualized domains that influence recovery (Chapter 1, p 12). The first domain encompasses individual, historical and clinical factors that include the proximal c linical, genetic, abuse/assault, and substance use factors that directly influence the indivi dual either currently or in the past. The second domain addresses factors th at influence recovery in the social context. Interaction with others is the thread that draws thes e factors together. For example, stigma, discrimination, empowerment, and social suppor t are factors that may influence recovery through direct interact ion between the individual and the environment. The final domain is the clinical and support services domain th at includes the access to services and types

PAGE 57

38 of services that may be associated with expe ctancies for recovery and choice of recovery strategy. Described in Chapter 3, this domain focuses less on which specific services are provided and focuses instead on the number of services and satisfaction with services. Each domain contains multiple constructs. These represent different levels of the social and personal environment. An i ndividual construct may represent micro (individual/familial), meso (bridging), or macro (community/societal) levels. A complete description of each construct will include an in dication of which level, or levels, that a construct represents. Domain One: Individual, Hi storical and Clinical Factors Individual and historical factors that may influen ce recovery expectancy and choice or utilization of recove ry strategies include diagnos is, current symptoms, age of onset, familial history of mental illness, histor y of trauma or abuse, treatment compliance, hospitalization history, and current use of substances as well as substance use history. Diagnosis Diagnosis is a micro-level construct that represents both specific symptoms and varying degrees of impairment in social, occupational, or educational functioning. Clinical manifestation, expected and objective impact on functioning, degree of anticipated recovery, and social sequelae are dissimilar for disorders typically associated with the SMI label. Less severe disorders have even greater variance in these areas. This section will briefly discuss these issues in re lation to schizophrenia, major depression and bipolar disorder.

PAGE 58

39 Schizophrenia. Previously noted, schizophrenia has been the only diagnosis consistently evaluated for its association with recovery when recovery is conceptualized beyond clinical improvement. Sympto ms of schizophrenia broadly include positive and negative clusters. Positive symptoms are experiences that are not supposed to be present, e.g. hallucinations and delusions. Negative sympto ms are defined as the inability to have experiences normally expected, for example when symptoms include fl at or inappropriate affect, anhedonia (inability to derive pleasure from what used to be or should be pleasurable) or apathy. Subtypes of schizophrenia include disorganized, paranoid, undifferentiated, catatonic, and residual. Diso rganized schizophrenia (also known as hebephrenia) is the most serious subtype and is likely to cause life long impairment and the need for custodial care. The negative symptoms and disorganized thought, speech, affect and attention process is quite severe and ther e is a relative little involvement of positive symptoms. This type of schizophrenia is ra rely treated in outpatie nt settings and there were no individuals in this st udy diagnosed with this subtype. Individuals with Paranoid schizophrenia usually has less cognitive impairment involved but is more likely to be associated with pervasive delusional be lief systems and hallucinations. Paranoid schizophrenia is amenable to treatment with psychotropic medication and cognitivebehavioral intervention after remittance of the delusional state. Marked psychomotor disturbance involving voluntary movement, posturing, and speech abnormalities (e.g., mutism or echoalia) is the primary characteristic of catatonic schi zophrenia. The rigid posturing may be replaced by a frenzied ex citement. Catatonia has decreased in

PAGE 59

40 prevalence over the last several decades and is rarely seen in outpatient settings. Psychotropic medications are effective in reducing the symptoms noted above. Undifferentiated schizophrenia meets criter ia for the primary characteristics of schizophrenia are present but the individual does not meet criteria for disorganized, paranoid, or catatonic subtypes. Residual schizophrenia is absent complete positive symptoms though attenuated symptoms may be present, e.g. eccentric behavior. Negative symptoms are still present. Describing the different s ubtypes of schizophrenia hopefully served as an indication of the complexity of schizophrenia. There is considerable debate still on the usefulness of subtyping. Inability to find a cause of schizophrenia may hinge, in part, on the low specificity of the subtypes and the study of unintentionally mixed groups in research that result in a systematic bias in th e data. Other typologies have been offered as options to increase the accur acy of subtypes including di viding schizophrenia into cortical, subcortical, and composite subtype s (Turetsky, et al., 2002). The cortical subtype has pervasive attentional and cogni tive deficits and few positive symptoms, similar to the disorganized t ype. The subcortical type has pervasive positive symptoms, multiple negative symptoms and dementia and appears similar to Huntingtons chorea. The composite type presents with features similar to both cortical and subcortical, but with less intensity. Memory deficits may also be useful in distinguishing between more accurate subtypes of schizophrenia (Barclay, 2002). Using the same subtyping as Turetsky, et al, (2002) above, Barclay (2002) found that fully 51% of study sample did not meet criteria for cortical or subcortical types, raising the question of the inclusiveness of other typologies.

PAGE 60

41 Studying schizophrenia has turned in the la st decade to a focus on predicting first onset and intervention for first onset cases. For first onset cases, adequate treatment will result in positive symptom remission for nearly 50% of cases and result in adequate social functioning for approximately 25% of cases. However, only approximately 14% of cases treated clinically have both symptom remission and adequate social functioning (Robinson, Woerner, McMeniman, Mendelowitz, & Bilder, 2004). Thought disorder associated w ith schizophrenia is fairly specific to the disorder. Neurocognitive dysfunction is present in schi zophrenia from the prodromal period to full remission in varying degrees. Typical for schizophrenia are deci sion making deficits including longer time to make decisions, impa ired risk adjustment, and impaired optimal decision making (choosing the right course of action) (Bell, 2001; Green, 2001). Individuals with schizophrenia may have difficulty utilizing available contextual information when processing social stimuli (Aghevli, Blanchard, & Horan, 2003). This could have direct consequences when studying the impact of do main 2 social factors such as social support and community connecte dness as outlined in brief in Chapter 1. Maintaining internal representations across time may also be impaired, inhibiting the ability to use information from the past in order to estimate future events (Kapur, 2003). Overor under-inhibition of behavior, motion perception impairment, inappropriate social distancing and stereotypic movements are additional cognitive-behavioral features of schizophrenia. Schizophrenia will impact on functioni ng level during active psychosis with residual effects in most cases (Falloon, He id, Roncone, Coverdale, & Laidlaw, 1998). Cognitive dysfunction certainly contributes to problems in overall functioning, but with a

PAGE 61

42 fair amount of variation acro ss individuals. The majority of impairment is in the beginning and middle of the life course of sc hizophrenia with less impairment noted later in life as the disease burns out. Longit udinal studies have demonstrated a natural remission for up to two-thirds of schizophr enia cases (Calabrese & Corrigan, 2005). Natural remission is not the same as active recovery, however, and active recovery is viewed as a way to accelerate remission and to increase and maintain gains in functioning (Ralph & Corrigan, 2005). Cognitive dysfunction and the resultant oddities in behavior are reasons that individuals with schizophrenia appear to pa y the highest social costs. Qualitative studies have especially captured the pain, isolation, and fear of individuals with schizophrenia, their re latives, and an uncomprehending public that is terrified at what is often interpreted as an assault on the very identify of a pe rson (Holzinger, Kilian, Lindenbach, Petscheleit, & Angermey er, 2003; Humberstone, 2002; Milliken & Northcott, 2003; Rungreangkulkij & Chesla, 2001; Williams & Collins, 2002). Major depression. Major depression (MD) is often measured in relation to multiple recurrences or relapses. A single occurrence is termed a ma jor depressive episode with any subsequent episodes indicating major depressive disorder (APA, 1994). A single MD episode has a high risk of being followed by a second at so me point during the life course. The overall risk of recurrence is 16% with each successive episode (Solomon et al., 2000). Helplessness, hopelessness, loss of energy, poor self-esteem, lack of self-efficacy, and disturbances in sleep and appe tite are typical for MD. Ep isodes of depression that are extreme may also include ha llucinations, thought disorder and pseudo-dementia (APA, 1994). Though MD is considered primarily a mo od disorder, it is important to note that

PAGE 62

43 there are cognitive-behavioral features of th e disorder as well. Arbitrary inferences (drawing conclusions in the ab sence of evidence), selective abstraction (focusing on the negative detail in a situation), overgeneralization (tendency to draw a general rule on the basis of an isolated incident and apply it in discriminately), and personalization (to relate external events to oneself without evidence) are characteristic for MD (Sacco & Beck, 1995). However, there is less residual cogni tive impairment compared to schizophrenia as the depression remits. For instance, access to dysfunctional schema (i.e. arbitrary inferences, personalization) is high when actively depressed while metacognition (the ability to think about how you ar e thinking) is low. For i ndividuals in partial remission access to dysfunctional schema remains high bu t metacognition is also high, allowing the individual to recognize the negative thought process. When fully remitted, access to dysfunctional schema shifts to low (health y) and metacognition remains high (Sheppard & Teasdale, 2004). The social sequelae alter slowly in depression compared to schizophrenia or other psychotic disorders. An i ndividual may go through severa l episodes of depression with reduced functioning before the SMI label is attached, whereas the la bel may be attached more or less automatically for a psychotic disorder like schizophrenia. Similarly, in psychosis residual compromise of functioning is expected while re mission of symptoms was automatically assumed to include full rest oration of functioning for depression. This assumption has recently changed with a growing emphasis on multiple recurrences of major depression and residual affects on gl obal functioning (Zegal, Person, & THase, 2003; Solomon et al., 2000).

PAGE 63

44 Research in recovery from major depressi on has isolated important similarities to the recovery from schizophrenia Chronicity of illness epis odes effects ov erall functional recovery with higher levels of functional impairment during and between episodes for chronic depression (Klein, Shankman, & Rose, 2006). Likewise, the course of schizophrenia is partly measured in terms of the number of ps ychotic episodes with greater number of episodes equating with lo wer functioning and decr eased probability of recovery. Further, one qualitative study a ttempting to isolate areas of importance for recovery from major depression found that th e areas of self-healing, managing the illness, receiving social support, a nd finding meaning were of primary importance (Sjarsater, Dencker, Bergbom, Haggstrom, & Fridlund, 200 3). Qualitative research targeting recovery and schizophrenia emphasize hope recovering the self, managing illness, maintaining relationships, and finding meaning in lif e (Davidson, 2003). Bipolar disorder. Bipolar disorder (BD) is a combination of manic disorder and major depression. Mania is a period of elevated mood that passes rationality an d effects cognition, decisions, judgment and reasonable selfcare (APA, 1994). Recurrence in bipolar disorder is much higher than other disord ers with 95% of individuals with a first occurrence of mania cycling in to recurrent bipolar disorder (Carlson, Bromet, & Sievers, 2000). Bipolar I disorder is more severe w ith one or more major depressive episodes and/or manic episodes. Bipolar II also has one or more MD episodes with one or more hypomanic episodes. Hypomania is a less se vere form of mania. There is some disagreement whether BD and MD are actually different disorders since there is some evidence that individuals with MD experien ce clusters of manic equivalent symptoms

PAGE 64

45 over the course of their illn ess but do not cro ss the diagnostic threshold to mania, suggesting a mood spectrum approach in place of mood categories (Ca ssano et al., 2004). Recovery from BD has focused a great deal on symptom management. Acquiring and maintaining skills for self-care are crucia l for BD because of the tendency to cycle through two distinct symptom clusters (Cutle r, 2001). Deficits in social support are especially problematic for BD, with full remission partially dependent on adequate support (Johnson, Lundstrom, Aberg-Wistedt, & Mathe, 2003). Bipolar disorder contributes significantly to disability, poverty, suicidal behavior and disrupted relationships (Judd & Akiskal, 2003). Combined, active pr oblems in these areas could interfere with recovery efforts. Compared to schizophrenia and even ma jor depression, bipola r disorder has had relatively little research attention targeted to recovery. Symptom recovery and effects of medication has predominated in the literatur e. Empowerment, hope and other recovery topics central to the consumer movement have not been investigated for bipolar disorder to any degree. Current Symptom Levels Current symptoms reflect more precisely the clinical picture of an individual compared to diagnosis. Like diagnosis, this is a micro level cons truct that describes psychiatric stability and is temporally more immediate than diagnosis, since the diagnostic label may have been first presen ted from the recent to the distant past. Diagnosis reveals what c linical manifestations could be present while an accurate description of sympto ms describes what is present. Depending to what degree an individual has remitted, current symptoms could be more relevant than diagnosis.

PAGE 65

46 Various investigators have incl uded measures that address symptoms while investigating recovery. For instance, Corrigan & Phelan ( 2004) utilized the Brief Psychiatric Rating Scale and Pevalin & Goldberg, (2003) administ ered the General Health Questionnaire to gain a general idea of psyc hiatric symptoms. Other rese archers choose to investigate more thoroughly the symptoms within specific diagnoses. For example, Gureje, Harvey, & Herrman (2004) explored the influence of self-esteem on psychosis and utilized the Diagnostic Interview for Psychosis to better understand psychotic symptoms within the sample. Similarly, Tait, Birchwood, & Trower (2003) used the Structured Clinical interview for the Positive and Negative Syndrome Scale to assess psychotic symptoms. Either approach would be valid, dependi ng on the goal of the research and the heterogeneity of the respondents. There is some evidence that specific symptoms that do not meet criteria for a mental disorder (sub-clinical or sub-syndr omal symptoms) influence recovery. Lower levels of depressive symptoms are related to a positive recovery orientation regardless of the psychiatric diagnosis (Res nick et al., 2004). Identifyi ng these symptom patterns is important for validating comorbid diagnoses and detecting unrecognized co-occurrences and their effects on recovery. For example, post-traumatic stress disorder (PTSD) is often diagnosed in general psychiatric samp les yet may also be undetected as a cooccurring disorder in individuals in ma ny psychiatric classifications (McFarlane, Bookless, & Air, 2001), and yet could still ha ve a profound impact on recovery. Overall, individuals with higher levels of current symptoms tend to endorse lower scores on measures of recovery strategies (Corrigan, Giffort et al., 1999).

PAGE 66

47 What is not known about different types of symptoms is to what degree they are associated with either the belief in the abil ity to recover or the choice of recovery strategy. To this point, no studies have b een found that address a broad selection of symptom patterns and recovery beliefs or choice in routes to recovery. This research will begin to close this gap. Age of Onset Age of onset is negatively correlated with recovery and is also a micro level factor. Early onset usually indicates a s horter period of normal development and less historical stability to inform compensation wh en symptoms remit. Onset before the age of 21 is correlated with greater barriers to recovery (Stephens, Richard, & McHugh, 1997). Early age of onset increases risk fo r developmental deviance and is associated with poor premorbid adjustment in childhood, and delays in speech, reading, and spelling (Vourdas, Pipe, Corrigall, & Frangou, 2003) Increased childhood psychopathology is related to reduced functional out come and level of remission that in turn affects the level of recovery (Carlson, Bromet, Driess ens, Mojtabai, & Schwartz, 2002). Information on age of onset age 21 or late r is not as conclusive for impact on course of illness or recovery. Schizophren ia, however, has an age of onset later for females, perhaps due to the protective factor of the female hormone oestradiol (RiecherRossler & Hafner, 2000). There is also a s econd spike for onset at approximately age 45 for females, presumably when there is a d ecline in oestradiol (A PA, 1994). Available information on age of onset is not conclusive This is partly due to inconsistent definitions of age of onset. Age of onset can be defined as the first age recalled or

PAGE 67

48 observed that burgeoning symptoms impact level of functioning. Onset may also be defined as the age of first diagnosis by a mental health professional. Trauma and Abuse Early experiences with child abuse impact on the onset and course of illness, with abuse associated with earlier onset of disord er, severity of course, suicide attempts and addictions for individuals with schizophren ia (Darves-Bornoz, Lemperiere, Degiovanni, & Gaillard, 1995). Trauma and abuse can be viewed both as a micro-level variable (individual affects of trauma experiences) and meso-level (necessity of interaction with someone in the immediate environment, often someone known to the victim). Repetition of sexual traumas has been observed in higher numbers for individuals with schizophrenia than bipolar di sorder, though both are elevated above prevalence levels in the general public (Darves-Bornoz et al., 1995) One estimate for lifetime physical or sexual assault for individuals with severe mental illness was nearly identical for females (86.8%) and males (86.7%) (Goodman et al., 2001). Lifetime sexual assault in the general population was estimated at 3.8% fo r males and 22.0% for women (Elliott, Mok, & Briere, 2004). For individuals with severe mental illness, one-year incidence of sexual assault was estimated at 7.6% males and 20.3% female. This climbed sharply to 36.7% male and 33.4% female when approximating both physical and sexual assault incidence rates. Experiences of childhood sexual or physical assault are n early identical in prevalence for males and females with seve re mental illness with approximately twothirds of the population reporting one or more events (Goodman et al., 2001). Specifically for major depression, 37% of those with major depression report sexual abuse compared to 23% without majo r depression (Cheasty, Clare, & Collins,

PAGE 68

49 1998). For individuals with any major mental disorder, early trauma frequently leads to undiagnosed stress reactions (e.g. post-traumatic stress disorder or PTSD) in a large proportion of individuals. For instance, undiagnosed PTSD was found in 43% of a sample that identified as experiencing at leas t one traumatic life event while only 2% had this diagnosis in their charts (Mueser et al ., 1998). Why such large segments of the SMI population are unidentified as trauma survi vors is unknown. One possi ble explanation is a lack of training for entrylevel providers in recognizing and being comfortable with inquiry into abuse issues. In a report from a recent eval uation, providers (n=83) either recognized that abuse experien ces were pervasive with the consumers they were assigned to or recognized no abuse at all (Walby, 2005). It is unlikely that i ndividuals with abuse experiences were accidentally selected into th ese groups and more likel y reflects provider ability and willingness to probe these sensitive areas. Prevalence of sexual and physic al assault is not only elev ated for individuals with severe psychopathology. Information is consiste nt that individuals with less severe forms of mental disorder (e.g., panic disorder, other anxiety disorder s, dysthymic disorder) have experienced more acts of abuse than non-ment ally ill individuals (Friedman et al., 2002; Hayden & Klein, 2001). In fact, one comm on theme across the abuse/mental illness literature is just how common these experiences are. Clearly, physical or sexual abuse is relate d to mental illness but what is less clear is why it has not been investigated as a factor relevant to recovery. While information on recovery from childhood abuse or adult assa ult is available, th e interaction between abuse, mental illness, and recovery self -efficacy and recovery activities is scarce (Alexander, Muenzenmaier, Dumont, & Ausla nder, 2005). However, there are beginning

PAGE 69

50 to be treatment methods designed to intervene with individuals with a SMI who have been traumatized (Fallot & Ha rris, 2002). Recovery requi res a degree of behavioral control, insight, and self-efficacy that may be disrupted by abuse experiences. An abuse history significantly increases the risk of suicide attempts, higher impulsivity, and aggression for individuals with mental illn ess (Brodsky et al., 2001) Noted earlier, comorbid disorders hamper the recovery proc ess and it has been dem onstrated that early abuse, especially sexual abuse, is linked to comorbid depression and anxiety more powerfully than either disorder category alone (Levitan, Rector, Sheldon, & Goering, 2003). The affects of abuse/assault potentially raise serious impediments to recovery that may affect both belief in recovery and how an individual structures his or her own recovery. Hospitalization History There is no evidence that hospitaliza tion history has any direct affect on expectation to recover or endorsement of r ecovery strategies when considered beyond the clinical benefits of hospitalization. However, hospitalization is an i ndicator that clinical stability is compromised, which in turn might affect efforts toward recovery. Hospitalization does not preclude recovery and may in fact be a reasonable compromise to accelerate needed clinical stability and could indeed be part of a recovery plan (Copeland, 2001). Conversely, hosp italization is described as a personal failure and an impediment to being accepted, appreciated and a barrier to meeting life goals by some individuals with mental illness (Walby, 2003b). Individuals with consistent patterns of hospitalization may also reflect more seri ous clinical and func tional impairment. Referring to the recovery section above, system atically coping with mental illness while

PAGE 70

51 fulfilling life goals is a common component of recovery definitions. An extensive history of hospitalizations could negativ ely impact meeting life goals a nd thus be an identifiable barrier to recovery efforts. However, th e lack of empirical evidence and the forced reliance on anecdotal evidence suggests that the impact of hospitalization on endorsement of recovery has not been adequately i nvestigated. The assumption of a negative association between recovery a nd hospitalization investigated in this study addresses this gap. Employment History and Current Employment Employment is another variable or construct that is micro-level in providing selfesteem, security and other factors and meso-lev el because of the inte ractional nature of most employment. Competitive employment, volunteering and other forms of employment is rare for individuals w ith a SMI, with only 14% reporting steady employment (Becker & Drake, 2003). Employ ment is helpful in reducing symptoms, raising self-esteem and increas ing a sense of attachment to the community (Evans & Repper, 2000; Gates, Akabas, & Zwelling, 2001). Stigma, fear, and insufficient skills are barriers to employment endorsed by many individuals with mental illness (Laudet, Magura, Vogel, & Knight, 2002). Once these barriers are overcome, however, there is general agreement from clinicians and consumers that employment is beneficial to recovery (Marwaha & Johnson, 2004; Warner & Mandiberg, 2004). Employment is normalizing for individuals with mental illness and helps to offset stigma (Laudet et al ., 2002). The literature does not illuminate the relationship between recove ry and extensiveness of work history or the relationship between work, recovery and i ndividuals with less severe forms of mental illness.

PAGE 71

52 Employment is, however, a goal that is c onsistently promulgated as a sign that stability has been achieved. Indeed, a primar y outcome measure for Florida that is used to evaluate programs working with severely mentally ill consumers is numbers of days worked. A minimum of 40 days per year per consumer is required to meet the employment criteria, justifying the block funds provided for an agency. Some providers feel this is artificial and does not particularly enhance either recovery or quality of life for the consumers, especially when employment quotas are not paired with sufficient rehabilitative and employ ment training (Monte, 2006). Current Use of Substances/Substance Use History Substance use for individuals with mental illness can range from casual use to cooccurring disorder. Substance use progresses differently for individuals with and without SMI including age of onset of serious substance use (usua lly younger for SMI), order of drug preference, length of use before addiction and other key areas (Gandhi, Bogrov, Osher, & Myers, 2003). Prevalence of substance use is higher in individuals with mental illness than non-mentally ill comparisons, whic h may be a function of or related to coping/self-soothing of disturbing symptoms and a much higher rate of abuse and assault experiences for individuals with mental illness (Gearon, Bellack, Rachbeisel, & Dixon, 2001). Treatment ranges from adjunct supportiv e counseling for individuals engaged in casual use to multiple component therapy in cluding harm reduction, treatment in stages, motivational interviewing, cognitive-behaviora l interventions, and modified 12-step selfhelp groups for individuals with substance addiction and a SMI (Rachbeisel, Scott, & Dixon, 1999).

PAGE 72

53 Substance use may dilute or negate th e effects of psychotropic medications prescribed for various purposes. Drugs and alcohol reduce motivation to succeed and may interfere with the personal growth and self-direction inherent in the recovery paradigm. Self-help groups and 12-step programs tailored to SMI (i.e. the Double Trouble in Recovery Program) embraces many of the key components of the recovery process (Magura et al., 2003). Recovery is still possible for indi viduals with SMI and that abuse substances, but there is little information whether the addition of substance problems changes expectancies for recovery or the individual strategies for recovery. Recovery from mental illness was preced ed by recovery from substance abuse and was motivated, in part, by the substance abuse recovery models. However, there has been no systematic comparison of recovery from substance abuse and recovery from mental illness. Evidence-based practices targeted to mental illness that, at least on the surface, advocate for recovery principles, ofte n focus on substance use recovery for those clients dual diagnosed (e.g., assertive commun ity treatment) with claims of better outcomes (McHugo, Drake, Teague, & Xie, 1999). However, there is also evidence supporting standard case managements positive impact on substance use that is equal to the more expensive, intensive, and perh aps disempowering, m odels like assertive community treatment (Essock, Mueser, Drak e, Covell, McHugo, Frisman, et al., 2006). Familial History of Mental Illness Having a family member with a mental disorder increases risk of onset of illness for individuals through vertical transmission (parent to chil d or more generally older generation to younger generation) via gene tic predisposition or through learning opportunities. Horizontal transmission (within the same generation, for instance sibling

PAGE 73

54 to sibling in some but not all pairs) through learning can also occur. For example, probands of individuals with sc hizophrenia were shown to be at higher risk for onset of schizophrenia than probands with relatives diagnosed with acute and transient psychotic disorders (ATPs) (Das, Malhotra, & Basu, 1999). Early onset of a major mental disorder is associated with higher risk for having a parent with a similar, though not necessarily identical, disorder. The risk of having a pa rent with schizophrenia is approximately 25% for a child with early onset schizophrenia compared to 1.5% for comparison subjects (Fogelson, Neuchterlein, Asarnow, Payne, & Subotnik, 2004; Nicolson et al., 2003), with an adjusted summary odds ratio of 4.59 (95%CI = 2.41-8.74) (van Os & Sham, 2003). Transmission within families of mental illness has been investigated to a significant degree for depression. Living with a depressed person and coping on a daily basis with the manifestations of depressi on are associated with family strain and disruption (Coyne, 1990). This burden manifests as a lack of intere st in social life, hopelessness, worrying and fatigue (Keitner, Miller, Epstein, et al., 1990). Various factors influence the childs internalizatio n/learning of depressi on including the childs temperament, the intensity of maternal affect how depression interf eres with attachment behavior, and the social learni ng of coercive processes in in teractions with a depressed parent (Schwoeri & Sholevar, 1994). Besides onset, familial history can also influence outcome of mental illness. For example, individuals with major depression we re significantly more lik ely to have a poor outcome if they had a familial history of schizophrenia, depression or suicide (Duggan, Sham, Minne, Lee, & Murray, 1998). Family relationships also have an affect on the

PAGE 74

55 level of social functioning, a sserting an indirect effect on clinical outcomes and possibly affecting recovery (Giron & Gomez-Beneyto, 2004). The relationship exists between family history of mental illness and recovery has not been investigated. Familial history of mental illness can be hypothesized to influence recovery by exposing the indi vidual to genetic and inte ractional experiences that influence the individuals belief in persona l recovery and what acceptable ways to recover are chosen. Witnessing family memb ers coping techniques, recovery attempts, successes and failures for compensating for ment al illness may in turn affect views on recovery. Again, however, this is all conjecture The first step will be to establish if an association exists between having a close relative with a mental illness and the expectations of recovery. This admittedly crude estimate will begin to answer this question and, if valid, can be followed up with more sophisticated investigation. Synthesis and conceptualization of the importance of domain 1 factors. A sizable percentage of the research into clinical recovery has targeted domain 1 factors, though consumer defined recovery has had limited inclusion of these factors outside of symptoms and diagnosis. Clinic al services targetin g recovery from the symptomatic and functional aspects of mental illness views these factors as critical to the treatment process. These are biological and ps ychological aspects of th e individuals that are treatment goals or objectiv es for clinical intervention. The recovery movement views factors in this domain as obstacles to recovery that are accepted for having occurred as part of the life story and not specifically empiricall y validated factors requiring intervention unless they are proven obstacle s to the recovery process. Recovery intervention can be clinical in nature provided by treatment pr ofessionals but has an equal

PAGE 75

56 chance of being consumer run services, info rmal intervention by comrades, alternative therapies, or any combination of the above. Symptoms drive the diagnostic process a nd specific diagnoses are related to severe mental illness. These, in turn, have been the focus of recovery research. The association between symptoms a nd diagnosis is not especially precise. Symptoms are not specific to disorders, with the exception of psychotic symptoms, which are far more likely to be associated with schizophreni a spectrum disorders. The degree that each individual is challenged by active symptoms w ill, in all likelihood, affect their belief in recovery (Mancini, Hardiman, & Lawson, 2005). The etiology of symptoms is a concern when considering the effects on the belief in recovery and the choice of strategies. Symptoms may represent a purely biological origin, an environmental origi n, or some combination. This is not a paraphrase of the nature versus nurture discussion. For instance, individuals that are sexually abused as children will often develop trauma related sy mptoms. These symptoms increase the risk for similar symptoms if a severe mental illne ss develops, as well as increasing the risk of the illness itself. Lack of employment due to cognitive impairments or poor social skills can lead to depressive symptoms that increa se the risk of mood and cognitive disorders over time. A major concern is the numb er of vulnerabiliti es generated through interaction of these various clinical factors th at influence the course of illness and affect recovery. Earlier age of onset and familial mental illness are part of the accumulation of genetic/biological and pathology modeling that leads to diagnosable illness (Sourander, Pihlakoski, Aromaa, Rautava, Helenius, & Si lanpea, 2006). Quality of familial support

PAGE 76

57 and the internalization of the label of mental illness affects assimilation of the new challenges of mental illness and thus recove ry as well (OBrien, Gordon, Bearden, Lopez, Kopelowicz, & Cannon, 2006). Further, the burden of additional family responsibilities, not understanding the symptoms and course of illness for close family member mental illness, and concern about their own and their fa milies future increases the risk of illness for children/adolescents and may reduce be lief in their own recovery potential (Valiakalayil, Paulson, & Tibbo, 2004). Expressed emotion has been investigated as a possible causal factor for mental illness and a risk for relapse and impedime nt to recovery (Hooley & Gotlib, 2000). Expressed emotion is a pattern of negative communication between individuals in close relationships, focused most often on the study of the family unit in regards to schizophrenia onset and relapse (Cutting & Docherty, 2000). The concept has elements of familial mental illness, abuse related experiences, symptoms (biologically and environmentally induced), and employment. Indeed, employment relationships can be stressful and negative sources of expressed emotion, increasing risk for exacerbation (Scheid, & Anderson, 1995). Experiences of child physical or sexual abuse contribute to the degree of clinical recovery attained and are important to consider within the additional context of hospitalizations, age of onset and familial mental illness. Vulnerability to abuse experiences is higher for individuals with me ntally ill parents, affecting monitoring of a childs health and safety and modeling of a ppropriate self-protectiv e behaviors are often impaired. Individuals with mental illness who also experienced childhood abuse have, on average, earlier onset of illness, elevated symptoms, and more lifetime hospitalizations

PAGE 77

58 (Schenkel, Spaulding, DiLillo, & Silverstein, 2005). Indeed, individuals with an abuse history have greater levels of impaired func tioning and elevated levels of hostility and suspiciousness (Shahar, Chinman, Sells, & Davidson, 2003). Acceptance of recovery, and by extension the combined consumer/clin ical perspective on recovery, requires a degree of illness management along with em powerment, trust, and connection to the community. These assets are challenged fo r abuse survivors who are also coping with some level of mental illness and may be a substantial barrier to identifying with recovery and deployment of recovery strategies. Finally, employment can be logically, and to a lesser extent, empirically linked with other domain 1 factors when considering recovery. Employment can be stressful and stress is known to exacerbate symptoms of mental illness. However, as Marrone and Golowka (2005) argue, so does unemploym ent, poverty and social isolation. Employment is a buffer to the negative e ffects of early onset, increased symptoms, a diagnostic label, and hosp italization (Warner & Mandber g, 2004). Conversely, earlier onset interferes with the establishment of a work history and lowers the potential benefits of a positive work environment. Work history has is negatively correlated with symptom levels and social functioning (Marwaha & Johnson, 2004). It is apparent that these clinical factors can increase or decrease th e negative impact of me ntal illness depending on timing, what experiences the individual has (e.g., abuse, poor employment), and what protective factors and potent ial resilience are present. Domain 2: Social Factors that Influence Recovery The second domain considers multiple aspect s of the social context that may be associated with recovery. Social sup port, social connectedness or integration,

PAGE 78

59 empowerment, stigma and discrimination will ea ch be reviewed for their potential impact on recovery. Sufficient information to guide this research is provided in the following discussion. Social support Social support has been identified as an important social component in physical and mental health as well as influential in th e recovery process. Social support is a mesolevel to micro-level concept that influences self-esteem and other important social and cognitive constructs via interaction with other individuals in the social environment. The following discussion expands on this statement by first looking briefl y at social support in relation to social networks. Next, multiple definitions of social support are considered and followed by an argument that social suppor t should be viewed as contextually based. Social support affects on gene ral health is touched on and then followed by social support for mental health and severe mental illness. Lastly, social support in the context of recovery is addressed in greater detail. Social support and social networks are cl osely linked in research for individuals with mental illness. The social networks of SMI individuals before their first admission are smaller and more conflicted than those of healthy individu als, and the continued loss of network members after multiple hospital admissions reduces even further the resources available to aid the patient returning to th e community (Nolan & Clancy, 1995). These same networks are deficient in maintaini ng the person in the co mmunity and preventing recurrent hospitalizations. Network members ar e frequently replaced with providers who offer support without reciprocity, further hi ndering the recovery of SMI individuals (Brunt & Hansson, 2002; Cresswell et al., 1992). This information is consistent across

PAGE 79

60 the literature and recent res earch indicates that network size and satisfaction are not significantly associated with psychiatric symptoms or desire to recover (Corrigan & Phelan, 2004). Social support as a concept has had many definitions and a lack of consensus on even the ingredients for a single definiti on (Williams, Barclay, & Schmied, 2004). This has hampered comparability of studies and the interpretability of findings. Some definitions have stood the test of time. For in stance, one of the most cited definitions is by Cobb (1976) who states Social support is de fined as information leading the subject to believe that he is cared for and loved, es teemed, and a member of a network of mutual obligation (p. 300). This definition is inte rpretive (leading the subject to believe), emotionally supportive (cared for and love d, esteemed), intera ctive or communal (member of a network) and reciprocal (mutual obligation). Anot her popular definition is by House, (1998) which states social support is an interp ersonal transaction involving one or more of the following: (1 ) emotional concern (liking, love, empathy), (2) instrumental aid (goods or services), (3 ) information (about the environment), or (4) appraisal (information relevant to self evaluation) (p 24). This definition is likely closest to the understand ing of social support in the soci al sciences as a more general form of help giving that can take discrete forms, depending what one person needs and another is willing to give. Another definition that is frequently cited is by Cohen & Syme, (1985), that states: Social support is defined as the resour ces provided by other persons. By viewing social support in terms of resources potential ly useful information or things we allow for the possibility that support may have negative as well as positive effects on health and

PAGE 80

61 well-being (p 33) This definition is more pragmatic and likens support to instrumental assistance (resources provided by other pers on), as well as know ledge (potentially useful information), and that support may not always be desired or welcome (negative as well as positive effects). This is one of the few definitions of social support that resonates with the Malone (1988) article di rectly investigating the social support and dissupport continuum. These representative definitions cover the emotional, instrumental, knowledge, appraisal, desirability of and response to social support (e.g., increased good will, access to goods). The individuals that provide these are not part of the de finitions noted above and require an understanding of the intersection between social support and social networks (Dickinson et al., 2002). Social ne twork theory recognizes that relationships exist in multiple contexts and that these may be positive or negative. Typical social network clusters include fa mily, friends, co-workers, and neighbors (Walsh, 1994; Watts, 1999). Other support networks that might be of special importance to individuals with mental illness include intimate partners or best friends and mental health providers (Macdonald, Hayes, & Baglioni, 2000; Meeks & Murrell, 1994). One consistent criticism of social support that is important to recognize in general and specifically to this research is that social support can also be vi ewed as contextually based, that is that social support will depend on the partic ipants involved, resources available, support needed, willingness to provide support, potential for reciprocity, extenuating circumstances and other factors (W illiams et al., 2004). Individuals with mental illness are more likely to need then be in a position to provide instrumental support, though needing instrumental support does not necessarily predict mental distress

PAGE 81

62 (Olstad, Sexton, & Sogaard, 1999). Another contex tually based aspect of social support, considering the limited social networks typi cal for individuals with mental illness noted above, are the clusters most likely to provide support. For individuals with mental illness, their family, friendships (often with other consumers of mental health services) and providers make up the majority of their social network and ar e the most likely to supply social support functions. This potentially is vital to understa nding the influence of social support on recovery through assessing a more realistic depicti on of social support. Social support and its effects on mental hea lth have been widely investigated. It is generally accepted that social support provid es a buffer to stress and enhances feelings of well being that are useful for protecting against depression, anxiety and even more serious and complicated mental disorders su ch as schizophrenia and major depression (Kawachi, & Berkman, 2001; Olstad, Sexton, & Sogaard, 200l; Sjar sater et al., 2002; Thoits, 1995). Conversely, lack of social support is a risk factor for onset and maintenance of common mental illnesses (Pre valin & Goldberg, 2003). Further, social support is of increased importance in areas with limited or difficu lt to access services (e.g., rural areas) (Letvak, 2002). When limiting consideration to individuals with severe mental illness, social support takes on additional meaning. Social support is important for severe mental illness because it is assumed to facilitate competence and coping (Buchanan, 1995). This is especially important for individuals w ith frequent hospitalizations because such individuals have smaller non-family networks and fewer intimate relationships (Lipton, Cohen, Fisher, & Katz, 1981). However, so cial support has addi tional value for SMI individuals because of its abil ity to ward off pervasive loneliness. Stigma and exclusion

PAGE 82

63 puts these individuals at risk for accepting almost any support as better than no support at all. Stigmatizing interactions are especia lly powerful in predicting low quality of life while social support is related to higher qua lity of life (Yanos, Primavera, & Knight, 2001). Fragile self-esteem and willingness to acce pt one-down roles puts them at risk for exploitation. It is no wonder that common and close social relationships ar e frequently viewed as both harmful and protective/s upportive (Green, Hayes, Dickinson, Whittaker, & Gilheany, 2002). For women with low income, limited resources, and mental illness, social support may increase role strain by placing too high a demand on the individual and interfering with coping (Kawachi & Berkman, 2001). The interaction between active psychotic or depressive symptoms with perception of social support shows less satisfaction with social suppor ts that have objectively not changed, furthering isolation (Skarsater, Agren, & Dencker, 2001). Thus, th ere is both peril and safety in social supports relationship with symptoms of mental illness, which leads next into a brief discussion on the role of so cial support with recovery. The role of social support in recovery has only recently been investigated and is probably one of the least understood areas for mental health improvement. Recovery has been shown to be positively associated w ith self-esteem, social support, empowerment and quality of life and is inve rsely related to age and sympto ms (Corrigan, Giffort et al., 1999). The relationship appears recipro cal between recovery and support and empowerment with a key role for self-esteem as a mediator between life satisfaction and symptoms (Markowitz, 2001). Low social suppor t increases the risk for relapse and the interruption of the recovery process may also inhibit rec overy self-efficacy (Pevalin & Goldberg, 2003). Small to moderate corre lations were found in a study that used a

PAGE 83

64 measure of social support and a measure of recovery and demonstrated that social support is related to a process orient ation for recovery (Corrigan & Phelan, 2004). However, this study did not investigate whether social s upport had different predictive capacity for diverse recovery pathways. In summary, social support is important for protection from stress and in limiting the impact of mental disease. Social support is important for all levels of mental illness, though the weight of evidence lies towards st udies of the SMI. How social support affects recovery is poorly understood and this study begins to fill this gap. While social support appears to be necessary for health and possibly for rec overy, the degree of perceived integration could be related as well, the s ubject of the next section. Social Integration or Co mmunity Connectedness Social integration became a part of th e mental health concept system when Durkheim found that a lack of percepti on of community connection significantly increased the risk of suicide (Alaszewski & Manthorpe, 1995). Community connectedness/social integration is experien ced individually at th e micro-level but is often investigated as a macro-level phenom enon. This bond to the community is based on attachment (the degree an individual f eels he or she belongs to a community as evidenced by the desire to maintain ties to society) and regulati on (the degree that an individual believes in and is governed by th e norms, values and beliefs of society) (Berkman & Glass, 2000). Regulatory functions of social integration provide for social control, an area of considerable concern for individuals with severe mental illness who may feel the affect of involuntary hospita lization or incarceration. Thus, social integration is a process that is in part feeli ng a sense of belonging to society; with society

PAGE 84

65 doing its part by providing a structure that an individual f eels is worth belonging to. Durkheims work is important for an addi tional reason, when he introduced the concept of alienation. Alienation results when an i ndividual does not feel either governed by or part of society with potential far-reaching ramifications for mental health and behavioral self-regulation (Berkman & Glass, 2000). The focus on social integration does not de tract from the importance of social support. Social cohesion is a macro-structur al condition (at the leve l of culture) that is the glue that binds individuals to society (Berkman & Glass, 2000). Because of its distance from the individual, social integra tion will not have as direct an influence on behavior as social support, but it does influence the provisi on of social support. For instance, an individual with mental illness that is living on the pe riphery of society may invoke support through a genuine need while living within the boundaries of social regulation. If this same indi vidual violates social conve ntions, in effect disrupting attachment and regulatory boundaries via bizarre or unpredictable behavior, the provision of support may reduce or eliminate, or be s ubstituted with powerf ul regulatory forces such as involuntary hospitaliz ation. This in turn may increase alienation and reduce reasonable attachment to the community. While this cycle is neither inevitable nor unstoppable, it does help to explain the lack of trust and connection some individuals with mental illness feel to their community and society in general (Kawachi & Berkman, 2000). The relationship between trust and social connection is important to this work. Trust is often viewed as a component of soci al capital, along with reciprocity and mutual aid (Kawachi & Berkman, 2000). Interpersonal tr ust is a broader concept in the literature

PAGE 85

66 and is synonymous with a general trust in so ciety and other citizens and is not a measure of level of trust in friends, providers, family members or intimate partners. Positive social support implies trust in these relations hips. Trust in others in the community is another way of saying that an individual is connected to their community. Thus, the expectation of general community support is a macro level form of social support (Stroul, 1989) and is important for community adap tation (Clinton, Lunney, Edwards, Weir, & Barr, 1998) and social bonding (Grusky, Ti erney, Manderscheid, & Grusky, 1985). Trust and social connectedness may work together to increase attachment to society and internalization of societal norms and beliefs, in other words trust and social connectedness are aspects of soci al integration. In this rese arch social connectedness is viewed more as community connectedness, an attachment to the immediate structural, organizational, and interactional aspects of the local community in which the individuals with mental illness (whether severe, moderate or mild in degree) live. This is a more personal and direct level of conn ection that falls short of cons idering the effects of social (macro) integration. This level is of greater importance to individuals with mental illness as they are tied to their communities more dire ctly via support services and mental health providers than they are tied to greater societ y (Nelson et al., 2001b). This is not to imply that individuals with mental illness lack c oncern for or connection to greater society. However, the pressing needs of managing the illness do force a self-preservative focus on the immediate community (Nelson, Lord, & Ochocka, 2001a). A focus on community connectedness has a pragmatic aspect as well. Community connectedness is potentially an impo rtant aspect of recovery, though there is little direct crossover between the literature on social integrati on/cohesion/connectedness

PAGE 86

67 and recovery. There is a reci procal nature between feeli ng connected and being accepted within the community. Feeling connected in creases concern for hygiene, appearance and others aspects of self-care th at in turn increases acceptanc e in the community. Connected individuals are more likely to be successful in employme nt (Nagle, Ccook, & Polatajko, 2002) and in social relations (Sorgaard et al., 2001). The interaction between social connectedness, trust, and social support is al so relevant to employment, with individuals with a SMI who have high levels of connect edness, trust and strong social support from family and friends more likely to be employe d halfor full-time (Evert, Harvey, Trauer, & Herrman, 2003). Work, social connectedness, and trust are all considered important aspects of recovery Ralph & Corrigan, 2005). In summary, social connectedness and pe rceived trust are potentially important aspects of recovery that may affect recovery self-efficacy and preferred route of recovery. However, feeling connected and attached also provides a certain degree of control, autonomy and choice. These are important as pects of empowerment, the subject of the next section. Empowerment Empowerment emerged, much like recove ry, in response to a mental health system viewed as inflexible and demean ing. Empowerment predated recovery and received a great deal of attention as an alternative model of ps ychiatric intervention (Byrne et al., 1999; Luckst ed, 1997; Segal, Silverman, & Temkin, 1995). Empowerment is based on the principle that consumers can take control of their lives, make their own decisions, decide which treatment alternatives are in their best interest, and reduce their dependency on providers (Dickerson, 1998).

PAGE 87

68 Empowerment is a multi-level construct th at has been investigated at the microlevel, meso-level, and macro-level (Kruge r, 2000a; Linhorst, Hamilton, Young, & Eckert, 2002; Nelson et al., 2001a; Rappaport, 1995; Zimmerman, 2000). A key reason for empowerment development is, not surp risingly, powerlessness experienced by individuals with disabilities (Fitzsimons, 2002). Thus, when describing and measuring empowerment, the concept of power gain ed and powerlessness avoided is common (Rogers, Chamberlin, Ellison, & Crean, 1997). Power is considered to exist with individuals, organizations and policy making/resource distribution bodies. Thus, empowerment is an ecological construct that can be measured at individual, organizational, and community levels (Rappaport, 1987). At the individual level, empowerment th eory investigates the degree in which individuals reach a state of psychological empowerment (Zimmerman & Warschausky, 1998). However, the systemic nature of the empowerment construct has no levels operating in isolation. Indi vidual empowerment is thus affected by empowerment at organizational and community levels. This is of primary importance in mental health as personal empowerment could be restricted or at best difficult in the face of a disempowering system of treatment provision or a community that refuses to allocate resources for mental health recovery. Concep tually, this is importa nt to avoid blaming the consumer for avoiding employment or other socially accepted choices. Success is increased if multi-level systems are mutually empowering and working in unison, and not in conflict. Psychological empowerment as a construct is intrapersonal, interactional, and behavioral (Zimmerman, 2000; Zimmerman, Is rael, & Schulz, 1992). The intrapersonal

PAGE 88

69 component most relevant to this research is perceived control and self-efficacy. Perceived control considers degree of control in all levels of inter action (e.g., control in relationships, work or education envi ronments, community groups, or political advocacy). Self-efficacy is a related concept that measures the degree a person believes he or she is able to reach desired outcomes Together, these influence self-esteem, the degree that an individual holds self-love and self-respect, which then influences the confidence to become involved in social relationships, employment, and other social exchanges. The interaction component involves how individuals relate to and think about their environment (Zimmerm an & Warschausky, 1998). Critical thinking, problem solving and decision-making are important aspe cts of this component. Several mental disorders, especially disorders typical to the SMI, often compromise the very skills needed to successfully empower oneself in relation to othe rs. Rehabilitation based on empowerment can help offset these potential deficits. The behavioral component is comprised of the specific actions taken by an individual to participate in community activities, join organizations, and address issues of power, resources, and politics. Participating in organized activities increas es self-esteem, confidence and efficacy. Participation also reduces bore dom, provides goals and instills direction in lives that, if unemployed and impoverished, are often ch allenged in finding a purpose. Empowerment encompasses both processes and outcomes. At the individual level, helping others and receiving help ar e examples of empowerment processes that foster a sense of control w ith critical awareness as a potentially critical outcome. Providing opportunities for orga nization members to develop skills and excel in their

PAGE 89

70 careers is an organizational level empowerment process. Actual provision of resources to facilitate employee growth is an organizat ional outcome. At the community level, providing equal access to resour ces and freedom to share ideas are processes that could culminate with citizens well trained in em powerment policy development and influence (Zimmerman & Warschausky, 1998). Providers of mental health services should ideally personify the organizational level of empowerment. When evaluating pr oviders, it is important to query whether providers are interested in empowerment, wh at their opinion of empowerment as a form of service delivery is, and what competen cies are required in order to provide empowerment-based services. However, this ty pe of inquiry is still quite rare for the field of mental health. Providers validat e empowerment abstractly but are generally unsure what the concept means. Individual empowerment is the best understood and most comfortable for providers (Acker son & Harrison, 2000). Organizational and political empowerment is, in turn, uncomforta ble concepts for provi ders. Organizational empowerment is viewed within the typical organizational hierarchy as something supplied from the top down, which runs counter to the principles of empowerment, while political empowerment is not considered part of what a provider should be involved in (Fawcett et al., 1996). Indeed, empowerment seeks to diminish hierarchy and increase voice for all participants. Such a concept is of ten confusing to providers that become lost in how to implement empowerme nt based services in an of ten rigid hierarchy based on measuring production of individual employees in order to solidify remuneration for the financial health of the agency.

PAGE 90

71 Empowering services ensure choices, consumer final word, flexibility and the ability for service changes while simultaneou sly being aware of dealing with forces impacting at the macro level. However, what constitutes an empowering service is often different between consumers and providers Staff tends to focus on medication compliance, skills of daily living, and access to help. Consumers do not negate these areas as unimportant but also focus more on improved access to material goods and improvement in general health (Meddings & Perkins, 2002). Consumer perception is important because quality of life and reduced sy mptoms are related to perceptions of met needs for consumers (Roth & Crane-Ross, 2002). Interventions that develop skills or abilitie s are considered of prime importance to consumers (Corrigan, 1997). Preventing loss of control and s ubsequent reliance on inpatient facilities is of hi gher importance to consumers than staff (Yurkovich, Smyer, & Dean, 1999). Outpatient services that are po tentially coercive, (i.e. Assertive Community Treatment), are viewed as disempowering by some, but not all, consumers (Spindel & Nugent, 2000). Indeed, strengths-based counse ling approaches used to collaboratively develop client goals have the capacity to increase personal empowerment (White, 2002). Becoming involved in empowerment serv ices requires that providers work through loss of boundaries, join with the lives of mentally ill people, and explore new definitions of professional identity (Byrne, 1999). This is probably the single greatest barrier to developing and su ccessfully implementing empowerment based services. The same argument is often made for recovery base d services as well. Th is redefinition of the working relationship between consumer and pr ovider to one of equal voice runs counter to the training of the majority of mental health professionals. Though not empirically

PAGE 91

72 substantiated, it makes sense to assume that the current trai ning of some mental health professionals is an additional barrier to establishing empowerment and recovery based services. Some researchers/providers have argued that empowerment be a standard outcome when planning with consumers who are disabled and that additional or new training programs should be developed to fost er both the therapeutic recovery based skills but also to help redefine the internalized id entity of the helping professional to include facilitator and t eam member (Dempsey & Foreman, 1997). Competencies for empowerment-based services are among many of the skills required for providing treatment or other servi ces for individuals with mental illness. A list of clinical competencies that support empowerment a nd rehabilitation was developed by Young and colleagues (Young, Forquer, Tran, Starzynski, & Shatkin, 2000). Categories of competencies include clinic ian-client relationship, initial and ongoing assessment, treatment, family and support system, social and cultural factors and resources and coordination of care. One ot her category, rehabilitation and empowerment, is pertinent to this discussion. This category contains seven competen cies that in are in many ways beliefs and values. (1) optimism: believe in the potential for growth and improvement. Have the skills to help the client restore or sustain hope and a sense of the future. (2) holistic approach: Be able to view the client as a whole person and to see beyond the illness. Have the skills to elicit the individuals personal experiences and worldview. (3) Goals: Be able to help the client identify and organize personal goals in the areas of learning, work, leisure, and living. Know how to support the clients unique pace toward goal achievement (4) Education: Be able to educate

PAGE 92

73 the client about mental illness, medica tions, and rehabilitation. Be able to communicate the value of reha bilitation and medication trea tment to clients. (5) Rehabilitation: Be skilled in using curre nt psychosocial/psychiatric rehabilitation approaches. Be able to teach goal-se tting and problem-solving skills and living, social, and illness self-management skills. Be able to help the client gain employment, education, and/or meaningful activity (when desired). (6) Client self-advocacy: Know how to create opportuni ties for the client to take optimal responsibility for his or her own life. Be able to fost er and support self-advocacy. (7) Natural supports: Be able to provide fl exible types and intens ities of services. Know how to help the clie nt meet changing needs and goals and transition from clinical to natural support s. (Young et al., 2000, p. 327). These competencies can move the field fo rward in promoting empowerment through the redefinition of what a helping professional do es and who a helping professional is. The emphasis, however, is still on providers providing and consumers accepting. Balancing the power differential is conspicuously absent from this list and almost unconsciously reinforces the professional identity still formalized in traditional therapy training programs. These are a good place to start, howev er, in assisting individuals with mental illness to become more empowered. Empowerment and recovery The relationship between empowermen t and recovery is complex. Many consumers who advocate for recovery strongl y promote empowerment-oriented services as part of the process (Deeg an, 2001). Noted above, empowe rment oriented services is more then an addition of new treatment type but requires a fundamental reassessment of

PAGE 93

74 organizational culture in service settings. The traditional and powerful roles of provider and consumer (or, more in line with the medical model, therapist and patient) are challenged to change. What fr ightens providers is the thoug ht of abandoning the years of training invested in becoming the helping prof essional that they envision themselves to be (Walby, 2005). There is also concern that the consumers are fundamentally incapable of making good choices and managing their illnesses and their lives (Walby, 2003b). This is often a projection of what a consumers life should be from the perspective of the trained professional and not an empowering/re covery based perspec tive that assists the individual to formulate goals based on curre nt strengths and capacities that promotes agency and growth toward new goals. Empowerment is often viewed as a goal in itself (Bassman, 2001) and there is still a lack of a clear picture on the relationsh ip between the constructs of recovery and empowerment other than that they are correlated (Corriga n, Giffort et al., 1999; Resnick et al., 2004). Correlations are sufficiently weak to indicate that differe nt constructs are in fact being measured. It is likely that empowerment is one aspect or component of recovery and is not synonymous with rec overy. Empowerment is necessary but not sufficient for recovery, hence the inclusion of other social, clinical and services factors in this research effort. It was noted in Chapter 1 that employment is considered a goal for and even a definition of recovery and empowerment has b een linked to employment as well. This area of overlap between the constructs together view employment as a source of personal control, self-esteem, challenge and socializa tion (Provencher et al., 2002). An additional area of overlap between empowerment and re covery is the increase in personal power

PAGE 94

75 (Corrigan, 2002). Behaviorally, personal power is enacted as the capacity make choices, live with consequences, assertively lobby to have needs met without becoming aggressive or imposing power on others, and recognizing ones limits. A sense of powerlessness is often described in the literature via firs t person accounts and qualitative studies of recovery (Davidson, 2003; Deegan, 1998: Sa yce, 1998). Often viewed as a deliberate attempt to marginalize the individual and make them disappear, the interactions that foster powerlessness are consistently de humanizing. Empowerment and recovery actively work against such interac tions to promote personal power. Empowerment and recovery often interact qualitatively in the life experience of the consumer. A consistent argument of indi viduals in recovery is that their life experience is often dismissed or mined for th e purpose of establishing the course of the illness without being viewed as a potential source of resources and strength (McGruder, 2001; Walby, 2003a). Recovery is strength based and focusing on past successes by concentrating on what went well then can transfer action steps to current problems, increasing a sense of control and empowerment. Empowerment can improve recovery in other ways as well. Environments that provide a variety of nonmedical interventions as well as meaningful work produce greater levels of empowerment and longevity of treat ment gains (Nikelly, 2001). Recovery is enhanced by treatment choices and employme nt. Leadership training for mentally ill individuals has recovering and empowering eff ects (Bentley, 2000). Leadership provides the opportunity to work outside the self and to give when one usually is receiving, increasing self-esteem, considered a s ub-construct of empowerment (Rogers, Chamberlin, Ellison, & Crean, 1997; Sega l, Silverman, & Temkin, 1995).

PAGE 95

76 Different research methodologies, speci fically participatory action research, empower individuals with a SMI through build ing partnerships betw een researchers and respondents (Rempfer & Knott, 2001). Suffici ent training allows consumers to become consumer-providers, further increasing empowerment and improving recovery (Bledsoe, 2001). In fact, consumers working with a co nsumer case manager have at least equal success in clinical outcomes (Chinman, Ro senheck, Lam, & Davidson, 2000) while the consumer case manager assumes a working cultu re of being there with the consumer compared to professional case managers who were more task oriented, which is preferred by the consumer in treatment (Paulson et al., 1999). Empowerment is a related facet of re covery and is multilevel, important for consumers in its own right, hope producing and can be part of service provision. Empowerment is also related to stigma, what the Surgeon General of the United States views as the greatest impediment to psychiat ric rehabilitation and provision of services (DHHS, 1999). Empowering individuals to meet life goals despite stigma is a benefit to recovery as well (Gingerich, 1998). Thus, stig ma and discrimination is the subject of the next section. Stigma and Discrimination Individuals with mental illness are likely to confront stigma, prejudice, stereotypes and discrimination multiple times throughout their life course (Wahl, 1999). Stigma is based on a perceived difference be tween self and other that is discrediting (Goffman, 1963), operates at multiple levels (community, familial, individual) (Hinshaw & Cicchetti, 2002), and has negative affects on identity, self-esteem, socialization, and experiences of acceptance and belongi ng (Angermeyer & Maatschinger, 2003;

PAGE 96

77 Blascovich, Mendes, Hunter, & Lickel, 2000; Camp, Finlay, & Lyons, 2002; Crocker, 1999; Crocker & Quinn, 2000; Thompson, Noel & Campbell, 2004). Link & Phelan, (2001) define stigma, as the co-occurren ce of its components-l abeling, stereotyping, separation, status loss, and disc rimination-and further indicate[s] that for stigmatization to occur, power must be exercised (p. 365). S tigma is, in effect, the interaction of multiple concepts that are rooted in power inequities. The most relevant of these concepts for this work is discrimination. Discrimination is the behavioral manifestation of stigma (Dovidio, Major, & Crocker, 2000). For exampl e, the action of refusing employment for an individual due to their mental health status is an act of discrimination. Discrimination is a common experience for many individuals with disadvantaged individuals more likely to experience it, and is strongly associat ed with mental health (Kessler, Mickelson, & Williams, 1999). There is no single theory of stigma. Various theories have been developed to explain stigma while other theo ries have been applied to th e general concept of stigma. For the former, attribution theory (Co rrigan, 2000; Corrigan et al., 2000; Corrigan, Rowan et al., 2002) and modified labeling theory (Camp et al., 2002; Link, 2001; Link, Cullen, Struening, Shrout, & Dohrenwend, 1989) were developed directly to explain stigma and its effects. Social comparis on theory (Finlay, Dinos, & Lyons, 2001) and social constructionism (Stangor & Crandall, 2000) have further deve loped the concept of stigma. For this work, modified labeli ng theory, attribution theory, and social comparison theory are all potential lenses for interpreting the affect of stigma on recovery.

PAGE 97

78 Partly due to the effect of stigma, labe ling an individual with mental illness leads to negative consequences that may reinforce or even cause mental illness. This is the initial conceptualization of la beling and its relationship to stigma (Scheff, 1974). Early critics of labeling theory felt that the label lead to treatment and other helpful outcomes and that stigma was overstated or even in consequential (Rosenfield, 1997). Modified labeling theory moved away from the direct e ffects of the label and states that labeling may not lead directly to mental disorder but does cause negative consequences through the perception of the label in the public mind and in the mind of the person with mental illness. Thus, individuals may suffer consequenc es unrelated to their actual behavior but based on their expected behavi or (Link, 1987). Modified labe ling theory helped to move stigma out of the label and into a broader social context. Within this context are individuals that make judgments about other individuals percei ved to be different. These judgments are cognitively based and require a theory that addresses the vital component of the cognitive process in understanding stigma. Attribution theory addresse s stigma from a primarily cognitive perspective and traces signaling events rela ted to stigma, mediating c ognitions, and the emotional responses that guide the choice of engaging in discriminatory behavior (Corrigan, 2000). Signals that promote cognition can result from direct interaction between two individuals that prompts attendance toward something that might be unusual or alarming, for instance an individuals appearance or behavior may signa l that something is out of the ordinary. If behavior is inappropriate to a setting or indicates symp toms commonly believed to be mental illness, then cognitive appraisal may commence. Other signals may be less obvious, for instance some type of skill deficit. If an individual is observed to have poor

PAGE 98

79 social skills or lack abilit ies deemed necessary for a particular work or academic environment and this is attributed to me ntal illness, then this could result in discriminatory behavior. Even more distal, if an individual is behaving and appears normal, yet they are suffering from a mental illness and this label becomes known, the cognitive-emotion-behavior path to discrimination may be trigge red as well. This path is essential to attribution theory. Cognitively, st ereotypes may be triggered that lead to the common emotional responses of pity, fear, or anger towards an individual with mental illness. Inappropriate stereotypes are misconcep tions about the mentally ill that lead to beliefs such as all mentally ill are dang erous, childlike, or rebels (Farina, 1998). Behaviorally, as noted, discrimination may be the ultimate response to this chain. Another key factor that influences this attribution chain is the relationship of causality to discriminatory beha vior. Causal attributions a ffect how stigma is generated, experienced, and acted upon (Corrigan, 2000) A stable causal attribution has traditionally been placed on severe mental illn ess, fostering the belief that mental illness is implacable, degenerative, and has ultimate negative results of lifetime hospitalization, suicide or premature death (K raepelin, 1919). Controllability of causal attribution is a related cognitive process that distributes re sponsibility for mental illness based on the reason for onset. If mental illness is believed to be the result of moral failure or lack of self-control through engagement in risk be haviors (e.g., substance abuse), then the individual is responsible for their own plight and will be viewed with contempt and discriminated against. Similarly, if the indivi dual is not engaging in what is believed to be appropriate responses to cope with and lim it the affect of the illness, this too will increase the likelihood of discriminatory responses.

PAGE 99

80 Social distance has been investigated with in and outside of attribution theory. Social distance is a proxy measure of stigma that examines how close a relationship an individual is willing to have with a menta lly ill individual (Angermeyer, Matschinger, & Corrigan, 2004). This relationship is often mediated by perceived dangerous which, in turn, is mediated by the amount of personal contact that an individual has had with mentally ill persons (Alexander & Link, 2003). Indeed, personal contact has been found to be an effective means of reducing stigmatiz ing beliefs toward individuals with mental illness, an area investigated exclusively betw een those with and w ithout mental illness (Corrigan, Rowan, et al., 2002). However, an area that labeling theory, attribution theory, and social distance do not address is how stigma ma y be perpetuated within the mentally ill population. Self-knowledge and self-esteem are pa rtially based on our perceptions and comparisons with others in our environment. Self-esteem is considered a necessary component for recovery and empowerment. However, individuals with a SMI are likely to have more challenges to their self-estee m through the combined insults of symptoms, labels and stigma. The literatu re indicates that self-esteem and, to a lesser degree, selfknowledge has been investigated for individual s with a severe mental illness (Blankertz, 2001). Social comparison, a potential key aspect pertaining to current levels of selfesteem and the maintenance of self-esteem, ha s been relatively ignored when considering mental illness. There has been limited invest igation into the relationship between mental illness and social comparison, with the greate st investment looking at depression (Ahrens & Alloy, 1997). Additionally, only one study has addresse d social comparisons and schizophrenia (Finlay et al., 2001).

PAGE 100

81 Mettee & Smith, (1977) provide a useful definition of social comparison, suggesting that it is a theory about our quest to know ourselves, about the search for self-relevant information and how people gain self-knowledge and discover reality about themselves (p. 72). In other words, peopl e will compare themselves to others to understand the self and to compare the self in order to make judgments about the self. Thus, individuals make social comparisons for three primary reas ons self-evaluation, self-improvement, and self-enhancement. Self-evaluation was the focus of the original formulation of social comparison theory (Festinger, 1954). The motivation for selfevaluation stemmed from the need for accuracy in our self-knowledge. Additionally, the unidirectional drive upward was a component of the original theory and hypothesized that a person would seek to compare themselves with those that they consider better in a task or trait for self-improvement. Social comparison theory has grown regarding self-improvement but has remained targeted on evaluation of tasks, tr aits and improvements generally unrelated to issues of mental illness. The third motivation, self-enhancement, more directly relates to self-esteem and, through the me diating effect of self-esteem to recovery. Downward comparisons, the process of comparing ourselves to those worse off in order to enhance self-esteem and feelings of well-being, are part of the self-enhancement comparison process. It is unknown whether those with a serious mental illness utilize downward social comparisons to the same degree as thos e without a mental illness or whether their comparison target are others with a mental illness or not. However, the investigation by (Finlay et al., 2001), does suggest that individuals with schizophrenia compare within the population of mentally ill indi viduals and also, actually to a greater degree, outside the

PAGE 101

82 population. The majority of social compar isons were downward, intended to increase self-esteem, and involved dimensions rela ted to mental illness (for within group comparisons) and not related to mental illness (between group comparisons). The complexity of theory surrounding stigma helps to highlight its potential influence on recovery. The consequences of stigma can be categorized as clinical symptom affects, internalized (e.g., self-esteem affects) and external, or discriminatory, affects. Stigma has been found to have a lasting effect on individuals, even when symptoms and functioning improve (Link, Struening, Rahav, Phelan, & Nuttbrock, 1997). For individuals with schizophrenia, both positive and negative symptoms have the affect of increasing stigma, with negative symptoms having a more consistent impact (Penn, Kolmaier, & Corrigan, 2000). Howe ver, positive symptoms correlate more strongly with beliefs in dangerousness and de sire for avoidance (S chumacher, Corrigan, & DeJong, 2003). Concern over stigma can lead to avoidance of social interactions and isolation, increasing sy mptoms of depression and lowering self-esteem (Perlick et al., 2001). It has also been documented that indi viduals with fewer concerns about stigma are associated with having less depressive symptoms (and the concomitant affect on negative cognitive schema), younger age and gr eater satisfaction w ith mental health (Pyne, Bean, & Sullivan, 2001). Research into the internalization of st igma has demonstrated the loss in selfesteem and self-efficacy that results from s tigma, particularly in narrative accounts of those afflicted with mental illness (Che rnomas, Clarke, & Chisholm, 2000; Knight, Wykes & Hayward, 2003). Other studies, however have challenged th e inevitability of low self-esteem (Camp et al., 2002; Finlay et al., 2001). Individuals will, at times, have

PAGE 102

83 the capacity to reject the negative societal me ssages and protect their self-esteem, and this ability may be contextually based (Crocker, 1999 ). The extent of this protective ability may rest, in part, on their beliefs and opinions regarding the mentally ill before their own onset of mental illness. Long before individuals become consumers, they have been socialized into what it means to have a me ntal illness, with many competing messages of attribution, fear, dangerousness, and the ability to recover (Link, 1987). This information becomes a filter for the self once the individual begins his or her own journey into mental illness and, hopefully, recovery. Strong messages of derision, fear, and contempt will interfere with self-esteem and self-efficacy. More temperate internal messages may result in less of an assault on the self. External affects, manifested as discrimination, impact socially (e.g. relationships), and in employment, education, and medical care. The two areas most pertinent for this review are social and employment effects of discrimination. Relationships for individuals with mental illness are heavily weighted towards others with mental disorders, with over 80% reveali ng that they lost all or the majority of their friends when they found out that the respondent had been diagnosed with a mental illness (Walby, 2003a). Investigations into social distan ce have found that approximately 20% of the non-mentally ill feel that they would be unable to maintain a friendship with an individual diagnosed with schizophrenia (likel y an underreport due to social desirability), 50% feel they would be unable to room wit h, and over 75% said they would not marry an individual with schizophrenia (Stuart & Arboleda-Florez, 2001). Refusal of employment is one of the most common and potentially harmful forms of discrimination. In one study, over 30% of individuals survey ed revealed that they had

PAGE 103

84 been turned away for a job they were qualif ied for due to revelation of mental illness sometimes to very often (Wahl, 1999). Indi viduals with mental illness are the least preferred disability type for hiring when co mpared to physical or sensory disabilities (Gouvier, Sytsma-Jordan, & Mayville, 2003). Ment al illness is often invisible and this affects the decision to disclose when making an application. The decision to disclose is complex and often is the choice of riski ng shaming oneself through disclosure and rejection, or increasing stress and anxiet y through actively telling a non-truth on an application (Dalgin & Gilbride, 2002). Consumers and providers often view employment, accurately, as an aid to recovery, or as a sign of stabil ity in recovery, yet 86% of individuals with a severe mental illn ess remain unemployed (Laudet et al., 2002; Warner & Maindiberg, 2004). There appears, then, to be a rela tionship between stigma and recovery. However, the relationship is ill-defined and under researched and of ten uses proxies for recovery, (e.g., employment, in place of direct investigation of stigma affects on recovery efficacy or recovery strategies). Stigma resi stance may be a factor in recovery and is often conceptualized as responsive to an individual level inte rvention. However, countering stigma often requires a combin ation of individual and community level interventions. Advertising of well-known indi viduals who successfully cope with mental illness and local community education are common in community interventions (Vaughan & Hansen, 2004). Interventions th at increase self-esteem, self-efficacy and personal empowerment are crucial for count ering self-stigma on the individual level (Corrigan & Calabrese, 2005). Maintaining control, accepting ones personal value and becoming more active in the community are ac tion steps individuals can take to counter

PAGE 104

85 the affects of self-stigma (Co rrigan, Calabrese, et al., 2002). What is interesting about these anti-stigma suggestions is that they are very close to the suggestions made for enhancing recovery. This may indicate that decreasing stigma is vital to recovery. Synthesis and conceptualization of th e importance of Domain 2 factors. Having reviewed social support, comm unity connectedness, empowerment and stigma, or more accurately stigma resistance, it is reasonable to state that they share common ground. Together, these constructs coun teract isolation, enhance assertiveness, increase trust, assist in reciprocity between individual and community, increase selfesteem and may even enhance the ability fo r competitive employment. The recovery literature touches on many of these same attri butes, yet there is litt le empirical evidence of a crossover. Though some recovery rese arch has touched independently on some of these factors collectively or in isolation, to this writers knowledge, no study has included all of them when investigating recovery expect ancies or choice of recovery strategies. For instance, social support a nd its impact on recovery was investigated in isolation and found that social support was associated with qu ality of life, which in turn was associated with recovery (Corrigan & Phelan, 2004). The potential effects of stigma and empowerment were not included in the Corri gan & Phelan study, which could potentially alter the studies conclusions. Empowerment has the most empirical research in relati on to recovery. Empowerment requires choice and choice is a key aspect of recovery (Paulson, Post, Herinckx, & Risser, 2002). Empowerment was a logical choice for inclusion in this study with a methodological impr ovement of measuring social support and stigma at the same time. This relationship has been adva nced theoretically, but has limited empirical

PAGE 105

86 support. The one study found that includes empowerment, social support and recovery investigated concurrent valid ity between these co nstructs and found m odest associations (Corrigan, Giffort et al., 1999). Finally, stig ma and recovery, noted in the previous section, have common ground in stigma resistan ce strategies. Stigma resistance will be one aspect of stigma investigated in this wo rk with the ability to see its contribution to recovery along with social support, empowerment, other soci al factors and the factors from the two other domains described in this ch apter. This turns our attention to the final domain, the services domain. In the followi ng section, the literature on different types of psychiatric and support services are reviewed in relation to recovery. Domain 3: Service Factors that Influence Recovery Services that promote mental health can be argued to be strategies for recovery and not factors that influence strategies for recovery. There is no doubt that services are important for recovery to some consumers and will even be considered a primary influence for recovery. However, not all cons umers agree that services are sufficient or even necessary for recovery and may instea d be viewed more as illness management (Mueser et al., 2002; Tait et al ., 2003). Services are more often targeted to clinical symptoms and restoration of functioning than to the less tangib le aspects of the consumer model of recovery (e.g., choice, satisfaction, qual ity of life). Thus, for this research, the domain of clinical and support services are viewed as possible asso ciates to recovery strategies and may influence so me strategies more than others. Also for this research, specific services are of less importance than se rvice availability, frequency of contact and consumer satisfaction with services. Howeve r, before measuring service and its impact on recovery (see Chapter 3), which services are actually provided will be briefly

PAGE 106

87 described. What follows is a discussion of th e services often provide d to individuals with mental illness with accompanying discussion on their impact on recovery. Many consumers have identified positive, flexible, non-coerci ve, and relevant services as key to moving them along the pa th of recovery (Walby, 2003b). There is evidence that services support recovery, though which services are more likely to influence different strategies for recovery is poorly understood. Case management, medication and residential treatment are examples of more traditional services. Supportive services (e.g.. supported employment and supported living) have a closer philosophical agreement with th e recovery philosophy. A key aspect of services is the consumer belief that appropriate services are available and accessible. Satisfaction with services may also be related to recovery. The partnering agency for this research project has a wide variety of potential services potentially pertinent to recovery and is in fact moving toward a recovery team approach (Ellis-Lang, 2005). Services that could impact recovery can be broken into traditional and supportive services, with an additional type being consumer run services. Traditional services include case management, outpatient thera py, medication services, substance abuse treatment, inpatient hospitalization, day treatm ent or partial hospita lization programs, and residential treatment. Together, these servic es support the notion of the individual as a patient. There tends to be less choice and flexibility in these services, with the exception of some types of case management, which may have an inverse relationship to recovery. Individuals with a SMI are likely to be involved in medication services and case management. Some of these individuals are involved in more traditional outpatient therapy and/or substance abuse treatment fo r a co-occurring addiction. Consumers that

PAGE 107

88 are more clinically or functionally compromised may participate in day treatment or partial hospitalization programs (Macchi, Peretti, & Barabino, 2996; Pittman, Parsons, & Peterson, 1990). Individuals who have not been successful in traditional outpatient settings or who have difficulty with psychotropic medication will at times be accepted in residential treatment programs (Bola & Mosh er, 2003; Fassino et al ., 2004). Medication and case management are the only traditional services that are addressed in the recovery literature, albeit minimally. Atypical antipsychotics are more a nd more the treatment of choice for schizophrenia and are often viewed as a way to enhance reco very (Glick, Murray, Vasudevan, Marden, & Hu, 2001; Littrell & Litt rell, 1998). Recovery in this sense is clinical recovery and often falls short of the consumer vision of recovery (Fisher, 2003). Medication is viewed by some individuals with a SMI as imperative for recovery, but there is a strong preference fo r integrated care that includes assistance in daily living, financial assistance, skills training, shared decision making and empowerment along with the medication component (Malm, Ivarsson, Allebeck, & Falloon, 2003). It should be remembered that medication is targeted to symptom control and consumers and other stakeholders consistently value functional outcomes, (e.g., improvement in social or occupational functioning, over symptom contro l). For medication, consumers are more concerned about side effects than other stakeholders, (e.g., family members, and policy makers) (Shumway et al., 2003). Antidepressants are best prescribed in conjunction with psychotherapy or some other appropriate tr eatment. Though medications for depression are generally efficacious, they are only worthw hile if taken appropria tely and do little to reduce the risk of further episodes if disc ontinued (Hollon, Thase, & Markowitz, 2003).

PAGE 108

89 Recovery from depression appears to be pa rtially linked to recurrence, with each recurrence increasing risk for another one, while the risk declines with the duration of recovery (defined as time episode free) (Solomon et al., 2000). Case management has traditionally served to promote continuity of services and adherence to treatment plan s (Crawford, de Jonge, Freeman, & Weaver, 2004). There are different types of case management that may influence recovery differentially. An exhaustive review of schools of case management is unnecessary for this review. However, there are some that emphasize autonomy and empowerment (strengths based) and others that are more control orient ed (assertive community treatment), and understanding the differences will help to illu minate the diversity of case management. Strengths based case management has a low emphasis on service integration and a high level of consumer input, with as much contro l wielded by the consumer as possible. This empowering approach is parall el to consumer-defined rec overy. When compared to traditional case management, strengths based service receivers have a greater reduction in need for care, less time spent in hospitals, a nd are more satisfied with care. However, there is also no difference in clinical or social outcomes (Bjorkman, Hansson, & Sandlund, 2002). Assertive community treatment (ACT) is a team approach to case management that occurs in the community and has a st rong emphasis on maintaining residence in the community, taking of medication as prescribed and keeping all treatme nt or rehabilitation oriented appointments (Nieves, 2002). A ssertive Community Treatment teams typically consistent of, at a minimum, a psychiatrist, nurse, social worker and case manager. Assessment of ACT demonstrates that it does increase treatment compliance (Jorgensen

PAGE 109

90 et al., 2000), but not necessarily reduce hospi talization or homelessness (Clarke et al., 2000), though this is not consistent in the lite rature (Salkever et al., 1999). Additionally, there is a great deal of concern that ACT and similar programs are overly coercive and disempowering (Spindel & Nugent, 2000). Finall y, housing is consistently an issue with mentally ill individuals and another prim ary task of case management, from all approaches, is to help indi viduals retain or obtain housing (Newman, 2001). There is little doubt that stable housi ng is a necessity for recover y, with housing assistance of primary importance to individuals with mental illness and their families. Supported programs include supporte d housing, supported employment, and supported education. In genera l, supported programs are more likely to be viewed as consumer oriented and helpful to recovery. Supported programs often offer more choices and are empowering. Housing and employme nt are two key conc rete components of recovery from consumer perspectives (Honey, 2004; Srebnik, Liningston, Gordon, & King, 1995). Supported housing includes settings with 24-hour on-site staff, apartments where staff visit, and homes or apartments with no on-site services or visitations but with financial rent assistance. Consumers in high supervision sites are more likely to be older, completed less education, unemployed, and diagnosed with schizophrenia (Friedrich, Hollingsworth, Hradek, Friedrich, & Culp, 1999). However, this is contradicted by other research that found high-intensity settings tended toward younger, hospital referred, and comorbid addicted individuals (Lipton, Si egel, Hannigan, Samuels, & Baker, 2000). Consumers tend to desire less supervision, a natural preference for independence, while family members favor greater levels of supervision.

PAGE 110

91 Supported housing was developed with community integration in mind (Carling, 1990). The reader will recall that integrati on/reintegration are concepts similar to recovery. However, there is evidence that individuals with SMI do not add to their support systems by living in supported housing but may in fact replace family and friends with staff and other residents (Goering et al., 1992). Increased community contact and support systems are cited as examples of succes sful recovery and services that interfere with this could be counterproductive. Additionally, peaceful surroundings are also conducive to recovery. Research indicates that residents wi th comorbid substance abuse problems or who was medication noncompliant were more likely to be disruptive and ultimately unsuccessful in supported housi ng (Grunebaum, Aquila, Portera, Leon, & Weiden, 1999). Supported employment is another service th at fits the spirit of consumer-defined recovery and is viewed as empowering (B ellamy & Mowbray, 1998). There are several models of supported employment, with the Individual Placement and Support (IPS) model and the Menu Approach (MA) the two most common. The IPS model focus is to return consumers to community-based, co mpetitive, integrated employment with unlimited follow-up provided to ensure succ ess. The MA approach is centered on providing work experiences to as many of the consumers at an agency as possible with a menu of options and with less emphasis on returning to competitive employment (Auerbach, 2001). Whichever approach taken, consumers have a strong desire to work, provided that it is equitable, does not threat en other benefits until they are comfortable and confident that the employment is safe and beneficial, and that it is supplemented by

PAGE 111

92 health and social services since many of the jobs for the mentally ill are entry level at best (Boardman, Grove, Perkins, & Shepherd, 2003). Challenges to supported employment incl ude implementation, job attendance and neurocognitive deficits. Implementation has been difficult from the perspective of taking programs that are academically derived and translating them into multiple real world settings. Success is predicated on foresight and careful planning while also having sufficient support personnel available, es pecially during the beginning of a job opportunity (Handler, Doel, Henry, & Lucca, 2003). Job attendance can be improved with careful selection of empl oyees for particular settings, adequate support services, and through innovative use of consumer employees. For example, entrance into a job site may be enhanced with the promise of no days absent. This could be accomplished by assigning more than one consumer to work at the setting, with a primary and one or more secondary employees that will provide the employer service if the primary worker is unable to attend. Attendance is also enhan ced through matching the consumer with a job that is interesting to him or her and th rough recognition that fear of relapse and socioeconomic realities are important factor s for seeking and maintaining employment (Mueser, Becker, & Wolfe, 2001; Nagle er al., 2002). Individuals with more severe positive a nd negative symptoms as well deficits in cognitive functioning (attention, psychomoto r speed, verbal learning and memory, and executive functions) are more likely to be challenged in employment, supported or otherwise, but can excel with proper suppor t (McGurk & Mueser, 2004). Factors that interact with cognitive defi cits to enhance or impede employment success include previous work history, amount of government financial assistance, and involvement in

PAGE 112

93 sheltered work activity before competitive employment (McGurk, Mueser, Harvey, LaPuglia, & Marder, 2003). The final support service is supported education. Supported education programs provide financial and academic support to individuals with mental illness who want training, usually in Community College or University settings. Services are individualized and based on client preference s while also incorporating mental health treatment as needed (Becker & Drake, 2003). Supported education, like supported employment, requires organizati on of community stakeholders that are interested in assisting individuals with me ntal illness. However, unlike other forms of supported assistance, supported education takes place in th e context of a large bureaucracy that is simultaneously providing services to possibly thousands of other individuals. Many of the individuals in supported employment may require assistance per the Americans with Disabilities Act, further increasing the burden on the educational setting. Because of these constraints, it is necessa ry to plan carefully for all contingencies and not to overextend resources (Mowbray, Gutierrez, Bellamy, Sziilvagyi, & Strauss, 2003). The unique needs of the setting and the consumer make supported education a challenge, but also of great value to recover y. To enhance recovery, individuals should be encouraged to stay engaged in productiv e activity outside of the school setting, to maintain contact with their support system, a nd to seek encouragement for their academic goals from their supports as well (Collins, Mowbray, & Bybee, 2000). Consumer run services are rare in conj unction with mental health settings and comprise the third area of service, with traditional and supporte d services, that may influence recovery. The notab le exception to its rare st atus is drop-in centers and

PAGE 113

94 clubhouses. Drop-in centers are generally owned by a mental health center but are completely staffed by consumers. Individuals that work in drop-in centers will benefit from empowerment and leadership training. This training often includes skills training for communication, problem solving, deescalation, and management skil ls, all of which could be potentially beneficial to recovery efforts (Bentley, 2000). The clubhouse model, a combination of psychosocial rehabilitati on and supported employment, has been found to enhance empowerment, a potentially key aspect of recovery (Dickerson, 1998). Autonomy, responsibility, lead ership and, in many cases, pa rticipation in professional training are self-enhancing and important for recovery. For individuals not employed by these settings, attending consumer run centers can lower isolation, increase socialization, and allow the consumer to practice social skills among compassionate and understanding individuals. A link between accessibility and satisfaction of services with recovery seems logical. The reality is that there are limited resources available for traditional and supported services and drop-in centers or clubhouses are often very crowded. Access is facilitated with appropriate triage that links the consumer most in need with an appropriate service. For measurement purposes it is easier to measure what services a consumer did receive compared to what serv ices they should have received. For this work, measurement of number of services received will be part of the data collection with full acknowledgment that each consumer may not receive all the services they require due to tremendous budgetary pressures. Satisfaction with services is correlated with availability and flexibility of services and with quality of life (Dixon, Goldber g, Lehman, & McNary, 2001; Perese, 1997).

PAGE 114

95 Understanding differing perceptions of h ealth between providers and consumers increases satisfaction (Berghofer, Schmidl, R udas, Steinger, & Schmitz, 2002). This is conceptually linked to the findi ng that services that understand consumer expectations are linked to higher levels of satisfaction (Noble, Douglas, & Newman, 2001). Finally, services that assist with instrumental n eeds such as benefits and occupation are considered more satisfying (Secker et al., 2002 ). The language used in these studies, such as flexibility, understanding perceptions and expectations, and assistance with instrumental needs, is identic al to language used recovery, bridging, theoretically, the two literature bases. To summarize, services that are recovery oriented will focus on the social roles and needs of consumers and not just their sy mptoms. These services will be strength oriented and will include the support systems important to the individual consumer, (e.g. friends or family members). Services that are recovery oriented will also provide hope and empowerment through allowing real choice s by the consumer and the services will be flexible enough to honor indi vidual goals and aspirations (Repper & Perkins, 2003). Synthesis and conceptualization of the importance of Domain 3 factors. The services outlined in this section are typical of the services offered in many full service mental health centers. Optimal services provision will provide access to all three areas of services (tra ditional, supported, consumer run). Taken alone, this domain may have an impact on recovery efficacy and pathway choice. However, for the majority of the services, their impact on recovery is hypothetical since there is little to no empirical research that intersects with the re covery literature. A strength of this study will be its ability to de termine if the number of services a respondent is currently engaged

PAGE 115

96 in, the amount of contact time with providers or perceived satisfaction with services is associated with recovery expectations. Services may have their greatest imp act on recovery when considered in conjunction with social and indi vidual factors. Services have been shown to increase empowerment and social support, two key components of the social domain (Hess & Mercer, 2001; Roth & Crane-Ross, 2002). Serv ices are also targeted to elimination of symptoms, and symptoms are demonstrated to be a barrier to recovery. These brief examples again help to illuminate the clear advantage of this study for the investigation of recovery: the investigation of multiple factors indicated in the theoretical or empirical literature to be associated with recovery. The process of this investig ation is the focus of Chapter 3.

PAGE 116

97 Chapter 3 Methodology This chapter begins with an expansion of the research design outlined briefly in Chapter 1. This is followed by discu ssion of the target population, inclusion and exclusion criteria and the samp le. Next, the research ques tions are re-stated with the addition of the research hypotheses linked to each question. After that, the setting in which the research was completed and the da ta collection instruments and tools that access the constructs investigated in this study are detailed. Fi nally, the methods and a priori analysis plan are described. Study Design This study is an ecologic, nonexperimental design with retros pective features. Multiple questionnaires/surveys were used to collect data from two samples of clinical populations. A number of factors that may influence the recovery process were culled from the literature and a multi-domain model of recovery influences was generated (see Figure 1 introduced in Chapter 2, page 8). Th is, in turn, led to the development of a model of hypothesized linkages between the va riables that constitute the three domains (see Figure 2, page 14). Dependent variables were generated through analysis of responses to three instruments completed by the participants Single question indicators targeting expectation of recovery were measured via the Recovery Expectancy Checklist. The Recovery Assessment Scale (RAS) and the Personal Vision of Recovery Questionnaire

PAGE 117

98 (PVRQ) were designed by their respective au thors (Corrigan, Salzer, Ralph, Sangster, & Keck, 2005; Borowitz-Ensfield, 1998) to access recovery strategies utilized by mentally ill individuals. The RAS and PVRQ are recently developed and, though there is information on external validity, neither inst rument has been utilized with multiple samples, nor have both instruments been administered to the same sample. These considerations and the newness of the recove ry construct, in general, informed the decision not to utilize, by de fault, the measurement models proposed by Corrigan, et al., (2005) and Borowitz-Ensfield, (1998). Instea d, the data from the RAS and PVRQ were subjected to a factor analysis to identify a measurement model for recovery strategies from the pooled questions. The analysis sec tion later in this chap ter provides greater detail on this procedure. All criterion vari ables are nominal or quasi-interva l (Hatcher & Stepanski, 1994). Quasi-interval is utilized to acknowledge that social science related research often treats scales as if there are equal distances betw een intervals when this is assumed and not proven true (e.g., the Personal Vision of Recovery Questionnaire). However, although this controversy continues to exist, this st udy follows the lead of the field and treats survey scales as interval (e.g., a differen ce between 30 and 40 is assumed to have the same value as the difference between 60 and 70). Potential correlates are generate d through one self-report/demographic instrument, six survey/questi onnaires, and one chart abstr action tool. Variables are nominal, interval and ratio in scale. There are a total of 51 correlates (independent variables) that represent the larges t main effect model examined.

PAGE 118

99 Population/Sample This study investigated two samples of clinical populations. These were representative of populations with severe me ntal illness (designated the SMI sample) and mild to borderline severe ment al illness being treated in an outpatient setting (designated the OP sample). Specific inclusion/exclusion criteria were used to determine if an individual was eligible to participate in the study. Following the statement of inclusion/exclusion criteria, the sampling frame, sampling method, and description of the samples follow. Finally, a brief descripti on of the power analys is used to ensure sufficient sample size is offered. Inclusion/Exclusion Criteria Inclusion criteria. Age 18-65 (SMI and OP samples); All clients of the partnering mental h ealth center who were diagnosed with schizophrenia, other schizophrenia spectrum disorders (e.g., paranoid disorder or schizoaffective disorder), bipolar disord er, major depression or diagnosed with a severe example of another disorder (e.g., post-traumatic stress disorder) and who met the criteria for severe mental illness (SMI sample); and All clients who did not meet the criteria for the SMI sample were eligible for the OP sample. It is important to note that diagnosis alone does not distinguish SMI status. There may have been crossover in diagnoses between the two samples, but the prevalence of more severe diagnos es was expected to be higher in the SMI sample than in the OP sample.

PAGE 119

100 Exclusion criteria. Any individual <18 years of age or 65 (SMI and OP sample); Any client who was hospitaliz ed or incarcerated (SMI and OP sample) at the time of participant recruitment; Any client who was experiencing sufficient instability that they were unable to participate in the study. The representa tive of the partnering agency who was most involved with the individual case made this determination (SMI or OP sample); and Any client diagnosed exclusively with a substance abuse or addiction disorder (OP sample). Sample Description and Procedures Sample 1: Individuals with Severe Mental Illness The target population for the first sample was consumers of mental health services with a severe mental illness (e.g., schizophrenia spectrum disorders, bipolar disorder, major depression and schizoaffective disorder), and with impairment in social, academic, and/or occupational functioning. The sampling frame was all individuals labeled as SMI who were receiving services at the partnering mental health ce nter that are not currently inpatient or symptomatically impaired to wh ere participation is unrealistic, and who are willing to participate. Thus, to maximize the variance available, all individuals with a severe mental illness who meet inclusi on/exclusion criteria were eligible for participation. The sample population was estimated to be between 2,200 and 2,400

PAGE 120

101 individuals served at the partnering mental health center. Table 1 summarizes distributions for this sampling fr ame by gender and race/ethnicity. Systematic sampling was used to select participants. This involved assigning a number to each eligible participant, and th en randomly selecting a single digit number and using that number to contact every nth person. The number randomly chosen was four and every fourth person was added to a list until 300 potential participants were selected. In order to preserve confidentialit y, the names of the reserve participants were not known to the researcher until needed for contact. Table 1 Gender/Race/Ethnicity of Individuals with a S evere Mental Illness Served at Partnering Community Mental Health Center Category % Sex Male 37 Female 63 Race American Indian / Alaskan Native 1 Asian <1 African-American 3 Hawaiian / Pacific Islander <1 Multi-Racial 1 Other 6 White 89 Ethnicity Cuban <1 Haitian 1 Mexican <1 Non-Hispanic 95 Puerto Rican 1 Other Hispanic 2 Note: Total N = 2,354

PAGE 121

102 Sample 2: Individuals with Mild to Borderline Severe Mental Illness Attending Outpatient Services The second target population was comprise d of individuals who were currently engaged in outpatient therapy for psychiatri c distress who are less se vere and do not meet the criteria for SMI. There are approximately 1,000-1,200 outpatient clients who meet inclusion/exclusion criter ia at the mental health center. As exclusion criteria, participants with strictly substance related issues (i.e., alcohol or drug abuse or dependence) were not eligible to participate. Data constraints at the partnering agency did not allow them to generate a gender/race/ethnicity breakdown of this sampling frame. To contact individuals, two lis ts of eligible clients we re generated by the mental health center (SMI and OP samples). Th e 175 names for each potential sample were given to the lead researcher. Each client was paired with their primary provider on the list. Each provider was contacted to ensure that the client met in clusion and exclusion criteria. The agency had asked that c ontact be carried out primarily through the researcher due to agency personnel limitations and task requirements. Each potential client received an invitation letter to the study (see Appendi x C). This letter briefly outlined the study purpose, client and resear cher obligations, and the incentive for participation. Additional names and number matches were made available as clients either refused or were ineligible. Each letter was followed up with a phone call to arrange an appointment time. The process of contacting and beginning da ta collection was considerably slowed when it became evident that the contact info rmation for the clients was not up-to-date. This required contacting providers and leaving letters for contacts at appointments. This

PAGE 122

103 slowed the process but did not interfere with the sample selection or randomization. The initial list for the SMI sample included a number of individuals who had moved out the area, discontinued treatment, and two who were deceased. The agency was asked to restrict their list to active clients and an a ppropriate list was generated. The providers recommended no contact for ten individuals, and ten more did not meet inclusion and exclusion criteria. A total of sixteen client s refused to participate. Eleven of these refused during the scheduling phone contact. None of the eleven clients opted to fill out the demographic short-form over the phone and only four clients did so in person. The low level of refusal and even lower level of participation for filli ng out the short history form did not allow a comparison between participants and non-participants. A summary of sample refusals and ineligibility is provi ded in Table 2. The final sample size met the power requirements (see next sectio n) of 175 SMI and 175 OP individuals. Table 2 Summary of Contact Refusals and Ineligibil ity by Sample Total Contacted Refused Provider Designated Ineligible Inclusion / Exclusion Ineligible SMI sample 195 6 10 4 OP sample 186 10 0 6 Power and Sample Size Estimation As stated, each sample was comprised of 175 participants, for a total of 350. This number was selected following a sample size/power analysis for multivariate statistics, specifically for multiple linear regression (Milton s procedure) and for logistic regression or its variants (e.g., probit analysis) (B rent, Mirelli, & Thompson, 1993; Cohen, 1988).

PAGE 123

104 (see analysis section below for details on a priori statistical models). For the Milton procedure, the following was considered in computation, with a final sample minimum of 175 participants per sample: Probability of a Type I error = .05; A main effects model with 65 total terms; Two-tailed analysis; Estimated partial correlation for the least significant variable to be detected = .003; and Estimated r-square (proportion of varian ce explained by the main effects models) = .15. The planned multivariate analyses included logistic regression, multiple linear regression, and probit analysis For a sample size of 175 per sample and 65 terms, logistic regression and probit an alysis are sufficiently powered with a Type I error = 0.05 and a Type II error = 0.2, to detect an odds rati o of 1.9 or larger. With 55 total terms, this improves to detecting an odds ratio of 1.6 or larger. Research Questions and Hypotheses Research questions (RQ) and hypotheses (Hyp) are logical extensions of the current state of the literatur e. Hypotheses listed are a priori RQ1: To what degree does illness severity influence beliefs in recovery for the individual? Hyp1.1: The OP sample will endorse that th ey will recover from mental illness to a greater extent than the SMI sample.

PAGE 124

105 Hyp1.2: The OP sample will endorse complete recovery as possible more than the SMI sample. Hyp1.3: The SMI sample will endorse no expectation of recovery more than the OP sample. RQ2: Are individual, social or service factors associated with recovery expectancy? RQ2.1: Are individual factors associated with recovery expectancy? Hyp2.1a: Less severe diagnoses will be associated with higher recovery expectancy. Hyp2.1b: Lower somatization symptoms reported by participants will be associated with higher recovery expectancy. Hyp2.1c: Lower obsessive-compulsive symptoms reported by participants will be associated with higher recovery expectancy. Hyp2.1d: Lower interpersonal sensitivity re ported by participants will be associated with higher recovery expectancy. Hyp2.1e: Lower symptoms of depression reported by participants will be associated with higher recovery expectancy. Hyp2.1f: Lower symptoms of anxiety reported by participants will be associated with higher recovery expectancy. Hyp2.1g: Lower hostility reported by partic ipants will be associated with higher recovery expectancy. Hyp.1h: Lower phobic anxiety reported by pa rticipants will be associated with higher recove ry expectancy.

PAGE 125

106 Hyp2.1i: Lower paranoid ideation reported by participants will be associated with higher recovery expectancy. Hyp2.1j: Lower psychoticism reported by participants will be associated with higher recove ry expectancy. Hyp2.1k: Lower total number of symptoms reported by participants will be associated with higher recovery expectancy. Hyp2.1l: Absence of a comorbid substance use disorder will be associated with higher recovery exp ectancy for participants. Hyp2.1m: No psychiatric hospitali zation in the last year will be associated with higher recovery exp ectancy for participants. Hyp2.1n: Lower numbers of lifetime psychiatric hospitalizations will be associated with higher recovery expectancy for participants. Hyp2.1o: Older age at onset of disorder will be associated with higher recovery expectancy for participants. Hyp2.1p: Absence of familial mental illness in first-degree relatives will be associated with higher recove ry expectancy for participants. Hyp2.1q: Absence of familial mental illness in extended family members will be associated with higher reco very expectancy for participants. Hyp2.1r: Being currently employed will be associated with higher recovery expectancy for participants. Hyp2.1s: Greater number of years worked will be associated with higher recovery expectancy for participants.

PAGE 126

107 Hyp2.1t: Absence of prescribed anti-psychotic medication will be associated with higher recovery expectancy for participants. Hyp2.1u: Absence of prescribed anti-depressant medication will be associated with higher recovery expectancy for participants. Hyp2.1v: Absence of prescribed anti-mani a medication will be associated with higher recovery exp ectancy for participants. Hyp2.1w: Absence of prescribed anti-anxiety medication will be associated with higher recovery expectancy for participants. Hyp2.1x: Absence of any other prescribed psychotropic medication will be associated with higher recove ry expectancy for participants. Hyp2.1y: Absence of child sexual abuse will be associated with higher recovery expectancy for participants. Hyp2.1z: Absence of child physical abuse will be associated with higher recovery expectancy for participants. Hyp2.1aa: Absence of adult sexual assault will be associated with higher recovery expectancy for participants. Hyp2.1ab: Absence of adult physical assault will be associated with higher recovery expectancy for participants. RQ2.2: Are social factors associated with recovery expectancy? Hyp2.2a: Lower feelings of alienation will be associated with higher recovery expectancy for participants. Hyp2.2b: Lower participant endorsement of mental illness stereotypes will be associated with higher reco very expectancy for participants.

PAGE 127

108 Hyp2.2c: Lower number of discrimination experiences will be associated with higher recovery exp ectancy for participants. Hyp2.2d: Lower endorsement of social withdrawal will be associated with higher recovery exp ectancy for participants. Hyp2.2e: Greater endorsement of stigma resistance will be associated with higher recovery exp ectancy for participants. Hyp2.2f: Greater support through intimate pa rtner or a best friend will be associated with higher recovery expectancy for participants. Hyp2.2g: Greater support through family members will be associated with higher recovery exp ectancy for participants. Hyp2.2h: Greater support via mental hea lth providers will be associated with higher recovery exp ectancy for participants. Hyp2.2i: Greater support through friendship will be associated with higher recovery expectancy for participants. Hyp2.2j: Greater feelings of being conn ected to the community will be associated with higher recovery expectancy for participants. Hyp2.2k: Greater trust in the motivation of others will be associated with higher recovery expectancy for participants. Hyp2.2l: Higher ratings of self-esteem a nd self-efficacy will be associated with higher recovery exp ectancy for participants. Hyp2.2m: Higher ratings of personal power will be associated with higher recovery expectancy for participants.

PAGE 128

109 Hyp2.2n: Greater involvement in the co mmunity will be associated with higher recovery expectancy for participants. Hyp2.2o: Greater confidence in personal control over the future will be associated with higher recovery expectancy for participants. Hyp2.2q: Higher ratings of righteous ange r will be associated with higher recovery expectancy for participants. RQ2.3: Are service factors associat ed with recovery expectancy? Hyp2.3a: Total number of services will be positively associated with higher recovery expectancies for participants. Hyp2.3b: The average number of contact hours per month will be positively associated with higher recovery expectancies for participants. Hyp2.3c: Satisfaction level with services will be positively associated with higher recove ry expectancies for participants. RQ3: Are individual, social or service factors associated with recovery strategies? RQ3.1: Are individual factors associated with recovery strategies? Hyp3.1a: Less clinically severe diagnos es for participants will be associated with greater endorsement for a recovery strategy. Hyp3.1b: Lower somatization symptoms reported by participants will be associated with greater endorsement for a recovery strategy. Hyp3.1c: Lower obsessive-compulsive symptoms reported by participants will be associated with greater endorsement for a recovery strategy. Hyp3.1d: Lower interpersonal sensitivity re ported by participants will be associated with greater endorsement for a recovery strategy.

PAGE 129

110 Hyp3.1e: Lower symptoms of depression reported by participants will be associated with greater endorsement for a recovery strategy. Hyp3.1f: Lower symptoms of anxiety reported by participants will be associated with greater endorsement for a recovery strategy. Hyp3.1g: Lower hostility reported by partic ipants will be associated with greater endorsement for a recovery strategy. Hyp3.1h: Lower phobic anxiety reported by participants will be associated with greater endorsement for a recovery strategy. Hyp3.1i: Lower paranoid ideation reported by participants will be associated with greater endorsement for a recovery strategy. Hyp3.1j: Lower psychoticism reported by participants will be associated with greater endorsement for a recovery strategy. Hyp3.1k: Lower total number of symptoms reported by participants will be associated with greater endorsement for a recovery strategy. Hyp3.1l: Absence of a comorbid substance use disorder will be associated with greater endorsement for a r ecovery strategy for participants. Hyp3.1m: No psychiatric hospitali zation in the last year will be associated with greater endorsement for a r ecovery strategy for participants. Hyp3.1n: Lower numbers of lifetime psychiatric hospitalizations will be associated with greater endorsement for a recovery strategy for participants. Hyp3.1o: Older age at onset of disorder wi ll be associated with greater endorsement for a recovery strategy for participants.

PAGE 130

111 Hyp3.1p: Absence of familial mental illness in first-degree relatives will be associated with greater endorsement for a recovery strategy for participants. Hyp3.1q: Absence of familial mental illness in extended family members will be associated with greater endo rsement for a recovery strategy for participants. Hyp3.1r: Being currently employed will be associated with greater endorsement for a recovery strategy for participants. Hyp3.1s: Greater number of years worked w ill be associated with greater endorsement for a recovery strategy for participants. Hyp3.1t: Absence of prescribed anti-psychotic medication will be associated with greater endorsement for a recovery strategy for participants. Hyp3.1u: Absence of prescribed anti-depressant medication will be associated with greater endorsement for a recovery strategy for participants. Hyp3.1v: Absence of prescribed anti-mani c medication will be associated with greater endorsement for a r ecovery strategy for participants. Hyp3.1w: Absence of prescribed anti-anxiety medication will be associated with greater endorsement for a recovery strategy for participants.

PAGE 131

112 Hyp3.1x: Absence of any other prescribed psychotropic medication will be associated with greater endorsement for a recovery strategy for participants. Hyp3.1y: Absence of child sexual abuse will be associated with greater endorsement for a recovery strategy for participants. Hyp3.1z: Absence of child physical abuse will be associated with greater endorsement for a recovery strategy for participants. Hyp3.1aa: Absence of adult sexual assault will be associated with greater endorsement for a recovery strategy for participants. Hyp3.1ab: Absence of adult physical assault will be associated with greater endorsement for a recove ry strategy for participants. RQ3.2: Are social factors associated with recovery expectancy? Hyp3.2a: Lower feelings of alienation will be associated with higher recovery expectancy for participants. Hyp3.2b: Lower participant endorsement of mental illness stereotypes will be associated with higher reco very expectancy for participants. Hyp3.2c: Lower number of discrimination experiences will be associated with higher recovery exp ectancy for participants. Hyp3.2d: Lower endorsement of social withdrawal will be associated with higher recovery exp ectancy for participants. Hyp3.2e: Greater endorsement of stigma resistance will be associated with higher recovery exp ectancy for participants.

PAGE 132

113 Hyp3.2f: Greater support through intimate pa rtner or a best friend will be associated with higher recovery expectancy for participants. Hyp3.g: Greater support through family members will be associated with higher recovery expectancy for participants. Hyp3.2h: Greater support via mental hea lth providers will be associated with higher recovery exp ectancy for participants. Hyp3.2i: Greater support through friendship will be associated with higher recovery expectancy for participants. Hyp3.2j: Greater feelings of being conn ected to the community will be associated with higher recovery expectancy for participants. Hyp3.2k: Greater trust in the motivation of others will be associated with higher recovery expectancy for participants. Hyp3.2l: Higher ratings of self-esteem a nd self-efficacy will be associated with higher recovery exp ectancy for participants. Hyp3.2m: Higher ratings of personal power will be associated with higher recovery expectancy for participants. Hyp3.2n: Greater involvement in the co mmunity will be associated with higher recovery expectancy for participants. Hyp3.2o: Greater confidence in personal control over the future will be associated with higher recovery expectancy for participants. Hyp3.2q: Higher ratings of righteous ange r will be associated with higher recovery expectancy for participants. RQ3.3: Are service factors associat ed with recovery expectancy?

PAGE 133

114 Hyp3.3a: Total number of services w ill be associated with higher recovery expectancies for participants. Hyp3.3b: The average number of contact hours per month (averaged over one year of service) will be associated with higher recovery expectancies for participants Hyp3.3c: Satisfaction level with services will be associated with higher recovery expectancies for participants. RQ4a: Does the expectation of recovery mediate the relationship between individual, social and service factors and recovery strategies? Hyp4a.1: Participants expectation of rec overy will mediate the relationship between individual factors and recovery strategies. Hyp4a.2: Participants expectation of rec overy will mediate the relationship between social factors a nd recovery strategies. Hyp4a.3: Participants expectation of rec overy will mediate the relationship between service factors a nd recovery strategies. RQ4b: Does the expectation of recovery mode rate the relationship between individual, social and service factors and recovery strategies? Hyp4b.1: Participants expectation of rec overy will moderate the relationship between individual factors and recovery strategies. Post Hoc: Post hoc analyses will target spec ific variables that are found to be statistically and clinically relevant to recovery Hyp4b.2: Participants expectation of rec overy will moderate the relationship between social factors a nd recovery strategies.

PAGE 134

115 Post Hoc: Post hoc analyses will target spec ific variables that are found to be statistically and clinically relevant to recovery Hyp4b.3: Participants expectation of rec overy will moderate the relationship between service factors a nd recovery strategies. Post Hoc: Post hoc analyses will target spec ific variables that are found to be statistically and clinically relevant to recovery RQ5: Does severity of mental illness moderate the relationship between individual, social and service factors and recovery strategies? Hyp5.1: Severity of mental illness will moderate the relationship between individual factors and recovery strategies for participants. Post Hoc: Post hoc analyses will target spec ific variables that are found to be statistically and clinically relevant to recovery Hyp5.2: Severity of mental illness will mode rate the relationship between social factors and recovery stra tegies for participants. Post Hoc: Post hoc analyses will target spec ific variables that are found to be statistically and clinically relevant to recovery Hyp5.3: Severity of mental illness will mode rate the relationship between service factors and recovery stra tegies for participants. Post Hoc: Post hoc analyses will target spec ific variables that are found to be statistically and clinically relevant to recovery Study Setting The partnering agency (The Harbor Beha vioral Health Care Institute), a large community mental health center serving over 20,000 individuals annua lly in 34 locations

PAGE 135

116 across three counties, generous ly provided access to clients for this research. Data collection occurred primarily in three clinic locations. For the majority of study participants (89%), the researcher travel ed to the agency location closest to the participant. Home or apartment visits acco mmodated the remaining participants. Data collection was completed on an individual basis with either the principal investigator or research assistant present and assisting each participant. A private office or meeting room was provided for data co llection in the agency and home visits were all completed free of interruption by family members. Data Collection Procedures This section provides details of the study methods and procedures in temporal order. Four sub-sections describe pre-data collection procedures, data collection procedures, theoretical constr ucts (in brief) and the inst ruments relevant to each construct, and post-data collec tion (data analysis) procedures. Pre-data collection procedures Institutional Review Board (IRB) authorization The community mental health center partner and the researcher negotiated full access for the study requirements. The study was reviewed and sanctioned by both the University of South Florida IRB and from the IRB that permits research at the Har bor (Morton Plant Mease Bay Care IRB). Training of research assistants Four research assistan ts were trained for this project. Training procedures and materials are detailed in Appendix D, but are briefly described here. Three research assistants we re undergraduate students at the University of South Florida, and one was a student at Pasco Hernando Community College. The one student from the Community College was unabl e to complete the project. Each student

PAGE 136

117 volunteered and gave 6-12 hours a week toward da ta collection. Before the start of data collection, each student completed the NIH training for human subject protection. Study procedures were offered in five training se ssions, each lasting three hours. The topics covered in the training include d: ethics and human subj ect protection, confidentiality, emergency procedures if the consumer bega n to show signs of active symptoms (e.g., crying or anger), consent proced ures and use of the informed consent checklist, use of the incentive receipt form, familiarity with each data collection instrument, procedures for when the consumer was willing and capable of filling out the forms for him or herself, procedures for when the interviewer read ea ch question to the consumer, monitoring each instrument as it is completed to intercept missing data, use of the oversized print laminated response cue cards that correspond with each collection tool, provision of the incentive, and when/where to drop off comp leted forms to the lead researcher. Data collection procedures Participant contact Contact occurred both forma lly and informally. Contact formally via a letter sent to the partic ipants residence or hand delivered by the participants primary provider occurred in 75% of cases. Informally, providers were aware of which of their clients had been randomly selected. Some providers informed their clients before letters of invitation could be personalized and delivered to the provider. Similarly, some providers phoned thei r clients before letters were available and the client would in turn cont act the researcher via phone to schedule an appointment. In most cases, follow-up phone contacts to schedu le were completed within four days of when the letter was received by the participant.

PAGE 137

118 Confidentiality and data storage Data collection requires assurance of confidentiality and appropriate data storage. Data are stored in the office of the researcher in a locked cabinet. Computer files are password prot ected. Consent forms are filed separately to ensure confidentiality and to verify that the full consent process was used. Data collection forms only have the participant number, no names, and are stored sequentially by particip ant number. Multiple backups of all computer data and analysis files have been maintained and encrypted. Informed consent and verification of inclusion/exclusion criteri a Participants had the option of reading the consent form and aut horization themselves or having it read to them. Ninety-three percent (93%) chose to read the surveys on their own. Informed consent for individuals with a mental disorder is somewh at controversial. The SMI population especially is at risk for cognitive de ficits that could impair their ability to understand and fully consent to the process, t hough similar deficits may be present in the OP sample. Because of this, each subject from both samples was provided an independent capacity assessment through the fo llowing procedure: The subjects had the consent form explained to them one point at a time. Additionally, a brief set of questions concerning the most salient points (e.g., right of refusal, right to quit at any time) was developed (see Informed Consent Checklist in Appendix C), and consent was not considered complete until the subject answer ed each question corre ctly and sequentially (Carpenter et al., 2000; Wirshing, Wirshing, Ma rder, Liberman, & Mintz, 1998). Results of the independent capacity assessment are pr ovided in Table 3. The majority (97%) of individuals in the SMI sample understood the consent process and passed the independent

PAGE 138

119 capacity assessment on the first attempt. A ll the OP sample participants passed on the first attempt. Table 3 Summary of Independent Capacity Assessment Results Attempts to Pass Consent SMI Sample OP Sample First Attempt 97% 100% Second Attempt 2% Third Attempt 1% Incentives During the informed consent procedure, the incentive was explained again. Each participant was given an incentive of $10.00 for participating after completion of the data collection process. In order to track the monetary outlay, each participant signed a form that they have received the $10.00 incentive. Data collection Once informed consent was completed data collection began. Ordering of instruments is important, with l onger instruments being used first followed by shorter instruments and inst ruments that require personal information being used last after rapport is established. Complete desc riptions of the instruments listed below are provided in the next section. The following sequence was used with each individual: 1. Symptom Checklist 90-Revised (SCL-90-R). This is the longest instrument and also asks questions that consumers of clinical services are accustomed to, increasing comfort. 2. Support and Community Connectedness Scale (SCCS) 3. Recovery Expectation Checklist (REC) 4. Recovery Assessment Scale (RAS)

PAGE 139

120 5. Personal Vision of Recovery Questionnaire (PVRQ) 6. Empowerment Scale (ES) 7. Internalized Stigma of Me ntal Illness Scale (ISMI) 8. Service Satisfaction Questionnaire (SSQ) 9. Background and History form. The researcher provided all supplies. No ted above, the participants had the option of having the researcher read the instrument questions aloud. If this option was taken, a laminated cue card with responses for each inst rument was used to he lp the client select responses on the appropriate scale. The amount of time for data collection (from introducti on to paying of incentive) was noted for each individual. Data collecti on took, on average, 68 minutes to complete. The range of time was 27 minutes to 340 minutes Only three individuals took over three hours to complete their participation. Four individuals needed between two and three hours. Subjects were encouraged to take a brief break if they became fatigued and were allowed to stop and resume at a different tim e. Of the 350 participants, less than one percent (n=2) chose to stop a nd resume on a different day. There may be some concern that this was an inordinate amount of time for an individual with mental illness to participate. However, qualitative studies us ing in-depth interviews with mentally ill participants have averaged 60-90 minutes per interview without any negative effect on the participant (Auquier, et al., 2003; Wal by, 2003a, 2003b). No individual complained of fatigue and there were no adve rse events during data collection.

PAGE 140

121 The total span of time for the data collection was eighteen weeks. The researcher offered to collect data on weekends and eveni ngs for clients who might indicate that this is the most convenient time, but no pa rticipants requested this option. Dependent Variables: Recovery Expectancies and Strategies Recovery is defined as both the belief that an individual will overcome mental illness, through suppression or compensation, and the strategies or processes an individual uses to enhance recovery. The c onstruct of recovery is operationalized using both expectancies and strategies in this rese arch. The expectation of recovery can be measured directly. However, the measurement of recovery strategies requires consensus on what strategies for recovery are. Identify ing and describing recove ry strategies is only beginning (Borowitz-Ensfield, 1998; Corrigan & Ralph, 2005; Corrigan et al., 2005). This research effort identifies recovery stra tegies by combining items from two recovery surveys and using factor analysis to identif y strategies. Recovery was assessed using three instruments. 1. Recovery Expectation Checklist (REC). Th is instrument is unique in that it provides data as a dependent variable fo r some analyses and as a mediating or moderating variable for ot her analyses. The REC data as mediator or moderator will be explained in the appropriate section below. 2. Recovery Assessment Scale (RAS) 3. Personal Vision of Recovery Questionnaire (PVRQ) Recovery Expectation Checklist The recovery efficacy checklist is an original instrument that measures recovery expectancy by the participant indicating if they believe they will be able to recover from

PAGE 141

122 mental illness and to what degree. This is a four question instrument that, in addition to the direct inquiry of recovery belief, also inquires into th e what recovery means to the participant and what resources or supports they feel are impor tant to their recovery (see Appendix A for a formatted version of this a nd all other instruments described in this Chapter). The first two questions measure whether recovery is subjectively experienced as possible for each individual and to what degr ee they expect to recovery. The responses are used to assess belief level in recovery for the total sample and to measure differences between the SMI and OP samples. Question 3 of the REC provides the option of endorsing seven options that matc h key aspects of recovery (e.g., I will be able to work full time). The fourth question is a simple ch ecklist derived from a literature review that lists specific relationships, hab its, or activities that were c onsidered recovery enhancing. The list is not viewed as comprehensive and there are several spaces for additional responses. Questions three and four are additional information for a more complete understanding of the recovery construct. Recovery Assessment Scale (RAS) The RAS is a 41-item survey that measures strategies important to recovery from mental illness (Corrigan, et al., 1999). The instrument was created in cooperation with consumers of mental health service and was designed to identify factors/pathways important to recovery. A recent paper inves tigating the structure of the RAS identified five principal components generated from 24 of the 41-item inventory (Corrigan, Salzer, Ralph, Sangster, & Keck, 2004).

PAGE 142

123 One area of concern is that the authors deviated from what might be considered best practice when confirming the factor stru cture of a new instrument. They first used principal components analysis (PCA), in place of exploratory factor analysis (EFA), and followed the PCA with a confirmatory factor analysis (CFA). Essentially, only factor analysis may be used to identify the f actor structure underlying a set of variables (Hatcher, 1994, p. 69). In PCA, the component s extracted are linear combinations of the observed variables, whereas in factor anal ysis, the observed variables are linear combinations of the underlying factors (Hatch er, 1994, p. 70). Further, total variance (combined from common and unique variance) detected in PCA is different then factor analysis, which detects common variance al one (Isaac & Michael, 1995). Thus, it is inappropriate to follow a PCA with a CFA sin ce they are measuring different aspects of variance. If items are generated that perf ectly measure constructs, common variance and total variance are identical and the choice be tween PCA and EFA as an analytic approach would be irrelevant. Since perfect items were not generated, the difference between common and total variance is salient for this an d other research. However, the errors in measurement addressed here do not affect the outcome of this research since the factor structure described by the author is not used in this analysis. Psychometric properties of the RAS include an overall test-retest reliability factor of r = 0.88 and internal consistency = 0.93. The results of the PCA and CFA are presented in Table 4. Concurrent validity was measured through correlations with various psychosocial variables with signifi cant concurrence identified for self-esteem, empowerment self-orientation, size of support network and quality of life. Concurrence was not demonstrated with empowerment, community-orientation, satisfaction with

PAGE 143

124 Table 4 Factor Structure for Recovery Assessment Scale Factor Loading PCA* Factor Loading CFA* Component 1: Personal Confidence and Hope Fear doesnt stop me from living the way I want to 0.46 0.54 I can handle what happens in my life 0.49 0.67 I like myself 0.68 0.72 If people really knew me, they would like me 0.61 0.59 I have an idea of who I would like to become 0.34 0.64 Something good will eventually happen 0.49 0.65 I am hopeful about my future 0.52 0.74 I continue to have new interests 0.41 0.70 I can handle stress 0.40 0.59 Component 2: Willingn ess to Ask for Help I know when to ask for help 0.74 0.76 I am willing to ask for help 0.81 0.76 I ask for help when I need it 0.81 0.82 Component 3: Goal and Success Orientation I have a desire to succeed 0.70 0.53 I have my own plan for how to stay or become well 0.49 0.68 I have goals in life that I want to reach 0.76 0.70 I believe I can meet my current personal goals 0.52 0.79 I have a purpose in life 0.50 0.75 Component 4: Reliance on Others Even when I dont care about myself, other people do 0.68 0.58 I have people I can count on 0.67 0.72 Even when I dont believe in myself, other people do 0.76 0.76 It is important to have a variety of friends 0.34 0.54 Component 5: No Domination by Symptoms Coping with mental illness is no longer the main focus of my life 0.74 0.59 My symptoms interfere less and less with my life 0.75 0.87 My symptoms seem to be a problem for shorter periods of time each time they occur 0.68 0.65 Note: *PCA = Principal Components Analysis ; *CFA = Confirmatory Factor Analysis

PAGE 144

125 support network, vocabulary or education. Thes e findings were in th e predicted direction and supported the authors claim of a valid and reliable instrument. Individual scale alpha-scores were in the middle to high range. The items for this scale were developed to reflect the four domains of recovery postulated by Ralph (2000) that include (1) internal fact ors, (2) self-managed care, (3) external factors, and (4) empowerment. A brief descri ption and Cronbachs alpha (i nternal consistency proxy for test-retest reliability) are provided for each of the factors of the RAS: (1) Personal confidence and hope : = 0.87. This factor is composed of nine items. The items reveal a future orient ation that is hopeful and also reflects liking the self and being comfortable that others would like the person as well. The questions also describe the ability to cope with stress and fear. (2) Willingness to ask for help : = 0.84. The authors voice concern about this factor since the wording of the three que stions that comprise the factor are obvious variants on a theme. The th ree questions address whether an individual knows when to ask for help, is willing to do so, and has a history of doing so if needed. (3) Goal and success orientation : = 0.82. The five items comprising this factor describe success, purpose, meaning and goal accomplishment. (4) Reliance on others: = 0.74. This factor encompasses four items. The name of the factor evokes a willingness or history of relying on others for assistance. However, the wording of th e items targets believing that others care about or believe in the respondent a nd that the individual recognizes that they have people they can count on.

PAGE 145

126 (5) No domination by symptoms : = 0.74. The three items in this factor describe the individuals perception of symptoms and if symptoms are a major point of consideration or contention in the respondents life. Personal Vision of Recovery Questionnaire (PVRQ). The PVRQ is a 24-item survey that measures recovery strategies for mental illness (Borowitz-Ensfield, 1998). The survey wa s created through discussion and assistance from consumers and revealed five factors related to recovery. Exploratory factor analysis, Cronbachs alpha and construct va lidation through correlation with associated measures were used to develop and confirm th e instrument. The final array of factors is consistent with firstperson accounts, quali tative research, and other documentation of the recovery process and is presented in Table 5. Models testing a single versus multiple f actors indicate that ther e is not one factor that sufficiently accounts for the variance, de monstrating that the re covery construct was multidimensional. The five-factor model, the final model, accounted for the most variance. The author used multiple indices of goodness of fit to evaluate the instrument. These included the scaled comparative fit in dex (CFI), which should exceed 0.9; the root mean square approximation (RMSEA), which should be 0.05; the Satorra-Bentler chisquare, which should be approximately e qual to the degrees of freedom; and, the standardized root mean squared resi dual (RMSR), which should also be 0.05 (Borowitz-Ensfield, 1998, pg 23-25). For th e five-factor (final) model, the CFI = 0.94, the RMSEA = 0.02, the Satorra-Bentler chi-sq uare = 249.77 (degrees of freedom = 234), and the RMSR = 0.08.

PAGE 146

127 Table 5 Factor Structure for Personal Vision of Recovery Questionnaire Factor Loading Factor 1: Recovery through Support Self-help groups are important to my recovery .68 Family support is important for my recovery .63 Helping others is part of my recovery .56 Recovery means becoming satisfied with my life .49 I am responsible for my own recovery .46 Support from mental health professiona ls is important for my recovery .42 Factor 2: Recovery through Personal Challenges Recovery involves finding new meaning in my life .68 People who expect very little of me interfere with my recovery .64 Side effects from my medication ma ke it harder for me to recover .54 Recovery means I will be free of symptoms .47 Recovery means getting mo re control of my life .41 Factor 3: Recovery through Professional Assistance At times, treatment against my wishes is necessary for my recovery. For example, involuntary hospita lization, forced medication, or community probate .70 The cause of mental illness is not important for my recovery .64 Recovery means I will not be mentally ill anymore .54 Being diagnosed correctly is necessary for my recovery .43 Factor 4: Recovery thro ugh Action and Help-Seeking Sticking up for clients rights is a part of my recovery .72 Asking for help is a part of my recovery .53 I know people who are recovering fr om problems similar to mine .49 Having something meaningful to do is important for my recovery .46 Support from a special person, such as a spouse or partner, is important for my recovery .32 Factor 5: Recovery through Affirmation Hope is important for my recovery .78 Spirituality is a part of my recovery .50 I am convinced that medication can help me to recover .49 Recovery means my symptoms will be easier to control .39

PAGE 147

128 It should be noted that the PVRQ scal es have a lower degree of internal consistency then is usually noted in validated instruments. The author hypothesizes that the low internal consistency likely reflected th e fluid nature of the recovery construct. The definition and components of recovery are not used consistently in the literature and there is preliminary evidence that the con cept of recovery alte rs over time (Andresen, Oades, & Caputti, 2003). The participants in the study that devel oped the PVRQ were likely in different stages of recovery and responded to th e question content differently due, in part, to where they lie on a continuum of recovery. Measurement of Potential Correlates (Independent Variables) Across Multiple Domains Many factors that may enhance or impede recovery were introduced in Chapter 2. It was not feasible to measure ev ery one of these potentially valuable factors in this single study. However, it was possibl e to measure the factor s with the strongest theoretical links to recovery. Individual m easurement tools are discussed as they pertain to a domain. Domain 1: Individual, Historical and Clinical Factors There is considerable information a bout diagnosis and, to a lesser degree, symptoms, and their effects on recovery. Anecd otal and clinical lite rature suggests that familial mental illness, abuse history, age of onset and other factors may also be related to recovery expectancy and recovery strate gies. The following three instruments were used to tap key components of this domain. 1. Background Characteristics, Related Asp ects of History and Demographics a. Background Short Form 2. Symptom Checklist 90-Revised (SCL-90R)

PAGE 148

129 3. Chart Abstraction Tool Background Characteristics, Related Asp ects of History and Demographics. Individuals were asked to complete a b ackground form with items targeting age, gender, education, income, employment, curr ent living situation, history of familial mental illness, substance use history, abuse/trauma history, age of onset, employment and other factors. Due to the sensitive nature of some of the questions, several areas are investigated via the chart ab straction tool (see below), both as a validity check and to respect the wishes of individual participants who may find it uncomfortable to discuss some areas. Only four individuals refused to answer questions, all of which were abuse history related. Information from the clinical charts and recorded in the abstraction tool verified that all four individua ls were severely abused as children and were reluctant to address this issue in therapy. Background Short Form This form is an abridged version of th e background form described above. This form has a few key demographic questions that individuals who refuse to participate were asked to complete. This information was in tended to be compared against full study participants to check for systematic differences between participants and nonparticipants. However, noted in Tabl e 2 above, only 16 individuals refused to participate. Of these 16 individuals, onl y four completed the background short form, which did not provide sufficient information for a comparison. The Symptom Checklist 90-Revised (SCL-90-R) The SCL-90-R is a 90-item self-report symp tom inventory designed to reflect the psychological symptom patterns of communit y, medical, and psychiatric participants

PAGE 149

130 (Derogatis, 1994). The SCL-90-R is a measure of current symptom status. Measurement of present symptom levels is important as th e effect of symptom severity and types of symptoms have not been sufficiently correlated with recovery. Further, other studies of recovery have not correlated symptoms with diagnosis. The nine subscales and three composite scales examine for a breadth of potential symptomatology. Each item is scaled on a five-point scale with the ra nge of answers from not at all (0) to extremely (4). Participants were asked to i ndicate how much discomfort that problem has caused you during the past seven days including the da y of administration (Derogatis, 1994). The symptom dimensions that comprise 83 of the 90 questions in the SCL-90-R include somatization, obsessive-compulsive, interp ersonal sensitivity, depression, anxiety, hostility, phobic anxiety, pa ranoid ideation, and psychotic ism. A tenth dimension investigates points of relevant c linical interest that do not fit into the symptom categories. This instrument has been used in numerous research and clinical settings with a considerable amount of information relating to both validity and reliability accrued. There have been numerous applications of th is survey to various types of psychiatric populations with analysis of reliability thr ough test-retest and coeffi cient alpha. Table 6 reproduces the results from a selection of these studies. The instrument is reliable across different types of psychiatric populations and for varying leng ths of time between administrations. One potential limitation is th at several studies with different populations and utilizing factor analysis ha ve found more or less factors than the nine factors detailed by the originator. However, there was not uni formity in principal component analysis vs. factor analysis, procedures used for samp le selection, or othe r methodological and analytic choices. These inc onsistencies may, in part, account for this discrepancy.

PAGE 150

131 Derogatis, (1994) provided evidence for the external validity of the SCL-90-R using three approaches. The internal struct ure of the instrument was tested through a factor analytic procedure. Factorial invariance the cons tancy in the composition of a dimension as one changes significant subject pa rameters such as age, gender or social class was examined through correlati ons across symptom dimensions by gender (Derogatis, 1994). Finally, converg ent and discriminant validity the measure of interest correlates highly with other instruments known to measure the same construct or correlates poorly with instruments that meas ure dissimilar constructs was confirmed by comparing the SCL-90-R with the Minnesota Multiphasic Personality Inventory (MMPI) and the Middlesex Hospital Qu estionnaire (Der ogatis, 1994). The MMPI clinical scales were suppleme nted with the Wiggins content scales (Graham, 2005) (e.g., psychoticism, manife st hostility) and Tryons cluster scales (Derogatis, 1994) (e.g., anxiety, autism). The convergent validity ranges from 0.42 between the paranoia scales of the MMPI and the SCL-90-R to 0.75 between the Wiggins depression scale and SCL-90-R depression di mension. Divergent validity also had a considerable range and in only one case was the divergent correlation for another factor higher than the convergent (SCL-R-90 pa ranoia and MMPI paranoi a = 0.42; SCL-R-90 psychoticism and MMPI paranoia = 0.48). Se veral other studies have added to the validity literature on the SCL-90 (Carpenter & Hittner, 1995; Evenson, Holland, Mehta, & Yasin, 1980; Hafkenscheid, 1993; Raut er, Leonard, & Swett, 1996; Vassend & Skrondal, 1999). These studies support the vali dity of the instrument and are especially consistent for the to tal pathology scale.

PAGE 151

132 Table 6 Reliability Estimates for the SCL-90-R Original Reliability and Subsequent Estimates Original Reliability Estimates Subsequent Reliability Estimates Dimension a Coef Test-Retest (r 11 ) Coef e Coef f Coef g Coef h Coef i Study 1 b Study 2 c Study 2 c Study 3 d Pap-Pen h Computer h Ref h P/S i PC i Ref i SOMA .86 .88 .68 .8 6 .92 .87 .79 .90 .79 .70 .86 .83 .70 O-C .86 .87 .70 .85 .75 .81 .78 .84 .8 5 .75 .86 .85 .75 I-S .86 .84 .81 .83 .79 .84 .79 .87 .8 4 .76 .87 .84 .76 DEP .90 .90 .75 .82 .96 .89 .82 .92 .8 6 .83 .90 .90 .83 ANX .85 .88 .80 .80 .56 .88 .82 .88 .8 8 .75 .87 .87 .75 HOS .84 .85 .73 .78 .86 .76 .75 .73 .5 9 .62 .78 .78 .62 PHOB .82 .89 .77 .9 0 .82 .89 .72 .92 .86 .51 .86 .80 .51 PAR .80 .79 .83 .86 .89 .80 .67 .65 .7 6 .63 .78 .80 .63 PSY .77 .80 .77 .84 .78 .76 .62 .77 .7 6 .65 .78 .79 .65 GSI .96 .97 .98 .94 a SOMA = somatization; O-C = obsessive-compulsive; I-S = interper sonal sensitivity; DEP = depression; ANX = anxiety; HOS = hosti lity; PHOB = phobic anxiety; PAR = paranoia; PSY = psychoticism; GSI = global severity index b N = 209 symptomatic volunteers (Derogatis, Rickels, & Rock, 1976 ) (original estimate) c N = 103 psychiatric outpatients (Horowitz et al., 1988); el apsed time between tests = 10 weeks (original estimate) d N = 94 heterogeneous psychiatric outpatients (Derogatis, Rickels, & Rock, 1976); elapsed time between tests = 1 week (original estimate) e N = 451 acute psychiatric inpatients (Holcomb, Adams & Ponder, 1983) f N = 437 short-stay psychiatric inpatients (Hafkenscheid, 1993) g N = 484 general population sample (Argentina) (Bonicatto, et al., 1997) h N = 32 for paper-pencil (Pap-Pen) and computer (matched psychosomatic outpatients),and 1000 for reference (Ref) (Schmitz, et al ., 1999) i N = 2425 Psychosomatic outpatients (P/S), 447 primary care sample (PC), 1006 reference (Ref) (Germany); (Schmitz, et al., 2000)

PAGE 152

133 Chart Abstraction Tool The chart abstraction was used to va lidate information provided through the demographic/history taking form and as a check for missing data. Data the abstraction tool collected included: age, education, current employment, past employment, current services, past services, medications, trauma (child and adult), primary psychiatric diagnosis, secondary psychiatric diagnosis, substance use/abuse dia gnosis, age of onset, hospitalizations in last year, hospitalizations lifetime, and familial history of mental illness. Domain 2: Social Factors that Influence Recovery Domain 2 addresses several social constr ucts with potential re levance to recovery. The constructs include social support, co mmunity connectedness, trust, empowerment, stigma and discrimination. These constructs are theoretically base d and were introduced in Chapter 2. There are three instruments that operationalize these constructs. These include: 1. Support and Community Connectedness Survey (SCCS), which provides data on social support, community connectedness and trust 2. Empowerment Survey (ES) 3. Internalized Stigma of Me ntal Illness Scale (ISMI) Support and Community Connectedness Survey (SCCS). The SCCS was designed specifically for this study. Full details of item selection and development are included in Appendix B. Two rounds of data collection were completed. The first round comprised 125 i ndividuals who took the original survey.

PAGE 153

134 Each question was accompanied with a blank area to comment on the question, make suggestions for wording, or a nything else the participant fe lt was needed. After updating the survey, 350 different individuals complete d the survey and a factor analysis was completed. The survey was designed to a ssess six different ar eas of support and community connection: support from an in timate partner, support from family, support from friends, support from providers, community connection and level of general trust. Thus, a six factor model was forced and the items were assessed for theoretical continuity. The factor analys is of the original 80-item i nventory reduced the items to a six-factor 34 total item surv ey that measure clusters of social support (e.g., family, friends), connection to the community and a related scale for general trust in others, consistent with item design and sub-scale prediction. Approximately 38% of the variance was explained via the six-factor, 34 item surve y. The six-factors of the final instrument and their factor loadings are provided in Table 7. In addition, one-factor, five-factor and se ven-factor models were generated to evaluate changes in factor lo adings. A one-factor model was forced to assess whether the instrument was measuring a single, univers al, construct. The one-factor solution explained 11% of the variance with 16 items loading on the single factor. This single factor encompassed questions from fam ily, significant other, and friend support questions, capturing a general support construct. Five a nd seven factor models were generated after the six-factor model to explor e changes in the factor structure and to see whether any of the theorized six-factors combined or separated. Items remained consistent in the fiveand seven-factor models with the ex ception of the three questions addressing trust in government, which broke out as its own factor in the seven-factor

PAGE 154

135 Table 7 Factor Structure for Social Support and Community Connectedness Survey Factor Loading Factor 1: Support Via Intimate Partner or Best Friend My partner (or best friend) and I support each other equally .65 My intimate partner (or best friend) helps me in many ways .77 I can always turn to my partner (or best friend) for advice if I am confused .79 My partner (or best friend) praises me and cheers for me when I accomplish something .83 I wish my partner (or best friend) would try harder not to hurt my feelings .49 Helping me feel good is what my partner or best friend does best .76 When I do something wrong, my partner (or best friend) points it out to me .77 When my back is in a corner, I can count on my partner (or best friend) to support me .83 My friends spend time with my intimate partner (or best friend) and I .64 My partner (or best friend) gives me things or helps me do things to make my life easier .77 Factor 2: Support Via Family I have a close relationship with my family .62 My family meets many of my needs .66 When I am sad or feeling blue, I can always turn to my family .80 If I need money or help with a bill, my family almost always gives it to me .66 If I need help fixing or making something, my family helps me .83 If I need to know something my family usually has the answer .75 Factor 3: Support Via Provider My provider helps me with my emotional stability .60 If I am making a mistake, my provider will always point it out .75 Sometimes I feel that my provider does not take the time to hear me .49 My friends know who my provider is .46 If I need to know something, I ask my provider .71 Factor 4: Support Via Friends My friends help me feel good about myself .62 If I make a mistake, my friends point it out so I will correct it .60 Most of my friends help me in whatever way I need .59 My closest relationships usually last for two years or more .39 Factor 5: Community Involvement or Connectedness I volunteer my time to organizations when I can .57 The more I give to the community, the more I want to .56 I keep myself informed on community issues .64 I am one of the first to contribute to community projects or concerns .73 I feel I give back to the community for what I take .64 Factor 6: Trust in Motivation of Others There are only a few people I completely trust .47 People are trustworthy, but society is not .40 Most people will do the wrong thing if they know they will not be caught .74 Most of the time people are looking out for themselves, not trying to be helpful to others .58

PAGE 155

136 model. The government questions were not of sufficient strength in the six-factor model to be included. Further, the wording and c ontent of the questions were repetitive, calling into question whether they represented an underlying concept of importance or reflected bias in question development. Reliability was assessed using both test -retest correlations with a 6-8 week interval between administrations and Cronbach alpha for internal consistency. The first 150 participants to complete the test were invited to retest, and 100 participants completed the second administration. The test-r etest reliability usi ng the Pearson Product Moment coefficient for the total test was 0.72 (p .01). Chronbach alpha for the test was 0.84. Further, total score on the total score a nd the sub-scales were compared for age, gender, and diagnosis with no difference detected, suggesting that th e instrument does not favor any specific clinical group that provided data. The six scales with their eigenvalues (E), test-retest (T) and Alpha coefficients ( ) are: (1) Support via intimate partner or best friend : This scale measures support from the most intimate non-blood related relationship of the individual and has items that address emotional, informational and inst rumental support, expectation of support and dissupport (Malone, 1988). Individuals wi th mental illness are less likely to have an intimate partner but often endorse one or more best friends (E = 7.86; T = 0.60 (p .01); = 0.74). (2) Support via family : Like the previous scale, this scale measures emotional, instrumental and informational support. Individuals with mental illness tend

PAGE 156

137 toward limited social networks with family members playing a larger role at more points in their lives (Wals h, 1994) (E = 3.78; T = 0.80 (p .01); = 0.89). (3) Support via provider : This scale investigates the support from the providers that the participant has relationships with. Th is is unique to this instrument and directly measures emotional, informa tional and dissupport from the provider to the consumer. Provider support is a substi tute for lack of support from other areas for some individuals with mental illness and has been conspicuously absent from the empirical literature but relatively consistent in qualitative and narrative research (E = 2.25; T = 0.57; (p .01); = 0.72). (4) Support via friends : The final support scale measures support from friends other than the best friend/intimate partner. Emotional and instrumental support is represented in specific qu estions (e.g., money when needed or having someone to talk to when needed) and one question that targets longevity of relationships (e.g., do most relationships last two or more years). Many consumers become limited in friendships as they are forced to de al with frightening changes, disturbing symptoms and stigmatization (Rizzo, 2002). For many of them, their friendships are often with other consumers. Howeve r, this remains a consistent area of support for mentally ill individuals (E = 1.95; T = 0.58 (p .01); = 0.73). (5) Community involvemen t or connectedness : This scale measures the degree that an individual feels they c ontribute to their community and are informed about their community. Support is given as well as received and this scale is a measure of how much the person feels they give back in a more generalized fashion (E = 2.58; T = 0.41 (p .05); .58). Test-retest and coefficient alpha may be

PAGE 157

138 depressed compared to the support scal es because of social desirability preferences. (6) Trust in motivation of others : This scale investigat es the degree that the participant feels others are motivated from altruism versus selfishness. Individuals with mental illness often have to hide their illness from others, especially others who can levy consequences, e.g. employers. Secrecy can be projected outward as a sense of perv asive distrust (E = 1.57; T = 0.45 (p .01); .60). Lack of trust has been identified as a barrier to recovery (Fisher, 2003). Empowerment Scale (ES). The ES is a scale designed with the a ssistance of consumers of mental health services to measure the construct of empo werment (Rogers et al., 1997). The survey consists of 28 items that comprise the tota l empowerment score and five sub-scales. Factor loadings and item listing for each subscale is provided in Table 8. A principal components analysis and oblique rotation extr acted the five factor s, accounting for 54% of the variance (Rogers, Chamberlain, Ellison, & Crean, 1997). Internal consistency of the empowerment scale was .86. A replic ation study using principal components analysis, though with an or thogonal (varimax) rotation, conf irmed the factor structure detected by Rogers, et al (1997) as well as a similar level of internal consistency ( = .85) (Wowra & McCarter, 1999). For each item, a four-point scale (str ongly agree, agree, disagree, strongly disagree) is used to endorse what the participant feels is the most relevant level of response for each item. The test has a high degree of internal consistency ( = 0.86). Construct validity was es tablished by comparing empowerment scores to demographic variab les, number of community activities that the participant

PAGE 158

139 Table 8 Empowerment Scale Factor Loadings Factor Loading Factor 1: Self-esteem and Self-Efficacy I generally accomplish what I set out to do .79 I have a positive attitude about myself .74 When I make plans, I am almost certain to make them work .72 I am usually confident about the decisions I make .70 I am often able to overcome barriers .56 I feel I am a person of worth, at least on an equal basis with others .47 I see myself as a capable person .46 I am able to do things as well as most other people .41 I feel I have a number of good qualities .41 Factor 2: Power-Powerlessness I feel powerless most of the time .69 Making waves never gets you anywhere .66 You cant fight city hall .66 When I am unsure about something, I usually go along with the group .66 Experts are in the best position to decide what people should do or learn .63 Most of the misfortunes in my life were due to bad luck .62 Usually, I feel alone .60 People have no right to get angry ju st because they dont like something .43 Factor 3: Community Activism and Autonomy People have a right to make their own decisions, even if the are bad ones .68 People should try to live their lives the way they want to .64 People working together can have an effect on the community .62 People have more power if they join together as a group .53 Working with others in my community can help to change things for the better .52 Very often a problem can be solved by taking action .41 Factor 4: Optimism and Control Over the Future People are limited only by what they think possible .78 I can pretty much determine what will happen in my life .62 I am generally optimistic about the future .58 Very often a problem can be solved by taking action .42 Factor 5: Righteous Anger Getting angry about something is often the first step toward changing it .73 People have no right to get angry just because they dont like something .52 Getting angry about something never helps .48 Making waves never gets you anywhere .40

PAGE 159

140 engages in, level of involvement with trad itional mental health services, level of employment (number of hours and pay status), quality of life, social support, and selfesteem. Addressing these one at a time, demographic variab les were not significantly related to empowerment, suggesting that there is no specific gender, race/ethnicity or age bias inherent in the instrument. The numb er of community activities had a small but significant relationship to empowe rment. This was hypothesized a priori to be correlated with empowerment and was validated. An inverse relationship was found between traditional psychiatric services and empowerme nt. Empowerment was slightly elevated for individuals who were employed compared to individuals who were not, but this did not reach significance between any employment groups. Quality of life, social support, and self-esteem were all signifi cantly related to empowerment. A separate analysis was completed to s ee if the ES could be reduced to the two factors of selfand community-orientations to empowerment. The scale did successfully yield two consistent factors that could be in terpreted as selfand community oriented. However, the authors of the study did not pr ovide any indication if all 28 items were retained in this analysis or which items fit the categories (Corrigan, Faber, Rashid, & Leary, 1999). The sub-scales of the ES and thei r corresponding Eigenvalue (E) include: (1) Self-esteem and self-efficacy : This scale has items th at measure feelings of worth, capabilities to accomplish goals, per ceived ability to overcome obstacles and perseverance (E=6.85). (2) Power and powerlessness: Items assessing locus of control, assertiveness, feelings of being controlled or small in comparison to organizations and society,

PAGE 160

141 and willingness to voice opi nions counter to the gr oup comprise this scale (E=3.48). (3) Community activism and autonomy : This scale measures level of belief in group activities and compares the ability to effect change in group versus individual action. This scale also measures the de gree an individual be lieves that a person should have the right of self-determinati on regardless of negative consequences (E=2.13). (4) Optimism and control over the future : This scales items directly measure the degree an individual feels he or she controls his or her own destiny and the degree that this destiny is e xpected to be positive (E=1.50). (5) Righteous anger: Empowering actions are sometim es stimulated by feelings of intense and justifiable anger. This scale measures this aspect of empowerment targeting motivationa l energy (E=1.12). Internalized Stigma of Me ntal Illness Scale (ISMI) The ISMI is a recently developed 29-item instrument that assesses psychological and subjective affects of stigma (Ritsher, Otlingam, & Grajales, 2003). Internalized stigma is the shame, devaluation, withdrawal and secrecy that result from internalizing negative stereotypes (Corriga n, 1998). This scale was developed and its psychometric properties validated on participants similar to the SMI and OP samples proposed in this study. Items were grouped a priori into the five theoretical areas, generating five subscales and forcing a five-factor model. Fifty-five items we re reduced to 29-items through factor analysis (Table 9). Items with lo w item-total correlations (< 0.40) were dropped from consideration. The authors of the inst rument report good reliability and validity

PAGE 161

142 Table 9 Internalized Stigma of Ment al Illness Factor Loadings Factor Loading Factor 1: Alienation I feel out of place in the world because I have a mental illness .54 Having a mental illness has spoiled my life .67 People without mental illness co uld not possibly understand me .39 I am embarrassed or ashamed that I have a mental illness .56 I am disappointed in myself for having a mental illness .85 I feel inferior to others who dont have a mental illness .67 Factor 2: Stereotype Endorsement Stereotypes about the ment ally ill apply to me .36 People can tell that I have a me ntal illness by the way I look .36 Mentally ill people tend to be violent .68 Because I have a mental illness, I need ot hers to make most decisions for me .55 People with mental illness cannot live a good, rewarding life .54 Mentally ill people s houldnt get married .30 I cant contribute anything to society because I have a mental illness .42 Factor 3: Discrimination Experience People discriminate against me because I have a mental illness .52 Others think that I cant achieve much in life because I have a mental illness .66 People ignore me or take me less seriously just because I have a mental illness .51 People often patronize me, or treat me like a child, just because I have a mental illness .65 Nobody would be interested in getting cl ose to me because I have a mental illness .41 Factor 4: Social Withdrawal I dont talk about myself much because I dont want to burden others with my mental illness .37 I dont socialize as much as I used to because my mental illness might make me look or behave weird .62 Negative stereotypes about mental illne ss keep me isolated from the normal world .39 I stay away from social s ituations in order to prot ect my family or friends from embarrassment .40 Being around people who dont have a me ntal illness makes me feel out of place or inadequate .51 I avoid getting close to people who dont have a mental illness to avoid rejection .55

PAGE 162

143 from their preliminary testing. Reliability evidence, measured via test-retest and Cronbachs alpha are provided with each sub-scale listed below. Construct validity was established by pr edicting whether items of the ISMI would, in general, positively or negatively co rrelate with items from other instruments (convergent and discriminant validity). Add itionally, items from the ISMI were factor analyzed with items from each other scale. If the relationship was expected to be positive, then the items from each instrument were expected to load on two separate factors but have a positive correlation, indica ting the measurement of distinct but related concepts. If the relationship was predicted to be negative, the items from the two surveys would be expected to have a clear divi sion and to have a negative correlation. Convergent validity was assessed via the Cent er for Epidemiological Studies-Depression scale (CES-D) and the Personal Devaluat ion and Discrimination Scale (PDDS). Discriminant validity was examined using the Rosenberg Self-Esteem Scale (RSES), the Empowerment Scale (ES), the Personal Empow erment Scale (PES), and the Recovery Expectancy Scale (RAS). As predicted the ISMI was positively associated with the CED-D (r = 0.35, p 0.01) and the PDDS (r = 0.53, p .01). The ISMI was negatively associated with the RSES (r = -0.52, p 0.01), the ES (r = -0.52, p 0.01), the PES (r = 0.34, p 0.01), and the RAS (r = -0.49, p 0.01). The scale provides a total internalization score as well as scores on the subscales of alienation, stereotype endorsement, discrimination experience, soci al withdrawal and stigma resistance. Measure of internal consistency ( ) and test-retest reliability (T) is provided for each subscale.

PAGE 163

144 (1) Alienation : Measures level of disenfranchisement, i.e. having a spoiled identity (Goffman, 1963) ( =0.79, T=0.68). (2) Stereotype endorsement : Measures how much the sample agrees with or has internalized stereotypes a bout mentally ill people ( =0.72; T=0.94). (3) Discrimination experience : Captures how the participants feel that others currently treat them ( =0.75; T=0.89). (4) Social withdrawal: This subscale gives an estimate of the degree that stigmatization causes a withdrawal from social interaction ( =0.80; T=0.89). (5) Stigma resistance: This subscale measures the ability to resist or be unaffected by internalized stigma ( =0.58; T=0.80). These five items are reverse coded for sake of data entry and subsequent analysis and thus provide a validity check. Domain 3: Service Factors that Influence Recovery This domain requires the least involvement from the participant. The participant provided information relevant to this do main via one instrument and additional information was gathered in the chart abstraction form introduced in Domain 1 above. Service Satisfaction Questionnaire This brief questionnaire was developed for this research effort. Services that the participant could be receiving at the partnering agen cy are listed. The participant indicated which services they are currently involved in, how many hours of contact per month for each service, length of time in each service, satisfaction level, and whether they feel that particular serv ice contributes to their recove ry. Together, these variables assess the service domain, Domain 3, described in Chapter 2. These questions are not

PAGE 164

145 related to form an underlying construct, thus th ey were not subjected to factor analysis or other psychometric examination. Service types (e.g., case management or assertive community treatment) versus service frequenc y are considerably more researched for their influence on recovery (Cunningham, Wolbert, Grazia no, Slocum, 2005; Mueser, Torrey, Lynde, Singer, Drake, 2003). Services in general are not consistent with the consumer recovery movement and have at times, drawn criticism for being disempowering or controlling (Spindel & Nugent, 2000). For these reasons, service satisfaction, number of services, and service contact hours were investigated in place of service type. The first indicator is the number of serv ices they currently receive (total number of services scale). The second indicator totals the number of hours per month of service (service contact scale). The third indicator is level of satisfaction with services, calculated by adding the satisfaction scores (satisfaction scal e) and dividing by the total number of services. This is a 6-point scale fr om very satisfied to very dissatisfied. The final indicator is the recovery relevancy score that totals th e number of services that the participant feels is relevant to their rec overy. As this had not been piloted, it was impossible to tell whether the total number of services would actually differ from the recovery relevant services. If identical, then the recovery relevancy indicator would be discarded. This, in fact, occurred, with the services satisfaction and recovery relevance scales highly correlated (r = .98, p .0001). Thus, the relevancy score was dropped from any further analysis. Supported housing offers a special meas urement challenge. Supported housing can include assistance via governmentally subsidized housing where the consumer is

PAGE 165

146 financially assisted in thei r independent living. The partnering agency also has housing owned by the agency that consumers are entitle d to live in for vari ous lengths of time, with varying levels of supervision. Suppor ted housing was coded as follows: 0=no supported housing, 1=government assisted housing but not agency housing, 2=agency housing with no to minimal supervision, and 3=agency housing with moderate to constant supervision. This information comp rises the housing scale. Individuals living in high levels of supervision from agency staff could claim 8-24 hour s of contact per day with staff, artificially inflating the service contact scale. The housing scale corrects for this. Initial analysis of this variable indicates that there is little ac tual variance. Nearly all individuals in the SMI a nd all of the individuals in the OP samples are living independently. Due to the lack of variance, this variable was also dropped from further analysis. Moderating and Mediating Variables. First introduced in Chapter 2, Figure 1 (p age 8) describes the model used in this research effort. The model demonstrates that recovery expectancy acts as a dependent measure when considering the affects of the individual, social and service domains (see recovery section above, specifically the Recovery Expectancy Checklist (REC)). However, recovery expectancy was also inve stigated as a mediating factor between the individual, social and services domains and recovery strategies. Finally, the moderating effect of recovery expectancy on the relati onship between the three domains and recovery strategies was investigated. Data from th e REC was used to investigate mediating and moderating affects. Question 1 from th e REC (reproduced below) was used as a

PAGE 166

147 mediating and moderating variable for eval uating research questions 4a and 4b (see analysis procedures below for details): (1) Do you think that recovery from mental illness will ever be possible for you? Figure 1 also presents severity of ment al illness as a potential moderating factor. Severity was measured as a dichotomous vari able with (1) representing the SMI sample and (2) representing the OP sample. Membersh ip in either group indicates some level of clinical symptoms and potential deficits in functioning. For many of the individuals, especially in the SMI sample, this label may mean a longterm struggle with mental illness. This struggle is accompanied by numerous experiences that remind the individual that they are mentally ill. These experiences could accrue over time and affect the relationship between indivi dual, social and service f actors and recovery strategy. Data Analysis Procedures Table 10 lists each variable by name, whet her dependent, independent or control, instrument from which the variable is derived, and level of measurement (e.g., nominal, continuous). Data Entry, Cleaning and Verification A database was constructed with appropriate delimiters to help avoid out of range data entry errors. Each form entered was followed by a quick visual check to assure that the numbers are entered correctly. In addition, a randomly selected group of particip ants (10%) were re-e ntered to check for discrepancies and to estimate total data entry error. Thus, 35 records were reentered into a separate database. Each record was then compared to the original record via SAS. Data entry mistakes were found in two records, each on one question. Thus, approximately 0.24% of data was wrongfully entered for 10% of randomly selected

PAGE 167

148 Table 10 Variable Descriptions Variable Name Instrument Measurement Level Control Variables Age Background and History/Chart Abstraction Continuous Gender Background and History/Chart Abstraction Dichotomous Income Background and History/Chart Abstraction Categorical Education Background and History/Chart Abstraction Categorical Dependent Variables Recovery Expectation Will you recover? Recovery Expectancy Checklist Dichotomous To what degree will you recover? Recovery Expectancy Checklist Continuous Will you completely recovery? Recovery Expectancy Checklist Dichotomous Will you never recover? Recovery Expectancy Checklist Dichotomous Recovery Strategy Effective illness management RAS and PVRQ* Continuous Positive Future Orientation RAS and PVRQ* Continuous Meaningfulness, Personal Control and Hope RAS and PVRQ* Continuous Recognizing Support RAS and PVRQ* Continuous Help Seeking RAS and PVRQ* Continuous Symptom Eradication RAS and PVRQ* Continuous

PAGE 168

149 Table10 (cont.) Independent Variables Domain 1: Personal, Historical, Clinical Factors Clinical Somatization symptoms Symptom Checklist 90-Revised Continuous Obsessive-compulsive symptoms Symptom Checklist 90-Revised Continuous Interpersonal sensitivity symptoms Symptom Checklist 90-Revised Continuous Depression symptoms Symptom Checklist 90-Revised Continuous Anxiety symptoms Symptom Checklist 90-Revised Continuous Hostility symptoms Symptom Checklist 90-Revised Continuous Phobic anxiety symptoms Symptom Checklist 90-Revised Continuous Paranoid ideation symptoms Symptom Checklist 90-Revised Continuous Psychoticism symptoms Symptom Checklist 90-Revised Continuous Severity of mental illness (SMI or OP sample) B ackground and History/Chart Abstraction Dichotomous Diagnosis Background and History/Chart Abstraction Categorical Substance Diagnosis Background and History/Chart Abstraction Dichotomous Hospitalized in the last 12-months Background and History/Chart Abstraction Dichotomous Hospitalizations in lifetime Background and History/Chart Abstraction Continuous Age of onset Background and History/Chart Abstraction Continuous Medication: Anti-psychotic Background and History/Chart Abstraction Dichotomous Medication: Anti-depressant Background and History/Chart Abstraction Dichotomous Medication: Anti-manic Background and History/Chart Abstraction Dichotomous Medication: Anti-anxiety Background and History/Chart Abstraction Dichotomous Medication: Other psychotropic Background and History/Chart Abstraction Dichotomous Employment Currently employed Background and History/Chart Abstraction Dichotomous Total years employed Background and History/Chart Abstraction Continuous

PAGE 169

150 Table 10 (cont.) Trauma/Abuse Child sexual abuse Background and History/Chart Abstraction Dichotomous Child physical abuse Background and History/Chart Abstraction Dichotomous Adult sexual assault Background and History/Chart Abstraction Dichotomous Adult physical assault Background and History/Chart Abstraction Dichotomous Familial Familial mental illness: Nuclear Background and History/Chart Abstraction Continuous Familial mental illness: Extended Background and History/Chart Abstraction Continuous Domain 2: Social Facto rs that Influence Recovery Stigma Alienation Internalized Stigma of Mental Illness Continuous Stereotype endorsement Internalized S tigma of Mental Illness Continuous Discrimination occurrence Internalized Stigma of Mental Illness Continuous Social withdrawal Internalized Stigma of Mental Illness Continuous Stigma resistance Internalized Stigma of Mental Illness Continuous Social Support/Community Connectedness/Trust Support via intimate partner or best friend Support and Community Connectedness Survey Continuous Support via family Support and Community Connectedness Survey Continuous Support via provider Support and Community Connectedness Survey Continuous Support via friends Support and Community Connectedness Survey Continuous Trust in motivation of others Support and Community Connectedness Survey Continuous Community involvement or connectedness Support and Community Connectedness Survey Continuous

PAGE 170

151 Table 10 (cont.) Empowerment Self-esteem and self-efficacy Empowerment Scale Continuous Power and powerlessness Empowerment Scale Continuous Community activism and autonomy Empowerment Scale Continuous Optimism and control over the future Empowerment Scale Continuous Righteous anger Empowerment Scale Continuous Domain 3: Service Fact ors that Influence Recovery Total number of services Service Satisfaction Questionnaire Continuous Service contact scale Service Satis faction Questionnaire Continuous Satisfaction scale Service Satisfaction Questionnaire Continuous Note: RAS = Recovery Assessment Scale; PVRQ = Personal Vision of Recovery Questionnaire

PAGE 171

152 records. These errors were corrected. Concern that data entry error may affect results prompted the research to select another 35 re cords (not duplicating from the first random selection) and the same process was complete d. Again, two records with one error each was located and corrected. Estimated data en try error is less than one-quarter of onepercent and should not have a heavy distortion affect on results. Also, the delimiters built into the database did not allow for accidental entry of an ove rly large or small data point. Data cleaning is a two-step proce ss of detection and correction. Several procedures were used to detect errors in th e data: frequencies, descriptive statistics, scatterplots and histograms. Frequencies help to locate the in accurate data among the entered variables. This was a quality check to make sure that the database delimiters were working accordingly. Frequency investigation of all questions and variables indicated no out of range vari ables on any variable. Crosstabulations (frequencies of the variable x variable type) help to learn the data and prepare for future analyses. Histograms and scatterplots are the easiest ways to detect outliers and errors in distribution. Descriptive stat istics via Univariate proce dures in SAS provide means, medians, range and standard deviations. This information is useful if, for instance, the standard deviation is found to be higher than the mean this indicates some extreme values have somehow been entered that may or may not be appropriate. For certain variables (e.g., total number of hospitalizations), this proved to be true. However, chart abstractions verified that these large data points were, in fact, accurate. One or two individuals with many hospita lizations increased the mean and standard deviation. Similar to frequencies, Univariate proce dure were run on each variable as well to continue assessing and learning the data. Finally, depending on the distribution of a

PAGE 172

153 particular variable, one or even many variable s may be standardized if it appears that the distribution is having an inflat ed affect on an analysis (e.g., conversion to standardized tscores, z-scores or LOG scores). This wa s assessed via skewness and kurtosis (provided in the Univariate procedure) for each variable. A risk for any analysis is too much missing data. Befo re data collection, treatment of missing values was stated to depend on the importance of the variable and the amount missing. Large amounts of missing data may have forced the removal of that variable from analysis. Sm all levels of missing data w ould likely be corrected via imputation of means or other appropriate values. However, a strength of this study is that there is virtually no missing data nega ting need for further consideration. Factor Analysis : In the discussion on dependent variable measures and recovery, the reader will recall that two separate surv eys provide responses for possible recovery strategies, the Personal Vision of Recovery Questionnaire (PVRQ) and the Recovery Assessment Scale (RAS). Both of these surv eys were developed to capture the construct of recovery and were developed in simila r ways. Each survey was developed in cooperation with individuals afflicted with mental illness, each are on a five-point identical scale, they had sim ilar test-retest time intervals and used similar methods for validation. However, because the recovery construct is still only beginning to be investigated, the questions developed and the scales derived from various factor analytic procedures share some similarities but are still markedly different. Further, as noted before, the process of following principal com ponents analysis with confirmatory factor analysis was questionable when developing th e RAS, necessitating the use of the 41-item versus 24-item survey. Finally, how the recove ry construct is operati onalized is different

PAGE 173

154 between the two surveys. The RAS appears to examine more internal processes and does not reference recovery in the questions. The PVRQ, in contrast, references recovery in each question and appears to address more interpersonal processes. Potential recovery strategies are inves tigated by combining the items from the RAS and PVRQ for factor analysis. A factor analysis requires a minimum sample of 100 participants or, better, five tim es the number of participants per question (Hatcher, 1994). There are a total of 65 questions between th e two surveys with an ideal number of participants at 325. Noted ear lier, this study requires a minimum of 350 participants to power the linear and logistic regression com ponents, providing sufficient numbers for the factor analysis. The a priori assumption of factor analysis is that any question may be associated with any factor. However, the number of f actors predicted was informed by theory and the factor structures presented in the devel opment of the PVRQ and RAS. Thus, a fivefactor or six-factor model was predicted to provide the most comprehensive explanation of the recovery construct. A promax ro tation was chosen to capture shared and unique variance. Eigenvalues above 1.00 and scree plot s were used to evalua te the findings each time through the procedure. When forced into a five-factor model the fifth factor was difficult to interpret with some factors posi tive and some negative. A six-factor model better fit the theory and also removed interp retive obstacles. A brie f discussion of each factor (dependent variable) is offered belo w and the factor loadings and items for each variable are summarized in Tabl e 11. Variables were eliminated if they did not load on a factor at an initial stringent .40 level. Additionally, any variable that loaded on more than one factor at the.40 level or above would be scrutinized. However, no item met the

PAGE 174

155 Table 11 Factor Loadings and Items for Dependent Variables Factor Numbers and Loadings Item # Item Description 1 2 3 4 5 6 Factor 1: Effective Illness Management 7 I understand how to control the symptoms of my mental illness 0.71 ns ns ns ns ns 8 I can handle it if I get sick again 0.75 ns ns ns ns ns 9 I can identify what triggers the symptoms of my mental illness 0.66 ns ns ns ns ns 10 I can help myself become better 0.40 ns ns ns ns ns 11 Fear doesnt stop me from living the way I want to 0.58 ns ns ns ns ns 13 There are things that I can do that help me deal with unwanted symptoms 0.49 ns ns ns ns ns 14 I can handle what happens in my life 0.62 ns ns ns ns ns 18 Although my symptoms may get worse, I know I can handle it 0.85 ns ns ns ns ns 27 Coping with my mental illness is no longer the main focus of my life 0.48 ns ns ns ns ns 28 My symptoms interfere less and less with my life 0.65 ns ns ns ns ns 29 My symptoms seem to be a problem for shorter periods of time each time they occur 0.45 ns ns ns ns ns 34 I know what helps me get better 0.39 ns ns ns ns ns 36 I can handle stress 0.50 ns ns ns ns ns Factor 2: Positive Future Orientation 1 I have a desire to succeed ns 0.67 ns ns ns ns 3 I have goals in life that I want to reach ns 0.82 ns ns ns ns 4 I believe I can meet my current personal goals ns 0.48 ns ns ns ns 5 I have a purpose in life ns 0.66 ns ns ns ns 12 I know that there are mental health services that do help me ns 0.40 ns ns ns ns 15 I like myself ns 0.47 ns ns ns ns 16 If people really knew me, they would like me ns 0.46 ns ns ns ns 20 I have an idea of who I want to become ns 0.69 ns ns ns ns 22 Something good will eventually happen ns 0.49 ns ns ns ns 24 Im hopeful about my future ns 0.65 ns ns ns ns 25 I continue to have new interests ns 0.50 ns ns ns ns 26 It is important to have fun ns 0.46 ns ns ns ns

PAGE 175

156 Table 11(cont.) Factor 3: Meaningfulness, Personal Control and Hope 40 It is important to have a variety of friends ns ns 0.38 ns ns ns 41 It is important to have healthy habits ns ns 0.35 ns ns ns 45 Recovery means becoming satisfied with my life ns ns 0.53 ns ns ns 46 Hope is important for my recovery ns ns 0.54 ns ns ns 47 Being diagnosed correctly is necessary for my recovery ns ns 0.66 ns ns ns 49 Sticking up for clients rights is a part of my recovery ns ns 0.52 ns ns ns 50 Having something meaningful to do is important for my recovery ns ns 0.74 ns ns ns 51 Helping others is part of my recovery ns ns 0.53 ns ns ns 52 Asking for help is a part of my recovery ns ns 0.41 ns ns ns 54 Recovery means my symptoms will be easier to control ns ns 0.53 ns ns ns 57 Recovery means getting more control of my life ns ns 0.57 ns ns ns 64 Recovery involves finding new meaning in my life ns ns 0.41 ns ns ns 65 Support from mental health professionals is important for my recovery ns ns 0.45 ns ns ns Factor 4: Recognition of Support 6 Even when I dont care about myself, other people do ns ns ns 0.65 ns ns 37 I have people I can count on ns ns ns 0.72 ns ns 39 Even when I dont believe in myself, other people do ns ns ns 0.80 ns ns 7 Family support is important for my recovery ns ns ns 0.46 a ns ns Factor 5: Help Seeking 30 I know when to ask for help ns ns ns ns 0.64 ns 31 I am willing to ask for help ns ns ns ns 0.84 ns 32 I ask for help when I need it ns ns ns ns 0.83 ns Factor 6: Symptom Eradication b 14 Recovery means I will be free of symptoms ns ns ns ns ns 0.58 20 Recovery means I will not be mentally ill anymore ns ns ns ns ns 0.65 Note: ns = not significant; a = though significant, this item was dropped due to a substantial decrease in Alpha; b = only two items loaded on this factor and will not be treated as a composite variable. Item number 14 is pa rticularly relevant in the recovery and was treated as a sing le item dependent variable

PAGE 176

157 multi-factor loading criteria. Lastly, a factor would be retained if it had a minimum of three or more items. It must be noted that the factors spli t almost perfectly across surveys, meaning items from the PVRQ and RAS were not combined within factors, raising concern that a method effect versus substantive variance wa s being detected. There is further concern that the lack of common variance (~42.0% of cumulative va riance explained across the six factors) calls into question whether a single construct is present. Additional factor analysis was completed stratifying by instrume nt and sample to c onfirm the six-factor model. There was no substantial difference in the amount of variance explained or items loading on the same or different factors. Furt her, because subset factor analysis violated the necessary n for the proce dure, Bartletts test of sphe ricity was used to verify stability of the procedure. Finally, items w ith a loading of 0.35 or higher were included in additional factor analyses and assessed whether they affected inte rnal consistency and were theoretically consistent. If they contributed to in ternal consistency and were theoretically constant they were retained in the factor. The five strategies consisting of multiple items were calculated by adding the number from each item and dividing by the tota l number of items in that construct. The six factors are 1. Effective illness management : Thirteen questions with a factor loading range of 0.39 to 0.85 encompass this strategy. This factor captures the most variance (22.26%) of the recovery construct and has an ei genvalue of 14.47. Cronbachs Alpha is 0.88. Handling stress, predicting symptom onset, co ntrolling symptoms, and the perception of impairment are included in this strategy.

PAGE 177

158 2. Positive future orientation : This strategy consists of 12 questions with a factor loading range of 0.40 to 0.82 and a Cronbachs Alpha of 0.88. The eigenvalue is 4.01 and the percent of variance explained is 6.17. This strategy comprises looking forward, a positive future outlook, forming goals, and having meaning in life. 3. Meaningfulness, personal control, and hope: This strategy is harder to define due to a wider range of concepts contained w ithin the questions. Hope, a key concept of recovery, satisfaction in life, and being dia gnosed accurately comprise this factor. An accurate diagnosis leads to be tter treatment and greater sy mptom relief which, in turn, may increase hope and lead to general life satisfaction. Key aspects of clinical intervention are present in this factor as well, including accurate diagnosis and support from professional providers. The factor load ing for this strategy has a range of 0.35 to 0.74. The eigenvalue is 2.47 and the variance ex plained is 3.81%. Cronbachs Alpha is 0.83. 4. Recognizing support : Four questions grouped together to delineate this aspect of recovery. Factor loadi ngs ranged from .46 to .80 and the Cronbachs Alpha is 0.76. However, when observing the item individual contribution to internal consistency, one question ( PVRQ question 7, Family support is important for my recovery) reduced Cronbachs Alpha by approximately 0.06 a nd was only tangentially theoretically connected. The decision was made to reduce th is factor from four to three items raising Cronbachs Alpha to 0.82. The eigenvalue is 2.33 and this factor explains 3.58% of the variance. This strategy descri bes the ability to recognize that support is available and that others hold a positive opin ion of the participant.

PAGE 178

159 5. Help seeking : The three questions in this strategy have an eigenvalue of 2.07 and explains 3.12% of the variance. The fact or loadings range from 0.64 to 0.84 with a Cronbachs of .85. The strategy targets seeki ng help and the willingness to ask for help. 6. Symptom Eradication : The final factor violated the criteria of a minimum of three items per factor and wa s initially considered for e limination. One of the two questions (PVRQ question 20, Recovery means I will not be mentally ill anymore) was similar to the Recovery Expectancy Check list question 1 (Do you th ink that recovery from mental illness will ever be possible for you?), reinforcing the decision to eliminate the factor. However, the second question (P VRQ question 14, Recovery means I will be free of symptoms) addresses a controversial aspe ct of the recovery literature, specifically the degree that a clinical focus should be maintained when targeting recovery. The decision was made to keep this question as the sixth strategy t hough it encompasses only one item. Univariate Analyses Univariate analyses were used to explore the data, check for inconsistencies, look at distri butions, and make decisions. Each question for each survey was reviewed. All variables were checked for violations of distribution and for outliers. Bivariate Analyses A priori analyses were completed at the .05 level and familywise error rate were not adjusted for in a priori analyses, though a Bonferonni correction is used for post hoc analyses, including moderating mode l evaluation (research questions 4b, and 5. This rule applies to all bivariate and multivariate analyses. Bivariate analyses were used to explore the data and to understand the relationships between depe ndent-dependent, dependent-i ndependent, and independentindependent variables. Both univariate and bivariate analys es helped to better understand

PAGE 179

160 the data and to draw an initial portrait of an individual engaged in recovery and one who is not. The choice of bivariate statistic depe nded on the measurement level of the variable. Chi-square assessed the relationship between two nominal (categorical) level variables. The independent sample t-te st was used for a two-level categorical independent variable and a continuous de pendent variable. The one-way ANOVA was used for three or more level categorical independent variables and a continuous dependent variable. Finally, many bivariate analyses were completed with the Pearson Product Moment Correlation. Correlation analysis is also helpful in the first steps of identifying multicollinearity. Thus, a ppropriate analyses (chi-square, ANOVA, independent sample t-tests) were conducte d on all dependent and independent variables to assure that there is sufficient autonomy within the dependent set and the correlate set of variables, each according to its measurement attributes. I turn next to research question 1 (RQ1) as all of the hypotheses in RQ1 can be answered via bivariate analysis. Research Question 1: To what degree does i llness severity influence beliefs in recovery for the individual? The next step of the analysis addresses research question (RQ1) one and related hypotheses. Each hypothesis is reproduced fo llowed by where the da ta is obtained and the type of analysis used. Hyp1.1: The OP sample will endorse that they will recover from mental illness to a greater extent than the SMI sample The data for this question are taken from question one of the Recovery Efficacy Check list (REC), Do you think that recovery

PAGE 180

161 from mental illness will ever be possible for you? The question is dichotomous (yes/no), requiring a chi-square analysis. Hyp1.2: The OP sample will endorse complete recovery as possible more than the SMI sample Question #2 from the REC provide s the data (To what degree do you think you will eventually recove r?) and a chi-square analysis was used to check for hypothesized associations. Hyp1.3: There will be no significant difference between samples in whether they endorse that there is no degree of recovery expected The data for this question is also taken from question two of the REC. The que stion is dichotomous (yes/no), requiring a chi-square analysis. Multivariate Analyses Multivariate analysis included logistic regression, ordinary least squares (OLS) regression, and pr obit analysis. These are discussed in the order presented to outline the multivariate analysis plan and each step of the analysis is linked directly to the relevant research questions and hypotheses. The goal of an analysis is to find the most parsimonious model where the greatest level of variance is explained with the least amount of variab les. Before outlining the analysis plan, however, a brief discussion of influential da ta points is required. Influential data points are outliers with high leverage that influence regression coefficients through changing the slope or in tercept of the regre ssion line. Diagnostic statistics are used to ascertain the extent the predicted values of the dependent variable or the parameter estimates change when an obser vation is eliminated from the analysis. Three diagnostic statistics were used to detect influential observations. Hat values measure leverage in regression models by summarizing how far the independent

PAGE 181

162 variables fall from their mean (Agresti and Finlay, 1997). Since the sample is greater than 200 (n > 200), Hat values greater than two times the mean were evaluated as possibly too influential. Cooks D, a summary index of an individual observations influence on regression coefficients, is the s econd diagnostic statistic. Cooks D values above 2.0 were flagged for review of influence on the model. Finally, Studentized Residuals, an index for detecting outliers, were also evaluated, utiliz ing the same cutpoint of above 2.0 as Cooks D. If any observation violated criteria for influence on two of three of the diagnostic tests they were elim inated before evaluating main effect model results (Agresti and Finlay, 1997). In addition to the diagnostics just di scussed, multicollinearity was assessed. Multicollinearity is an unacceptably larg e degree of correlation among independent variables. Substantial multicollinearity makes conclusions regarding the degree of association between the corre lated independent variables and the dependent variable potentially incorrect. Muliticollinearity may re sult in unstable regression coefficients and standard errors that make it difficult to de tect the degree of association between each independent and dependent variable (Zar, 1999) Indicators of multicollinearity include regression coefficients that do not appear to be sensible or have the wrong sign. Thus, using SAS, tolerance was assessed for th e predictor variables when examining for multicollinearity. Tolerance is an estimate of the proportion of variance of predictor variables not overlapping w ith other individual variab les. Following statistical convention, the tolerance absolute minimum is 0.2, meaning that a minimum level of independence required is 20%, or in other words, that 20% of the variable does not overlap.

PAGE 182

163 Research Question 2: Are individual, social, or service factors associated with recovery expectancy? Logistic regression was used when the crit erion variable is dichotomous, in this case with the Recovery Expectancy Checklis t (REC) providing the data for comparison of the two samples. Models were fit with the dependent variable being question one from the REC that address if recovery is expected for each participant. For research question 2.1, whether individua l factors are associated with recovery expectancy, the following progression was used to enter variables into logistic regression models. First, control variab les (gender, age, income, edu cation) were entered. Control variables remained the same for all multivariate models in this analysis. After the control variables, the independent vari ables were entered in the foll owing order: Subscales from the SCL-90-R, other clinical variables in a block (substa nce diagnosis, hospitalization variables, age of onset, medication variables); employment variables; family history variables; and finally abuse/assault variab les. Finally, diagnoses were added one diagnosis at a time (and then removed) to check the individual contribution of each diagnosis and to see if any di agnosis was significant. For e ach analysis that involved diagnosis, the following order was entere d: schizophrenia, bipolar disorder, schizoaffective disorder, major depression, a nxiety disorder, mild/moderate depression, adjustment disorder, and other. Research question 2.2 (hypotheses 2.2a-2.2q) addresses the question of whether social factors are associated with the recovery process. Again, logistic regression was used with question #1 of the REC providi ng the dichotomous dependent variable. Following entry of the control variables, subsca les from the ISMI (Internalized Stigma of

PAGE 183

164 Mental Illness) were entered in a bloc k (alienation, stereotypes, discrimination experiences, social withdrawal, and stigma resistance). This was followed by the subscales from the SCCS (Support and Co mmunity Connectedness Survey) (support via intimate partner or best friend, support via family members, support via mental health providers, support through friendship, community connectedness, and trust in others), also in a block. Finally, the block of empowerment survey (ES) variables (self-esteem and self-efficacy ratings, power-powerless ness, community activism and autonomy, optimism and control over the future, and righteous anger) we re entered. Again, entry of variables in this order directly parallels the corresponding hypotheses. Research question 2.3 (hypotheses 2.3a-2.3e) addresses the question of whether service factors are associated with the recovery process. Following entry of the control variables, the first independent variable to be entered was total number of services. This was followed, in order, by average number of contact hours per month with a provider, and, finally, satisfaction with services. Once data analysis for the three sub-rese arch questions (2.1-2.3) was completed, a main effects model was generated. First, the final model from RQ2.1 began the process of building the main effects model. Variables from RQ2.2 and then RQ2.3 were then added until all variables had been entered. Finally, each diagnosis variable was again entered one at a time. This satisfied rese arch question 2 and allowed what may be the first look at what happens wh en variables from all three domains are directed at the question of whether or not recovery is expect ed by individuals with mental illness. Once this analysis was completed, the next step required OLS (ordinary least squares) regression and ta rgets Research Question #3.

PAGE 184

165 Research Question 3: Are individual, social or service factors associated with recovery strategies? Noted previously, the dependent variable s for research question 3 are the six recovery strategy subscales generated via factor analysis. It is nave to assume that the complexity of each participants life relegate s the use of only one strategy. Indeed, a single participant may endorse none, some or a ll of the strategies as relevant to their recovery. To discover potentially relevant information more in keeping with the lived reality of the mentally ill population, it is important to consider how highly endorsed combinations of strategies can be detected and then analyzed for associations with individual, social, and service factors. Probit analysis is a technique that allows combinations of dependent variables to be investigated as a single multi-level dependent variable. Prior to the analys is it was hoped that recovery st rategies would cluster to a degree that would make probit analysis us eful. Thus, for research question 3, the dependent variables were to be of two sets: (1) the individual stra tegies generated via factor analysis; and (2) the highly endorsed combination strategies. The following approach was used to determine strategy endorsement for each participant. A cutpoint of above the median was used to state that an individual endorses a particular pathway. The median was ar bitrarily chosen after consultation with individuals who serve the menta lly ill since a review of th e literature yielded no guidance (Ellis-Lang, 2005). Each participant was c oded with a one (1) for each pathway they endorse and zero (0) if they do not. Stat istical software (SAS, v.9.1) identified combinations of pathways endorsed by the pa rticipants. Out of 66 potential combinations (which includes endorsing zero strategies above the median point), 63 of the

PAGE 185

166 combinations were endorsed. Some stra tegy combinations had only one individual meeting that combinations criterion. Twenty -eight (28) individua ls sanctioned all six strategies above the median and 68 did not endorse any strategy above the median. Again after consultation, the cutpoint was ra ised to 80% endorsement and this reduced the number of clusters of research strategies to 36. This indicates a wide variation of strategy use and also is overly cumbersome for including all combinations in a probit analysis. Thus, though theoretica lly of interest, pragmatically the analysis could not be completed. Probit analysis was discontinued for the rest of the analytic process. Analysis of research question 3.1, whethe r individual factors ar e associated with recovery strategies, (hypotheses 3.1a-3.1af) utilized the following progression to enter variables into OLS regression models. First, control variables (gender, age, income, education) were entered. Af ter the control variables, th e independent variables were entered in the following order with carefu l tracking of changes in betas (OLS) and significance for each addition. Subscales fr om the SCL-90-R was added as the first independent variable. After symptoms, the vari ables were entered in the following order: substance diagnosis, hospitalization variable s, age of onset, familial mental illness variables, employment variables, medica tion variables, and fi nally abuse/assault variables. As before, diagnoses were added last, one at a time. Research question 3.2 (hypotheses 3.2a-3.2q) addresses the question of whether social factors are associated with the recovery process. Following entry of the control variables, subscales from the ISMI (Internalized Stigma of Mental Illness) were be entered (alienation, stereotypes, discrimination experiences, social withdrawal, and stigma resistance). This was followed by the subscales from the SCCS (Support and

PAGE 186

167 Community Connectedness Survey) (support via intimate partner or best friend, support via family members, support via mental he alth providers, support through friendship, community connectedness, and trust in others). Finally, the ES subscales (Empowerment Survey) (self-esteem and self-efficacy rati ngs, power-powerlessness, community activism and autonomy, optimism and control over the fu ture, and righteous anger) were entered. Research question 3.3 (hypotheses 3.3a-3.3e) addresses the question of whether service factors are associated with the recovery process. Following entry of the control variables, the first independent variable to be entered was total number of services. This was followed, in order, by average number of contact hours per month with a provider, and satisfaction with services. Once data analysis for the three sub-rese arch questions (3.1-3.3) was completed, a main effects model was generated. First, the final model from RQ3.1 was employed to begin the process of building the main effects model. Variables were added to this model by following the progression from RQ3.2 and then RQ3.3 until all variables were entered. This satisfies research question 3 and displa ys what happens when variables from all three domains are directed at each research strategy individually and what factors are associated with these choices. Following this analysis, the next step required OLS regression once again to help detect whether recovery expectances mediate or moderate the effects of correlates for each strate gy, also targeting research question 3. Research Question 4a: Does the expectat ion of recovery mediate the relationship between individual, social and servic e factors and recovery strategies? A mediator is a variable that explains the relation between an independent and dependent variable (Baron & Kenny, 1986; Hombeck, 1997). Thus, the mediator is the

PAGE 187

168 mechanism through which the independent variab le influences the dependent variable by reducing the variance explained by an original (already presen t) variable in magnitude and/or significance (Baron & Kenny, 1986). However, in a crosssectional study, the best that can be said is that the independent variable infl uences the mediator, which, in turn, influences the dependent variable, or outcome (Holmbeck, 1997). Furthermore, the effect of the mediating variable may wholly, or in part, influence the relationship between the independent and dependent variable. If the addition of a new variable reduces the magnitude of the parameter estimate for the original variable and changes the significance level to non-significant, than the new variable fully mediates the effect of the original variable. If the magnitude of the original vari able is reduced but the original variable remains significant, then the new variable is said to partially mediate the original variable. A potentially important mediator operating between individual, social, and service factors and recovery is recovery expectancy. The reader will recall that analysis of data to answer research question two utilized reco very expectancy as the dependent variable. Here, recovery expectancy is placed on th e path between the three domains (individual, social, and service) and the individual and combined recovery strategies. The first questions from the Recovery Expectancy Check list (REC) was used in this capacity and is reproduced below. 1. Do you think that recovery from ment al illness will ever be possible for you? Mediating relationships for each question were assessed separately using the following procedure. First, changes in associations were assessed for individual/clinical factors and recovery strategies by adding the results from REC question 1, assessing whether or not

PAGE 188

169 recovery is possible, to the individual models developed to answer RQ3.1 (Are individual factors associated with recovery strategies?) Variables that were significant in those models were monitored for changes in magnit ude and significance to assess the degree of influence of recovery expectancy (satisfying hypothesis 4a.1). The same procedure was used to assess th e association of recovery expectancy on the relationship between social factors and recovery strate gies (RQ3.2). By assessing whether recovery is expected on the relationship between social factors and recovery strategies, information to evaluate hypothesis 4a.2 was generated. Similarly, the relationship between service factors and recovery strategies (individual and combined) can be mediated by whether recovery is e xpected (RQ3.3). Thus utilizing the same procedure a third time provide s the information to assess hypotheses 4a.3. Finally, to directly address RQ4a, assessing the medi ating influence of REC question 1 on the relationship between individual, social, and se rvice factors and rec overy strategies, the same process was used to assess the fu ll model developed to satisfy RQ3. Research Question 4b: Does the expectat ion of recovery moderate the relationship between individual, social and servic e factors and recovery strategies? To this point, multivariate analyses were completed to detect relevant associations between the independent vari ables and recovery strategi es (main effects), and to investigate the mediating effect of recovery expectancy. The current research question and the next explore the moderating influence of recovery expectancy and the label of illness severity. A moderating variable effects the direction or strength of a relationship between independent and dependent variable s. The nature of the impact on the dependent variable by the independent variab le varies by the value of the moderating

PAGE 189

170 variable (Baron & Kenny, 1986). Where as a mediator is additive in its effect, moderators are multiplicative or synergistic. Moderating effects were assessed for Question 1 from the REC (whether recovery is expected) for individual, social, and service affects (hypotheses 4b.1-4b.3) and th en for the full model (RQ4b). The choice of which variables to assess in a moderating relationship with recovery expectancy was made after the results of the main effect model were assessed. The independent variables significantly associated with the six recovery strategies are not uniform. The specifics of these relationships are fully detailed in chapter four. Though there was no independent va riable significantly associat ed with all the recovery strategies, there were several that were significantly associated with three or four strategies. These important va riables were selected to asse ss for moderating effects with recovery expectancy. Research Question 5: Does severity of mental illness moderate the relationship between individual, social and service fa ctors and recovery strategies? Analysis was targeted to examine the mode rating effect of inclusion in the SMI or OP samples and specific independent variables on recovery strategies The label of SMI is a proxy for a disease pattern that is exp ected to profoundly affect the functioning level and psychiatric status of an individual compar ed to those without the label. This study was designed to capture the se verity continuum by surveyi ng those with the SMI label (SMI sample) and those without (OP sample), with the understanding th at the variation in symptoms and functional capacity is expected to be great within each sample and even greater between them. This was operationalized by creation of the va riable severity of mental illness) through assigning a to all in dividuals in the SMI sample and a to

PAGE 190

171 all individuals in the OP sample. Combined interaction variables were created with specific independent variables using the same procedure briefly outlined in the discussion of research question 4b above and more fully described in chapter 4. In the psychiatric literature a label of SMI (severe mental illness) is indicative of a serious diagnostic label (e.g., schizophrenia) as well as ongo ing compromise in one or more areas of functioning, and that these indivi duals are more severe than those typically seen in general outpatient therapy (Kessl er & Zhao, 1999; Kessler, Zhao, Katz, Kouzis, Frank, Edlund, & Leaf, 1999). A potentially prob lematic assumption for this research is that the individuals were pr operly categorized as SMI and non-SMI. Confidence that categorization was accurate was enhanced th rough (1) the primary clinician of each participant certified them as SMI or non-SMI, (2) this was ve rified by the lead researcher, an experienced clinician, in most cases via direct contact with th e individual, and (3), each participants clinical chart was reviewed and verified for appropriate placement in the SMI or OP sample. The moderating affects of illness severity are assessed separately for the individual, social, and service domains ( hypotheses 5.1-5.3) and then for the combined domains (main effects model, research quest ion 5) in the same way as described for research 4b above. This concl udes the discussion of the analysis plan. However, due to the large number of multivariate analyses conducted a summary of all multivariate statistical models by resear ch question and hypotheses is provided in Table 12.

PAGE 191

172 Table 12 A Priori Multivariate Models Research Question Hypotheses Range Model # Model Label Analysis Dependent Variable Independent Variables RQ2.1 1 Control Logistic Regression Recovery Expectancy Age, gender, income, education Hyp2.1b-11 2 Symptoms Logistic Regression Recovery Expectancy Control + somatization, OCD, interpersonal sensitivity, depression, anxiety, hostility phobic anxiety, paranoia, psychoticism Hyp2.1j-1r 3 Other clinical Logistic Regression Recovery Expectancy Control + symptoms + substance dx, total hospitalizations, hospitalization in last year, age of onset, medication use (antipsychotic, anti-depressant, anti-manic, anti-anxiety, other psychotropic) Hyp2.1s-1t 4 Employment Logistic Regression Recovery Expectancy Control + symptoms + other clinical + currently employed, total years employed Hyp2.1u-1v 5 Family history Logistic Regression Recovery Expectancy Control + symptoms + other clinical + employment + nuclear family MI history, extended family MI history Hyp2.1w-1z 6 Abuse Logistic Regression Recovery Expectancy Control + symptoms + other clinical + Employment + Family history + child sexual abuse, ch ild physical abuse, adult sexual assault, adult physical assault Hyp2.1aa 7-14 Diagnosis Logistic Regression Recovery Expectancy Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis (schizophrenia, bipolar, schizoaffective, major depression, anxiety, depression, adjustment, other)* RQ2.2 Hyp2.2a-2.2e 15 Stigma Logistic Regression Recovery Expectancy Control + alienation, stereotype endorsement, discrimination occurrence, social withdr awal, stigma resistance

PAGE 192

173 Table 12 (cont.) Research Question Hypotheses Range Model # Model Label Analysis Dependent Variable Independent Variables Hyp2.2f-2.2k 16 Social support Logistic Regression Recovery Expectancy Control + stigma + support via intimate partner/best friend, support via family, support via provider, support via friends, community involvement, trust Hyp2.2l-2.2q 17 Empowerment Logistic Regression Recovery Expectancy Control + stigma + social sup port + self-esteem/self-efficacy, power/powerlessness, commun ity activism and autonomy, optimism and control, righteous anger RQ2.3 Hyp2.3a-3c 18 Current services Logistic Regression Recovery Expectancy Control + service variables (total number of services, total contact hours, average satisfaction level) RQ2 19-26 Main effects Logistic Regression Recovery Expectancy Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + current services + contact h ours + service satisfaction + diagnosis* RQ3 (RQ3.1-3.3) Hyp3.1a-1aa Hyp3.2a-2q Hyp3.3a-3c 27-52 Control Main effects OLS Regression Illness Management Same progression as models 1-26 (Full model = Control + symptoms + other clinical + employment + family history + abuse + stigma + social suppor t + empowerment + services + diagnosis) RQ3 (RQ3.1-3.3) Hyp3.1a-1aa Hyp3.2a-2q Hyp3.3a-3c 53-78 Control Main effects OLS Regression Future Orientation Same progression as models 1-26 (Full model = Control + symptoms + other clinical + employment + family history + abuse + stigma + social suppor t + empowerment + services + diagnosis*) RQ3 (RQ3.1-3.3) Hyp3.1a-1aa Hyp3.2a-2q Hyp3.3a-3c 79-104 Control Main effects OLS Regression Meaning and Hope Same progression as models 1-26 (Full model = Control + symptoms + other clinical + employment + family history + abuse + stigma + social suppor t + empowerment + services + diagnosis*)

PAGE 193

174 Table 12 (cont.) Research Question Hypotheses Range Model # Model Label Analysis Dependent Variable Independent Variables RQ3 (RQ3.1-3.3) Hyp3.1a-1aa Hyp3.2a-2q Hyp3.3a-3c 105-130 Control Main effects OLS Regression Recognizing Support Same progression as models 1-26 (Full model = Control + symptoms + other clinical + employment + family history + abuse + stigma + social suppor t + empowerment + services + diagnosis*) RQ3 (RQ3.1-3.3) Hyp3.1a-1aa Hyp3.2a-2q Hyp3.3a-3c 131-156 Control Main effects OLS Regression Help Seeking Same progression as models 1-26 (Full model = Control + symptoms + other clinical + employment + family history + abuse + stigma + social suppor t + empowerment + services + diagnosis*) RQ3 (RQ3.1-3.3) Hyp3.1a-1aa Hyp3.2a-2q Hyp3.3a-3c 157-182 Control Main effects OLS Regression Symptom Eradication Same progression as models 1-26 (Full model = Control + symptoms + other clinical + employment + family history + abuse + stigma + social suppor t + empowerment + services + diagnosis*) RQ4.a Hyp4.a.1 183-190 Mediating Domain 1 OLS Regression Illness Management Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + recovery expectation RQ4.a Hyp4.a.2 191 Mediating Domain 2 OLS Regression Illness Management Control + stigma + social sup port + empowerment + recovery expectation RQ4.a Hyp4.a.3 192 Mediating Domain 3 OLS Regression Illness Management Control + service variables + recovery expectation RQ4.a 193-200 Mediating Main Effects OLS Regression Illness Management Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + current services + contact h ours + service satisfaction + diagnosis* + recovery expectation RQ4.a Hyp4.a.1 201-208 Mediating Domain 1 OLS Regression Future Orientation Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + recovery expectation

PAGE 194

175 Table 12 (cont.) Research Question Hypotheses Range Model # Model Label Analysis Dependent Variable Independent Variables RQ4.a Hyp4.a.2 209 Mediating Domain 2 OLS Regression Future Orientation Control + stigma + social sup port + empowerment + recovery expectation RQ4.a Hyp4.a.3 210 Mediating Domain 3 OLS Regression Future Orientation Control + service variables + recovery expectation RQ4.a 211-218 Mediating Main Effects OLS Regression Future Orientation Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + recovery expectation RQ4.a Hyp4.a.1 219-226 Mediating Domain 1 OLS Regression Meaning and Hope Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + recovery expectation RQ4.a Hyp4.a.2 227 Mediating Domain 2 OLS Regression Meaning and Hope Control + stigma + social sup port + empowerment + recovery expectation RQ4.a Hyp4.a.3 228 Mediating Domain 3 OLS Regression Meaning and Hope Control + service variables + recovery expectation RQ4.a 229-236 Mediating Main Effects OLS Regression Meaning and Hope Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + recovery expectation RQ4.a Hyp4.a.1 237-244 Mediating Domain 1 OLS Regression Recognizing Support Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + recovery expectation RQ4.a Hyp4.a.2 245 Mediating Domain 2 OLS Regression Recognizing Support Control + stigma + social sup port + empowerment + recovery expectation RQ4.a Hyp4.a.3 246 Mediating Domain 3 OLS Regression Recognizing Support Control + service variables + recovery expectation

PAGE 195

176 Table 12 (cont.) Research Question Hypotheses Range Model # Model Label Analysis Dependent Variable Independent Variables RQ4.a 247-254 Mediating Main Effects OLS Regression Recognizing Support Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + recovery expectation RQ4.a Hyp4.a.1 255-262 Mediating Domain 1 OLS Regression Help Seeking Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + recovery expectation RQ4.a Hyp4.a.2 263 Mediating Domain 2 OLS Regression Help Seeking Control + stigma + social sup port + empowerment + recovery expectation RQ4.a Hyp4.a.3 264 Mediating Domain 3 OLS Regression Help Seeking Control + service variables + recovery expectation RQ4.a 265-272 Mediating Main Effects OLS Regression Help Seeking Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + recovery expectation RQ4.a Hyp4.a.1 273-280 Mediating Domain 1 OLS Regression Symptom Eradication Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + recovery expectation RQ4.a Hyp4.a.2 281 Mediating Domain 2 OLS Regression Symptom Eradication Control + stigma + social sup port + empowerment + recovery expectation RQ4.a Hyp4.a.3 282 Mediating Domain 3 OLS Regression Symptom Eradication Control + service variables + recovery expectation RQ4.a 283-290 Mediating Main Effects OLS Regression Symptom Eradication Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + recovery expectation

PAGE 196

177 Table 12 (cont.) Research Question Hypotheses Range Model # Model Label Analysis Dependent Variable Independent Variables RQ4.b Hyp4.b.1 291-298 Moderating Domain 1 OLS Regression Illness Management Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis + select variables x recovery expectation RQ4.b Hyp4.b.2 299 Moderating Domain 2 OLS Regression Illness Management Control + stigma + social sup port + empowerment + recovery expectation + select variab les x recovery expectation RQ4.b Hyp4.b.3 300 Moderating Domain 3 OLS Regression Illness Management Control + service variables + select variables x recovery expectation RQ4.b 301-308 Moderating Main Effects OLS Regression Illness Management Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x recovery expectation RQ4.b Hyp4.b.1 309-316 Moderating Domain 1 OLS Regression Future Orientation Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x recovery expectation RQ4.b Hyp4.b.2 317 Moderating Domain 2 OLS Regression Future Orientation Control + stigma + social sup port + empowerment + recovery expectation RQ4.b Hyp4.b.3 318 Moderating Domain 3 OLS Regression Future Orientation Control + service variables + select variables x recovery expectation RQ4.b 319-326 Moderating Main Effects OLS Regression Future Orientation Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x recovery expectation RQ4.b Hyp4.b.1 327-334 Moderating Domain 1 OLS Regression Meaning and Hope Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x recovery expectation

PAGE 197

178 Table 12 (cont.) Research Question Hypotheses Range Model # Model Label Analysis Dependent Variable Independent Variables RQ4.b Hyp4.b.2 335 Moderating Domain 2 OLS Regression Meaning and Hope Control + stigma + social sup port + empowerment + recovery expectation RQ4.b Hyp4.b.3 336 Moderating Domain 3 OLS Regression Meaning and Hope Control + service variables + select variables x recovery expectation RQ4.b 337-344 Moderating Main Effects OLS Regression Meaning and Hope Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x recovery expectation RQ4.b Hyp4.b.1 345-352 Moderating Domain 1 OLS Regression Recognizing Support Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x recovery expectation RQ4.b Hyp4.b.2 353 Moderating Domain 2 OLS Regression Recognizing Support Control + stigma + social sup port + empowerment + recovery expectation RQ4.b Hyp4.b.3 354 Moderating Domain 3 OLS Regression Recognizing Support Control + service variables + select variables x recovery expectation RQ4.b 355-362 Moderating Main Effects OLS Regression Recognizing Support Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x recovery expectation RQ4.b Hyp4.b.1 363-370 Moderating Domain 1 OLS Regression Help Seeking Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x recovery expectation RQ4.b Hyp4.b.2 371 Moderating Domain 2 OLS Regression Help Seeking Control + stigma + social sup port + empowerment + recovery expectation

PAGE 198

179 Table 12 (cont.) Research Question Hypotheses Range Model # Model Label Analysis Dependent Variable Independent Variables RQ4.b Hyp4.b.3 372 Moderating Domain 3 OLS Regression Help Seeking Control + service variables + select variables x recovery expectation RQ4.b 373-380 Moderating Main Effects OLS Regression Help Seeking Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x recovery expectation RQ4.b Hyp4.b.1 381-388 Moderating Domain 1 OLS Regression Symptom Eradication Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x recovery expectation RQ4.b Hyp4.b.2 389 Moderating Domain 2 OLS Regression Symptom Eradication Control + stigma + social sup port + empowerment + recovery expectation RQ4.b Hyp4.b.3 390 Moderating Domain 3 OLS Regression Symptom Eradication Control + service variables + select variables x recovery expectation RQ4.b 391-398 Moderating Main Effects OLS Regression Symptom Eradication Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x recovery expectation RQ5 Hyp5.1 399-406 Moderating Domain 1 OLS Regression Illness Management Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x illness severity RQ5 Hyp5.2 407 Moderating Domain 2 OLS Regression Illness Management Control + stigma + social support + empowerment + select variables x illness severity RQ5 Hyp5.3 408 Moderating Domain 3 OLS Regression Illness Management Control + service variables + select variables x illness severity

PAGE 199

180 Table 12 (cont.) Research Question Hypotheses Range Model # Model Label Analysis Dependent Variable Independent Variables RQ5 409-416 Moderating Main Effects OLS Regression Illness Management Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x illness severity RQ5 Hyp5.1 417-424 Moderating Domain 1 OLS Regression Illness Management Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x illness severity RQ5 Hyp5.2 425 Moderating Domain 2 OLS Regression Illness Management Control + stigma + social support + empowerment + select variables x illness severity RQ5 Hyp5.3 426 Moderating Domain 3 OLS Regression Illness Management Control + service variables + select variables x illness severity RQ5 427-434 Moderating Main Effects OLS Regression Illness Management Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x illness severity RQ5 Hyp5.1 435-442 Moderating Domain 1 OLS Regression Future Orientation Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x illness severity RQ5 Hyp5.2 443 Moderating Domain 2 OLS Regression Future Orientation Control + stigma + social support + empowerment + select variables x illness severity RQ5 Hyp5.3 444 Moderating Domain 3 OLS Regression Future Orientation Control + service variables + select variables x illness severity RQ5 445-452 Moderating Main Effects OLS Regression Future Orientation Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x illness severity

PAGE 200

181 Table 12 (cont.) Research Question Hypotheses Range Model # Model Label Analysis Dependent Variable Independent Variables RQ5 Hyp5.1 453-460 Moderating Domain 1 OLS Regression Meaning and Hope Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x illness severity RQ5 Hyp5.2 461 Moderating Domain 2 OLS Regression Meaning and Hope Control + stigma + social support + empowerment + select variables x illness severity RQ5 Hyp5.3 462 Moderating Domain 3 OLS Regression Meaning and Hope Control + service variables + select variables x illness severity RQ5 463-470 Moderating Main Effects OLS Regression Meaning and Hope Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x illness severity RQ5 Hyp5.1 471-478 Moderating Domain 1 OLS Regression Recognizing Support Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x illness severity RQ5 Hyp5.2 479 Moderating Domain 2 OLS Regression Recognizing Support Control + stigma + social support + empowerment + select variables x illness severity RQ5 Hyp5.3 480 Moderating Domain 3 OLS Regression Recognizing Support Control + service variables + select variables x illness severity RQ5 481-488 Moderating Main Effects OLS Regression Recognizing Support Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x illness severity RQ5 Hyp5.1 489-496 Moderating Domain 1 OLS Regression Help Seeking Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x illness severity

PAGE 201

182 Table 12 (cont.) Research Question Hypotheses Range Model # Model Label Analysis Dependent Variable Independent Variables RQ4.b Hyp4.b.2 497 Moderating Domain 2 OLS Regression Help Seeking Control + stigma + social support + empowerment + select variables x illness severity RQ4.b Hyp4.b.3 498 Moderating Domain 3 OLS Regression Help Seeking Control + service variables + select variables x illness severity RQ5 499-506 Moderating Main Effects OLS Regression Help Seeking Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x illness severity RQ5 Hyp5.1 507-514 Moderating Domain 1 OLS Regression Symptom Eradication Control + symptoms + other clinical + employment + family history + abuse/assault + diagnosis* + select variables x illness severity RQ5 Hyp5.2 515 Moderating Domain 2 OLS Regression Symptom Eradication Control + stigma + social support + empowerment + select variables x illness severity RQ5 Hyp5.3 516 Moderating Domain 3 OLS Regression Symptom Eradication Control + service variables + select variables x illness severity RQ5 517-524 Moderating Main Effects OLS Regression Symptom Eradication Control + symptoms + other clinical + employment + family history + abuse + stigma + social support + empowerment + services + select variables x illness severity Note: Diagnosis is a series of 8 dichotomous variables. Entering them all into a model violates least square and logistic s olutions and cannot be interpreted. Thus, the diagnosis variable is entered last, one at a time, to assess affects of individual diagnoses. This creates a series of domain 1 models and main effects models

PAGE 202

183 Chapter 4 Results This chapter presents results based on the analysis plan and methods outlined in Chapter 3. I begin with a description of th e respondents and provide relevant univariate results to underscore important differences between the severely mentally ill (SMI) and less severe (OP) participants. Differences ar e also highlighted that distinguish those who feel they will recover from mental illness and those that do not. This segmentation by recovery expectancy traverses the SMI and OP samples and illustrates that low expectation of personal recovery is common for those with and without a label of severe mental illness. Additional bivariate relationships are examined that includes bivariate investigation of dependent variables (recovery expectancy and rec overy strategies) and independent variables (Domain 1: clinical historical, Domain 2: social, Domain 3: service). Discussion of bivariate analysis results is structured, in part, to follow the research questions and hypotheses. The goal was to inte rmingle the numeric results with the story of the participants. Too often, research targ eting individuals with mental illness neglects the essential humanity of the participan ts. Thus, all effort was made to avoid unintentional marginalization by focusing on the liv ed context. Further, not all bivariate analyses are displayed or referenced in this chapter. Bivariate analyses were completed between all control, dependent, and independent variables (dom ains 1-3) as part of the process of learning the story. However, only the results relevant to better understanding

PAGE 203

184 of the population and the relationships inve stigated in the res earch questions are displayed in this chapter. Tw o additional procedural notes are offered: first, all bivariate and multivariate statistics accompanied by a pvalue use the same scale and alphanumeric designations consistently thr oughout analyses to indicate si gnificance. Thus, an a indicates that p 0.05, b 0.01, c 0.001 and d 0.0001. Second, for t-tests, when variances are unequal, the Satterthwaite sta tistic (SAS/STAT, Version 9.1) was used in place of the pooled statistic fo r equal variances. Following bivariate analysis multivariate analysis results are presented and are or ganized by research que stion and hypothesis. Study Participants Demographic Characteristics Respondents provided information on age, gender, education and income. A targeted description of this information a nd any differences between samples (SMI and OP) will help to deepen understanding of th e individuals with mental illness that participated in this research a nd the population they re present. The first descriptor is age, a continuous variable, with re sults presented in Table 13. The average age for the pooled sample of 350 is 41 years. The SMI population is slightly older at 42.5 years compared to the OP sample at 39.6 and the range is similar between the samples. The age difference is significant, t(348) = 2.41, p .05), although less than a three year difference is of little theoretical or practical conseque nce. There is also a si gnificant difference between age and recovery expectancy. Individuals that do not endorse recovery (n = 134, M = 41.5, SD = 9.0) are significantly older then those that do expect to recover (n = 216, M = 38.1, SD = 10.8), t(348) = p .004.

PAGE 204

185 Table 13 Mean and Standard Deviation of Demographic Variable Age n Mean (s.d.) Lower Range Upper Range SMI 175 42.5 (10.7) 20 64 OP 175 39.6 (11.8) 18 63 Total 350 41.1 (11.3) 18 64 Gender, education and income are displaye d together in Table 14. Approximately two-thirds (69%) of the part icipants are female. The total clinical population of the partnering agency is also approximately twothirds female, see Table 1, page 103). Thus, random selection would be expected to favor the female gender and, as predicted, did so. There is no significant difference between sa mples for gender distri butions. Similarly, there are no significant differences between the SMI and OP samples for education or income distributions. Limited formal educa tion and low income le vels are common. The vast majority have a high school or less th an high school education. For income, most individuals have $15,000 or less household income per year. This is representative of mental health consumers in community mental health centers in general and is another layer of challenge that these individuals face each day. The following three sections help to incr ease appreciation of the respondents and provide further insight into their history, clin ical status, beliefs, relationships, and view on services. These sections are not directly linked to the resear ch questions but are necessary to help strengthen understanding a nd empathy for individua ls with differing levels of mental illness. The research que stions described in Chapter 3 are focused on

PAGE 205

186 recovery and not, until the final research que stion (RQ 5), concerned with severity of illness. However, in order to identify importa nt factors associated with recovery, it is imperative to have a greater understanding of the individuals that participated in this study and that are striving for recovery. Table 14 Gender, Education, and Income for SMI, OP and Total Sample SMI OP Total n =175 n =175 N = 350 % % % Gender Male 34 28 31 Female 66 72 69 Education < High School 19 22 21 High School 43 34 39 1-2 yrs. College 27 31 29 2-4 yrs. College 10 11 10 5+ yrs. College 1 1 1 Income 0-10,000 57 46 51 10,001-15,000 21 25 23 15,001-20,000 10 11 11 20,001-25,000 4 5 5 25,001-30,000 2 4 3 30,001-40,000 3 3 3 40,001-50,000 2 2 2 50,001-75,000 1 2 1 75,0011 2 1

PAGE 206

187 Univariate Assessment of Clinical and Histor ical Factors (Independent Variable Domain 1) by Sample: Furthering U nderstanding of the Respondents Individuals diagnosed with a mental illness have varied histories, experiences, and symptom patterns that differ due to severity of illness with possible associations to recovery. Domain 1 independent variables include diagnosis, symptoms, age of onset, substance abuse, hospitalization, medication use, familial mental illness, employment, and abuse history. The following discussion wi ll highlight differences and similarities between samples. The SMI sample was categorized per in clusion criteria with diagnoses of schizophrenia, bipolar disorder, schizoaffec tive disorder, and majo r depression (Table 15). Other diagnoses (e.g., seve re obsessive compulsive disord er) could have qualified as severe mental illness provided there was accompanying impairment of social, occupational, or educational functioning. Documentation of such impairment is necessary for the SMI label. Random sampli ng did select four i ndividuals outside the four main diagnoses for inclusion in the SMI sample and discussions with the individuals clinicians indi cated that these individuals met criteria for inclusion. Diagnosis is a consistent point of debate in the recovery literatu re and the decision was made to keep all diagnostic categories even if they are few in number. The SMI sample contains the majority of i ndividuals with disorders that affect the thought process and are characterized by hallu cinations and delusi ons (schizophrenia: SMI = 88% and schizoaffective disorder: SMI = 87%). Unexpectedly, there are nearly equal numbers of individuals with bipolar diso rder in sample categories, 52% in the SMI sample and 48% in the OP sample. The two individuals diagnosed with anxiety in the

PAGE 207

188 SMI sample were diagnosed with a severe obsessive-compulsive di sorder and one other individual with complex post-traumatic stress disorder. All individua ls diagnosed with a depressive disorder that is not major de pression are in the OP sample as are all individuals with an adjustment disorder. Fi nally, the diagnostic ca tegory of other has 87% of individuals in the OP sample. Thes e include individuals diagnosed with adult attention deficit disorder, pe rsonality disorder, Tourettes syndrome, head injury, and other neurological disorder. The two indivi duals in the SMI sample who are in the other diagnostic categor y were diagnosed with dissociativ e identity disorder and a head injury related psychotic disorder, respectively. Table 15 Proportion of Domain 1 Primary Diagnoses by Sample Primary Diagnosis SMI n = 175 OP n = 175 Total N = 350 % % % Schizophrenia 88 11 15 Bipolar disorder 52 48 38 Schizoaffective 87 13 11 Major depression 44 56 16 Anxiety Disorder 6 94 9 Depression (not major) 0 100 5 Adjustment disorder 0 100 2 Other 13 87 4

PAGE 208

189 Table 16 contains the results of the Symptom Checklist 90-Revised (SCL-90-R) symptom scales by sample. Each symptom cate gory has a range of 0 to 4 for severity. Table 16 Univariate Statistics (Means and Standard Devia tions) for Domain 1 Symptom Scales by Sample Symptom Scales SMI OP Total n=175 n=175 N =350 Mean (s.d.) Mean (s.d.) Mean (s.d.) Psychosomatic 1.50 (0.95) 1.60 (0.89) 1.53 (0.92) OCD 1.92 (0.97) 1.86 (1.03) 1.87 (1.00) Interpersonal Sensitivity 1.65 (0.95) 1.62 (1.06) 1.63 (1.01) Depression 1.83 (0.97) 1.91 (1.07) 1.87 (1.02) Anxiety 1.62 (0.98) 1.61 (1.09) 1.61 (1.03) Hostility 1.07 (0.95) 1.25 (1.0) 1.16 (0.98) Phobia 1.28 (1.04) 1.14 (1.07) 1.21 (1.06) Paranoia 1.79 (1.0) 1.73 (1.05) 1.76 (1.02) Psychoticism 1.34 (0.90) 1.13 (0.83) 1.23 (0.87) Global Severity Index 1.58 (0.82) 1.57 (0.86) 1.58 (0.84) The total symptom severity levels for the tw o samples are markedly similar (SMI =1.58; OP = 1.57), measured by the global severity i ndex (GSI). The GSI is a summary score that indicates the overall symptom level of a respondent with scores closer to four indicating greater cl inical involvement. The three symptoms with the highest aver ages for the total sample are depression: obsessive compulsive sympto ms, and paranoia. Depressi on is ubiquitous in mental

PAGE 209

190 illness and often implicated as a barrier to recovery (Corrigan & Ralph, 2005). Though the SMI sample has higher levels of obs essive compulsive sy mptoms and paranoia compared to the OP sample, the similarity in intensity is surprising. This is particularly perplexing for paranoia considering paranoia is often associated w ith thought disorder and yet the OP sample has fa r less thought disorder (schizophrenia a nd schizoaffective disorder) diagnosed. Conversely, the OP sample has a higher level of depression. As noted in chapter 2, depression is the symptom most investigated for its effect on recovery, yet it can be seen that for this sample that many symptoms may be implicated in recovery and symptoms alone do not distinguish between those labeled severely mentally ill and those that are not. Other clinical variables investigated in clude age of onset, hospitalization history, substance abuse history (classified as whet her the respondent was ever diagnosed with dependence or abuse), medication use, abuse/assault history and familial mental illness. Results are displayed in Table 17 and Table 18. Age of onset can be defined as the first time a person is officially diagnosed with a mental disorder or when the individual can first remember being symptomatic to a point of concern. For this study, beginning of symptoms was used as the point in time for age of onset. The mean age of onset for the entire sample is 21.2 years (SD =10.8) (Table 17). Individuals from the SMI sample had an earlier average age of onset co mpared to the OP sample. Psychiatric hospitalization in the previ ous 12-months was reported by 34% of the respondents (Table 18). The SMI sample had a much higher average number of lifetime hospitalizations then the OP sample (Table 17). Substance a buse is a critical issue in the

PAGE 210

191 recovery from mental illness and can be an additional risk for hospitalization. Cooccurring substance use disord ers were noted in 34% of participants (Table 18). Table 17 Univariate Means and Standard Deviations fo r Continuous Domain 1 Independent Variables (Age of Onset, Hospitalizations, Years Employed, Familial Mental Illness) by Sample SMI n = 175 OP n = 175 Total N = 350 Mean (s.d.) Mean (s.d.) Mean (s.d.) Age of Onset 19.8 (9.9) 22.7 (11.5) 21.2 (10.8) Hospitalizations in Lifetime 10.1 (13.8) 3.3 (5.1) 6.7 (10.9) Total years employed 12.1 (10.7) 15.6 (10.3) 13.9 (10.6) Mental illness: nuclear family 2.3 (2.0) 2.4 (2.2) 2.3 (2.1) Mental illness: extended family 0.8 (1.0) 0.8 (1.1) 0.8 (1.1) Medication was frequently prescribed to i ndividuals in both clinical samples as an avenue to clinical recovery. For instance, 76% of the total sample is prescribed antidepressant medication (Table 18). Anti-anxiety medication is prescribed the second most frequently at 52%. When comparing SMI to OP samples, the SMI sample is prescribed more medication of all types. This is most marked in anti-psychotic medication and other psychotropic medications. Familial mental illness, a potential contri butor to mental illness onset, was broken into nuclear family members (parents and siblings) and extended family (all others) (Table 18). Individuals in th e OP sample have a slightly hi gher level of nuclear family

PAGE 211

192 mental illness. There is no difference in the number of extended family members between the two samples. Employment is freq uently equated with or used as a proxy for Table 18 Proportion of Categorical Domain 1 Independent Variables (Hospitalization, Employment, Substance Use, Medication, and Child Abuse) by Sample SMI n=175 OP n=175 Total N=350 % % % Hospitalized in last year 35 33 34 Currently Employed 11 31 21 Co-occurring substance diagnosis 34 34 34 Anti-psychotic medication 63 25 44 Anti-depressant medication 79 73 76 Anti-manic medication 30 25 27 Anti-anxiety medication 57 48 52 Other psychotropic medication 40 25 32 Child sexual abuse Recurrent 34 34 34 Single incident 9 8 8 Child physical abuse Recurrent 42 35 39 Single incident 1 0 1 Adult sexual assault Recurrent 31 23 27 Single incident 6 6 6 Adult physical assault Recurrent 36 36 36 Single incident 4 3 4

PAGE 212

193 successful recovery. Only 21% of the pooled sample are currently employed (Table 18). Eleven percent of the SMI sample is employed while the OP sample is employed at 31%. The final group of variables in the clin ical-historical domain (domain 1) deals with abuse and assault experiences (Table 18) Recurrent child sexual abuse (CSA) was experienced by 35% of the full sample. Recurrent child physical abuse (CPA) for the total sample is 39%, with 27% reporting a dult sexual assault (ASA) and 36% adult physical assault (APA). The prevalence of CS A and APA is nearly identical between the SMI and OP samples for both single episode and recurrent abuse. For both recurrent CPA and ASA, the prevalence is higher fo r the SMI population. Recovery is often complicated by experiences of abuse and individuals with severe mental illness frequently have greater prevalence of a buse (Alexander, Muenzenmaier, Dumont, & Auslander, 2005). A final procedural not e regarding abuse targets the number of categories defining abuse experiences. Because of the low numbers of individuals with single incident abuse, categories of single-in cident and recurrent abuse were collapsed into an experienced abus e category. Thus, abuse variables are converted to dichotomous for all subsequent biva riate and multivariate analyses. Univariate Assessment of Social Factors (I ndependent Variable Domain 2) by Study Sample: Furthering Underst anding of the Respondents Domain 2 is comprised of three sets of variables or experiences that target the social/interactional areas of stigma, social support and community connection, and empowerment. Each area will be discusse d separately and the univariate results for domain 2 are summarized in Table 19.

PAGE 213

194 Table 19 Univariate Domain 2 Means and Standard D eviations for Stigma, Social Support, and Empowerment by Sample SMI n=175 OP n=175 Total N=350 Mean (s.d.) Mean (s.d.) Mean (s.d.) Stigma Alienation 2.51 (0.61) 2.42 (0.62) 2.47 (0.61) Stereotype endorsement 2.01 (0.48) 1.93 (0.46) 1.97 (0.47) Discrimination occurrence 2.51 (0.63) 2.31 (0.61) 2.41 (0.62) Social withdrawal 2.42 (0.55) 2.29 (0.63) 2.36 (0.59) Stigma resistance 2.14 (0.47) 2.17 (0.45) 2.16 (0.46) Social Support and Community Connection Partner or best friend support 4.17 (1.11) 4.12 (1.24) 4.17 (1.11) Family Support 3.80 (1.51) 3.63 (1.54) 3.80 (1.51) Provider Support 4.22 (0.78) 3.88 (0.91) 4.22 (0.78) Friends Support 4.10 (1.12) 4.12 (1.17) 4.10 (1.12) Community involvement 3.63 (1.17) 3.34 (1.16) 3.63 (1.17) Trust 3.15 (0.98) 2.99 (0.92) 3.15 (0.98) Empowerment Self-esteem and self-efficacy 2.86 (0.53) 2.90 (0.53) 2.88 (0.53) Power and powerlessness 2.41 (0.42) 2.55 (0.42) 2.48 (0.43) Community activism and autonomy 3.15 (0.43) 3.16 (0.38) 3.15 (0.41) Optimism and control over the future 2.81 (0.49) 2.82 (0.43) 2.81 (0.46) Righteous anger 2.43 (0.55) 2.49 (0.52) 2.46 (0.54)

PAGE 214

195 Stigma is believed to have a powerfu l influence on recovery and general well being for individuals with mental illness (see chapter 2 for a detailed discussion). Stigma is investigated via five scal es that correspond with sub-cons tructs of stigma and include alienation, stereotype endorsement, discrimi nation occurrence, soci al withdrawal, and stigma resistance (Ritsher, Otilingam, & Graj ales, 2003). The range for all responses is from 1-5, with higher numbers indicating more negative, stigmatizing, experiences or effects for all scales except for stigma resistance where a higher number indicates a greater ability to resist the negative effects of stigma. For the total sample, alienation is the scale with the highest mean. The SM I sample scores higher on the scales of alienation, stereotype endorsement, discrimi nation occurrence, and social withdrawal compared to the OP sample. Conversely, the OP sample has a higher mean score for stigma resistance. Consequently, individuals wi th severe mental illness are more likely to report greater impact of stigma in their lives and appear to have less coping or resistance capacity. Important relationships that could potent ially provide social and instrumental support include a partner/best friend, family members, provider(s), and friends. Results from scales matching these relationship cluste rs are displayed in Ta ble 19. Related to these are two additional scales of community involvement and trust. Each scale can range from 1-6, with higher numbers indicating more support, community involvement, or general feelings of trust. The highest sc ore for the total sample is on the partner or best friend support scale and the least is trus t. For nearly all comparisons between the SMI and OP samples the SMI sample reports higher levels of support, community

PAGE 215

196 involvement, and trust. The one exception is where the OP sample reports a slightly higher level of support from friends compared to the SMI sample. The final domain 2 construct addresses empowerment. The constructs that encompass empowerment include self-esteem /self-efficacy, power and powerlessness, community activism and autonomy, optimism and control over the fu ture, and righteous anger. The measurement range is 1-5, with high er scores indicating greater levels of selfesteem, powerfulness, etc. The results for th e empowerment scales are also presented in Table 19. Community activism and autonomy is the most highly endorsed empowerment construct. The OP sample score slightly hi gher on all empowerment scales compared to the SMI sample. The most marked difference is in power/powerlessness where the OP sample appears to experience a greater sense of pe rsonal power or control in their lives. Univariate Assessment of Service Factors (Independent Variable Domain 3) by Study Sample: Furthering Underst anding of the Respondents Domain 3 addresses service factors th at include variables for the number of services received per month, number of cont act hours per month, and average satisfaction level with services (Table 20). For many, rec overy involves successful involvement with the right combination of therapeutic and support services. The poverty or near poverty income levels, inconsistent transportati on, and other obligations are obstacles to successful involvement in a comprehensive trea tment plan. The strain and cost incurred to arrive at and engage in se rvices is often taken for grante d by providers. A service is a specific clinical encounter (e.g., individual therapy, cas e management, psychiatric

PAGE 216

197 medication appointment). The mean number of services in a month for the pooled sample is 2.4 (SD = 1.2). Mean contact hours are 7.1 (SD = 7.7) per month and the average satisfaction level is quite high, 5.1 (SD = 1.0) on a six-point scale with six indicating very satisfied. Av erage satisfaction is identical for the SMI and OP samples. However, average number of services is higher for the SMI populat ion compared to the OP population. Total contact hours al so are greater for the SMI sample. Table 20 Univariate Means and Standard Deviations for Domain 3 Independent Variables: Service Factors by Sample SMI n=175 OP n=175 Total N=350 Mean (s.d.) Mean (s.d.) Mean (s.d.) Total number of services 3.1 (1.1) 1.7 (0.9) 2.4 (1.2) Total contact hours 9.5 (9.2) 4.7 (4.8) 7.1 (7.7) Average satisfaction 5.1 (0.9) 5.1 (1.1) 5.1 (1.0) Results of Comprehensive Bivariate Analysis This section details the re sults of bivariate analyses relevant to the research questions detailed in Chapter 3. First, demogr aphic factors are inves tigated in relation to recovery expectancy and r ecovery strategies. Second, recovery expectancy is investigated between sample s in response to research question 1. Next, sample differences in recovery strategies are targ eted followed by an investigation of the

PAGE 217

198 relationship between recovery expectancy a nd recovery strategies Finally, bivariate results that respond to research questions 2 (a ssociations with recovery expectancy) and 3 (associations with recovery strategies) are summarized. Associations between Age, Gender, Education, Income and Recovery The understanding of what influences r ecovery remains limited. The following describes the contribution of demographic f actors to the understanding of the target population through investigating the associatio ns between demogra phic variables (e.g., age, gender, education and income) and recovery expectancy and strategies. Recovery expectancy is not signi ficantly associated with gender, 2 (1, N = 350) = 0.42, ns, education, 2 (3, N = 350) = 4.04, ns, or income, 2 (3, N = 350) = 3.40, ns. However, recovery expectancy is signifi cantly associated with age, t(348) = 2.99, p .003. Thus, individuals that endorse recovery are signi ficantly younger (M = 39.7, SD = 11.9 years) compared to those who do not expe ct to recover (M = 43.3, SD = 10.1 years). The relationship between age a nd the recovery strategies is displayed in Table 21. Correlations are susceptible to sample size a nd often provide a highl y significant p-value with a moderate or mild magnitude correla tion. The significant relationship between effective illness management (r(348) = -0.12, p .05), positive future orientation (r(348) = -0.15, p .01), recognizing support (r(348) = -0.13, p .01), and help seeking (r(348) = 0.10, p .05) illustrates the need to exam ine magnitude. These are not strong correlations. However, it is significant that for three of four asso ciations, use of the strategies increases as age decreases. These findings may reflect the need for management skills and support during the stage of illness near or at least closer to onset.

PAGE 218

199 Table 21 Pearson Correlations for Age by Recovery Strategies Effective illness management Positive future orientation Meaningfulness, personal control and hope Recognizing support Help seeking Symptom eradication Age -0.12 a -0.15 b 0.02 -0.13 b 0.10 a -0.02 Note: a p .05; b p .01; c p .001; d .0001 Younger individuals, in tur n, would naturally have a c oncern about what they are experiencing and the effect on their future, esp ecially if changes in plans are required to adapt to the illness. Gender, income, and education are associat ed with recovery strategies via the independent sample t-test and one-way ANOVA statistics in Table 22. Gender is significantly associated only with strategy one, effective illness management, t(348) = 3.06, p .01). Male respondents, on average, en dorse this strategy more than female respondents. This is not i ndicative of a general trend of male strategy use; however, since female respondents score higher on three strategies and males on the other three. Income is significantly associated with thr ee strategies: effective illness management (F(3,347) = 3.25, p .05), positive future orientation (F(3,347) = 3.75, p .05) and meaningfulness, personal cont rol and hope (F(3,347) = 3.41, p .05). Due to the low numbers in upper income categories, the f our highest income categories were collapsed into a $20,001 or more annual income category. The mean score in the three significant relationships (effective illness management, positive future orientation, and meaningfulness, personal control and hope) in the lowest three income levels are relatively similar for the three st rategies, and the signif icant relationship is

PAGE 219

200 Table 22 Independent Sample T-tests and One-Way ANOVA Results of the Relationship Between Gender, Income, Education and Recovery Strateg ies Effective illness management Positive future orientation Meaningfulness personal control and hope Recognizing support Help seeking Symptom eradication N Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Ge nder ome tion Male 109 3.48 (0.71) 3.06 b 4.05 (0.60) 1.71 4.16 (0.49) -1.15 3.83 (0.91) 0.72 3.94 (0.89) -0.58 3.27 (1.24) 0.96 Female 241 3.22 (0.75) 3.92 (0.67) 4.22 (0.44) 3.75 (0.99) 4.00 (0.81) 3.13 (1.24) N Mean (s.d.) F Mean (s.d.) F Mean (s.d.) F Mean (s.d.) F Mean (s.d.) F Mean (s.d.) F Inc 0 10k 180 3.26 (0.69) 3.25 a 3.93 (0.64) 2.75 a 4.20 (0.43) 3.41 a 3.73 (0.91) 2.29 3.96 (0.79) 1.18 3.27 (1.22) 2.49 10,001-15k 79 3.28 (0.77) 3.90 (0.65) 4.19 (0.45) 3.62 (1.01) 3.89 (0.90) 3.30 (1.27) 15,001-20k 37 3.15 (0.77) 3.94 (0.70) 4.06 (0.56) 3.94 (1.01) 4.03 (0.79) 2.86 (1.25) 20,000154 3.58 (0.85) 4.20 (0.64) 4.36 (0.44) 4.01 (1.00) 4.16 (0.90) 2.87 (1.21) Educa < H.S. 73 3.24 (0.82) 0.76 3.92 (0.77) 0.84 4.22 (0.43) 0.26 3.74 (1.07) 1.49 3.99 (1.01) 1.74 3.48 (1.21) 4.14 b H.S. 136 3.37 (0.70) 4.03 (0.58) 4.18 (0.49) 3.89 (0.87) 3.97 (0.81) 3.21 (1.26) 2-years college 102 3.25 (0.70) 3.95 (0.61) 4.21 (0.44) 3.72 (0.88) 4.10 (0.69) 3.10 (1.20) > 2 year college 39 3.31 (0.89) 3.86 (0.78) 4.25 (0.46) 3.55 (1.22) 3.74 (0.88) 2.64 (1.20) Note: a p .05; b p .01; c p .001; d .0001 H.S. = high school education

PAGE 220

201 likely due to the increase in strategy use in th e fourth, and highest, income level. The significance of the association might be an artifact of the decision to collapse income categories and the significance must be considered with caution. Similar to the handling of income data, low numbers of individuals with higher education led to collapsing the top three education cat egories into one category (education greater than 2 years of college). Education is significan t associated only with symptom eradication (F(3, 347) = 4.14, p .01). Belief in symptom eradication is utilized more by individuals with lower edu cation with less endorsement noted with each step in educational level. One explanati on is that higher edu cation achievement may expose the individual to the accepted expl anation of the chronic nature of many psychiatric symptoms and thus decrease the emphasis on symptom eradication. Overall, there is no readily discernibl e pattern of association betw een education and recovery strategy. Sample Differences in Recovery Expectancy Approximately two-thirds of the 350 partic ipants believed that they will recover from mental illness (62%) while one-third did not (38%). The OP sample participants were more likely to believe that they will recover (73%), compared to the SMI sample (50%). The results are summarized in Tabl e 23. For the total sample, individuals are more likely to endorse complete recovery compared to endorsing that they will not recover at all, with the OP sample, in gene ral, having greater recovery expectations. More specifically, individuals in the OP sample are appr oximately three times more

PAGE 221

202 likely to sanction complete recovery and are only half as likely to believe that they will not improve past there current level of health and functioning. These results pertain directly to research question 1 and the three accompanying hypotheses. Table 23 Chi-square Results for Recovery Ex pectation Variables by Sample SMI OP Total % % % 2 Recovery Expectation Yes 50 73 62 19.35 d No 50 27 38 Complete Recovery Expected Yes 13 20 16 3.54 No 87 80 84 No Recovery Expected Yes 7 3 5 3.75 No 93 97 95 Note: a p .05; b p .01; c p .001; d .0001 variables in domains 1-3 and/ or may mediate the relationshi p of predictor variables and recovery strategies. Multivariate analytic procedures (logistic regression and OLS regression) are used to assess these relations hips. However, the relationship between the belief that one will recover and severity of illness, dichotomized by the absence or presence of the SMI label, can be assessed via chi-square analysis. There is a strong and significant differ ence between the two samples on recovery expectancy (Table 23). Those who do not beli eve they will recover are much more likely

PAGE 222

203 to be in the SMI sample (50%) co mpared to the OP sample (27%), 2 (1, N = 350) = 19.35, p .0001). This directly responds to hypothesis one of research question one (RQ1: Hyp1.1). The accompanying hypotheses to research question 1 are reproduced in Table 24. Table 24 Summary of Support for Hypotheses 1.1 to 1.3 Hypothesis Statement Supported? Hyp1.1 The OP sample will endorse that they will recover from mental illness to a greater extent than the SMI sample. Yes Hyp1.2 The OP sample will endorse complete recovery as possible more than the SMI sample. No Hyp1.3 The SMI sample will endorse no expectation of recovery more than the OP sample. Yes The second hypothesis (Hyp1.2) states that the OP sample will expect complete recovery to a greater degree. The data does not support this hypothesis, 2 (1, N = 350) = 3.54, p 0.06, though it just misses significance. The third hypothesis (Hyp1.3), representing no difference between samples fo r endorsing complete lack of recovery, is marginally supported, 2 (1, N = 350) = 3.75, p = 0.053. The borderline significance for hypotheses 1.2 and 1.3 suggests that there might be differences in beliefs of complete recovery/non-recovery and that further investigation would not be wasted effort.

PAGE 223

204 Differences in Recovery Strategies be tween Samples and Recovery Expectations Table 25 displays the mean and standa rd deviation of each recovery strategy stratified by sample (SMI and OP). Chapte r 3 provides an initia l description of each strategy. There is no significan t or consistent direction of relationship due to sample membership, a proxy for severity of mental illness. Symptom eradication is the least endorsed strategy for the SMI and OP samples. This may reflect, in part, that this variable is a single item, not a composite of several recovery items. Both samples endorse the strategy of finding meaning, persona l control, and hope to a greater degree then the other strategies. However, in gene ral, there is little difference between the samples in the extent that they endorse any of the recovery strategies. Table 25 Independent Sample T-test Results, Means, and Standard Deviations for Recovery Strategies (Continuous Depende nt Variables) by Sample SMI n = 175 OP n = 175 Total N = 350 Recovery Strategy Mean (s.d.) Mean (s.d.) Mean (s.d.) t-test Effective illness management 3.31 (0.70) 3.29 (0.79) 3.30 (0.75) 0.34 Positive Future Orientation 3.92 (0.67) 4.01 (0.64) 3.96 (0.65) -1.27 Meaningfulness, Personal Control and Hope 4.19 (0.48) 4.22 (0.44) 4.21 (0.46) -0.51 Recognizing Support 3.75 (0.92) 3.79 (1.01) 3.77 (0.96) -0.41 Help Seeking 4.02 (0.83) 3.95 (0.84) 3.98 (0.84) 0.70 Symptom Eradication 3.13 (1.28) 3.21 (1.21) 3.17 (1.24) -0.60

PAGE 224

205 A different story emerges when examin ing differences in recovery strategy endorsement by recovery expectancy. Immedi ately evident is that for each strategy, respondents who believe they will recover tend to endorse the recove ry strategies to a greater degree (Table 26). These associati ons are statistically significant for every strategy except symptom eradication. This is a potentially valuable finding when considered in the context of the minimal information on recovery expectancy in the literature and the drive to identify functional and accurate recovery strategies. Of equal importance to the consistent direction of rela tionships across strategies is the magnitude of associations, with most re lationships leaving little room for chance as a reason for the associations. Thus, the results suggest that it would be useful to consider recovery expectancy when developing programs designed to improve use of recovery strategies. Table 26 Independent Sample T-test Results for Recovery Strategies by Recovery Expectation Recovery Strategy Expect to Recover Yes Mean (s.d.) (n = 216) No Mean (s.d.) (n = 134) t-test Effective illness management 3.46 (0.71) 3.05 (0.73) -5.22 d Positive future orientation 4.11 (0.61) 3.74 (0.66) -5.31 d Meaningfulness, personal c ontrol and hope 4.27 (0.46) 4.12 (0.44) -3.22 c Recognizing support 3.94 (0.90) 3.51 (1.01) -4.07 d Help seeking 4.08 (0.81) 3.83 (0.86) -2.67 b Symptom eradication 3.27 (1.19) 3.01 (1.32) -1.86 Note: a p .05; b p .01; c p .001; p d .0001

PAGE 225

206 Bivariate Analysis for Research Question 2: Are Individual, Social or Service Factors Associated with Recovery Expectancy? Research question 2.1: Are individual factors associated with recovery expectancy? The following analyses investigate the re lationship between recovery expectancy and each of the predictor variables from Domain 1. Research question 2.1 is specific to Domain 1 and Tables 27-30 display the results of the bivariate analyses. The relationship of diagnosis to recovery expectancy has clin ical and recovery implications. The chisquare relationships are summarized in Table 27. Individuals with bipolar disorder are significantly less likely to believe they will recover (55% endorse recovery compared to 66% of the rest of the sample). Individuals that are depressed but are not severe enough to be labeled with major depression are significantly more likely to believe they will recover (84%) compared to other sample members (60. All i ndividuals with an adjustment disorder, as would be expect ed, endorse recovery (Fishers Exact, p .05). Unexpectedly, neither of the diagnostic cate gories that usually indicate impairment in cognitive functioning (schizophrenia and schi zoaffective disorder) was significantly associated with recovery expectancy. Even more surprising when vi ewed in the context of the recovery literature, individuals dia gnosed with schizophrenia are more apt to endorse recovery (69%) compared to the re st of the sample (60%) (Torgalsboen, 2005).

PAGE 226

207 Table 27 Chi-square Results of Recovery Expectancy by Diagnosis Schizophrenia Bipolar Disorder Schizoaffective Disorder Major Depression Recovery Expectancy Yes % No % 2 Yes % No % 2 Yes % No % 2 Yes % No % 2 Yes 69 60 1.21 55 66 4.61 a 55 62 0.75 58 62 0.34 No 31 40 45 34 45 38 42 38 Anxiety Disorder Depression (not Major) Adjustment Disorder Other Disorder Recovery Expectancy Yes % No % 2 Yes % No % 2 Yes % No % 2 Yes % No % 2 Yes 67 61 0.38 84 60 4.30 a 100 61 *F a 73 61 0.89 No 33 39 16 40 0 39 27 39 Note: a p .05; b p .01; c p .001; p d .0001; *F = Fishers Exact stat istic used due to cell size

PAGE 227

208 The associations between symptom s cales and recovery expectancy are summarized in Table 28. Every symptom scale except for hostility is significantly associated to recovery expectancy. All associ ations are in the same direction, with higher levels of symptoms experienced by individuals who do not expect to recover. Depression, in the recovery literature is the symptom most often linked to recovery failure (Corrigan & Ralph, 2005), has the largest difference between those that sanction recovery and those that do not. Global severity of symptoms is second and strongly illustrates the impact of multiple symptoms on recovery. Next is psychoticism (psychotic symptoms), paralleling the recove ry literature that rec ognizes thought disorder as a challenge to recovery (Spaniol, We wlorski, Gagne, Anthony, 2005; Torgalsboen, 2005). Having detected a strong associa tion between symptoms and recovery expectancy, the next set of associations to investigate are the continuous measures domain 1 variables (age of onset, lifetime hospitalizations, years employed and familial mental illness). These associations are summarized in Table 29 by means of the independent sample t-tests. Lifetime hos pitalizations are si gnificant though it is impossible to tell with this information whethe r hospitalizations lead to lack of recovery expectancy or vice versa. Two other significant associatio ns are noted, both addressing familial mental illness. The stronger associati on is with nuclear fam ily mental illness. Individuals that lack expectancy have mo re immediate family members with mental illness, t(348) = 3.19, p .001, as well as extended familial mental illness, t (348) = 2.61, p .01).

PAGE 228

209 Table 28 Independent Sample T-test Results of Recovery Expectancy by Symptom Scales Recovery Expectancy Psychosomatic Obsessive Compulsive Interpersonal Sensitivity Depression Anxiety Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Recovery No 1.84 (0.92) 5.12 c 2.44 (0.94) 5.47 c 1.95 (0.98) 4.78 c 2.27 (0.97) 6.15 c 2.00 (0.99) 5.81 c Recovery Yes 1.34 (0.87) 1.66 (0.98) 1.44 (0.98) 1.62 (0.97) 1.51 (0.99) Hostility Phobic Anxiety Paranoia Psychoticism Global Severity Index Recovery No 1.31 (1.01) 1.34 1.59 (1.03) 5.58 c 20.9 (1.00) 4.94 c 1.57 (0.90) 5.94 c 1.91 (0.80) 6.14 c Recovery Yes 1.07 (0.94) 0.97 (1.00) 1.56 (0.99) 1.03 (0.78) 1.37 (0.79) Note: a p .05; b p .01; c p .001; p d .0001, Recovery NO, n=134; Recovery YES, n = 216.

PAGE 229

210 Table 29 Independent Sample T-test Results for R ecovery Expectancy and Domain 1 Variables: Age of Onset, Hospitalizations, Empl oyment, and Familial Mental Illness Variable Recovery NO n = 134 Recovery Yes n = 216 Mean (s.d.) Mean (s.d.) t-test Age of onset 21.8 (15.7) 21.3 (10.8) 0.28 Lifetime hospitalizations 8.4 (13.9) 5.6 (8.5) 2.12 a Total years employed 15.1 (10.3) 13.1 (10.8) 1.66 Mental illness: Nuclear family 2.8 (2.1) 2.1 (2.0) 3.19 c Mental illness: Extended family 1.0 (1.2) 0.7 (1.0) 2.61 b Note: a p .05; b p .01; c p .001; d .0001 The remaining variables from domain 1 are examined in Table 30. Chi-square analysis was used to detect important rela tionships between recovery expectancy and hospitalization in the last year, current em ployment, substance abuse history, medication use, child abuse and adult assault. Several of the associations are significant. First, a substance abuse history is significantly more likely in i ndividuals with no recovery expectancy. The additional st ruggle for recovery noted for individuals with a substance dependence history emphasized in the literature parallels this finding (White, Boyle, & Loveland, 2005). Four of five medication variables have si gnificant relationships with recovery. Anti-depressant medication is the on e that is not related to expectancy and may reflect the acceptance of this medication both in the general culture and this sample (76% of the individuals in this study are prescrib ed anti-depressant medi cation). Anti-anxiety

PAGE 230

211 medication has the strongest link with recove ry expectancy with nearly twice as many individuals who have no recove ry expectancy taking anti-anx iety medications than not. This is a potentially interesting finding that reinforces the earlier association wherein all three anxiety symptom scales are significantly higher for those not endorsing recovery (Table 28). Anti-manic medication is prescr ibed more to non-reco very individuals and appears to be an indicator for more inte nse pathology. Anti-psychotic medication and other psychotropic medication are also si gnificantly related to recovery. Table 30 Chi-square Results for Recovery Expectancy and Hospitalization in Last Year, Current Employment, Substance Abuse, Medication, Abuse and Assault History Variable Recovery NO (n = 134) Recovery YES (n = 216) 2 % Yes % Yes Hospitalized in last year 37 32 0.88 Currently Employed 16 24 2.91 Co-occurring substance use 40 30 3.84 a Anti-psychotic medication 51 40 3.67 a Anti-depressant medication 81 73 3.39 Anti-manic medication 37 22 9.11 c Anti-anxiety medication 66 44 15.60 d Other Psychotropic 39 28 4.22 a Child sexual abuse 48 38 2.96 Child physical abuse 46 35 4.63 a Adult sexual assault 42 27 8.65 c Adult physical assault 42 37 1.35 Note: a p .05; b p .01; c p .001; p d .0001

PAGE 231

212 The final variables addressed in domain 1 ar e child abuse and adult assault history. The association between recovery expectancy and ch ild sexual abuse is su rprising in its lack of significance considering the prevalence of sexual abuse and the effect on recovery noted in the literatur e (Goodman, Rosenberg, Mueser, Drake, 1997; Wexler, Lyons, Lyons, Mazure, 1997). Howeve r, child physical abuse and adult sexual assault are both significantly rela ted to recovery expectancy. Research question 2.2: Are social factors associated with recovery expectancy? Domain 2 variables of stigma, so cial support, and empowerment were investigated via independent sample t-test analysis to ascertain their relationship to recovery expectancy. These findings are su mmarized in Table 31. Regarding stigma, each scale (alienation, stereotype endorsem ent, discrimination occurrence, social withdrawal, and stigma resistan ce) are significantly associated with lack of recovery expectation. Alienation has the most pow erful association with discrimination occurrence second. Alienation resulting from discrimination, though not supported with cross-sectional data, makes conceptual sense. Table 31 also summarizes the results from independent sample t-tests between recovery expectancy and social support variab les. Individuals who feel they will recover report significantly higher s upport from a partner or best friend, family members, and friends. This reinforces the l iterature that suggests, in general, support is necessary for recovery (Rogers, Anthony, & Lyass, 2004). Co mmunity involvement is higher for those endorsing recovery as well. There is no difference in perceived provider support, suggesting that strictly a c linical support func tion is not required for recovery

PAGE 232

213 Table 31 Independent Sample T-test Results, Means and Standard Deviations, for Domain 2 Variables (Stigma, Social Support, & Empowerment) and Recovery Expectancy Variable Recovery NO (n = 134) Recovery YES (n = 216) Mean (s.d) Mean (s.d.) t-test Stigma Alienation 2.80 (0.59) 2.34 (0.59) 5.07 c Stereotype endorsement 2.08 (0.46) 1.90 (0.47) 3.51 b Discrimination occurrence 2.59 (0.60) 2.30 (0.62) 4.24 c Social withdrawal 2.51 (0.57) 2.26 (0.59) 3.79 b Stigma resistance 2.27 (0.46) 2.09 (0.44) 3.75 b Social Support and Community Connection Partner or best friend su pport 3.86 (1.24) 4.32 (1.11) -3.58 c Family Support 3.40 (1.57) 3.91 (1.47) -3.08 b Provider Support 4.03 (0.94) 4.06 (0.80) -0.39 Friends Support 3.88 (1.25) 4.25 (1.05) -2.88 b Community involvement 3.10 (1.20) 3.45 (1.14) -2.30 a Trust 3.04 (0.92) 3.09 (0.97) -0.46 Empowerment Self-esteem/ self-efficacy 2.72 (0.56) 2.98 (0.49) -4.67 c Power/powerlessness 2.37 (0.40) 2.55 (0.43) -3.81 b Community activism and autonomy 3.10 (0.41) 3.19 (0.40) -1.99 a Optimism and control over the future 2.72 (0.47) 2.87 (0.44) -2.99 b Righteous anger 2.41 (0.57) 2.49 (0.52) -1.46 Note: a p .05; b p .01; c p .001; p d .0001 expectation or that provider support may in fluence other aspects of recovery then expectancy. The final set of variables for domain 2 independent variables is the empowerment cluster. Self-esteem / self-efficacy have the most powerful association with those endorsing recovery having a higher mean score (Table 31). For all sub-constructs of

PAGE 233

214 empowerment the pattern is the same, expectatio n of recovery is associated with greater empowerment. Only righteous anger does not reach significance in the analysis. The overall pattern of domain 2 va riables should be emphasized. Those endorsing recovery recognize le ss stigmatization and experi ence more support and more empowerment in nearly every instance. Previous investigations of what comprises recovery have assessed this point to a degree, but this curr ent analysis is unique in its emphasis on recovery expectancy (Corriga n, Giffort, Rashid, Leary & Ihemoa, 1999). Due to the limitations of cross-sectional data, it cannot be determined whether less support and empowerment leads to reduced expe ctation of recovery or the reverse. Research Question 2.3: Are service factors associated with recovery expectancy? Domain three is comprised of three servic e related variables: number of current services, average contact hours per month, a nd service satisfaction. Table 32 summarizes this information. Those who do not endorse recovery have more services and more contact time, but neither indicator reached si gnificance. Service satisfaction did reach significance, with individuals that state th ey will recover having significantly more satisfaction with services. This suggests q uality over quantity in service provision and likely the necessity of a good rapport. What is not considered in this analysis are specific services, (e.g., case management, supported employment, outpatient therapy, and their relationship to recovery expectancy). Such an analysis could be cons idered a logical next step in investigating services in relation to recovery.

PAGE 234

215 Table 32 Independent Sample T-test Results for Recovery Expectancy and Domain 3 Service Variables Recovery Expectancy Variable NO (n = 134) YES (n = 216) Mean (s.d) Mean (s.d.) t-test Total number of services 2.57 (1.20) 2.33 (1.19) 1.78 Total contact hours 7.49 (8.30) 6.84 (7.32) 0.76 Average satisfaction 4.93 (0.88) 5.20 (0.98) -2.40 b Note: a p .05; b p .01; c p .001; p d .0001 Table 33 summarizes the hypotheses linke d to research questions 2.1-2.3 and whether the result is supported in bivariat e analysis. Bivariate relationships are not representative of the complex interactive real ity of these concepts, taken in isolation as they are. This is a crude representation designed to understand the relationship of key constructs in a limited way, and will be more fully explicated in multivariate analysis. Thus, this display is repeated when consid ering multivariate results and, taken together, illustrates how each predictor affects recovery expectancy alone and in combination with other variables.

PAGE 235

216 Table 33 Summary of Bivariate Results for Research Question 2 Hypotheses Hypothesis Statement Supported? Hyp2.1a Less severe diagnoses will be associated with higher recovery expectancy. Hyp2.1 Lower somatization symptoms reported by participants will be associated with higher recovery expectancy. Hyp2.1c Lower obsessive-compulsive symptoms reported by participants will be associated with higher recovery expectancy. Hyp2.1d Lower interpersonal sensitivity reported by participants will be associated with higher recovery expectancy. Hyp2.1e Lower symptoms of anxiety reported by participants will be associated with higher recovery expectancy. Hyp2.1f Lower symptoms of depression reported by participants will be associated with higher recovery expectancy. Hyp2.1g Lower hostility reported by particip ants will be associated with higher recovery expectancy. Hyp.1h Lower phobic anxiety reported by participants will be associated with higher recovery expectancy. Hyp2.1i Lower paranoid ideation reported by participants will be associated with higher recovery expectancy. Hyp2.1j Lower psychoticism reported by participants will be associated with higher recovery expectancy. Hyp2.1k Absence of a comorbid substance use disorder will be associated with higher recovery expectancy for participants. Hyp2.1l Lower numbers of lifetime psychiatric hospitalizations will be associated with higher recovery expectancy for participants. Hyp2.1m No psychiatric hospitalization in the last year will be associated with higher recovery expectancy for participants. Hyp2.1n Older age at onset of disorder will be associated with higher recovery expectancy for participants. Hyp2.1o Absence of prescribed anti-depressant medication will be associated with higher recovery expectancy for participants. Hyp2.1p Absence of prescribed anti-psychotic medication will be associated with higher recovery expectancy for participants. Hyp2.1q Absence of prescribed anti-mania medication will be associated with higher recovery expectancy for participants. Hyp2.1r Absence of prescribed anti-anxiety medication will be associated with higher recovery expectancy for participants. Hyp2.1s Absence of any other prescribed psychotropic medication will be associated with higher recovery expectancy for participants. Hyp2.1t Being currently employed will be associated with higher recovery expectancy for participants. Hyp2.1u Greater number of years worked will be associated with higher recovery expectancy for participants.

PAGE 236

217 Table 33 (cont.) Hypothesis Hypothesis Hypothesis Hyp2.1v Absence of familial mental illness in first-degree relatives will be associated with higher recovery expectancy for participants. Hyp2.1w Absence of familial mental illness in extended family members will be associated with higher recovery expectancy for participants. Hyp2.1x Absence of child sexual abuse w ill be associated with higher recovery expectancy for participants. Hyp2.1y Absence of child physical abuse w ill be associated with higher recovery expectancy for participants. Hyp2.1a Absence of adult sexual assault will be associated with higher recovery expectancy for participants. Hyp2.1aa Absence of adult physical assault w ill be associated with higher recovery expectancy for participants. Hyp2.2a Lower feelings of alienation will be associated with higher recovery expectancy for participants. Hyp2.2b Lower respondent endorsement of mental illness stereotypes will be associated with higher recovery expectancy for participants. Hyp2.2c Lower number of discrimination experiences will be associated with higher recovery expectancy for participants. Hyp2.2d Lower endorsement of social withdrawal will be associated with higher recovery expectancy for participants. Hyp2.2e Greater endorsement of stigma resistance will be associated with higher recovery expectancy for participants. Hyp2.2f Greater support through intimate part ner or a best friend will be associated with higher recovery expectancy for participants. Hyp2.2g Greater support through family members will be associated with higher recovery expectancy for participants. Hyp2.2h Greater support via mental health providers will be associated with higher recovery expectancy for participants. Hyp2.2i Greater support through friendship will be associated with higher recovery expectancy for participants. Hyp2.2j Greater feelings of being connected to the community will be associated with higher recovery expectancy for participants. Hyp2.2k Greater trust in the motivation of others will be associated with higher recovery expectancy for participants. Hyp2.2l Higher ratings of self-esteem and self-efficacy will be associated with higher recovery expectancy for participants. Hyp2.2m Higher ratings of personal power will be associated with higher recovery expectancy for participants. Hyp2.2n Greater involvement in the community will be associated with higher recovery expectancy for participants. Hyp2.2o Greater confidence in personal control over the future will be associated with higher recovery expectancy for participants. Hyp2.2p Higher ratings of righteous anger will be positively associated with higher recovery expectancy for participants.

PAGE 237

218 Table 33 (cont.) Hypothesis Hypothesis Hypothesis Hyp2.3a Total number of services will be positively associated with higher recovery expectancies for participants. Hyp2.3b The average number of contact hours per month will be positively associated with higher recovery expectancies for participants. Hyp2.3c Satisfaction level with services will be associated with higher recovery expectancies for participants. This next set of bivariate analyses re places the dichotomous recovery expectancy variable, with continuous level recovery strate gies. The recovery strategies investigated include (1) effective illness managemen t; (2) positive future orientation; (3) meaningfulness, personal control, and hope; (4), recognizing support; (5), help seeking; and (6) symptom eradication (a single item variable). The association between each strategy and the predictor vari ables contained in domain 1 (clinical/historical factors), domain 2 (social factors), and domai n 3 (service factors) are addressed. Research Question 3: Are Individual, Social or Serv ice Factors Associated with Recovery Strategies? Research question 3.1: Are individual factors associated with recovery strategies? Associations between the recovery stra tegies and primary clinical diagnosis suggest that the first strategy, effective illness management, is significantly associated only with adjustment disorder (Table 34). Ad justment disorder is greatly influenced by stressful occurrences and environmental fact ors for both onset and illness course. The transient and less severe nature of the illness would lend itself to less complex

PAGE 238

219 Table 34 Independent Sample T-test Results for Recovery Strategies and Domain 1 Diagnosis Variables Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom Eradication Diagnosis n Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Schizophr enia olar tive sion rder Yes 51 3.40 (0.63) -1.04 3.95 (0.70) 0.21 4.08 (0.53) 2.14 a 3.97 (0.76) -1.88 4.10 (0.73) -1.25 3.35 (1.34) -1.13 No 299 3.28 (0.77) 3.97 (0.65) 4.23 (0.44) 3.74 (0.99) 3.96 (0.85) 3.14 (1.23) Bip Yes 132 3.21 (0.80) 1.76 3.92 (0.69) 0.89 4.25 (0.43) -1.50 3.74 (1.01) 0.43 3.97 (0.84) 0.29 3.11 (1.28) 0.76 No 218 3.36 (0.71) 3.99 (0.63) 4.18 (0.47) 3.79 (0.94) 3.99 (0.83) 3.21 (1.22) Schizoaffec Yes 38 3.28 (0.69) 0.22 4.04 (0.58) -0.79 4.30 (0.46) -1.38 3.80 (0.93) -0.17 4.10 (0.69) -0.88 2.92 (1.17) 1.32 No 312 3.31 (0.75) 3.95 (0.66) 4.19 (0.46) 3.77 (0.97) 3.97 (0.85( 3.20 (1.25) Major Depres Yes 55 3.32 (0.77) -0.24 3.84 (0.66) 1.58 4.13 (0.46) 1.32 3.51 (1.11) 2.17 a 3.99 (0.80) 0.46 3.49 (1.15) -2.21 a No 295 3.30 (0.74) 3.99 (0.65) 4.22 (0.46) 3.82 (0.93) 3.93 (1.01) 3.11 (1.25) Anxiety Diso Yes 33 3.20 (0.70) 0.80 4.04 (0.54) -0.72 4.09 (0.39) 1.51 3.71 (0.80) 0.41 3.92 (0.83) 0.47 2.82 (1.07) 1.72 No 317 3.31 (0.75) 3.96 (0.67) 4.22 (0.46) 3.78 (0.98) 3.99 (0.84) 3.21 (1.26)

PAGE 239

220 Table 34 (cont.) Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom Eradication Diagnosis n Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Depression (not MD) rder ther Yes 19 3.44 (0.70) -0.84 3.97 (0.64) -0.03 4.24 (0.33) -0.33 3.58 (1.04) 0.90 3.89 (0.80) 0.48 3.32 (1.11) -0.52 No 331 3.29 (0.75) 3.96 (0.66) 4.20 (0.47) 3.78 (0.96) 3.99 (0.84) 3.16 (1.25) Adjustment Diso Yes 7 3.95 (0.64) -2.31 a 4.42 (0.37) -1.85 4.31 (0.39) -0.60 4.00 (0.75) -0.63 4.24 (0.71) -0.81 4.00 (1.00) -1.79 No 343 3.29 (0.74) 3.96 (0.66) 4.20 (0.46) 3.77 (0.97) 3.98 (0.84) 3.15 (1.24) O Yes 15 3.47 (0.79) -0.90 4.26 (0.57) -1.77 4.41 (0.57) -1.78 4.53 (0.65) -3.16 b 3.78 (0.90) 0.97 2.80 (1.37) 1.18 No 335 3.29 (0.75) 3.95 (0.66) 4.20 (0.45) 3.74 (0.96) 3.99 (0.83) 3.12 (1.24) Note: a p .05; b p .01; c p .001; d p .0001

PAGE 240

221 management and more success. The second strategy, positive future orientation, is not significantly associated with any diagnostic category. Fi nding meaning, control and hope (strategy 3) is signifi cant only with schizophrenia. Indi viduals with schizophrenia report significantly less sanctioning of this strategy and is repres entative of decreased hope and meaningfulness pervasive in indi viduals with this diagnosis. Individuals can only capita lize on support if they r ecognize it. The fourth strategy, recognizing support, is significant for major depr ession, with those diagnosed having less recognition of support then other participants. Those categorized as other diagnoses have the opposite association with higher levels of support recognition then other respondents. The fifth strategy of he lp seeking has no significant associations and symptom eradication is signi ficant only for major depres sion wherein those without major depression are significantly more li kely to believe that symptoms will be eliminated at some point. Pearson correlations are reported as estimates of the relationship between recovery strategies and symptom scale da ta collected via the Symptom Checklist 90Revised (Table 35). Every association is nega tive. This indicates that endorsement of recovery strategies decreases as symptoms increase. The uniformity of this finding is noteworthy but the magnitude varies across strategies. Symptoms are most strongly associated with the first strategy of effec tive illness management. Each relationship between symptom and strategy is significan t and of moderate intensity. Illness management, in part, focuses on symptom ma nagement. Increased symptoms might mean either a failure to utilize the strategy or a lack of skills for effective management.

PAGE 241

222 Table 35 Pearson Correlation Results for Recovery Strategies and Symptom Scales Effective illness management Positive future orientation Meaningfulness personal control, and hope Recognizing support Help seeking Symptom eradication Somatization -0.34 d -0.28 d -0.00 -0.21 d -0.08 -0.05 Obsessive compulsive -0.49 d -0.37 d -0.05 -0.24 d -0.15 b -0.05 Interpersonal Sensitivity -0.45 d -0.40 d -0.03 -0.29 d -0.14 b -0.01 Depression -0.56 d -0.49 d -0.05 -0.36 d -0.21 d -0.06 Anxiety -0.53 d -0.39 d -0.09 -0.28 d -0.18 c -0.07 Hostility -0.30 d -0.21 d -0.06 -0.16 b -0.12 a -0.02 Phobic Anxiety -0.47 d -0.39 d -0.14 b -0.26 d -0.14 b -0.06 Paranoia -0.44 d -0.33 d -0.07 -0.27 d -0.14 b -0.07 Psychoticism -0.42 d -0.38 d -0.05 -0.24 d -0.16 b -0.06 Global Severity Index -0.53 d -0.43 d -0.06 -0.31 d -0.17 c -0.06 Note: a p .05; b p .01; c p .001; d p .0001 The second strategy, positive future orient ation, has an identical but less intense relationship with symptoms compared to effective illness management. Disturbing, disheartening and possibly chronic sympto ms demand attention and interrupt goal making and future focus. It is logical that symptoms would disrupt attention on future goals. The third strategy (meaningfulness, personal control, and hope), however, is inconsistent with the literature with as sociations between symptom and strategy

PAGE 242

223 significant only for phobic anxiety. Recovery literature focuses a great deal on hope and often addresses, directly or obliquely, the disruptive effect of symptoms on hope and living a meaningful life (Lysaker, Buc k, Hammoud, Taylor & Row, 2006; Resnick, Rosenheck & Lehman, 2004). This result might be explained by the wording of some of the items (e.g., I have a desire to succeed; I have goals in life I want to reach) that access whether a particular thought or desire is present without any mention of whether the respondent views these as possible or whet her there are obstacles. The literature supports, and it is reasonable to assume, that a desire to succeed can exist even though symptoms are present that makes it seem impossible to achieve success. Recognizing support (strate gy 4) returns to the earlie r pattern with mild to moderate significant negative associations be tween the strategy and each symptom scale. Many individuals who are actively symptomatic feel isolated and without necessary support. If symptoms are associated with poor recognition of support then this may suggest another problem in obtaining and maintaining support for those with mental illness. Nonetheless, it cannot be dismissed that the problem, in some cases, is true abandonment by the support system for thos e who are symptomatic because of fear, disgust, or blame. The fifth strategy (h elp seeking) has nega tive though less intense associations with each symptom scale. Activ e symptoms appear to reduce help seeking behavior. Chronic or acute symptoms can in terfere with the ability to acknowledge the need for help and to follow through with a request. Finally, symptom eradication (strategy six) is not significan tly related to any symptom scale. This is rational as the belief in a symptom free life would be hard to maintain in the f ace of active symptoms.

PAGE 243

224 Other Domain 1 variables were correlated with recovery stra tegies including age of onset, lifetime hospitalizations, total years employed and familial mental illness (Table 36). Starting with age of onset, the only significant association is with symptom eradication. Specifically, late r onset is weakly associat ed with belief in symptom eradication. Having a longer peri od of time before onset provides a stronger sense of self and history that is disease free and may act as a base and a remembered time of nonillness that the person believes they will eventually see again. The number of lifetime hospitalizations is not significantly associated with any of the recovery strategies excep t for symptom eradication. Th is is unexpected considering hospitalization is viewed as a negative occurrence for many clients, though the recovery movement has consistently viewed hospitaliza tion and relapse as another part of the Table 36 Pearson Correlation Results for Recovery Strategies and Age of Onset, Lifetime Hospitalization, Years Employed, and Familial Mental Illness Effective illness management Positive future orientation Meaningfulness personal control and hope Recognizing support Help seeking Symptom eradication Age of Onset -0.04 0.02 0.02 0.00 0.02 0.13 a Hospitalizations in Lifetime -0.09 -0.08 -0.01 -0.04 0.05 -0.15 b Total years employed -0.11 a -0.12 a -0.04 -0.16 b -0.01 -0.11 a Mental illness: nuclear family -0.12 a -0.10 0.08 -0.08 -0.04 -0.18 c Mental illness: extended family -0.08 -0.10 0.04 -0.10 -0.06 -0.14 b Note: a p .05; b p .01; c p .001; d p .0001

PAGE 244

225 recovery journey. However, examining the di fference in recovery strategy endorsement between individuals who were and were not hos pitalized in the 12-months prior to data collection (Table 37) finds that effective il lness management is sanctioned less by those who were hospitalized compared to those that were not, as is positive future orientation. Thus, how recently an individual was hos pitalized may have more impact on implementation of recovery strategies then the number of hospitalizations. Total years employed correlated with r ecovery strategies provides additional counterintuitive findings (Table 36). Albeit of minimal magnitude, every strategy is negatively correlated with total years empl oyed suggesting that employment reduces commitment to recovery strategies. This is difficult to explain unless employment is indicative of some status that does not require rec overy strategies, thus the association would be non-relevant. When comparing those that are curr ently employed to those who are not on recovery strategies (Table 37) effective illness management and positive future orientation are endorsed significantly more by those that are employed. This might indicate a more specific a nd immediate association between employment and recovery strategies. Addressing Table 36 a final time, fa milial mental illness is weakly associated with two recovery strategies. Effective illness management is negatively associated with nuclear family mental illness and may be a response to poor role modeling, that is growing up with a parent or sibling that is mentally ill may not, in some case, provide direction in how to handle ones own onset of symptoms, though th is is speculative. In a similar fashion, symptom eradication is negatively associated with familial mental illness,

PAGE 245

226 Table 37 Independent Sample T-test Results for Research Strategies and Hospitalization in Last Year, Employment, Substance Abuse, Medica tion and Abuse/Assault Variables Effective illness management Positive future orientation Meaningfulness, personal control and hope Recognizing support Help seeking Symptom eradication n Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Hospitalized in last year NO 230 3.38 (0.76) 2.71 b 4.03 (0.62) 2.71 b 4.20 (0.45) -0.44 3.77 (0.94) 0.05 3.98 (0.83) -0.22 3.05 (1.22) 0.69 YES 120 3.02 (0.71) 3.83 (0.71) 4.22 (0.47) 3.77 (1.01) 4.00 (0.85) 2.87 (1.29) Currently Employed NO 276 3.25 (0.72) -2.52 b 3.92 (0.64) -2.45 b 4.19 (0.45) -1.22 3.76 (0.95) -0.65 4.01 (0.80) 0.95 3.14 (1.26) -0.87 YES 74 3.49 (0.81) 4.13 (0.70) 4.26 (0.47) 3.84 (1.03) 3.89 (0.97) 3.28 (1.19) Co-occurring substance use NO 231 3.26 (0.77) -1.64 3.91 (0.69) -2.28 a 4.20 (0.47) -0.24 3.76 (0.94) -0.23 3.98 (0.86) -0.08 3.22 (1.28) 0.94 YES 119 3.39 (0.70) 4.07 (0.57) 4.21 (0.44) 3.79 (1.02) 3.99 (0.80) 3.08 (1.18) Anti-psychotic medication NO 195 3.36 (0.81) 1.72 4.01 (0.67) 1.52 4.23 (0.45) 1.36 3.81 (1.01) 0.73 4.00 (0.85) 0.45 3.21 (1.21) 0.65 YES 155 3.23 (0.66) 3.90 (0.64) 4.17 (0.47) 3.73 (0.91) 3.96 (0.82) 3.12 (1.29)

PAGE 246

227 Table 37 (cont.) Effective illness management Positive future orientation Meaningfulness personal control and hope Recognizing support Help seeking Symptom eradication n Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Anti-depressant medication NO 84 3.51 (0.77) 2.91 b 4.18 (0.60) 3.55 c 4.22 (0.49) 0.40 3.86 (0.99) 0.91 4.01 (0.88) 0.35 3.37 (1.26) 1.67 YES 266 3.24 (0.73) 3.90 (0.66) 4.20 (0.45) 3.75 (0.96) 3.97 (0.82) 3.11 (1.23) Anti-manic medication NO 254 3.36 (0.75) 2.47 b 4.01 (0.65) 2.14 a 4.20 (0.46) 0.02 3.83 (0.94) 1.73 4.02 (0.82) 1.35 3.21 (1.24) 0.91 YES 96 3.14 (0.72) 3.84 (0.65) 4.20 (0.47) 3.63 (1.02) 3.88 (0.88) 3.07 (1.26) Anti-anxiety medication NO 167 3.53 (0.74) 5.74 d 4.05 (0.69) 2.37 a 4.21 (0.50) 0.10 3.88 (0.95) 2.03 a 4.05 (0.83) 1.37 3.29 (1.24) 1.67 YES 183 3.09 (0.69) 3.89 (0.61) 4.20 (0.42) 3.67 (0.97) 3.93 (0.84) 3.07 (1.24) Other psychotropic NO 237 3.32 (0.80) 0.91 4.00 (0.67) 1.51 4.22 (0.45) 1.19 3.79 (1.00) 0.52 3.98 (0.85) -0.20 3.17 (1.24) -0.06 YES 113 3.25 (0.62) 3.89 (0.62) 4.16 (0.48) 3.73 (0.90) 4.00 (0.81) 3.18 (1.26) Child sexual abuse NO 203 3.42 (0.78) 3.48 c 4.02 (0.67) 1.85 4.18 (0.48) -1.33 3.82 (0.94) 1.12 4.00 (0.87) 0.38 3.30 (1.21) 2.38 a YES 147 3.14 (0.68) 3.89 (0.63) 4.24 (0.43) 3.71 (0.99) 3.96 (0.80) 2.99 (1.27)

PAGE 247

228 Table 37 (cont.) Effective illness management Positive future orientation Meaningfulness personal control and hope Recognizing support Help seeking Symptom eradication n Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Mean (s.d.) t-test Child physical abuse NO 213 3.35 (0.77) 1.57 3.99 (0.65) 0.90 4.19 (0.46) -0.62 3.83 (0.96) 1.32 3.99 (0.82) 0.28 3.21 (1.24) 0.74 YES 137 3.22 (0.70) 3.93 (0.65) 4.22 (0.46) 3.69 (0.97) 3.97 (0.86) 3.11 (1.26) Adult sexual assault NO 234 3.39 (0.76) 3.18 c 4.00 (0.67) 1.47 4.21 (0.46) 0.19 3.82 (0.95) 1.34 4.01 (0.83) 0.92 3.30 (1.24) 2.85 b YES 116 3.12 (0.68) 3.89 (0.62) 4.20 (0.46) 3.68 (0.99) 3.93 (0.85) 2.91 (1.22) Adult physical assault NO 212 3.35 (0.76) 1.40 3.97 (0.68) 0.16 4.19 (0.47) -0.81 3.86 (0.91) 2.17 a 4.00 (0.83) 0.32 3.28 (1.25) 2.00 a YES 138 3.23 (0.73) 3.96 (0.62) 4.23 (0.45) 3.63 (1.02) 3.97 (0.86) 3.01 (1.22) Note: a p .05; b p .01; c p .001; d p .0001

PAGE 248

229 though this time with extended family as well as nuclear family. This suggests that belief in symptom eradication decreases with an increase in familial mental illness. Being diagnosed with a co-occurring substance use disorder is associated with greater endorsement of rec overy strategies for those with the substance diagnosis (Table 37). However, this is significant only for th e strategy of positive future orientation. This may reflect the emphasis on recovery that is considerably more established in substance abuse intervention than mental illness and may also indicate commonalities between mental health and substance abuse recover y, though there is littl e empirical research targeting both forms of recovery in comparison. Symptom eradication, though not significant, is the one strategy that those who do not have a substance abuse diagnosis sanction to a greater degree. This remains c onsistent with substance abuse recovery since the disease is considered lif e long and eradication is not one of the beliefs or goals, though management of the primary symptom, using substances, is a goal. The association between medication categ ory and recovery strategies indicates that those not taking medication have a higher mean score on almost every recovery strategy (Table 37). This does not reach sta tistical significance in each instance but the trend is consistent. Anti-psychotic and th e category of other psychotropic medication are not significantly related to any recove ry strategy. Non-use of anti-depressant medication is significant for the strategy of effective illn ess management. Clinically, adherence to a medication regimen is a very common, if not the most common, mode of symptom and disease management. This asso ciation, however, indicat es that effective illness management does not rely on medication. The simplistic bivariate relationship dos not offer the opportunity to assess the di rection of the relationship and the pattern

PAGE 249

230 noted could indicate that individuals with higher levels of symptoms require medication and that this is simply a proxy of the re lationship noted in Table 35 wherein every symptom scale was negatively associated with each recovery stra tegy. The pattern of non-medication use for effective illness management reaches significance for anti-manic medication and anti-anxiety medication as well. Medication use is associated with lowe r mean scores on recovery strategy 2, positive future orientation (Table 37). For anti-depressant medication, a significant association is noted, as it is for anti-man ic medication and anti-anxiety medication. The strategies of meaningfulness, personal control and hope (strategy 3), help seeking (strategy 4) and symptom eradication (strate gy 6) are not significantly associated with any medication category. Recognizing support (s trategy 4) is signifi cantly associated with anti-anxiety medication only. The last domain 1 variables to be investig ated are the abuse related variables. The trend is similar to the one noted in medication use; abuse survivors have lower scores, in general, on recovery strategies, and this rela tionship is significant in a few cases (Table 37). It is interesting that the recovery stra tegy of meaningfulness, personal control, and hope has the only negative associations. Though not significant, this means that those who did experience abuse, specifically child sexual abuse, child physic al abuse, and adult physical assault, have higher mean scores for strategy endorsement. This may reflect the clinical intervention typical for this population: the focus on finding hope, regaining personal control and building a life filled with meaning (Alexander, Muenzenmaier, Dumont, & Auslander, 2005). Survivors of child sexual abuse (CSA) have a significantly lower mean score compared to those who did not experience CSA on effective illness

PAGE 250

231 management. Symptom eradication is also significantly associated with CSA with CSA survivors sanctioning eradication less than th e non-abused. Child physical abuse is not significantly related to any of the rec overy strategies. Adult sexual assault (ASA) is significantly associated to the same two strategies as CSA: effective illness management a nd symptom eradication. The direction is identical for both associations, with survivors scoring lower means on both strategies. This may reflect some common underlying process since the correlation between experiencing CSA and ASA is significant, though only moderate in magnitude (r(348) = 0.47, p .0001). Adult physical assa ult is significant in asso ciation with recognition of support and symptom eradication. Symptom erad ication, thus, is significant for three of four abuse variables. Research question 3.2: Are social factors associated with recovery expectancy? The three social constructs or su b-domains (stigma, social support, and empowerment) are investigated for patterns of association with the recovery strategies using Pearson correlations. The results are displayed in Table 38. With only two exceptions all stigma subscales are negatively associated with recovery strategies. The magnitude of associations is low to moderate with the strongest correlations noted for effective illness management and positive future orientation across stigma subscales. Symptom eradication is the one subscale not significantly associat ed with any of the stigma subscales while the other five recovery strategies are significantly associated with every stigma subscale. This, as noted, is possibly due to symptom eradication being a

PAGE 251

232 Table 38 Pearson Correlation Results for Recovery Strategies and Domain 2 Variables (Stigma, Social Support, and Empowerment) Effective illness management Positive future orientation Meaningfulness, personal control and hope Recognizing support Help seeking Symptom eradication Stigma Alienation -.57 d -.47 d -.25 d -.43 d -.26 d -.05 Stereotype endorsement -.35 d -.33 d -.33 d -.22 d -.18 c .04 Discrimination occurrence -.44 d -.35 d -.16 d -.30 d -.18 c -.06 Social withdrawal -.52 d -.52 d -.26 d -.38 d -.25 d -.06 Stigma resistance -.55 d -.49 d -.39 d -.40 d -.32 d .02 Social Support Partner/best friend support .37 d .42 d .25 d .58 d .35 d .06 Support via family .26 d .32 d .18 c .53 d .20 d .05 Support via provider .24 d .34 d .26 d .28 d .34 d .11 a Support via friends .32 d .37 d .27 d .48 d .31 d .06 Community involvement .39 d .42 d .27 d .20 d .28 d .16 b Trust -.08 .03 .08 .08 .08 -.07

PAGE 252

233 Table 38 (cont.) Effective illness management Positive future orientation Meaningfulness, personal control and hope Recognizing support Help seeking Symptom eradication Empowerment Self-esteem/self-efficacy .63 d .73 d .34 d .45 d .36 d .18 c Power/powerlessness .27 d .31 d .07 .22 d .16 b -.17 c Community activism and autonomy .25 d .33 d .46 d .22 d .13 b .08 Optimism and control over future .46 d .45 d .39 d .39 d .22 d .23 d Righteous anger .01 .01 -.07 -.07 -.07 -.13 a Note: a p .05; b p .01; c p .001; d p .0001

PAGE 253

234 single item measuring a narrow recovery strate gy and the other strategies are composite variables that measure broader concepts. Social support, community connectedne ss and trust are, for the most part, positively associated with recovery strate gies (Table 38). Recognition of support (strategy 4) is significantly associated with increased strategy endorsement in all cases except for trust. All associations are mild to moderate in magnitude. The strongest associations are between the strategies of recognizing support and positive future orientation with social support variables. C onsidering that it may require the recognition of support to benefit from support, the magn itude of the associat ions are surprisingly weak since they are quit e conceptually linked. Though support is necessary for immediate needs it also is required for futu re needs and goals, thus the magnitude of these associations match expectations. Recove ry strategies are con ceptualized as active concepts, requiring behavioral as well as cognitive and emotional investment. Community involvement also requires action a nd is significantly asso ciated with all the recovery strategies. The last set of subscales addresses empow erment. In most cases, empowerment is associated with increased sanction of recovery strategies (Table 38). This is particularly true for self-esteem/self-efficacy and optimism and control over the future. Self-esteem is highly associated with effective illness ma nagement and positive future orientation. As noted previously, the cau sality of the association cannot be determined, so it is unclear whether effective illness management leads to greater self-esteem or vice versa. An additional possibility is that some other fact or is driving the rela tionship, which, however, is unsupported in this study, considering the link between self-esteem and illness

PAGE 254

235 management established in the recovery literature (Rusch, Lieb, Bohus, & Corrigan, 2006; Shahar & Davidson, 2003). The absence of association is sometimes relevant and, in the case of meaningfulness, person al control and hope (strategy 3) and power/powerlessness, this is tr ue. The expectation is that personal experience of power would associate positively with control and mean ing. It is unclear w hy this association is absent. Research question 3.3: Are service factors associated with recovery expectancy? Service factor Pearson correlations with r ecovery strategies are provided in Table 39. Total number of services is not significan tly associated with any recovery strategy. Total contact hours is significantly associated with positive future orientation and recognizing support. This indicates that a pos itive future orientati on decreases with the number of contact hours. Individuals with chronic mental i llness are often counseled to limit their lives to manage stress and thus re duce exacerbation of symptoms and this may Table 39 Pearson Correlation Results for Recovery Strategies and Domain 3 Service Variables Effective illness management Positive future orientation Meaningfulness, personal control and hope Recognizing support Help seeking Symptom eradication Total number of services .01 -.07 .02 .08 .04 -.09 Total contact hours .03 -.13 a -.03 .13 a .00 -05 Average satisfaction .31 d .35 d .13 a .40 d .30 d .01 Note: a p .05; b p .01; c p .001; d p .0001

PAGE 255

236 be reflected here. In cont rast, recognition of support in creases with contact hours, possibly reflecting the benefits of clinical intervention. Individuals with serious mental illness of ten appreciate the attention and nurturing support of providers and recognize this as supporting. Average satisfaction demonstrates the most consistent and significant relationshi p with recovery strategies. Satisfaction with services is positively corre lated with all recovery strategi es and is significant for all but symptom eradication. Client satisfaction is often a combination of adequate service and rapport. This is a positive finding for th e interface of recovery and clinical services. A summary of bivariate support for each of the hypothese s by recovery strategy is provided in Table 40. Similar to Table 32, this display is meant to provide a summary of two-variable covariance and is not meant as a final word on support for the hypotheses. This concludes bivariate analyses. Attention is next turned to multivariate analysis to more directly explore resear ch questions two through five.

PAGE 256

237 Table 40 Summary of Bivariate Results for Research Question 3 Hypotheses Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom eradication Hypothesis Statement Supported? Supported? Supported? Supported? Supported? Supported? Hyp3.1a Less severe diagnoses will be associated with higher recovery expectancy. Hyp3.1b Lower somatization symptoms reported by participants will be associat ed with higher recovery expectancy. Hyp3.1c Lower obsessive-compulsive symptoms reported by participants will be associated with higher recovery expectancy. Hyp3.1d Lower interpersonal sensitivity reported by participants will be associat ed with higher recovery expectancy. Hyp3.1e Lower symptoms of anxiety reported by participants will be associat ed with higher recovery expectancy. Hyp3.1f Lower symptoms of depression reported by participants will be associat ed with higher recovery expectancy. Hyp3.1g Lower hostility reported by participants will be associated with higher recovery expectancy. Hyp3.1h Lower phobic anxiety reported by participants will be associated with high er recovery expectancy. Hyp3.1i Lower paranoid ideation reported by participants will be associated with higher recovery expectancy.

PAGE 257

238 Table 40 (cont.) Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom eradication Hypothesis Statement Supported? Supported? Supported? Supported? Supported? Supported? Hyp3.1j Lower psychoticism reported by participants will be associated with high er recovery expectancy. Hyp3.1k Absence of a comorbid substance use disorder will be associated with higher recovery expectancy for participants. Hyp3.1l Lower numbers of lifetime psychiatric hospitalizations will be associated with higher recovery expectancy for participants. Hyp3.1m No psychiatric hospitalization in the last year will be associated with higher recovery expectancy for participants. Hyp3.1n Older age at onset of disorder will be associated with higher recovery expectancy for participants. Hyp3.1o Absence of prescribed anti-depressant medication will be associated with higher recovery expectancy for participants. Hyp3.1p Absence of prescribed anti-psychotic medication will be associated with higher recovery expectancy for participants. Hyp3.1q Absence of prescribed anti-mania medication will be associated with higher recovery expectancy for participants. Hyp3.1r Absence of prescribed anti-anxiety medication will be associated with higher recovery expectancy for participants.

PAGE 258

239 Table 40 (cont.) Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom eradication Hypothesis Statement Supported? Supported? Supported? Supported? Supported? Supported? Hyp3.1s Absence of any other prescribed psychotropic medication will be associated with higher recovery expectancy for participants. Hyp3.1t Being currently employed will be associated with higher recovery expectancy for participants. Hyp3.1u Greater number of years worked will be associated with higher recovery expectancy for participants. Hyp3.1v Absence of familial mental illness in first-degree relatives will be associated with higher recovery expectancy for participants. Hyp3.1w Absence of familial mental illness in extended family members will be associated with higher recovery expectancy for participants. Hyp3.1x Absence of child sexual abuse will be associated with higher recovery expectancy for participants. Hyp3.1y Absence of child physical abuse will be associated with higher recovery expectancy for participants. Hyp3.1z Absence of adult sexual assault will be associated with higher recovery expectancy for participants. Hyp3.1aa Absence of adult physical assault will be associated with higher recovery expectancy for participants. Hyp3.3a Lower feelings of alienation will be associated with higher recovery expectancy for participants. Hyp3.3b Lower respondent endorsement of mental illness stereotypes will be associat ed with higher recovery expectancy for participants.

PAGE 259

240 Table 40 (cont.) Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom eradication Hypothesis Statement Supported? Supported? Supported? Supported? Supported? Supported? Hyp3.3c Lower number of discrimination experiences will be associated with higher recovery expectancy for participants. Hyp3.3d Lower endorsement of soci al withdrawal will be associated with higher recovery expectancy for participants. Hyp3.3e Greater endorsement of stigma resistance will be associated with higher recovery expectancy for participants. Hyp3.3f Greater support through intimate partner or a best friend will be associated with higher recovery expectancy for participants. Hyp3.3g Greater support through family members will be associated with higher recovery expectancy for participants. Hyp3.3h Greater support via mental health providers will be associated with higher recovery expectancy for participants. Hyp3.3i Greater support through friendship will be associated with higher recovery expectancy for participants. Hyp3.3j Greater feelings of being connected to the community will be associat ed with higher recovery expectancy for participants.

PAGE 260

241 Table 40 (cont.) Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom eradication Hypothesis Statement Supported? Supported? Supported? Supported? Supported? Supported? Hyp3.3k Greater trust in the motivation of others will be associated with higher recovery expectancy for participants. Hyp3.3l Higher ratings of self-est eem and self-efficacy will be associated with higher recovery expectancy for participants. Hyp3.3m Higher ratings of personal power will be associated with higher recovery expectancy for participants. Hyp3.3n Greater involvement in the community will be associated with higher recovery expectancy for participants. Hyp3.3o Greater confidence in pe rsonal control over the future will be associated with higher recovery expectancy for participants. Hyp3.3p Higher ratings of righteous anger will be associated with higher recovery expectancy for participants. Hyp3.3a Total number of services will be associated with higher recovery expectancies for participants. Hyp3.3b The average number of contact hours per month will be associated with higher recovery expectancies for participants. Hyp3.3c Satisfaction level with services will be associated with higher recovery exp ectancies for participants.

PAGE 261

242 Multivariate Analysis An assessment of multicollinearity was completed prior to following the analysis plan outlined below. Multicollinearity was a ssessed with the tolerance statistic and one variable violated the preset cutpoint. Th e Global Severity Index (GSI) of the Symptom Checklist 90-Revised was be low the cutpoint of .20 and was dropped from further analyses. The GSI tolerance was .08, indica ting that it was highly correlated with, most likely, other symptom variables. Modifications to the Analysis Plan There were two general modifications to the analysis plan. Previously noted in Chapter 3, when attempting to generate th e composite dependent variable based on number of recovery strategies endorsed per respondent for probit analysis, too many combinations of recovery strategies were sanctioned. Sixty-th ree of a possible 66 combinations of recovery strategies were endorsed by one or more respondents. This indicates a wide variance in individual st rategy use, supporting the common statement in the consumer driven recovery literature that recovery is an idiosyncratic and personal process. From a pragmatic perspective, a de pendent variable with 63 levels was judged too imprecise and unwieldy to generate us able information and probit analysis was dropped from further consideration. The final modification included droppi ng two variables from the domain three service factors. The variables dropped were: (1) the contribution of services to recovery, and (2) the housing variable. The contribution to recovery variable was a dichotomous (yes/no) statement of whether services in gene ral contributed to rec overy. Nearly 99% of the respondents answered affirmatively. When considering the housing variable,

PAGE 262

243 approximately 94% of the respondents are li ving independently, again limiting variance, influencing the decision to drop the variable from further analysis. Overview This study focused on five overarching research questions that examine (1) whether there is a difference in recovery expectancy between the two samples (SMI and OP), targeted above in bivariate analysis, (2 ) what factors are asso ciated with recovery expectancy; (3) what factors are associated with each of the six defined recovery strategies; (4) whether recove ry expectancy mediates or moderates the relationship between independent variable domains 1-3 (c linical, social, and se rvice) and recovery strategies, and (5) whether membership in the SMI or OP samples (a proxy for severity of mental illness) moderates the relationship between variable domains 1-3 and recovery strategies. Research questions 2 and 3 are examined via a sequence of statistical models beginning with control variables, then domain 1 variables with accompanying hypotheses, followed by domain 2, and finishi ng with domain 3. Next, the effects of diagnostics are detailed follo wed by discussion of the full or main effects model. A restricted model containing only the indepe ndent variables that were significantly associated with the dependent variable for each research question assesses the degree in which non-significant variables co ntribute to the variance expl ained. This assumes that there are significant main effect results. A final table summarizing the results of the analysis for each dependent variable is provided for easier comparison across the different models.

PAGE 263

244 The same sequence noted above is used to examine the mediating and moderating effect of recovery expectancy, targeting rese arch questions 4a and 4b. Research question 5 also uses the same progression to exam ine moderating effects of illness severity. Focus on Recovery Expectancy Recovery expectancy, a dichotomous fo rced choice variable asking whether the respondent expects to recover from mental il lness or not, was examined using logistic regression. Research question 2, sub-quest ions 2.1-2.3, and hypotheses 2.1a-2.3c are evaluated in this section (see pages 107-111, Chapter 3). The logistic regression analysis was modeled on those who endorse recovery exp ectancy. The general goal was to locate factors associated with belief in recovery in hope that this information will assist in creating programs that enhance recovery e xpectancy and the active use of recovery strategies. Using the R-squa re option in SAS software, th e R-square was calculated to provide an estimate of variance explained fo r each model. The option was included to better estimate the usefulness of the results and to maintain consistency with OLS regression results addressi ng research questions 3-5. The control variables of age, gender, income, and education were entered simultaneously in a logistic regression model with the dichotomous dependent variable of recovery expectancy. The results are summarized in Table 41. Gender, income and education are not significantly related to recovery expect ancy, though income approaches significance. This suggests that higher income may be associ ated with slightly higher expectancy. Age is significan tly related with an odds ra tio (OR) of 0.97. The estimate for age is -0.03, indicating that the pr obability of positively endorsing recovery expectancy decreases by .03 w ith each increase in age. Thus, younger age is mildly

PAGE 264

245 related to increased recovery expectancy. This ma y indicate that younger respondents have less experience with both il lness and the system and thus have greater expectation of recovery, a statement reinforced in results from a pilot study for this research (Walby, 2003a). Table 41 Control Variable Logistic Regression Results Variable Odds Ratio (95% CI) p-value Age 0.97 (0.95-0.99) .01 Gender 0.87 (0.54-1.40) ns Income 1.16 (0.95-1.43) ns Education 1.02 (0.80-1.30) ns Note: n = 350; R 2 = .03 Research Question 2.1: Are Individual Fact ors Associated with Recovery Expectancy? Hypotheses 2.1a-2.1aa, pertai ning to domain 1 clinica l/historical factors, are addressed in isolation at this time. The resu lts of the logistic analysis are summarized in Table 42. The full model results at the end of this section provide a more comprehensive response. Lower scores on the depression scale ar e associated with increased recovery expectancy. The estimate for depression is 0.62, meaning there is a probability decrease in recovery expectancy of 0.62 for each incr ease in depression score. This mirrors previous findings of high depression levels associated with poor recovery outcomes (Corrigan & Ralph, 2005). Not being prescrib ed an antianxiety medication increases the odds of endorsing recovery, with an estimate of -0.63. Hence, being prescribed an

PAGE 265

246 anti-anxiety medication decrease s the probability of endorsi ng recovery by 0.63. This suggests a link between anxiety and expectancy though it should be noted that none of Table 42 Domain 1 (Clinical and Historical Fa ctors) Logistic Regression Results H 1 Variable Odds Ratio (95% CI) pvalue Age 0.99 (0.96-1.02) ns Gender 1.21 (0.64-2.31) ns Income 1.09 (0.85-1.39) ns Education 1.17 (0.82-1.52) ns 2.1b Somatization 0.83 (0.55-1.26) ns 2.1c Obsessive-compulsive 1.07 (0.63-1.81) ns 2.1d Interpersonal sensitiv ity 1.13 (0.84-2.42) ns 2.1e Anxiety 0.93 (0.55-1.62) ns 2.1f Depression 0.54 (0.31-0.94) .05 2.1g Hostility 1.19 (0.82-1.71) ns 2.1h Phobic anxiety 0.83 (0.56-1.23) ns 2.1i Paranoia 0.97 (0.60-1.56) ns 2.1j Psychoticism 0.77 (0.47-1.25) ns 2.1k Substance abuse history 0.71 (0.41-1.23) ns 2.1l Hospitalization history 0.98 (0.96-1.00) ns 2.1m Hospitalized in last year 1.36 (0.76-2.46) ns 2.1n Age of onset 0.99 (0.97-1.01) ns 2.1o Medication: Anti-depressant 1.54 (0.77-3.09) ns 2.1p Medication: Anti-psychotic 0.82 (0.47-1.43) ns 2.1q Medication: Anti-manic 0.71 (0.40-1.27) ns 2.1r Medication: Anti-anxiety 0.53 (0.30-0.93) .05 2.1s Medication: Other psyc hotropic 0.81 (0.46-1.40) ns 2.1t Current employment 0.93 (0.45-1.93) ns 2.1u Years of employment 0.99 (0.97-1.02) ns 2.1v Nuclear family mental illness 0.93 (0.75-1.15) ns 2.1w Extended family mental illness 0.92 (0.61-1.38) ns 2.1x Child sexual abuse history 0.92 (0.49-1.70) ns 2.1y Child physical abuse history 0.98 (0.54-1.77) ns 2.1z Adult sexual assault history 0.65 (0.32-1.33) ns 2.1aa Adult physical assault history 1.52 (0.79-2.91) ns Note: n = 350; R 2 = .20

PAGE 266

247 the anxiety symptom scales were significan t in this analysis and that psychiatric medications are often prescribed for reasons other then which they were originally developed. This domain explains approximately one-f ifth of the variance of the expectancy construct (R 2 = 0.20). Variance explained is affecte d, in part, by the number of variables entered into the model. In this case, thirty variables were entered and two were significantly associated with recovery expectancy. Research Question 2.2: Are Social Facto rs Associated with Recovery Expectancy? This section targets hypoth eses 2.2a-2.2q, again in isolation, to better understand the independent contribution of social factor s to recovery expectancy. The results are presented in Table 43. None of the indepe ndent variables are si gnificantly associated with recovery expectancy. Age, a control variable, is significantly associated with expectancy in this model and has the same direction and approximate magnitude as age did in the control variable model. It a ppears that social factors alone have limited association with expectancy. The model does explain approximately 14% of the variance without any clear influence by s ubgroups of social factors (e.g. stigma, social support, or empowerment). When the literatures focus on social factors in relation to recovery is considered, the lack of association between social factors and expectancy is puzzling. Research Question 2.3: Are Service Factors Associated with Recovery Expectancy ? The relationship between service factors and recovery expectancy is summarized in Table 44, targeting hypotheses 2.3a-2.3c. Service factors captur e relatively little of the expectancy variance (R 2 = 0.05), though age and average satisfaction with services are

PAGE 267

248 Table 43 Domain 2 (Social Factors) Logistic Regression Results H 1 Variable Odds Ratio (95% CI) p-value Age 0.97 (0.95-0.99) .05 Gender 1.06 (0.62-1.80) ns Income 1.09 (0.87-1.37) ns Education 0.95 (0.72-1.25) ns Stigma 2.2a Alienation 0.56 (0.29-1.11) ns 2.2b Stereotype endorsement 0.60 (0.29-1.27) ns 2.2c Discrimination experience 0.64 (0.34-1.20) ns 2.2d Social withdrawal 2.23 (0.97-5.09) ns 2.2e Stigma resistance 0.88 (0.43-1.78) ns Social Support and Related 2.2f Partner or best friend support 1.08 (0.81-1.44) ns 2.2g Family support 1.17 (0.98-1.39) ns 2.2h Provider support 0.86 (0.62-1.19) ns 2.2i Friend support 1.05 (0.78-1.41) ns 2.2j Community involvement 1.07 (0.84-1.35) ns 2.2k Trust 0.95 (0.73-1.24) ns Empowerment 2.2l Self-esteem/self-efficacy 1.70 (0.84-3.44) ns 2.2m Power/powerlessness 1.44 (0.72-2.87) ns 2.2n Community activism and autonomy 0.79 (0.39-1.59) ns 2.2o Optimism and control over the future 1.07 (0.54-2.13) ns 2.2p Righteous anger 1.33 (0.81-2.17) ns Note: n = 350; R 2 = .14 both significantly associated with expectan cy. Age follows the previous pattern of borderline significant, with younger age associ ated with expectancy endorsement. Average satisfaction with services is positively associated with expectancy, with greater treatment or service sa tisfaction increasing the odds of endorsing recovery expectancy by 34% compared to those with less satisfaction with services. It is interesting that the

PAGE 268

249 subjective satisfaction scale is more highly asso ciated with recovery expectancy then the objective scales measuring actual clinical and support contacts, favoring quality over quantity. Table 44 Domain 3 (Service Factors) Logistic Regression Results H 1 Variable Odds Ratio (95% CI) p-value Age 0.97 0.95-0.99 .01 Gender 1.06 0.62-1.80 ns Income 1.09 0.87-1.37 ns Education 0.95 0.72-1.25 ns 2.3a Total number of services 0.87 0.69-1.09 ns 2.3b Total service hours per month 1.00 0.97-1.04 ns 2.3c Average satisfaction score 1.34 1.07-1.68 .01 Note: n = 350; R 2 = .05 Research Question 2: Are Individual, Soc ial or Service Factors Associated with Recovery Expectancy?: Main Effects Model As discussed in Chapter 3, diagnostic st atistics (e.g., hat valu es, Cooks D) were used to detect influential observations. If two of three diagno stic tests were violated then an observation was removed from main effect analysis. A total of eight observations violated this criteria and we re dropped from the analysis presented next. Logistic regression results for control variables and the three independent variable domains are presented in Table 45. Immediatel y apparent is the lack of a ny significant associations in the full model, although approximately 25% of the variance is explai ned. Several of the variables are near significance and may exert greater influence on recovery expectancy

PAGE 269

250 Table 45 Main Effects: Logistic Regression Results H 1 Variable Odds Ratio (95% CI) p-value Age 0.98 (0.95-1.02) ns Gender 1.12 (0.56-2.25) ns Income 1.10 (0.83-1.44) ns Education 1.15 (0.81-1.63) ns Domain 1 2.1b Somatization 0.75 (0.46-1.21) ns 2.1c Obsessive-compulsive 1.10 (0.60-2.03) ns 2.1d Interpersonal sensi tivity 1.78 (0.98-3.18) ns 2.1e Anxiety 0.82 (0.45-1.50) ns 2.1f Depression 0.60 (0.31-1.16) ns 2.1g Hostility 1.16 (0.78-1.74) ns 2.1h Phobic anxiety 0.82 (0.52-1.29) ns 2.1i Paranoia 1.09 (0.62-1.90) ns 2.1j Psychoticism 0.80 (0.46-1.38) ns 2.1k Substance abuse history 0.66 (0.35-1.25) ns 2.1l Hospitalization history 0.98 (0.95-1.00) ns 2.1m Hospitalized in last year 1.52 (0.79-2.91) ns 2.1n Age of onset 1.01 (0.98-1.04) ns 2.1o Medication: Anti-depressant 1.31 (0.59-2.87) ns 2.1p Medication: Anti-psychotic 0.82 (0.43-1.57) ns 2.1q Medication: Anti-manic 0.80 (0.41-1.54) ns 2.1r Medication: Anti-anxiety 0.59 (0.32-1.10) ns 2.1s Medication: Other ps ychotropic 0.87 (0.47-1.60) ns 2.1t Current employment 0.85 (0.38-1.92) ns 2.1u Years of employment 1.00 (0.97-1.03) ns 2.1v Nuclear family mental illness 0.89 (0.70-1.13) ns 2.1w Extended family mental illness 0.88 (0.56-1.38) ns 2.1x Child sexual abuse history 1.02 (0.52-2.02) ns 2.1y Child physical abuse history 1.05 (0.54-2.03) ns 2.1z Adult sexual assault history 0.73 (0.34-1.58) ns 2.1aa Adult physical assault history 1.30 (0.63-2.65) ns Domain 2 Stigma 2.2a Alienation 0.48 (0.21-1.11) ns 2.2b Stereotype endorsement 0.45 (0.18-1.11) ns 2.2c Discrimination experience 0.80 (0.38-1.70) ns 2.2d Social withdrawal 2.48 (0.97-6.37) ns 2.2e Stigma resistance 1.06 (0.44-2.52) ns

PAGE 270

251 Table 45 (cont.) H 1 Variable Odds Ratio (95% CI) p-value Social Support and Related 2.2f Partner or best friend support 1.10 (0.78-1.55) ns 2.2g Family support 1.11 (0.89-1.38) ns 2.2h Provider support 0.86 (0.55-1.35) ns 2.2i Friend support 0.97 (0.68-1.36) ns 2.2j Community involvement 1.16 (0.87-1.55) ns 2.2k Trust 1.10 (0.78-1.55) ns Empowerment 2.2l Self-esteem/self-efficacy 0.97 (0.41-2.31) ns 2.2m Power/powerlessness 1.51 (0.63-3.59) ns 2.2n Community activism and autonomy 1.17 (0.51-2.67) ns 2.2o Optimism and control over the future 0.86 (0.38-1.94) ns 2.2p Righteous anger 1.38 (0.76-2.49) ns Domain 3 2.3a Total number of services 0.97 (0.71-1.34) ns 2.3b Total service hours per month 1.00 (0.96-1.05) ns 2.3c Average satisfaction score 1.26 (0.92-1.74) ns Note: n = 342; R 2 = .25 (e.g., interpersonal sensitivity and age of ons et). These might be considered starting points for future investigation of recovery expectancy. It appears that the variables chosen for this analysis are not, in gene ral, sufficiently relevant to expectancy. It is important to consider the variable s that were significant in the analysis of individual domains that were no longer significant in the fu ll model. Further, other variables approached significance in the full model that were clea rly unrelated in the individual domain analysis. Table 46 summarizes the models and provides an overview of the impact of each independent variable in association with expectancy. The single most consistent variable a ppears to be age, with younger respondents more likely to

PAGE 271

252 Table 46 Summary Table: Logistic Regressi on Results for Research Question 2 Model 1 Model 2 Model 3 Model 4 Model 5 Age Sex Income Education Psychosomatic Obsessive Compulsive Interpersonal Sensitivity Anxiety Depression Hostility Phobia Paranoia Psychoticism Substance History Lifetime Hospitalizations Hospitalized in Last Year Age of Onset Anti-Depressant Medication Anti-Psychotic Medication Anti-Manic Medication Anti-Anxiety Medication Other Psychotropic Medication Currently Employed Years Employed Lifetime Nuclear Family Mental Illness Extended Family Mental Illness Child Sexual Abuse Child Physical Abuse Adult Sexual Assault Adult Physical Assault Alienation Stereotype Endorsement Discrimination Occurrence Social Withdrawal Stigma Resistance Partner of Best Friend Support Family Support Provider Support Friends Support

PAGE 272

253 Table 46 (cont.) Community Involvement Trust Self-Esteem and Self-Efficacy Power and Powerlessness Community Activism and Autonomy Optimism and Control Over the Future Righteous Anger Total Number of Services Total Contact Hours Average Satisfaction M1 = Model 1: n = 350; Control variables M2 = Model 2: n = 350; Control variab les + Individual and historical factors M3 = Model 3: n = 350; Contro l variables + Social factors M4 = Model 4: n = 350; Control variables + Service factors M5 = Model 5; n = 342; Full (main effects) model (post diagnostics) sanction recovery. Depression and anti-a nxiety medication, signi ficant in domain 1 analysis, are no longer significant in the main effects model. Similarly, the domain 3 service variable average satisfac tion with services is no longer significant in main effects. Joining interpersonal sensitivity and age of onset in approaching significance are hospitalization history, years of employment, social withdr awal, and total service hours per month. The amount of variance explained in models for each individual domain and together in main effects is displayed in Fi gure 3. The most influential model for this analysis is the clinic al/historical, domain 1, variable s. This suggests that altering recovery expectancy may begin with clinical factors (e.g., seve rity of symptoms, types of medication prescribed), historical factors (e.g., age when symptoms began, number of

PAGE 273

Figure 3 Variance Explained by Each Domain of Recovery Expectancy Control Variables (R2 = .03) Domain 1: Clinical and Historical Factors (R2 = .20) Full Model R2 = .25 Recovery Expectancy Domain 3: Service Factors (R 2 = .05) Domain 2: Social Factors (R 2 = .14) hospitalizations, when the latest hospitalization occurred), and possibly cognitive or intrapsychic factors that were not investigated in this study. Social factors (domain 2) apparently cont ribute to expectancy as well, though this is through the combination of multiple variables in this domain. Social withdrawal is the only social variable that a pproaches significance. Number of contact hours per month also approaches significance (domain 3). A fu ture analysis will explore whether type of service (e.g., case management, supported liv ing) are associated with expectancy. 254

PAGE 274

255 The series of logistic regressions presen ted were targeted to respond to research question 2 and the individual hypotheses linked to the research question. Due to the large number of hypotheses it was more comprehensible and less cumbersome to explain each model without directly referring to each hypothesis in turn. However, it is important to directly address whether each hypothesis was supported or not. Thus, Table 47 summarizes whether sufficient support was f ound for each hypothesis in the same fashion that Table 33 addressed the questions, though at the bivariate level. As can be seen, no hypothesis was supported fo r research question 2. Table 47 Summary of Bivariate Results for Research Question 2 Hypotheses Hypothesis Statement Supported? Hyp2.1a The less clinically severe the pa rticipants diagnosis the more recovery expectancy will increase. Hyp2.1 Lower somatization symptoms reported by participants will be associated with higher recovery expectancy. Hyp2.1c Lower obsessive-compulsive symptoms reported by participants will be associated with higher recovery expectancy. Hyp2.1d Lower interpersonal sensitivity reported by participants will be associated with higher recovery expectancy. Hyp2.1e Lower symptoms of anxiety reported by participants will be associated with higher recovery expectancy. Hyp2.1f Lower symptoms of depression reported by participants will be associated with higher recovery expectancy. Hyp2.1g Lower hostility reported by particip ants will be associated with higher recovery expectancy. Hyp.1h Lower phobic anxiety reported by participants will be associated with higher recovery expectancy. Hyp2.1i Lower paranoid ideation reported by participants will be associated with higher recovery expectancy. Hyp2.1j Lower psychoticism reported by participants will be associated with higher recovery expectancy. Hyp2.1k Absence of a comorbid substance use disorder will be associated with higher recovery expectancy for participants. Hyp2.1l Lower numbers of lifetime psychiatric hospitalizations will be associated with higher recovery expectancy for participants.

PAGE 275

256 Table 47 (cont.) Hypothesis Hypothesis Supported? Hyp2.1m No psychiatric hospitalization in the last year will be associated with higher recovery expectancy for participants. Hyp2.1n Older age at onset of disorder will be associated with higher recovery expectancy for participants. Hyp2.1o Absence of prescribed anti-depressant medication will be associated with higher recovery expectancy for participants. Hyp2.1p Absence of prescribed anti-psychotic medication will be associated with higher recovery expectancy for participants. Hyp2.1q Absence of prescribed anti-mania medication will be associated with higher recovery expectancy for participants. Hyp2.1r Absence of prescribed anti-anxiety medication will be associated with higher recovery expectancy for participants. Hyp2.1s Absence of any other prescribed psychotropic medication will be associated with higher recovery expectancy for participants. Hyp2.1t Being currently employed will be associated with higher recovery expectancy for participants. Hyp2.1u Greater number of years worked will be associated with higher recovery expectancy for participants. Hyp2.1v Absence of familial mental illness in first-degree relatives will be associated with higher recovery expectancy for participants. Hyp2.1w Absence of familial mental illness in extended family members will be associated with higher recovery expectancy for participants. Hyp2.1x Absence of child sexual abuse w ill be associated with higher recovery expectancy for participants. Hyp2.1y Absence of child physical abuse w ill be associated with higher recovery expectancy for participants. Hyp2.1z Absence of adult sexual assault will be associated with higher recovery expectancy for participants. Hyp2.1aa Absence of adult physical assault w ill be associated with higher recovery expectancy for participants. Hyp2.2a Lower feelings of alienation will be associated with higher recovery expectancy for participants. Hyp2.2b Lower respondent endorsement of mental illness stereotypes will be associated with higher recovery expectancy for participants. Hyp2.2c Lower number of discrimination experiences will be associated with higher recovery expectancy for participants. Hyp2.2d Lower endorsement of social withdrawal will be associated with higher recovery expectancy for participants. Hyp2.2e Greater endorsement of stigma resistance will be associated with higher recovery expectancy for participants. Hyp2.2f Greater support through intimate part ner or a best friend will be associated with higher recovery expectancy for participants. Hyp2.2g Greater support through family members will be associated with higher recovery expectancy for participants. Hyp2.2h Greater support via mental health providers will be associated with higher recovery expectancy for participants.

PAGE 276

257 Table 47 (cont.) Hypothesis Hypothesis Supported? Hyp2.2i Greater support through friendship will be associated with higher recovery expectancy for participants. Hyp2.2j Greater feelings of being connected to the community will be associated with higher recovery expectancy for participants. Hyp2.2k Greater trust in the motivation of others will be associated with higher recovery expectancy for participants. Hyp2.2l Higher ratings of self-esteem and self-efficacy will be associated with higher recovery expectancy for participants. Hyp2.2m Higher ratings of personal power will be associated with higher recovery expectancy for participants. Hyp2.2n Greater involvement in the community will be associated with higher recovery expectancy for participants. Hyp2.2o Greater confidence in personal control over the future will be associated with higher recovery expectancy for participants. Hyp2.2q Higher ratings of righteous anger will be associated with higher recovery expectancy for participants. Hyp2.3a Total number of services will be associated with higher recovery expectancies for participants. Hyp2.3b The average number of contact hours per month will be associated with higher recovery expe ctancies for participants. Hyp2.3c Satisfaction level with services will be associated with higher recovery expectancies for participants. Summary of Research Question 2, Investigating Recovery Expectancy Recovery expectancy, the belief that one will recover from mental illness, was found to be strongly associated with recovery strategy endors ement in bivariate analysis, with individuals that do not expect to recover endorsing st rategies significantly less for all strategies except symptom eradication. However, in multivariate analysis using logistic regression that tested the associ ation between the dichotomous expectancy variable and the independent variables in domains 1-3, there were no significant relationships detected for direct effect s. Though the variance in the independent variables included in the mode ls did account for 25% of the variance of recovery expectancy, no significant asso ciations suggests that recovery expectancy is driven by

PAGE 277

258 factors not present in this analysis but is supported to a lesser degree by the variables included. Despite the lack of multivariate results, there were some intriguing findings in bivariate analysis that will be highlighted to better understand expectancy. Beginning with clinical variables, domain 1, a diagnos is of bipolar disord er has the strongest relationship to non-belief in recovery compared to other diagnos es. This is of interest considering the almost exclusive focus that schizophrenia has received in the research literature on recovery. Active symptoms are al so related to negative belief in recovery, with significant associations noted in nine of ten symptom scales. Psychiatric symptoms are often targeted with medication, so it is reasonable to find, in light of increased symptoms, an increased use of psychiatri c medications in all medication categories except anti-depressants by those indicating no e xpectation of recover y. Anti-depressants are prescribed to 81% of individuals not endorsing recovery and 73% of those that believe that they will recover, demonstrating that a prescription for antidepressants is ubiquitous, rendering differences irrelevant. Similarly, more lifetime hospitalizations and greater numbers of family members with a mental illness negatively associate with expectancy. Those lacking rec overy belief are also more likely to have been physically assaulted as a child and sexua lly assaulted as an adult. Those individuals that believe in re covery have significantly higher average scores on 14 of 17 domain 2 social scales th at target the construc ts of stigma, social support and empowerment. Thus, those that belie ve in recovery are less likely to report alienation, discrimination, and soci al withdrawal. They are more likely to express stigma resistance and to not endorse stereotypes of the mentally il l. Support from family and

PAGE 278

259 friends are significantly higher fo r those with recovery beliefs as is an active and positive connection to the community. The self-esteem optimism, sense of personal power and control, belief in community activism and personal autonomy are likewise notably elevated in those sanctioning recovery. Finally, for domain 3, only service satisfaction is mildly elevated in those believing in r ecovery compared to those that are not. As noted, these relationships are not replicated in multivariate analysis. These are preliminary results providing a fi rst look at recovery belief in relation to clinical, social, and service factors. Further research is required to bette r understand these factors in relation to expectancy. Thus, a summativ e statement on the results for recovery expectancy is that it is a complicated concep t that was not adequately modeled in this research. It is reasonable to assume that believing that one will recover is important to engaging in and succeeding in recovery. Furthe r investigation is required to determine factors significantly a ssociated with the belief in r ecovery and to understand their relationship with the recovery process. Focus on Recovery Strategies Analysis pertaining to research qu estion 3 employs the same independent variables with recovery strate gies as the new dependent vari ables in place of recovery expectancy. Ordinary least s quares (OLS) regression is utiliz ed with the continuous level recovery strategy dependent variables. However, because there are multiple strategies, this section will repeat the sequence of analys es used in the previous section six times. The six strategies examined include: 1. Effective illness management

PAGE 279

260 2. Positive future orientation 3. Meaningfulness, personal control, and hope 4. Recognizing support 5. Help seeking 6. Symptom eradication Research Question 3: Are Individual, Soc ial or Service Factors Associated with Recovery Strategy #1: Effe ctive Illness Management? Associations between control variables and effective illness management are summarized in Table 48. The m odel is signifi cant (F = 4.46, p .001) and accounts for approximately 4% of the variance for stra tegy 1. Age and income are just below significance with younger age and greater income contributin g slightly to strategy 1. Gender is significantly associated with st rategy 1, with male respondents endorsing strategy 1 to a greater degree. Research question 3.1: Are individual factors associ ated with effective illness management? Effective illness management entails ac tive involvement in control of symptoms and sequelae (e.g., level of active symptoms, stigma, identifying as a person with mental illness, involvement in treatment). Illness management may include professional services or therapy, adjunctive treatments and consumer led interventions. The following analysis examines what factors are important in associ ation with effective illness management for individuals with varying levels of diagnosed mental illness. Hypotheses 3.1a-3.1aa is addressed within this analysis as well. Op erating in isolation from social and service

PAGE 280

261 Table 48 Results of OLS Regression Testing the Association of Control Variables (Domain 1) on Recovery Strategy 1 (Effective Illness Management) Variable b SE t Age -0.01 0.00 -1.90 Gender -0.26 0.08 -3.03 c Income 0.07 0.04 1.88 Education 0.00 0.04 0.11 Note: a p .05; b p .01; c p .001; d .0001 n = 350; R 2 = .04; F = 4.46, p .01 factors, the model explains 37% of the vari ance of effective illness management and the model rejects the nul l hypothesis that R 2 = 0 in the population (F = 7.81, p .0001). The results are summarized in Table 49. The depression symptom score is associat ed with strategy 1, suggesting that the endorsement of this strategy decreases with higher levels of depression. Likewise, prescription of anti-depressant medication is also associated with strategy use, though in this case indicating increased endorsement of the strategy with anti-depressant use. This suggests that taking an anti-depressant is a part of active illness management and, when effective, would lower depression and possi bly increase endorseme nt of the strategy through offsetting the effect of depression. The effect is reversed for anti-anxiety medication, indicating decreased sanctioning of the strategy when prescribed anti-anxiety medication. The majority of recovery re search targeting symptoms has focused on depression. However, there is some s upporting evidence that anxiety is a key impediment to managing symptoms, presenting an obstacle to this recovery strategy

PAGE 281

262 Table 49 Results of OLS Regression Testing the Asso ciation of Clinical/H istorical Variables (Domain 1) on Recovery Strategy 1 (Effective Illness Management) H 1 Variable b SE t Age -0.00 0.00 -0.18 Gender -0.08 0.08 -0.93 Income 0.04 0.03 1.18 Education 0.02 0.04 0.54 3.1b Somatization 0.15 0.06 2.64 3.1c Obsessive-compulsive -0.04 0.07 -0.60 3.1d Interpersonal sens itivity 0.10 0.07 1.39 3.1e Anxiety -0.13 0.07 -1.88 3.1f Depression -0.34 0.07 -4.61 d 3.1g Hostility -0.02 0.05 -0.37 3.1h Phobic anxiety -0.09 0.05 -1.76 3.1i Paranoia -0.04 0.06 -0.58 3.1j Psychoticism 0.07 0.07 1.03 3.1k Substance abuse history 0.12 0.07 1.63 3.1l Hospitalization history -0.00 0.00 -1.03 3.1m Hospitalized in last year -0.03 0.08 -0.35 3.1n Age of onset -0.00 0.00 -1.11 3.1o Medication: Anti-depressant 0.17 0.09 1.97 a 3.1p Medication: Anti-psychotic -0.01 0.07 -1.41 3.1q Medication: Anti-manic -0.05 0.08 -0.59 3.1r Medication: Anti-anxiety -0.27 0.07 -3.61 c 3.1s Medication: Other psychotropic 0.04 0.07 0.50 3.1t Current employment -0.05 0.09 -0.50 3.1u Years of employment -0.01 0.00 -2.12 a 3.1v Nuclear family mental illness 0.01 0.03 0.47 3.1w Extended family mental illness -0.03 0.05 -0.58 3.1x Child sexual abuse history -0.06 0.08 -0.79 3.1y Child physical abuse history 0.02 0.08 0.32 3.1z Adult sexual assault history -0.11 0.09 -1.16 3.1aa Adult physical assault history 0.19 0.08 2.27 a Note: a p .05; b p .01; c p .001; d .0001 n = 350; R 2 = .37; F = 7.81, p .0001

PAGE 282

263 (Bruce, Yonkers, Otto, Eisen, Weisber g, Pagano, 2005; Simon, Otto, Wisniewski, Fossey, Sagduyu, Frank et al., 2004). Data was not collected on respondent rated efficacy of the medication so it is impossible to inve stigate this as a potential explanation. The lack of significance noted in the anxiety related symptom subscales (anxiety, phobic anxiety, obsessive compulsive) is inconsistent with the an ti-anxiety medication finding, though anxiety and phobic anxiet y do approach significance. Two other variables were significantly associated with effective illness management. Total years worked was significantly associated with the strategy. The direction of the association is inconsistent with previous research where work has been described as a form of illness management (Walby, 2005). However, it is unlikely that all individuals would think of employment as a form of illness management and might view it more as a result of successful illness management (outcome versus process). The final significant association is with adult physical assault. Assault was not defined during data collec tion and could reflect a st ranger assault (e.g. mugging), domestic violence or other form of violence. What can be stated is that increased assault experiences increases the like lihood of utilizing illness management. This might possibly reflect the traditional focus in abuse treatme nt on symptom management due to the high levels of depression and post-traumatic stress associated with abusive experiences. Research question 3.2: Are social factors associated with effective illness management? Hypotheses 3.2a-3.2q are target ed in this section, eval uating for significance the associations between stigma, social support and empowerment with illness management (Table 50). The full mode l is significant (F = 21.33, p .0001) and a significant portion

PAGE 283

264 of the variance is explained (R 2 = .54). Higher levels of alienation experiences were associated with lower endorsement of eff ective illness management. Likewise, a lower resistance to stigmatization messages and experiences also decreases endorsement of Table 50 Results of OLS Regression Testing the Association of Social Variables (Domain 2) on Recovery Strategy 1 (Effective Illness Management) H 1 Variable b SE t Age -0.00 0.00 -1.98 a Gender -0.08 0.06 -1.36 Income -0.01 0.03 -0.26 Education -0.02 0.03 -0.60 Stigma 3.2a Alienation -0.24 0.08 -3.08 c 3.2b Stereotype endorsement -0.06 0.09 -0.75 3.2c Discrimination experience -0.09 0.07 -1.30 3.2d Social withdrawal 0.06 0.09 0.68 3.2e Stigma resistance -0.33 0.08 -4.14 d Social Support and Related 3.2f Partner or best friend support 0.03 0.03 0.97 3.2g Family support 0.01 0.02 0.43 3.2h Provider support 0.06 0.04 1.48 3.2i Friend support 0.00 0.03 0.06 3.2j Community involvement 0.08 0.03 3.13 c 3.2k Trust -0.04 0.03 -1.15 Empowerment 3.2l Self-esteem/self-efficacy 0.42 0.08 5.20 d 3.2m Power/powerlessness -0.01 0.06 1.10 3.2n Community activism and autonomy -0.13 0.08 -1.65 3.2o Optimism and control over the future 0.13 0.08 1.66 3.2p Righteous anger 0.06 0.08 -0.17 Note: a p .05; b p .01; c p .001; d .0001 n = 350; R 2 = .54; F = 21.33, p .0001

PAGE 284

265 effective illness management. Social s upport and related factor s are significantly correlated with effective illness management for one construct only. Greater degree of community involvement increases effective illness management, but, when compared with the two stigma constructs noted a bove, community involvement does not increase effective illness management as much as aliena tion and lack of resist ance decreases it. The most powerful association between a social construct vari able and effective illness management is with the empowerment construct of self-esteem/self-efficacy. Individuals possessed of high self-esteem are more likely to sanction effective illness management. The cross-sectional nature of the study does not supply the direction of the relationship. An argument can be made for either direction: increased self-esteem leading to greater illness management or gr eater illness manageme nt leading to higher self-esteem. In summary, several associations of moderate to high magnitude reflect the relationship between social factors effectiv e illness management, again operating in isolation from clinical and service factors. Research question 3.3: Are service factors associat ed with effective illness management? Hypotheses 3.3a-3.3c was explored in th is analysis. Table 51 summarizes the information. Age and gender are both signifi cantly associated with effective illness management when modeled with service vari ables only. As age increases, effective illness management is slightly less likely to be endorsed. Males are more likely to engage in effective illness management in a ssociation with services factors. The one significant service factor is average satisfaction with services, indicating increased satisfaction is associated with increased endorsement of stra tegy 1. This is consistent

PAGE 285

266 with the recovery literature that indi cates a positive and empowering relationship increases recovery potential as well as adhe rence to clinical regimen (Tait, Birchwood, & Trower, 2003). Table 51 Results of OLS Regression Testing the Impac t of Service Variables (Domain 3) on Recovery Strategy 1 (Effective Illness Management) Variable b SE t Age -0.01 0.00 -2.26 a Gender -0.24 0.81 -2.94 b Income 0.05 0.03 1.44 Education -0.00 0.04 -0.01 3.3a Total number of services 0.01 0.04 0.32 3.3b Total service hours per month -0.00 0.01 -0.15 3.3c Average satisfaction score 0.23 0.04 5.93 d Note: a p .05; b p .01; c p .001; d .0001 n = 350; R 2 = .12; F = 7.91, p .0001 Research question 3: Are individual, social or serv ice factors associated with effective illness management ? Main effects model As discussed in chapter 3 and again a bove when describing logistic regression results, diagnostic statistics (e.g., hat values or Cooks D ) were used to detect influential observations. To remind the read er, if two of three diagnostic tests were violated then an observation was removed from main effect an alysis. A total of fifteen observations violated this criteria a nd were dropped from the anal ysis presented next. The main effect model combines domains 1-3 to assess association with effective illness management (Table 52). The adjusted R 2 for this model is 0.67, indicating that

PAGE 286

267 Table 52 Main Effects Results of OLS Regression Testi ng the Impact of Domain 1-3 Variables on Recovery Strategy 1 (Effective Illness Management) H 1 Variable b SE t Age -0.00 0.00 -1.32 Gender -0.04 0.06 -0.66 Income -0.03 0.02 -1.36 Education -0.00 0.03 -0.03 Domain 1 3.1a Anxiety diagnosis -0.20 0.09 -2.30 a 3.1b Somatization 0.05 0.04 1.10 3.1c Obsessive-compulsive -0.07 0.05 -1.28 3.1d Interpersonal sensi tivity 0.15 0.05 2.82 b 3.1e Anxiety -0.08 0.05 -1.51 3.1f Depression -0.10 0.06 -1.68 3.1g Hostility -0.03 0.04 -0.78 3.1h Phobic anxiety -0.05 0.04 -1.15 3.1i Paranoia -0.04 0.05 -0.75 3.1j Psychoticism -0.01 0.05 -0.22 3.1k Substance abuse history 0.06 0.06 1.02 3.1l Hospitalization history -0.00 0.00 -0.57 3.1m Hospitalized in last year -0.05 0.06 -0.93 3.1n Age of onset 0.00 0.00 0.12 3.1o Medication: Anti-depressant 0.01 0.07 0.19 3.1p Medication: Anti-psychotic -0.11 0.06 -1.82 3.1q Medication: Anti-manic -0.02 0.06 -0.27 3.1r Medication: Anti-anxiety -0.16 0.06 -2.99 b 3.1s Medication: Other psychotropic 0.02 0.05 0.51 3.1t Current employment -0.04 0.07 -0.60 3.1u Years of employment -0.00 0.00 -0.87 3.1v Nuclear family mental illness -0.00 0.02 -0.15 3.1w Extended family mental illness -0.01 0.04 -0.35 3.1x Child sexual abuse history -0.08 0.06 -1.30 3.1y Child physical abuse history 0.04 0.06 0.67 3.1z Adult sexual assault history -0.00 0.07 -0.08 3.1aa Adult physical assault history 0.10 0.06 1.63 Domain 2 Stigma 3.2a Alienation -0.07 0.07 -0.97 3.2b Stereotype endorsement -0.03 0.08 -0.42 3.2c Discrimination experience -0.05 0.06 -0.72 3.2d Social withdrawal 0.02 0.08 0.28 3.2e Stigma resistance -0.40 0.08 -5.31 d

PAGE 287

268 Table 52 (cont.) Variable b SE t Social Support and Related 3.2f Partner or best friend support 0.03 0.03 1.18 3.2g Family support 0.01 0.02 0.47 3.2h Provider support 0.01 0.04 0.20 3.2i Friend support 0.01 0.03 0.34 3.2j Community involvement 0.08 0.02 3.18 b 3.2k Trust 0.00 0.03 0.11 Empowerment 3.2l Self-esteem/self-efficacy 0.28 0.08 3.69 c 3.2m Power/powerlessness 0.04 0.07 0.55 3.2n Community activism and autonomy -0.09 0.07 -1.25 3.2o Optimism and control over the future 0.18 0.07 2.57 b 3.2p Righteous anger -0.02 0.05 -0.51 Domain 3 3.3a Total number of services 0.02 0.03 0.70 3.3b Total service hours per month 0.00 0.00 1.31 3.3c Average satisfaction score 0.03 0.03 0.92 Note: a .05; b .01; c .001; d .0001 n = 335 R 2 = .67; F = 14.64, p .0001 67% of the variance of the dependent variab le effective illness management can be explained by the variations in the independe nt variables of the full model. The null hypothesis that the variance captured is zero (R 2 = 0) is rejected (F = 14.64, p .0001). Anxiety related associations are again significant. A diagnosis of an anxiety disorder (cluster of diagnoses including post-traumatic stress disorder, generalized anxiety disorder, etc.) is nega tively associated with illness management. The prescription of anti-anxiety medications is also significantly negatively associated. Having the opposite effect, interpersonal sensitivity is posi tively associated with illness management.

PAGE 288

269 Whereas increased indicators of anxiety are associated with less effective management (though this cannot be interpreted as causal), increased feelings of inferiority, self-doubt, and inadequacy (key aspects of interpersonal sensitivity) are associated with greater endorsement of illness management. This c ould reflect a need for and willingness to accept direction and the desire to follow a management plan to offset uncomfortable affective states. Examining domain 2 (social) independent variables, poor stigma resistance is associated with decreased illness management. However, a small group of social and empowerment variables are posi tively associated with illne ss management. Involvement with the community increases illness manageme nt. Increased self-esteem/self-efficacy is associated with better illness management as is optimism and control over the future. None of the domain 3 (service) variables are significantly associated with illness management in the full model. When attempting to detect the most parsimonious model of association between a set of independent variables and a dependent variable, it can be usef ul to assess only the significant variables in a reduced or truncated model. Changes in R 2 can be evaluated to see how much of the full model variance may be associated with non-significant coefficients or that may even be suppressing variance if the limited model has an increase in R 2 Table 53 displays the results of the reduc ed model. Variance explained diminishes by 4%. Thus, most of the variance is related to the significant associations in the main effects model. Figure 4 summarizes the va riance explained by each domain, the main effect (full) model and the reduced model. The largest change is in the parameter estimate for interpersonal sensitivity. The magn itude is increased and the direction of the

PAGE 289

270 association changes from positive to negativ e. The regression coefficient for anxiety diagnosis drops slightly in magnitude in th e reduced model while the coefficient for antianxiety medication increases in size. The di rection of the coefficient for interpersonal sensitivity changes to the same direction as the other clinical variables, indicating that increased interpersonal sensitivity is associ ated with a decrease in effective illness management. With clinical vari ables assessed in isolation (Tab le 49) the effects of social and service variables were absent and interp ersonal sensitivity had a positive association with illness management. Community involvement is slightly less potent in the reduced or limited model and optimism is approximately the same, as evidenced by changes in their coefficients. Stigma resistance and self-esteem are both of greater magnitude. Limited stigma resistance may, in part, be a function of low self-esteem. The Pearson correlation between stigma resistance and self-esteem appears to reinforce this possibility Table 53 Reduced Model Results of OLS Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 1 (Effective Illness Management) Variable b SE t Interpersonal sensitivity -0.11 0.03 -4.30 d Anxiety diagnosis 0.08 0.08 -1.88 Medication: Anti-anxiety -0.21 0.05 -4.08 d Stigma resistance -0.52 0.06 -8.26 d Community involvement 0.07 0.02 3.02 b Self-esteem/self-efficacy 0.42 0.07 6.31 d Optimism and control over the future 0.07 0.07 2.56 b Note: a .05; b .01; c .001; d .0001 n = 335; R 2 = .63; F = 81.10, p .0001

PAGE 290

271 (r(348) = -0.48, p .0001). A summary of the models presented in Tables 48-53 is presented in Table 54. Age and gender are the two control variables w ith some consistency in association with effective illness management, with younger age and male gender associated with greater endorsement of the strategy. Depression in th e clinical model (Table 49) gives way to a more anxiety focus in the full model (Tab le 52). Employment and adult assault associations are no longer significant in the full model. Optimism and control over the future becoming significant in the main effect model is reasonable since effective illness management is in effect an investment in future health and goal attainment, free of debilitating symptoms and hospitalizations. Satisfaction with servic es drops out in the full model. Effective illness management can involve professional services as well as more consumer driven and holistic approaches Still, it is unexpected that no service variable would be significantl y associated with a strategy th at is potentially somewhat treatment dependent. Consistent between models are stigma resistance, community involvement, and self-esteem/self-efficacy. Noted above, high levels of self-esteem and low levels of stigma resistance appear to be inversely rela ted. It appears that self-esteem and stigma resistance are both associated with in creased illness management. Community involvement, positively associated with illn ess management, may reflect the consumer recovery literature addre ssing the need for connection and feeling a part of the community, even if not consistent wi th active engagement (Deegan, 2005).

PAGE 291

Figure 4 Amount of Variance Explained by Each Do main for Effective Illness Management Research Question 3: Are Individual, Soc ial or Service Factors Associated with Recovery Strategy #2: Positive Future Orientation? Control Variables (R2 = .04) Domain 1: Clinical and Historical Factors (R2 = .37) Main Effects Model R2 = .67 Reduced Main Effects Model R 2= .63 Effective Illness Management Domain 3: Service Factors (R 2 = .12) Domain 2: Social Factors (R 2 = .54) Associations between control variable s and positive future orientation are summarized in Table 55. The m odel is signifi cant (F = 3.48, p .01) and accounts for approximately 3% of the variance for strate gy 2. Age is significantly associated with positive future orientation, with younger res pondents sanctioning the strategy slightly more than older respondents. 272

PAGE 292

273 Table 54 Summary Table: Results of OLS Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 1 (Effective Illness Management) Model 1 Model 2 Model 3 Model 4 Model 5 Age Sex Income Education Psychosomatic Obsessive Compulsive Interpersonal Sensitivity Anxiety Depression Hostility Phobia Paranoia Psychoticism Anxiety Diagnosis Substance History Lifetime Hospitalizations Hospitalized in Last Year Age of Onset Anti-Depressant Medication Anti-Psychotic Medication Anti-Manic Medication Anti-Anxiety Medication Other Psychotropic Medication Currently Employed Years Employed Lifetime Nuclear Family Mental Illness Extended Family Mental Illness Child Sexual Abuse Child Physical Abuse Adult Sexual Assault Adult Physical Assault Alienation Stereotype Endorsement Discrimination Occurrence Social Withdrawal Stigma Resistance Partner of Best Friend Support Family Support

PAGE 293

274 Table 54 (cont.) Provider Support Friends Support Community Involvement Trust Self-Esteem and Self-Efficacy Power and Powerlessness Community Activism and Autonomy Optimism and Control Over the Future Righteous Anger Total Number of Services Total Contact Hours Average Satisfaction M1 = Model 1: n = 350; Control variables M2 = Model 2: n = 350; Control variab les + Individual and historical factors M3 = Model 3: n = 350; Contro l variables + Social factors M4 = Model 4: n = 350; Control variables + Service factors M5 = Model 5; n = 335; Full (main effects) model (post diagnostics) Table 55 Results of OLS Regression Testing the Impact of Control Variables on Recovery Strategy 2 (Positive Future Orientation) Variable b SE t Age -0.01 0.00 -2.42 a Gender -0.13 0.07 -1.69 Income 0.06 0.03 2.04 Education -0.01 0.04 -0.17 Note: a .05; b .01; c .001; d .0001 n = 350; R 2 = .03; F = 3.48, p .01

PAGE 294

275 Research question 3.1: Are individual factors associ ated with positive future orientation? Clinical/historical factors are significantly associ ated with strategy 23 (F = 4.95, p .0001) with an adjusted R 2 of 0.26 (Table 56). Like the first strategy investigated, this analysis targets hypotheses 3.1a-3.1aa, t hough specific to future orientation. Schizophrenia was the only diagnostic categor y significantly associated with positive future orientation, with individuals with a diagnosis of schizophrenia less likely then other diagnoses to endorse positive future orientation. Two symptom categories, depression and phobic anxiety are negatively associ ated with strategy 2. Individuals with a substance history are actually more likely to endorse a positive future orientation. Noted in the section on bivariate analysis, in dividuals with a substa nce history and that are in treatment at a mental health center have likely experienced substance abuse recovery services that are focused on the pr esent (one day at a time) and on future good works based on sobriety. Research question 3.2: Are social factors associated with positive future orientation? Social variables are significantly associated with positive future orientation (F = 32.02, p .0001) and account for 64% of the varian ce, while addressing hypotheses 3.2a3.2p (Table 57). Social withdraw al is significantly negatively associated with positive future orientation. Social contact and the support it engenders is an integral part of recovery, especially as described by the c onsumer recovery movement (Chadwick, 1997). Support variables were predicted to be positiv ely associated with strategy 2. However,

PAGE 295

276 Table 56 Results of OLS Regression Testing the Impac t of Clinical/Historical (Domain 1) on Recovery Strategy 2 (Positive Future Orientation) H 1 Variable b SE t Age -0.00 0.00 -0.63 Gender -0.01 0.08 -0.17 Income 0.04 0.03 1.28 Education -0.01 0.04 -0.46 3.1a Schizophrenia diagnosis -0.21 0.10 -2.02 a 3.1b Somatization 0.05 0.05 0.98 3.1c Obsessive-compulsive 0.05 0.07 0.73 3.1d Interpersonal sens itivity 0.00 0.07 0.02 3.1e Anxiety 0.08 0.07 1.25 3.1f Depression -0.37 0.07 -5.17 d 3.1g Hostility 0.00 0.05 0.11 3.1h Phobic anxiety -0.14 0.05 -2.72 b 3.1i Paranoia 0.05 0.06 0.75 3.1j Psychoticism -0.04 0.06 -0.57 3.1k Substance abuse history 0.14 0.07 2.09 a 3.1l Hospitalization history -0.00 0.00 -0.99 3.1m Hospitalized in last year -0.10 0.07 -1.34 3.1n Age of onset 0.00 0.00 0.09 3.1o Medication: Anti-depressant -0.07 0.08 -0.82 3.1p Medication: Anti-psychotic 0.01 0.08 0.13 3.1q Medication: Anti-manic -0.05 0.08 -0.71 3.1r Medication: Anti-anxiety 0.00 0.07 0.02 3.1s Medication: Other psychotropic -0.02 0.07 -0.34 3.1t Current employment -0.08 0.09 -0.85 3.1u Years of employment -0.01 0.00 -1.80 3.1v Nuclear family mental illness 0.01 0.03 0.34 3.1w Extended family mental illness -0.02 0.05 -0.42 3.1x Child sexual abuse history -0.04 0.08 -0.61 3.1y Child physical abuse history 0.02 0.07 0.26 3.1z Adult sexual assault history -0.00 0.09 -0.06 3.1aa Adult physical assault history 0.15 0.08 1.89 Note: a .05; b .01; c .001; d .0001 n = 350; R 2 = .26; F = 4.95, p .0001

PAGE 296

277 Table 57 Results of OLS Regression Testing the Association of Social Variables (Domain 2) on Recovery Strategy 2 (Positive Future Orientation) H 1 Variable b SE t Age -0.00 0.00 -2.35 a Gender 0.06 0.05 1.19 Income -0.00 0.02 -0.30 Education -0.03 0.02 -1.25 Stigma 3.2a Alienation 0.06 0.06 0.93 3.2b Stereotype endorsement -0.09 0.07 -1.31 3.2c Discrimination experience 0.10 0.06 1.85 3.2d Social withdrawal -0.21 0.07 -2.99 b 3.2e Stigma resistance -0.07 0.06 -1.12 Social Support and Related 3.2f Partner or best friend support 0.01 0.03 0.47 3.2g Family support 0.04 0.02 2.38 3.2h Provider support 0.13 0.03 4.41 d 3.2i Friend support -0.00 0.03 -0.04 3.2j Community involvement 0.04 0.02 1.92 3.2k Trust 0.01 0.02 0.43 Empowerment 3.2l Self-esteem/self-efficacy 0.69 0.06 11.09 d 3.2m Power/powerlessness 0.14 0.06 2.38 3.2n Community activism and autonomy 0.06 0.06 0.95 3.2o Optimism and control over the future -0.03 0.06 -0.49 3.2p Righteous anger 0.03 0.04 0.83 Note: a .05; b .01; c .001; d .0001 n = 350; R 2 = .64; F = 32.02, p .0001 only provider support was associated with posit ive future orientati on, again, however, in isolation from the other predicto r domains. The clinicians at the partnering agency have been exposed to the role recovery program in recent months. This has emphasized the

PAGE 297

278 recovery philosophy and may be biasing this data albeit in a way that is positive for the consumer. The emphasis is on planning for th e future and coping with current limitations while capitalizing on strengths. Hi storically, individua ls with mental illness, especially severe mental illness, have not enjoyed rela tionships that supported planning for careers, family, and other aspects of future most othe rs expect automatically (Yanos, Rosenfield, Horwitz, 2001). Providers often provide substitute support, though systematic investigation of provider support is negligible (Meeks & Murrell, 1994). Of the empowerment scales only self-esteem/self-efficacy is significantly associated with positive future orientation. Th is is by far the largest estimate for domain 2 variables. Self-esteem is linked to conf idence and the expectation of success and is logically linked with future orientati on (Rogers, Chamberlin, Ellison, & Crean, 1997; Wood, Heimpel, NewbyClark, & Ross, 2005). Research question 3.3: Are service factors associated with positive future orientation? Positive future orientation is significantly associated with service variables (F = 7.91, p .0001), accounting for 12% of the variance (Table 58). Average satisfaction score, as expected, is positively associated w ith future orientation. Counter to what was hypothesized, total service hours per month have an inverse relationship with positive future orientation, though the st rength of the association is we ak. A likely explanation is that the number of service hours increases with severity of illness. This apparent association may be a proxy for illness severit y, with severity known to be negatively correlated with a pos itive future outlook.

PAGE 298

279 Table 58 Results of OLS Regression Testing the Impac t of Service Variables (Domain 3) on Recovery Strategy 2 (Positive Future Orientation) Variable b SE t Age -0.01 0.00 -2.58 b Gender -0.12 0.07 -1.75 Income 0.05 0.03 1.56 Education -0.02 0.04 -0.66 3.3a Total number of services 0.03 0.03 1.04 3.3b Total service hours per month -0.02 0.00 -3.37 c 3.3c Average satisfaction score 0.25 0.03 7.60 d Note: a .05; b .01; c .001; d .0001 n = 350; R 2 = .12; F = 7.91, p .0001 Research question 3: Are individual, social or serv ice factors associated with positive future orientatio n? Main effects model Identical to the discussion of logistic regression and the analysis of recovery strategy one, diagnostic statisti cs were applied and evaluated for the main effects model targeting positive future orientation. Fifteen observations violated the cutpoint of the criteria identifying overly in fluential observations and were dropped from the subsequent analysis. The main effect model is significant (F = 17.72, p .0001) with an adjusted Rsquare of 0.71 (Table 59). Three symptom scales are associated with positive future orientation. Depression and phobic anxiety decrease endorsement of pos itive future orientation. Much like the confusing relationship in the main effects model for strate gy 1, involving interpersonal sensitivity, anxiety is positively associated with the positive future orientation recovery

PAGE 299

280 Table 59 Main Effects Results of OLS Regression Testi ng the Impact of Domain 1-3 Variables on Recovery Strategy 2 (Positive Future Orientation) H 1 Variable b SE t Age -0.00 0.00 -0.96 Gender 0.03 0.05 0.66 Income -0.03 0.02 -1.39 Education -0.05 0.02 -2.07 a Domain 1 3.1a Schizophrenia diagnosis -0.19 0.07 -2.68 a 3.1b Somatization -0.06 0.03 -1.78 3.1c Obsessive-compulsive 0.04 0.04 0.98 3.1d Interpersonal se nsitivity 0.07 0.04 1.61 3.1e Anxiety 0.12 0.05 2.73 b 3.1f Depression -0.11 0.05 -2.21 a 3.1g Hostility -0.00 0.03 -0.14 3.1h Phobic anxiety -0.12 0.03 -3.31 c 3.1i Paranoia -0.03 0.04 -0.60 3.1j Psychoticism -0.06 0.04 -1.42 3.1k Substance abuse history 0.10 0.05 2.05 a 3.1l Hospitalization history 0.00 0.00 0.15 3.1m Hospitalized in last year -0.10 0.05 -2.07 a 3.1n Age of onset 0.00 0.00 0.30 3.1o Medication: Anti-depressant -0.10 0.06 -1.78 3.1p Medication: Anti-psychotic 0.03 0.05 0.66 3.1q Medication: Anti-manic -0.02 0.05 -0.57 3.1r Medication: Anti-anxiety 0.08 0.05 1.79 3.1s Medication: Other psychotropic -0.02 0.05 -0.43 3.1t Current employment 0.04 0.06 0.63 3.1u Years of employment -0.00 0.00 -0.61 3.1v Nuclear family mental illness 0.02 0.02 1.20 3.1w Extended family mental illness -0.02 0.03 -0.74 3.1x Child sexual abuse history 0.02 0.05 0.37 3.1y Child physical abuse history 0.03 0.05 0.70 3.1z Adult sexual assault history 0.03 0.06 0.45 3.1aa Adult physical assault history 0.05 0.05 0.94 Domain 2 Stigma 3.2a Alienation 0.09 0.06 1.46 3.2b Stereotype endorsement -0.07 0.07 -1.02 3.2c Discrimination experience 0.09 0.05 1.71 3.2d Social withdrawal -0.18 0.07 -2.59 b 3.2e Stigma resistance -0.07 0.06 -1.16

PAGE 300

281 Table 59 (cont.) H 1 Variable b SE t Social Support and Related 3.2f Partner or best friend support 0.01 0.02 0.43 3.2g Family support 0.05 0.02 3.11 b 3.2h Provider support 0.10 0.03 3.25 c 3.2i Friend support 0.01 0.02 0.32 3.2j Community involvement 0.05 0.02 2.47 b 3.2k Trust 0.01 0.02 0.28 Empowerment 3.2l Self-esteem/self-efficacy 0.67 0.06 10.34 d 3.2m Power/powerlessness 0.04 0.06 0.65 3.2n Community activism and autonomy 0.09 0.06 1.43 3.2o Optimism and control over the future -0.06 0.06 -1.07 3.2p Righteous anger 0.02 0.04 0.56 Domain 3 3.3a Total number of services -0.01 0.02 -0.26 3.3b Total service hours per month -0.01 0.00 -1.61 3.3c Average satisfaction score 0.05 0.03 1.82 Note: a .05; b .01; c .001; d .0001 n = 335; R 2 = .71; F = 17.72, p .0001 strategy. This finding runs counter to phobic anxiety noted above, furthering the confusion. However, turning to the items in the SCL-90-R that combine for the anxiety scale, the majority appear to be written to capture panic disorder and the somatic symptoms that are the essential criteria for that disorder (e.g., h eart pounding or racing; spells of terror or panic) and generalized anxiety disorder (GAD) (e.g., feeling tense or keyed up; feeling so restless you couldnt sit stil l). Mild to moderate panic disorder is often internalized by clients as something to monitor and to have a plan for, not something as life altering as schizophreni a. The same is true for GAD though these

PAGE 301

282 symptoms are often more continuous and unsettling, but respond well to medication. Anti-anxiety medication is highly efficacious for the two disorders listed, and of the seventeen individuals with either panic diso rder or GAD, sixteen are prescribed antianxiety medication. Positive future orientation is negatively associated with being hospitalized in the last year. This result is in the hypothesized direction for this study. Even though every attempt is made to maintain individuals in the community, it can be difficult for some individuals who find themselves inpatient on a relatively regular basis. There is substantial variance in number of lifetime hos pitalizations for respondents in this study, ranging from zero to over one hundred. Desp ite the recovery movements attempts at normalizing hospitalization as a learning encounter that can aid in future recovery (Mueser, Corrigan, Hilton, Tanzman, Scha ub, Gingerich, et al., 2002; Xie, McHugo, Helmstetter, Drake, 2005), it is still considered a personal failure and disrupts both present day and future plans (Walby, 2003a). The final domain 1 significant association is substance abuse history, which is positively associated with strategy 2. This might reflect the experience and belief in recovery many individuals with substance diagnoses are exposed to through professional and consumer run interventions. The social variables (domain 2) appear to have a strong influence on the large degree of variance captured in this model, sim ilar to their impact in isolation noted in the discussion of Table 57. The stigma variable of social withdrawal is significantly negatively associated with positive future or ientation. As predic ted, support variables are, as a group, positively associated with strategy 2, three of wh ich are significant. Family support is weakly but significantly associated and provider support is also

PAGE 302

283 associated. Community involveme nt is also positively associated at a low magnitude. The only domain 2 empowerment va riable that is associated w ith future orientation in the main effect model is self-esteem/self-efficacy. This is a powerful association with endorsement of future orientation increas ing by 0.67 for each unit increase in selfefficacy. Identical to the analysis approach used with strategy 1, a limited model was created containing only the significant variables from the main effect model. The results are displayed in Table 60. All the variables remain significant and maintain the same direction. The amount of variance explained drops 3% for a total of 67%. Anxiety and phobic anxiety decrease in effect while depression gains strength in the association with Table 60 Reduced Model Results of OLS Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 2 (Positive Future Orientation) Variable b SE t Education -0.05 0.02 -2.42 b Anxiety 0.10 0.04 2.49 b Depression -0.10 0.04 -2.53 b Phobic anxiety -0.10 0.03 -3.18 c Schizophrenia diagnosis -0.27 0.06 -4.38 d Substance abuse history 0.10 0.04 2.18 a Hospitalized in last year -0.09 0.04 -2.11 a Social withdrawal -0.13 0.43 -3.11 b Family support 0.03 0.01 2.10 a Provider support 0.13 0.03 4.94 d Community involvement 0.05 0.02 2.51 b Self-esteem/self-efficacy 0.67 0.05 13.38 d Note: a .05; b .01; c .001; d .0001 n = 335; R 2 = .68; F = 61.46, p .0001

PAGE 303

284 strategy 2. A diagnosis of schi zophrenia is stronger in its association in the limited model (b = -0.27 compared to b = -0.21) and is the most powerful negative covariate. Substance abuse, hospitalization in the last year and community invo lvement are approximately of equal strength compared to the full model. Family support actually decreases in strength while social withdrawal, provi der support and self-esteem/se lf-efficacy gain strength in the association. Self-esteem/self-efficacy is the most powerful positive covariate with strategy 2 and is of equal strength in main effects or limited models (b = 0.67). Figure 5 displays the variance explaine d in each model detailed above. Social factors explain the most variance in the isolated models and appear to contribute the most in the main effects model as well. Clinical/historical factors are not insignificant, nor are service factors. However, there is more fluctuation in these domains when comparing number of significant contribu tors as well as magnitude. There appears to be minimal noise in the main effects model as the significant variables in the main effects model explain the majority of th e variance with only a minima l loss (3%) between the main effects and reduced models. Table 61 summarizes the models for positive future orientation. Age is the most influential of the control variables. For the majority, if a variable was significant in an isolated model, then it was likely significant in the main effects model. This is accurate for diagnosis of schizophrenia, depressi ve symptoms, phobic anxiety, substance use history, social withdrawal, pr ovider support, and self-esteem/self-efficacy. Anxiety,

PAGE 304

Figure 5 Amount of Variance Explained by Each Do main for Positive Future Orientation Control Variables (R2 = .03) Domain 1: Clinical and Historical Factors (R2 = .26) Main Effects Model R2 = .71 Reduced Main Effects Model R2 = .68 Positive Future Orientation Domain 2: Social Factors (R 2 = .64) Domain 3: Service Factors (R 2 = .18) being hospitalized in the last year, fam ily support and community involvement are significant in the main effects model but not in the isolated models Conversely, none of the service variables are signi ficant in the main effects m odel though total service hours in a month and average satisfaction were significant in isolation. 285

PAGE 305

286 Table 61 Summary Table: Results of OLS Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 2 (Positive Future Orientation) Model 1 Model 2 Model 3 Model 4 Model 5 Age Sex Income Education Psychosomatic Obsessive Compulsive Interpersonal Sensitivity Anxiety Depression Hostility Phobia Paranoia Psychoticism Schizophrenia Diagnosis Substance History Lifetime Hospitalizations Hospitalized in Last Year Age of Onset Anti-Depressant Medication Anti-Psychotic Medication Anti-Manic Medication Anti-Anxiety Medication Other Psychotropic Medication Currently Employed Years Employed Lifetime Nuclear Family Mental Illness Extended Family Mental Illness Child Sexual Abuse Child Physical Abuse Adult Sexual Assault Adult Physical Assault Alienation Stereotype Endorsement Discrimination Occurrence Social Withdrawal Stigma Resistance Partner of Best Friend Support Family Support Provider Support

PAGE 306

287 Table 61 (cont.) Friends Support Community Involvement Trust Self-Esteem and Self-Efficacy Power and Powerlessness Community Activism and Autonomy Optimism and Control Over the Future Righteous Anger Total Number of Services Total Contact Hours Average Satisfaction M1 = Model 1: n = 350; Control variables M2 = Model 2: n = 350; Control variab les + Individual and historical factors M3 = Model 3: n = 350; Contro l variables + Social factors M4 = Model 4: n = 350; Control variables + Service factors M5 = Model 5; n = 335; Full (main effects) model (post diagnostics) Research Question 3: Are Individual, Soc ial or Service Factors Associated with Recovery Strategy #3: Meaningfuln ess, Personal Control, and Hope? Hope and meaningfulness are cornerst ones of the recovery movement (Deegan, 2001; Lysaker, Buck, Hammoud, Taylor, & Roe, 2006; Repper & Perkins, 2003; Resnick, Rosenheck, & Lehman, 2004). Indeed hope is considered the most necessary component for personal recovery (Deegan, 2001; Walby, 2003b). The presence of hope and meaning provide guidance and purpose, supporting an individual to structure their time and plan ahead, whether it is for the next few minutes or few decades. The importance of hope to recovery and potential ly recovery program development adds a level of importance to detecting significant co variates. This is especially true in the

PAGE 307

288 current climate of reduction in services and tightening of spending on treating mental illness. Interventions, whether professional or consumer, which fosters hope and provides meaning has the potential to pa rtially offset systemic deficiencies. Utilizing the same analysis steps as the two previous recovery strategies, Table 62 confirms that there are no observed significan t associations present between the control variables of age, gender, income, and education and recove ry strategy 3. Table 62 Results of OLS Regression Testing the Impact of Control Variables on Recovery Strategy 3 (Meaningfulness, Personal Control, and Hope) Variable b SE t Age 0.00 0.00 0.49 Gender 0.06 0.05 1.11 Income 0.03 0.02 1.30 Education 0.00 0.03 0.01 Note: n = 350; R 2 = .00; F = 0.81, ns Research question 3.1: Are individual factors associated with meaningfulness, personal control, and hope? Table 63 summarizes the results of the OLS regression of strategy three on the domain 1 clinical/historical variables. The model is not significant, indicating that the null hypothesis of R 2 = 0 is supported (F = 1.08, ns). Approximately 1% of the variance for hope, meaning and personal control is expl ained by clinical factors. The lack of supporting evidence in the statistical model supports the recovery movements emphasis on non-clinical factors in relati on to this key aspect of r ecovery (Anthony, 1993; Deegan,

PAGE 308

289 Table 63 Results of OLS Regression Testing the Impac t of Clinical/Historical (Domain 1) on Recovery Strategy 3 (Meaningfulness Personal Control, and Hope) H 1 Variable b SE t Age 0.05 0.00 1.74 Gender 0.03 0.06 0.56 Income 0.03 0.02 1.38 Education 0.01 0.03 0.19 3.1b Somatization 0.07 0.04 1.62 3.1c Obsessive-compulsive 0.00 0.05 0.04 3.1d Interpersonal sens itivity 0.09 0.05 1.67 3.1e Anxiety -0.02 0.06 -0.32 3.1f Depression -0.03 0.06 -0.48 3.1g Hostility -0.02 0.04 -0.67 3.1h Phobic anxiety -0.12 0.04 -3.06 b 3.1i Paranoia -0.04 0.05 -0.88 3.1j Psychoticism 0.04 0.05 0.80 3.1k Substance abuse history 0.03 0.06 0.49 3.1l Hospitalization history -0.00 0.00 -0.57 3.1m Hospitalized in last year 0.07 0.06 1.20 3.1n Age of onset 0.00 0.00 0.13 3.1o Medication: Anti-depressant 0.00 0.07 0.01 3.1p Medication: Anti-psychotic -0.03 0.06 -0.60 3.1q Medication: Anti-manic 0.03 0.06 0.48 3.1r Medication: Anti-anxiety -0.00 0.06 -0.05 3.1s Medication: Other psychotropic -0.07 0.06 -1.18 3.1t Current employment 0.04 0.07 0.60 3.1u Years of employment -0.01 0.00 -2.37 a 3.1v Nuclear family mental illness 0.02 0.02 0.95 3.1w Extended family mental illness -0.02 0.04 -0.48 3.1x Child sexual abuse history 0.07 0.06 1.13 3.1y Child physical abuse history 0.02 0.06 0.35 3.1z Adult sexual assault history -0.06 0.07 -0.87 3.1aa Adult physical assault history 0.06 0.06 0.95 Note: a .05; b .01 n = 350 R 2 = .01; F = 1.08, ns

PAGE 309

290 2005). Two independent variables, phobic anxiety and years of employment are negatively associated with strategy 3. D ecreased endorsement of hope and meaning due to higher levels of phobic anxi ety is logically linked. The relationship between higher reported years of employment and decreased hop e is less clear. Taken in the context of what jobs were endorsed by the respondents, ho wever, it becomes apparent that nearly all the employed respondents are members of the working poor, putting in long hours for minimal pay and often not being eligible fo r state assistance due to just surpassing assistance cutpoints (21% currently employe d, 87% of the employed at or below $15,000 annual income, 99% in service or labor related jobs). Research question 3.2: Are social factors associ ated with meaningfulness, personal control, and hope? Table 64 summarizes the associations betw een domain 2 variables and strategy 3. Hypotheses 3.2a-3.2p are addressed in the analysis and the model is significant (F = 10.71, p .0001) and accounts for 36% of the variance. Female participants endorse the strategy significantly more than male participants. The is olated effects of domain 2 stigma variables indicates that higher endorseme nt of stereotypes and lower resistance to stigma are negatively associated with hope and meaning. There is a conceptual link between accepting societal ster eotypes of mental illness as relevant to the self (e.g., mentally ill individuals are dangerous, unpredic table, lazy, etc.) and poor resistance to the stigmatizing messages and behaviors that stem from the stereotypes. The single significant support variable is provider support, suggesting that hope and meaning increases in association with a positive relationship with a provider. There

PAGE 310

291 is not, as hypothesized, a significant positive relationship between self-esteem and hope. However, community activism and autonomy is positively associated. The items on this Table 64 Results of OLS Regression Testing the Im pact of Social Variables (Domain 2) on Recovery Strategy 3 (Meaningfulness Personal Control, and Hope) H 1 Variable b SE t Age 0.00 0.00 1.41 Gender 0.10 0.04 2.16 a Income -0.00 0.02 -0.32 Education -0.01 0.02 -0.42 Stigma 3.2a Alienation 0.03 0.06 0.59 3.2b Stereotype endorsement -0.20 0.06 -3.21 c 3.2c Discrimination experience 0.09 0.05 1.65 3.2d Social withdrawal -0.01 0.07 -0.10 3.2e Stigma resistance -0.16 0.06 -2.76 b Social Support and Related 3.2f Partner or best friend support -0.02 0.02 -0.65 3.2g Family support 0.01 0.01 1.02 3.2h Provider support 0.06 0.03 2.08 a 3.2i Friend support 0.04 0.02 1.70 3.2j Community involvement 0.02 0.02 0.89 3.2k Trust 0.02 0.02 0.71 Empowerment 3.2l Self-esteem/self-efficacy 0.03 0.06 0.57 3.2m Power/powerlessness -0.04 0.06 -0.67 3.2n Community activism and autonomy 0.35 0.06 5.95 d 3.2o Optimism and control over the future 0.15 0.06 2.60 b 3.2p Righteous anger -0.05 0.04 -1.29 Note: a .05; b .01; d .0001 n = 350; R 2 = .36; F = 10.71, p .0001

PAGE 311

292 scale encompass, as the scale name s uggests, a proactive relationship with the community that challenges the negative be liefs about the mentally ill. A related empowerment construct, optimism and control over the future, is also significantly associated with hope and meaning. Research question 3.3: Are service factors associated with meaningfulness, personal control, and hope? Service factors are not significantly a ssociated with hope and meaning (F = 1.59, ns), accounting for only 1% of the variance (Table 65). Average satisfaction with services (hypothesis 3.3c) is positively associated with hope and meaning suggesting again that the relationship or competent ment al health services may be important in relation to recovery. Table 65 Results of OLS Regression Testing the Impac t of Service Variables (Domain 3) on Recovery Strategy 3 (Meaningfulness Personal Control, and Hope) Variable b SE t Age 0.00 0.00 0.39 Gender 0.06 0.05 1.20 Income 0.02 0.02 1.12 Education -0.00 0.03 -0.04 3.3a Total number of services 0.03 0.03 1.12 3.3b Total service hours per month -0.00 0.00 -1.30 3.3c Average satisfaction score 0.06 0.02 2.61 b Note: b .01 n = 350 R 2 = .01; F = 1.59, ns

PAGE 312

293 Research question 3: Are individual, social or serv ice factors associated with meaningfulness, personal control, and hope? Main effects model The model, summarized in Table 66, that investigates the association between independent variables from domains 1-3 with re covery strategy 3 is significant (F = 7.27, p .0001), and accounts for 48% of the variance. The main effect model has a sample size of 334, signifying that 16 observations viol ated diagnostic criteria and were dropped from analysis. Several clinical variables that were not significant in the isolated model are now significant in the main effect model. A diagnosed anxiety disorder is negatively associated with the strategy of genera ting hope and meaning. While anti-anxiety medication is not significantly associated with strategy 3, which might be expected given the significant relationship with anxiety disorder s, the use of anti-psy chotic medication is. Recovery research has focused the most on ps ychotic disorders and the literature shows a clear association between sc hizophrenia spectrum diso rders and reduction in hope (Lysaker, et al., 2006, Lysaker, Campbell, & Johannesen, 2005). However, contrary to expectation, schizophrenia or schizoaffective di sorder were not significant in the model. Interpersonal sensitivity was significantly associated with strategy 3, though not in the hypothesized direction. The mechanis ms driving this association can only be conjectured and, if not a statis tical artifact, might indicate a careful planning for control and meaning to offset the negative emotions associated with interpersonal sensitivity. Child sexual abuse experiences is significantly associated but also in the direction other than what was hypothesized. This may re flect the emphasis on recovery and finding hope consistent with the treatm ent of abuse survivors. Ind eed, the evidence-based group

PAGE 313

294 Table 66 Main Effects Results of OLS Regression Testi ng the Impact of Domain 1-3 Variables on Recovery Strategy 3 (Meaningfulness Personal Control, and Hope) H 1 Variable b SE t Age 0.00 0.00 1.42 Gender 0.06 0.05 1.22 Income -0.02 0.02 -0.94 Education -0.02 0.02 -0.75 Domain 1 3.1a Anxiety diagnosis -0.24 0.07 -3.43 c 3.1b Somatization 0.01 0.03 0.44 3.1c Obsessive-compulsive -0.06 0.04 -1.44 3.1d Interpersonal sens itivity 0.11 0.04 2.76 b 3.1e Anxiety 0.02 0.04 0.52 3.1f Depression 0.06 0.04 1.33 3.1g Hostility -0.08 0.03 2.55 b 3.1h Phobic anxiety -0.04 0.03 -1.35 3.1i Paranoia -0.04 0.04 -0.97 3.1j Psychoticism 0.04 0.04 1.07 3.1k Substance abuse history -0.00 0.04 -0.02 3.1l Hospitalization history 0.00 0.00 0.75 3.1m Hospitalized in last year 0.01 0.05 0.19 3.1n Age of onset -0.00 0.00 -0.14 3.1o Medication: Anti-depressant -0.03 0.05 -0.56 3.1p Medication: Anti-psychotic -0.12 0.05 -2.58 b 3.1q Medication: Anti-manic 0.02 0.05 0.51 3.1r Medication: Anti-anxiety 0.06 0.04 1.40 3.1s Medication: Other psychotropic -0.05 0.04 -1.16 3.1t Current employment 0.06 0.06 1.14 3.1u Years of employment -0.00 0.00 -1.19 3.1v Nuclear family mental illness 0.03 0.02 1.80 3.1w Extended family mental illness -0.04 0.03 -1.40 3.1x Child sexual abuse history 0.10 0.05 2.08 a 3.1y Child physical abuse history 0.02 0.05 0.40 3.1z Adult sexual assault history -0.03 0.06 -0.49 3.1aa Adult physical assault history -0.03 0.05 -0.48 Domain 2 Stigma 3.2a Alienation -0.02 0.06 -0.40 3.2b Stereotype endorsement -0.20 0.06 -3.34 c 3.2c Discrimination experience 0.08 0.05 1.55 3.2d Social withdrawal -0.00 0.07 -0.05 3.2e Stigma resistance -0.16 0.06 -2.79 b

PAGE 314

295 Table 66 (cont.) H 1 Variable b SE t Social Support and Related 3.2f Partner or best friend support -0.01 0.02 -0.44 3.2g Family support 0.04 0.02 2.46 b 3.2h Provider support 0.07 0.03 2.55 b 3.2i Friend support 0.06 0.02 2.46 b 3.2j Community involvement 0.03 0.02 1.50 3.2k Trust 0.00 0.02 0.08 Empowerment 3.2l Self-esteem/self-efficacy 0.07 0.06 1.11 3.2m Power/powerlessness -0.03 0.06 -0.47 3.2n Community activism and autonomy 0.35 0.06 5.92 d 3.2o Optimism and control over the future 0.11 0.06 1.98 a 3.2p Righteous anger -0.06 0.04 -1.45 Domain 3 3.3a Total number of services -0.01 0.02 -0.36 3.3b Total service hours per month 0.00 0.00 0.06 3.3c Average satisfaction score 0.01 0.02 0.34 Note: a .05; b .01; c .001; d .0001 n = 334; R 2 = .48; F = 7.27, p .0001 method used at the partnering agency, which many of the respondents have been exposed to, has hope as its key criteria. Domain 2 social variables specific to stig ma are, in general, negatively associated with hope and meaning. Stereotype endorsement and stigma resistance are significantly associated. As hypothesized, support variables are also re lated to hope and meaning. Family support, provider support, and support from friends are positiv ely associated with meaning and hope. Support from a best frie nd or partner, though not significant, is negatively associated with strategy 3 and may re flect what previous research has revealed

PAGE 315

296 as the mixed burden and support that individu als with mental illness experience in close relationships (Walby, 2003a). Finally, e xploring the impact of empowerment, two variables, community activism and autonomy and optimism and contro l over the future are associated with increased meaning and hope. The final domain, services domain, are absent significant associations with strategy 3. Average satisfac tion, significant in the isol ated model, is no longer significant. When combined with social and cl inical factors, services are not associated with increased hope, again reinforcing the consumer recovery lit erature (Deegan, 2003). The results from the limited model that tested the associa tion between only the significant variables from the main effects model (Table 66 ) are presented in Table 67. The model is significant (F = 27.51, p .0001) and explains slightly more variance (49%) then the main effects model (48%). Figur e 6 provides a synopsis of the variance explained for each domain and the main effects (full and limited) models. The direction of association is the same and magnitude of a ssociation for each variable in the model is increased over the main effects. There are, however, no overwhelming associations contained in the model. The two most powerful associations could possibly be conceptually linked if the assumption that an anxiety diagnosis (t he strongest negative association) and community activism (the st rongest positive) act as counterpoints if anxiety is a limitation to activities asso ciated with community activism. When considering models separately, social factors explain far more variance then clinical or service factors. However, the emergence of clinical factors in the main effects model compared to the isolated model likely is a closer step to the lived reality of the

PAGE 316

297 Table 67 Reduced Model Results of OLS Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 3 (Meaningfulness, Personal Control, and Hope) Variable b SE t Anxiety Diagnosis -0.24 0.06 -3.69 c Interpersonal sensi tivity 0.12 0.02 5.08 d Hostility -0.09 0.02 -3.74 c Medication: Anti-psychotic -0.13 0.04 -3.44 c Child sexual abuse history 0.10 0.04 2.71 b Stereotype endorsement -0.17 0.05 -3.67 c Stigma resistance -0.15 0.05 -2.99 b Family support 0.03 0.01 2.52 b Provider support 0.09 0.02 3.89 d Friend support 0.06 0.02 3.14 c Community activism and autonomy 0.38 0.05 7.18 d Optimism and control over the future 0.13 0.05 2.75 b Note: a .05; b .01; c .001; d .0001 n = 334; R 2 = .49; F = 27.51, p .0001 respondents. A summary table (Table 68) is offered to better unde rstand the interaction of the isolated models in relation to the main effects model. Variables significant in early models are, except for average satisfaction wi th services and years employed, significant in main effects (phobic anxiety, stereot ype endorsement, provider support, community activism, and optimism), with similar magnitudes. Clinical factors emerge only in relation to social variables, suggesting a more complex relationship between domain 1 and 2 variables then can be expl ained in the current analysis.

PAGE 317

Figure 6 Amount of Variance Explained by Each Doma in for Meaningfulness, Personal Control, and Hope Research Question 3: Are Individual, Soc ial or Service Factors Associated with Recovery Strategy #4: Recognizing Support? Control Variables (R2 = .00) Domain 1: Clinical and Historical Factors (R2 = .01) Main Effects Model R2 = .48 Reduced Main Effects Model R2 = .49 Meaningfulness, Personal Control, and Hope Domain 2: Social Factors (R 2 = .36) Domain 3: Service Factors (R 2 = .01) Recognition of support means that the indi vidual is aware that they are not alone and that there are avenues for support whether or not they choose to access them or not. Actively seeking assistance (hel p seeking) is the next stra tegy (#5) evaluated. Some individuals with mental illness actively seek isolation, usually out of a concern that relationships or even social interaction is stressf ul and exacerbates symptoms (Marrone & 298

PAGE 318

299 Table 68 Summary Table: Results of OLS Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 3 (Meaningfuln ess, Personal Control, & Hope) Model 1 Model 2 Model 3 Model 4 Model 5 Age Sex Income Education Psychosomatic Obsessive Compulsive Interpersonal Sensitivity Anxiety Depression Hostility Phobia Paranoia Psychoticism Anxiety Diagnosis Substance History Lifetime Hospitalizations Hospitalized in Last Year Age of Onset Anti-Depressant Medication Anti-Psychotic Medication Anti-Manic Medication Anti-Anxiety Medication Other Psychotropic Medication Currently Employed Years Employed Lifetime Nuclear Family Mental Illness Extended Family Mental Illness Child Sexual Abuse Child Physical Abuse Adult Sexual Assault Adult Physical Assault Alienation Stereotype Endorsement Discrimination Occurrence Social Withdrawal Stigma Resistance Partner of Best Friend Support Provider Support

PAGE 319

300 Table 68 (cont.) Family Support Friends Support Community Involvement Trust Self-Esteem and Self-Efficacy Power and Powerlessness Community Activism and Autonomy Optimism and Control Over the Future Righteous Anger Total Number of Services Total Contact Hours Average Satisfaction M1 = Model 1: n = 350; Control variables M2 = Model 2: n = 350; Control variab les + Individual and historical factors M3 = Model 3: n = 350; Contro l variables + Social factors M4 = Model 4: n = 350; Control variables + Service factors M5 = Model 5; n = 334; Full (main effects) model (post diagnostics) Golowka, 2005; Walby, 2003b). However, the recovery movement strongly endorses the healing provided by support and views the abili ty to recognize when help is needed or companionship is desired and who best to c ontact as necessary abilities for those in recovery (Corrigan & Phelan, 2004; Johns on, Lundstrom, Aberg-Wistedt, & Mathe, 2003; Laudet, Magura, Vogel, & Knight, 2000). The control model for support recognition is displayed in Table 69. The model is significant (F = 2.65, .05) and accounts for 2% of the variance. Age is the variable that drives the model, with y ounger age associated with a higher likelihood of recognizing support. Younger individuals may naturally be in a sate of greater need for support

PAGE 320

301 Table 69 Results of OLS Regression Testing the Impact of Control Variables on Recovery Strategy 4 (Recognizing Support) Variable b SE t Age -0.01 0.00 -1.98 a Gender -0.08 0.11 -0.75 Income 0.08 0.05 1.82 Education -0.05 0.06 -0.89 Note: a .05 n = 350; R 2 = .02; F = 2.65, p .05 and may be in living situations that more naturally provide support. Income approaches significance with higher levels of income a ssociated with more support recognition. Research question 3.1: Are individual factors asso ciated with recognizing support? The results of the OLS regression m odel examining the relationship between domain 1 variables and recogni tion of support are summarized in Table 70. The model is significant (F = 2.83, p .0001, adjusted R 2 = .14). Depression has a significant effect in lowering the endorsement of support r ecognition. Anti-depressant medication approaches significance and has the opposite as sociation of depression symptoms. The diagnostic category of other has a positive asso ciation with strategy. Individuals in this category have a range of diagnoses including dissociative disorders, head injury, and other neurological disorders that may app ear more obvious to casual observation then individuals with other il lnesses. On the other hand, these disorders are more likely to be viewed as beyond the control of the responde nt and subsequently less susceptible to

PAGE 321

302 Table 70 Results of OLS Regression Testing the Impac t of Clinical/Historical (Domain 1) on Recovery Strategy 4 (Recognizing Support) H 1 Variable b SE t Age -0.00 0.01 -0.27 Gender 0.06 0.12 0.53 Income 0.06 0.05 1.27 Education -0.03 0.06 -0.57 3.1a Other diagnosis 0.74 0.25 2.95 b 3.1b Somatization 0.04 0.08 0.51 3.1c Obsessive-compulsive 0.19 0.11 1.71 3.1d Interpersonal sens itivity -0.01 0.11 -0.12 3.1e Anxiety -0.00 0.11 -0.02 3.1f Depression -0.45 0.11 -3.98 d 3.1g Hostility -0.01 0.07 -0.22 3.1h Phobic anxiety -0.08 0.08 -0.99 3.1i Paranoia -0.09 0.11 -.095 3.1j Psychoticism 0.12 0.10 1.16 3.1k Substance abuse history 0.05 0.11 0.47 3.1l Hospitalization history -0.00 0.00 -0.84 3.1m Hospitalized in last year 0.08 0.11 0.73 3.1n Age of onset -0.00 0.00 -0.10 3.1o Medication: Anti-depressant 0.23 0.12 1.75 3.1p Medication: Anti-psychotic 0.02 0.11 0.18 3.1q Medication: Anti-manic -0.11 0.12 -0.93 3.1r Medication: Anti-anxiety -0.09 0.11 -0.80 3.1s Medication: Other psychotropic -0.03 0.11 -0.29 3.1t Current employment -0.09 0.14 -0.62 3.1u Years of employment -0.01 0.01 -1.90 3.1v Nuclear family mental illness 0.04 0.04 1.13 3.1w Extended family mental illness -0.09 0.08 -1.08 3.1x Child sexual abuse history -0.00 0.12 -0.00 3.1y Child physical abuse history -0.03 0.12 -0.28 3.1z Adult sexual assault history 0.02 0.14 0.18 3.1aa Adult physical assault history -0.10 0.12 -0.79 Note: b .01; d .0001 n = 350 R 2 = .14; F = 2.83, p .0001

PAGE 322

303 societal stereotyping and endorsement (Hin shaw & Cicchetti, 2002; Phelan, 2002). This may, in turn, lead to greater recognition of support. Research question 3.2: Are social factors associated with recognizing support? The social factors contained in dom ain 2 explain 52% of the variance of recognizing support (F = 20.12, p .0001) (Table 71). Only one stigma scale, alienation, was significantly associated with recognition of support with more alienation associated with less recognition of support. Alienation is an e xpression of internalized stigma that reduces contact with others (Goffman, 1963). Alienation also has been linked to denial and development of an overly individualistic loner personality style (Mahoney-Holst, 2005). Support variables were expected to be highly associated with support recognition. Partner or best friend support and family support followed the expected pattern. Against prediction, provider support and support from friends were not significantly associated. Further, community involvement was negativel y associated with recognizing support. The assumption was that exposure to comm unity members and organizations would increase recognition of support. A possible explanation is community involvement could have been viewed as provision of support and not recognition of support needed by respondents. The empowerment variables were rela tively unassociated with recognition of support. Self-esteem is positively, though weakly, associated with support recognition. Self-esteem in previous st udies has been found to medi ate the relationship between negative emotion and the executive decisi on making functions of the self, which by extension could include the ability to r ecognize support and the w illingness to seek

PAGE 323

304 assistance when in need (Neiss, Stevenson, Sedikidas, Kumashiro, Finkel, & Rusbuilt, 2005). Table 71 Results of OLS Regression Testing the Im pact of Social Variables (Domain 2) on Recovery Strategy 4 (Recognizing Support) H 1 Variable b SE t Age -0.00 0.00 -1.67 Gender 0.06 0.08 0.69 Income -0.01 0.03 -0.27 Education 0.00 0.04 0.02 Stigma 3.2a Alienation -0.25 0.10 -2.42 b 3.2b Stereotype endorsement -0.01 0.11 -0.13 3.2c Discrimination experience 0.03 0.09 0.34 3.2d Social withdrawal -0.01 0.12 -0.06 3.2e Stigma resistance -0.16 0.10 -1.56 Social Support and Related 3.2f Partner or best friend support 0.25 0.04 5.66 d 3.2g Family support 0.21 0.03 7.94 d 3.2h Provider support 0.06 0.05 1.11 3.2i Friend support 0.03 0.04 0.78 3.2j Community involvement -0.08 0.04 -2.18 a 3.2k Trust 0.00 0.04 0.12 Empowerment 3.2l Self-esteem/self-efficacy 0.21 0.10 2.01 a 3.2m Power/powerlessness 0.05 0.10 0.46 3.2n Community activism and autonomy 0.01 0.10 0.08 3.2o Optimism and control over the future 0.11 0.10 1.06 3.2p Righteous anger 0.10 0.07 1.40 Note: a .05; b .01; d .0001 n = 350; R 2 = .52; F = 20.12, p .0001

PAGE 324

305 Research question 3.3: Are service factors associat ed with recognizing support? The results of OLS regression examining the relationship of support recognition and service variables are summarized in Table 72. Younger individu als are more likely to recognize support in this model. Satisfacti on with services are st rongly associated with recognition of support. The number of servi ces and contact hours per month were not significant. Table 72 Results of OLS Regression Testing the Im pact of Service Variables (Domain 3) on Recovery Strategy 4 (Recognizing Support) Variable b SE t Age -0.01 0.00 -2.82 b Gender -0.04 0.10 -0.37 Income -0.06 0.04 1.32 Education -0.05 0.05 -0.98 3.3a Total number of services 0.05 0.05 1.09 3.3b Total service hours per month 0.01 0.01 0.85 3.3c Average satisfaction score 0.38 0.05 8.03 d Note: b .01; d .0001 n = 350 R 2 = .18; F = 12.21, p .0001 Research question 3: Are individual, social or serv ice factors associated with recognizing support? Main effects model The main effects model targeting su pport recognition accounts for 60% of the variance (F = 10.98, p .0001). The results are displayed in Table 73. Diagnostic analysis of overly influe ntial scores resulted in the rem oval of thirteen observations from main effect analysis due to viol ation of preset cutpoints.

PAGE 325

306 Table 73 Main Effects Results of OLS Regression Testi ng the Impact of Domain 1-3 Variables on Recovery Strategy 4 (Recognizing Support) H 1 Variable b SE t Age -0.00 0.00 -1.11 Gender -0.03 0.09 -0.39 Income -0.02 0.03 -0.64 Education 0.01 0.04 0.25 Domain 1 3.1a Other diagnosis 0.39 0.18 2.14 a 3.1b Somatization 0.02 0.06 0.28 3.1c Obsessive-compulsive -0.02 0.08 -0.22 3.1d Interpersonal sens itivity 0.03 0.08 0.37 3.1e Anxiety -0.04 0.08 -0.49 3.1f Depression -0.07 0.08 -0.78 3.1g Hostility -0.02 0.05 -0.34 3.1h Phobic anxiety -0.01 0.06 -0.19 3.1i Paranoia 0.03 0.07 0.40 3.1j Psychoticism 0.02 0.08 0.31 3.1k Substance abuse history -0.04 0.08 -0.48 3.1l Hospitalization history 0.00 0.00 0.06 3.1m Hospitalized in last year 0.03 0.08 0.37 3.1n Age of onset -0.00 0.00 -0.40 3.1o Medication: Anti-depressant 0.11 0.10 1.16 3.1p Medication: Anti-psychotic -0.05 0.09 -0.59 3.1q Medication: Anti-manic -0.09 0.09 -1.00 3.1r Medication: Anti-anxiety -0.01 0.08 -0.18 3.1s Medication: Other psychotropic 0.11 0.08 1.34 3.1t Current employment -0.00 0.10 -0.03 3.1u Years of employment -0.00 0.00 -0.57 3.1v Nuclear family mental illness 0.07 0.03 2.15 a 3.1w Extended family mental illness -0.11 0.06 -1.95 a 3.1x Child sexual abuse history 0.08 0.09 0.87 3.1y Child physical abuse history -0.05 0.09 -0.56 3.1z Adult sexual assault history 0.03 0.10 0.34 3.1aa Adult physical assault history -0.03 0.09 -0.38 Domain 2 Stigma 3.2a Alienation -0.22 0.10 -2.15 a 3.2b Stereotype endorsement -0.01 0.11 -0.10 3.2c Discrimination experience -0.03 0.09 -0.29 3.2d Social withdrawal 0.04 0.12 0.34 3.2e Stigma resistance -0.11 0.11 -1.06

PAGE 326

307 Table 73 (cont.) H 1 Variable b SE t Social Support and Related 3.2f Partner or best friend support 0.22 0.04 5.09 d 3.2g Family support 0.23 0.03 8.02 d 3.2h Provider support -0.07 0.05 -1.32 3.2i Friend support 0.07 0.04 1.58 3.2j Community involvement -0.08 0.04 -2.08 a 3.2k Trust 0.05 0.04 1.28 Empowerment 3.2l Self-esteem/self-efficacy 0.06 0.11 0.52 3.2m Power/powerlessness 0.09 0.11 0.89 3.2n Community activism and autonomy 0.05 0.11 0.52 3.2o Optimism and control over the future 0.24 0.11 2.22 a 3.2p Righteous anger 0.07 0.07 0.94 Domain 3 3.3a Total number of services 0.06 0.04 1.39 3.3b Total service hours per month 0.00 0.01 0.40 3.3c Average satisfaction score 0.17 0.04 3.98 d Note: a .05; d .0001 n = 337; R 2 = .60; F = 10.98, p .0001 The domain 1 clinical/historical variables are weakly associated in the model with only other diagnosis and nuclear family mental il lness significantly positively associated with support recognition. Having close relatives that are mentally ill may model recognition of appropriate supports, setti ng the stage for help seeking. Conversely, extended family mental illness is negatively associated with support recognition. There is no evidence in the current analysis to explain why nuclea r and extended family members with mental illness would have the opposite association. One hypothesis is that extended family members may have had less direct effect and interaction with the respondent, generating

PAGE 327

308 stigma that was later internalized when the respondents mental illness developed (Goffman, 1963). Domain 2 social variables have the strongest association with the dependent variable. The stigma alienati on scale is negatively associat ed with support recognition. Partner, best friend support and family suppor t are positively associated and explain the bulk of the variance. Community involvement is negatively associated with support recognition. Optimism and control over the future is the one empowerment variable significantly associated with recognizing support. Optimis m and control accrue over time and it is reasonable that positive support cont ributed to the growth of these traits. A reciprocal process of support and optimism may then perpetuate recovery (Resnick, Rosenheck, & Lehman, 2004). Domain 3 variables contribute to explan ation of the relationship with support recognition chiefly through averag e satisfaction. Similar to the service model described in Table 72, satisfaction with services suggest s a comfort and attachment that could allow a person to be more open to the possibility of support, and thus recognizing it when it is present. The limited model (Table 74) is significant (F = 58.80, p .0001) and explains slightly more of the variance (61%) compared to the main effects model (60%). Each of the nine significant variables for main eff ects increased in degree of association and maintained direction of asso ciation. By maintaining, and even slightly surpassing, the variance explained, these variables appear to contribute the bulk of the associative power in relation to support recognition in this study. Partner or best friend support and

PAGE 328

309 Table 74 Reduced Model Results of OLS Regression Te sting the Impact of Domain 1-3 Variables on Recovery Strategy 4 (Recognizing Support) Variable b SE t Other Diagnosis 0.44 0.17 2.63 b Nuclear family mental illness 0.08 0.03 2.78 b Extended family mental illness -0.12 0.05 -2.26 a Alienation -0.32 0.06 -5.14 d Partner or best friend support 0.28 0.03 8.59 d Family support 0.23 0.02 9.44 d Community involvement -0.07 0.03 -2.31 a Optimism and control over the future 0.28 0.08 3.43 c Average satisfaction score 0.15 0.04 4.19 d Note: a .05; b .01; c .001; d .0001 n = 337; R 2 = .61; F = 58.80, p .0001 alienation gained the most in the limited model and would benefit from additional investigation into their effects. Figure 7 provides a graphic that brings together the contribution of each model to variance explained. Social factors contribute the most to variance explained with service and clinical factors approximately equal in magnitude. As noted, alienation and best friend support require additional attention to distinguish the mechanism(s) in which they influence support recognition. In addition, the processes that juxtapose the association between nuclear and extended family mental illness would benefit from systematic inquiry, as would the unexpected negative a ssociation between rec ognition of support and community involvement.

PAGE 329

Figure 7 Amount of Variance Explained by Each Domain for Recognizing Support Control Variables (R2 = .02) Domain 1: Clinical and Historical Factors (R2 = .14) Main Effects Model R2 = .60 Reduced Main Effects Model R2 = .61 Recognizing Support Domain 2: Social Factors (R 2 = .52) Domain 3: Service Factors (R 2 = .18) Significant associations remained relativ ely stable across models (Table 75). Clinical variable are often investigated in is olation from social and service factors. The changes noted in the clinical domain, and to a lesser extent in the so cial domain, suggest that this is an inappropriate practice. Fo r instance, depression was significant in the isolated clinical model in its association with su pport recognition, but th is effect was not significant after controlling for other variable s across domains. Similarly, familial mental 310

PAGE 330

311 Table 75 Summary Table: Results of OL S Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 4 (Recognizing Support) Model 1 Model 2 Model 3 Model 4 Model 5 Age Sex Income Education Psychosomatic Obsessive Compulsive Interpersonal Sensitivity Anxiety Depression Hostility Phobia Paranoia Psychoticism Other Diagnosis Substance History Lifetime Hospitalizations Hospitalized in Last Year Age of Onset Anti-Depressant Medication Anti-Psychotic Medication Anti-Manic Medication Anti-Anxiety Medication Other Psychotropic Medication Currently Employed Years Employed Lifetime Nuclear Family Mental Illness Extended Family Mental Illness Child Sexual Abuse Child Physical Abuse Adult Sexual Assault Adult Physical Assault Alienation Stereotype Endorsement Discrimination Occurrence Social Withdrawal Stigma Resistance Partner of Best Friend Support Family Support

PAGE 331

312 Table 75 (cont.) Provider Support Friends Support Community Involvement Trust Self-Esteem and Self-Efficacy Power and Powerlessness Community Activism and Autonomy Optimism and Control Over the Future Righteous Anger Total Number of Services Total Contact Hours Average Satisfaction M1 = Model 1: n = 350; Control variables M2 = Model 2: n = 350; Control variab les + Individual and historical factors M3 = Model 3: n = 350; Contro l variables + Social factors M4 = Model 4: n = 350; Control variables + Service factors M5 = Model 5; n = 337; Full (main effects) model (post diagnostics) illness was not significant when considering clinical variables only, though something within the combined model altered the association and familial illness was then significant. Further, the di rection of association was different depending on whether nuclear or extended family members illness was measured. Consistency across models lends support for the strength of an association as well. For this strategy, other diagnosis, aliena tion, partner or best friend support, familial support, community involvement, and servi ce satisfaction remained significant across models two through five. This information adds to the growing empirical evidence that different recovery strategies exist and that different factors are associated with them.

PAGE 332

313 Research Question 3: Are Individual, Socia l, or Service Factors Associated with Recovery Strategy #5: Help Seeking? The recovery movement has consisten tly acknowledged that active help seeking should be utilized when a problem that canno t be handled by the individual arises (Borg & Kristiansen, 2004). The strategy reduces isol ation and promotes tr ust. The holistic approach views help seeking as more then a clinical interv ention. Help seeking includes asking directions, acquiring assistance for household repairs and all other non-mental illness related needs and activities. Further, psychiatric assistance can be supplemented with consumer led interventions (Young, Chinma n, Forquer, Knight, Vogel, Miller, et al., 2005). A concern for help seeki ng is that the individual, if currently symptomatic to the point of impairment in cognition and/or judgm ent, may not make the correct decision or may ignore their need for assistance (Wild er-Willis, Shear, Steffen, & Borkin, 2002). The concept of recovery does not ignore this pos sibility, but does view it as a state that will, for the majority of individuals be transient and remit in time. The important role of help seeking in recovery is clear, though there is little empirical information that has identified fact ors that are strongly associated with the strategy of help seeking. He lp seeking was regressed against age, gender, income, and education (control variables, Table 76), with age as the on ly significant result. The estimate is small, but it indicates that help seek ing increases slightly with increasing age. The model is not significant and explains 1% of the variance.

PAGE 333

314 Table 76 Results of OLS Regression Testing the Impact of Control Variables on Recovery Strategy 5 (Help Seeking) Variable b SE t Age 0.01 0.00 2.35 b Gender 0.04 0.10 0.39 Income 0.07 0.04 1.79 Education -0.06 0.05 -1.17 Note: b .01 n = 350 R 2 = .01; F = 2.06, ns Research question 3.1: Are individual factors associated with help seeking? Domain 1 clinical/historical variable s were next evaluated for significant associations with help seeking (Table 77). The domain 1 mo del is not significant (F = 1.44, ns), and accounts for only 4% of the varian ce. For some individuals, disturbing or debilitating psychiatric symptoms and/or a history of disturbing experience (e.g., abuse, hospitalizations) would increase help seeking behavior. For others, symptoms such as severe depression may lead to hopelessness sufficient to prevent help seeking. The evidence from this study does not support the supposition. Age is again significant, with older individuals more likely to employ help seeking. One independent variable, depression (b is significantly associated w ith help seeking, suggesting that increased depression limits help seeking.

PAGE 334

315 Table 77 Results of OLS Regression Testing the Impac t of Clinical/Historical (Domain 1) on Recovery Strategy 5 (Help Seeking) H 1 Variable b SE t Age 0.02 0.01 2.96 b Gender 0.13 0.11 1.15 Income 0.07 0.04 1.70 Education -0.03 0.06 -0.60 3.1b Somatization 0.11 0.08 1.45 3.1c Obsessive-compulsive 0.03 0.10 0.32 3.1d Interpersonal sens itivity 0.07 0.10 0.74 3.1e Anxiety -0.04 0.10 -0.38 3.1f Depression -0.27 0.10 -2.64 b 3.1g Hostility 0.01 0.07 0.22 3.1h Phobic anxiety -0.06 0.07 -0.77 3.1i Paranoia 0.03 0.09 0.36 3.1j Psychoticism -0.04 0.09 -0.45 3.1k Substance abuse history 0.09 0.10 0.87 3.1l Hospitalization history 0.00 0.00 0.91 3.1m Hospitalized in last year 0.14 0.11 1.28 3.1n Age of onset -0.00 0.00 -0.77 3.1o Medication: Anti-depressant 0.07 0.12 0.60 3.1p Medication: Anti-psychotic -0.03 0.10 -0.32 3.1q Medication: Anti-manic -0.12 0.11 -1.07 3.1r Medication: Anti-anxiety -0.16 0.10 -1.57 3.1s Medication: Other psychotropic 0.01 0.10 0.07 3.1t Current employment -0.10 0.13 -0.84 3.1u Years of employment -0.00 0.00 -1.11 3.1v Nuclear family mental illness 0.02 0.04 0.51 3.1w Extended family mental illness -0.08 0.08 -1.03 3.1x Child sexual abuse history 0.06 0.11 0.57 3.1y Child physical abuse history 0.04 0.11 0.37 3.1z Adult sexual assault history -0.15 0.13 -1.13 3.1aa Adult physical assault history 0.05 0.11 0.42 Note: b .01 n = 350 R 2 = .04; F = 1.44, ns

PAGE 335

316 Research question 3.2: Are social factors associated with help seeking? Help seeking was next regressed on th e combined stigma, social support, and empowerment variables, resulting in a significant model (F = 6.61, p .0001), explaining Table 78 Results of OLS Regression Testing the Im pact of Social Variables (Domain 2) on Recovery Strategy 5 (Help Seeking) H 1 Variable b SE t Age 0.01 0.00 2.97 b Gender 0.13 0.09 1.50 Income 0.01 0.04 0.41 Education -0.04 0.05 -0.86 Stigma 3.2a Alienation -0.03 0.11 -0.23 3.2b Stereotype endorsement -0.04 0.12 -0.31 3.2c Discrimination experience -0.03 0.10 -0.31 3.2d Social withdrawal 0.07 0.13 0.51 3.2e Stigma resistance -0.21 0.11 -1.81 Social Support and Related 3.2f Partner or best friend support 0.08 0.05 1.60 3.2g Family support 0.00 0.03 0.09 3.2h Provider support 0.19 0.05 3.52 c 3.2i Friend support 0.03 0.05 0.63 3.2j Community involvement 0.03 0.04 0.86 3.2k Trust 0.05 0.04 1.13 Empowerment 3.2l Self-esteem/self-efficacy 0.37 0.12 3.24 c 3.2m Power/powerlessness 0.16 0.11 1.42 3.2n Community activism and autonomy -0.05 0.11 -0.13 3.2o Optimism and control over the future -0.07 0.11 -0.68 3.2p Righteous anger -0.09 0.08 -1.19 Note: b .01; c .001 n = 350; R 2 = .24; F = 6.61, p .0001

PAGE 336

317 24% of the variance (Table 78). Age again wa s significant with increased age associated with more help seeking. Defying prediction, non e of the stigma scales were significant, though impaired stigma resistance approached significance. Provider support was the only social support variable sign ificantly related to help seeking. The lack of significance for the other potential sources of help (signi ficant other, family, friends) may be due, in part, to the wording of the questions that comprise the dependent variable. The questions are non-specific, such that they do not re ference sources of he lp, only whether the individual is willing to seek and accept assi stance. Considering also that respondents took the surveys inside a mental health center, there may have been a context bias operating. Lastly, only self-esteem/self-e fficacy was significant for the empowerment construct. Thus, for each increase in se lf-esteem score, help seeking endorsement increases by 0.37, holding all other variables constant. Research question 3.3: Are service factors associated with help seeking? The next group of variables evaluated wa s the domain 3 service variables. The model was significant (F = 6.24, p .0001), and explains 10% of the variance (Table 79). Age was significant in this model once again. Number of services and contact hours were not associated with help seeking. However, satisfaction with services was strongly associated, with help seeking though the mechanism explaining this association is unclear. This finding has been consistent a nd likely, as stated pr eviously, reflects the safety and attachment of a positive therapeutic relationship that generates the trust needed to ask for and accept help from others as well as the benefits of competent service provision.

PAGE 337

318 Table 79 Results of OLS Regression Testing the Im pact of Service Variables (Domain 3) on Recovery Strategy 5 (Help Seeking) Variable b SE t Age 0.01 0.00 2.17 a Gender 0.06 0.09 0.61 Income 0.05 0.04 1.40 Education -0.07 0.05 -1.43 3.3a Total number of services 0.06 0.04 1.43 3.3b Total service hours per month -0.01 0.01 -1.57 3.3c Average satisfaction score 0.25 0.04 5.81 d Note: a .05; d .0001 n = 350 R 2 = .10; F = 6.24, p .0001 Research question 3: Are individual, social or serv ice factors associated with help seeking? Main effects model Using the same modeling plan as previous recovery strategies, the main effects model examines the associations between help seeking and the three independent variable domains. The results of the diagnostic an alysis for overly influential scores was the removal of seven observations due to cutpoint violations (n = 343 for the model). The model testing whether R 2 = 0 was significant, indicating that R 2 is not zero with a fair degree of confidence (F = 3.32, p .0001), accounting for 25% of the variance (Table 80). Age remains significant in this model at approximately the same magnitude and in the same direction. Higher levels of inte rpersonal sensitivity are associated with increased help seeking, representing the only si gnificant clinical va riable. The central aspect of interpersonal sensitivity, noted earlie r, is a pervasive lack of confidence, self

PAGE 338

319 Table 80 Main Effects Results of OLS Regression Test ing the Impact of Domain 1-3 Variables on Recovery Strategy 5 (Help Seeking) H 1 Variable b SE t Age 0.01 0.00 2.10 a Gender 0.16 0.10 1.57 Income 0.02 0.04 0.59 Education -0.01 0.05 -0.11 Domain 1 3.1b Somatization 0.02 0.07 0.31 3.1c Obsessive-compulsive 0.02 0.09 0.28 3.1d Interpersonal sensi tivity 0.20 0.09 2.18 a 3.1e Anxiety 0.03 0.09 0.35 3.1f Depression -0.12 0.10 -1.25 3.1g Hostility -0.01 0.06 -0.19 3.1h Phobic anxiety -0.09 0.07 -1.27 3.1i Paranoia 0.04 0.08 0.51 3.1j Psychoticism -0.07 0.09 -0.84 3.1k Substance abuse history 0.01 0.09 0.07 3.1l Hospitalization history 0.00 0.00 0.71 3.1m Hospitalized in last year 0.11 0.10 1.15 3.1n Age of onset -0.00 0.00 -0.37 3.1o Medication: Anti-depressant 0.03 0.11 0.29 3.1p Medication: Anti-psychotic -0.05 0.10 -0.47 3.1q Medication: Anti-manic -0.05 0.10 -0.46 3.1r Medication: Anti-anxiety -0.10 0.09 -1.02 3.1s Medication: Other psychotropic 0.01 0.10 0.13 3.1t Current employment -0.06 0.12 -0.48 3.1u Years of employment 0.00 0.00 0.38 3.1v Nuclear family mental illness 0.01 0.04 0.28 3.1w Extended family mental illness -0.07 0.07 -1.07 3.1x Child sexual abuse history 0.08 0.10 0.76 3.1y Child physical abuse history 0.02 0.10 0.19 3.1z Adult sexual assault history -0.04 0.12 -0.32 3.1aa Adult physical assault history -0.06 0.10 -0.58 Domain 2 Stigma 3.2a Alienation -0.02 0.12 -0.16 3.2b Stereotype endorsement 0.06 0.13 0.44 3.2c Discrimination experience -0.08 0.11 -0.74 3.2d Social withdrawal 0.02 0.14 0.15 3.2e Stigma resistance -0.17 0.13 -1.35

PAGE 339

320 Table 80 (cont.) H 1 Variable b SE t Social Support and Related 3.2f Partner or best friend support 0.06 0.05 1.09 3.2g Family support -0.01 0.03 -0.32 3.2h Provider support 0.14 0.06 2.19 a 3.2i Friend support 0.01 0.05 0.23 3.2j Community involvement 0.07 0.04 1.65 3.2k Trust 0.08 0.05 1.59 Empowerment 3.2l Self-esteem/self-efficacy 0.36 0.13 2.73 b 3.2m Power/powerlessness 0.25 0.12 2.05 a 3.2n Community activism and autonomy -0.65 0.13 -0.50 3.2o Optimism and control over the future -0.02 0.12 -0.14 3.2p Righteous anger -0.19 0.09 -2.20 a Domain 3 3.3a Total number of services 0.02 0.05 0.58 3.3b Total service hours per month -0.00 0.01 -0.41 3.3c Average satisfaction score 0.12 0.05 2.52 b Note: a .05; b .01 n = 343 R 2 = .25; F = 3.32, p .0001 consciousness and feelings of inferiority. Willingness to seek help may be a compensatory approach to meet needs with as little challenge to a fragile self as possible. Domain 2 social variables are moderately associated with help seeking. However, none of the stigma variables and only support from providers was signi ficantly associated with help seeking, consistent with the domain 2 model in isolation. Empowerment appears to be the construct most related to help seeking. Self-esteem/self-efficacy is positively associated with help seeking as is power/powerlessness. Positive self evaluations and experiences of personal power may ease the way to ask for help and also

PAGE 340

321 protect from negative self refl ections based on having to need help. Further, experiences of righteous anger are negatively associated with help seeking, suggesting that this expression of empowerment interferes with he lp seeking behavior. Righteous anger may lead to righteous reject ion of assistance. Finally, the services domain is significant only for average satisfaction. Next presented is the reduced model that contains only the variables that were significant in main effects (Table 81). The model is significant (F = 19.26, p .0001) and explains 27% of the variance, an increase over the main effect variance explained of 24%. Several changes are noted in these re sults compared to the main effects. Interpersonal sensitivity is no longer significant in this m odel. The direction remains the same but the magnitude is diminished. Power/powerlessness, an empowerment construct scale, is also no longer significant, but also reverses direction (b = -0.13, t = -1.62, ns). Table 81 Reduced Model Results of OLS Regression Te sting the Impact of Domain 1-3 Variables on Recovery Strategy 5 (Help Seeking) Variable b SE t Age 0.01 0.00 2.29 a Interpersonal sens itivity 0.08 0.04 1.73 Provider support 0.24 0.05 5.05 d Self-esteem/self-efficacy 0.54 0.09 6.11 d Power/powerlessness -0.13 0.08 -1.62 Righteous anger 0.29 0.10 2.82 b Average satisfaction score 0.10 0.04 2.30 a Note: a .05; d .0001 n = 343 R 2 = .27; F = 19.26, p .0001

PAGE 341

Righteous anger remains signifi cant but the degree is less and the direction is reversed, which now indicates that increased righteous an ger is associated with help seeking. Both provider support and self-esteem/self-efficacy substantially increase in magnitude. Figure 8 displays the contribution of each model towards explanation of variance. Help seeking appears to be heavily associated with social factors. However, the role of some of the social variables in association wi th help seeking is confused when comparing the main effects to the limited model. Note d earlier, the impact of clinical pressures Figure 8 Amount of Variance Explained by Each Domain for Help Seeking 322 Control Variables (R2 = .01) Domain 1: Clinical and Historical Factors (R2 = .04) Main Effects Model R2 = .25 Reduced Main Effects Model R2 = .27 Help Seeking Domain 3: Service Factors (R 2 = .10) Domain 2: Social Factors (R 2 = .24)

PAGE 342

323 would arguably justify help seeking behavior, though the da ta does not support this. The contribution of domain 3 is relegated to satisfaction with serv ices, though this is a powerful enough association to account for an appreciable amount of variance. The final analytic step for this recove ry strategy is to review and assess for consistencies across models (Table 82). Age is significant for models 1-5. All models suggest that increasing age is associated with help seeking. Clinical domain variables are generally unrelated to help s eeking. Depression is significant in the isolated model and interpersonal sensitivity in the full model, though both sign and magnitude for interpersonal sensitivity cha nges in the limited model. Pr ovider support, self-esteem and satisfaction with services are consistently associated with help seeking across models. Research Question 3: Are Individual, Soc ial or Service Factors Associated with Recovery Strategy #6: Symptom Eradication? Symptom eradication is a point of concern and even contention when discussing recovery. Individuals in the consumer moveme nt consistently ask that symptoms be less of a focus when discussing recovery, and go so far as to state that complete symptom relief is an artificial goal advocated by c linical services and is not necessary (Deegan, 2005). Further, empirical investigations of the effects of sympto ms on quality of life have found that many individuals are able to live fulfilling lives with some symptoms present (Lysaker, et al., 2006). Chapter 3 discussed how the dependent vari able recovery strategies were initially generated. The reader will r ecall that the sixth strategy, sy mptom eradication, is a one

PAGE 343

324 Table 82 Summary Table: Results of OL S Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 5 (Help Seeking) Model 1 Model 2 Model 3 Model 4 Model 5 Age Sex Income Education Psychosomatic Obsessive Compulsive Interpersonal Sensitivity Anxiety Depression Hostility Phobia Paranoia Psychoticism Substance History Lifetime Hospitalizations Hospitalized in Last Year Age of Onset Anti-Depressant Medication Anti-Psychotic Medication Anti-Manic Medication Anti-Anxiety Medication Other Psychotropic Medication Currently Employed Years Employed Lifetime Nuclear Family Mental Illness Extended Family Mental Illness Child Sexual Abuse Child Physical Abuse Adult Sexual Assault Adult Physical Assault Alienation Stereotype Endorsement Discrimination Occurrence Social Withdrawal Stigma Resistance Partner of Best Friend Support Family Support Provider Support

PAGE 344

325 Table 82 (cont.) Friends Support Community Involvement Trust Self-Esteem and Self-Efficacy Power and Powerlessness Community Activism and Autonomy Optimism and Control Over the Future Righteous Anger Total Number of Services Total Contact Hours Average Satisfaction M1 = Model 1: n = 350; Control variables M2 = Model 2: n = 350; Control variab les + Individual and historical factors M3 = Model 3: n = 350; Contro l variables + Social factors M4 = Model 4: n = 350; Control variables + Service factors M5 = Model 5; n = 343; Full (main effects) model (post diagnostics) item variable that was originally part of a tw o question factor that failed for inclusion as a multiple item strategy. However, the one item investigating the need for symptoms to be eliminated as necessary for recovery addresse d the issue of symptoms in recovery noted above. Because of the controversy, the deci sion was made to retain this single item variable and investigate it as a strategy to a ssess potential covariates of importance. In keeping with the analysis pattern displayed in the first five strategies, control variables are first assessed for associat ion with symptom eradication. Age, gender, income, and education (c ontrol variables) account for 4% of the variance of symptom er adication (F = 4.22, p .01). The results are summarized in Table 83. Both income and education are negatively associated with eradication, indicating that higher income and educati on are associated with less likelihood of

PAGE 345

326 endorsing symptom eradication. Income a nd education are highly correlated ( 2 (2, 350) = 3.45, p .0001) and may reflect greater exposur e to psychiatric literature and internalization of the stereotypic, and gene rally inaccurate, message that symptoms and mental illness are chronic conditions, much like diabetes, that requires constant care over the lifetime. Table 83 Results of OLS Regression Testing the Impact of Control Variables on Recovery Strategy 6 (Symptom Eradication) Variable b SE t Age 0.00 0.01 0.12 Gender -0.16 0.14 -1.12 Income -0.12 0.06 -2.08 a Education -0.23 0.07 -3.19 c Note: a .05; c .001 n = 350 R 2 = .04; F = 4.33, p .01 Research question 3.1: Are individual factors associated with symptom eradication? Symptom eradication was regr essed on domain 1 clinical/historical variables with a significant result (F = 1.89, p .01) that explained 7% of the variance (Table 84). Income and education, similar to the contro l variable analysis discussed above, are negatively associated with symptom eradication. Interpersonal sensitivity is positively asso ciated with symptom eradication. This is contrary to the predicted direction. Hypothetically, in creased sensitivity may be

PAGE 346

327 Table 84 Results of OLS Regression Testing the Impac t of Clinical/Historical (Domain 1) on Recovery Strategy 6 (Symptom Eradication) H 1 Variable b SE t Age 0.01 0.01 0.29 Gender 0.03 0.17 0.18 Income -0.14 0.06 -2.25 a Education -0.15 0.08 -1.98 a 3.1a Major depression diagnosis 0.44 0.19 2.28 a 3.1b Somatization -0.01 0.11 -0.06 3.1c Obsessive-compulsive 0.03 0.14 0.20 3.1d Interpersonal sensi tivity 0.39 0.14 2.72 b 3.1e Anxiety 0.01 0.15 0.09 3.1f Depression -0.14 0.15 -0.90 3.1g Hostility 0.05 0.10 0.55 3.1h Phobic anxiety -0.07 0.11 -0.69 3.1i Paranoia -0.23 0.13 -1.76 3.1j Psychoticism -0.03 0.14 -0.20 3.1k Substance abuse history -0.04 0.15 -0.25 3.1l Hospitalization history -0.01 0.01 -2.06 a 3.1m Hospitalized in last year 0.07 0.16 0.43 3.1n Age of onset 0.01 0.01 1.77 3.1o Medication: Anti-depressant -0.17 0.18 -0.93 3.1p Medication: Anti-psychotic 0.02 0.15 0.13 3.1q Medication: Anti-manic 0.12 0.16 0.71 3.1r Medication: Anti-anxiety -0.03 0.15 -0.20 3.1s Medication: Other psychotropic 0.01 0.15 0.09 3.1t Current employment 0.13 0.19 0.71 3.1u Years of employment -0.01 0.01 -0.96 3.1v Nuclear family mental illness -0.09 0.06 -1.60 3.1w Extended family mental illness 0.06 0.11 0.51 3.1x Child sexual abuse history -0.25 0.17 -1.53 3.1y Child physical abuse history 0.27 0.16 1.72 3.1z Adult sexual assault history -0.16 0.19 -0.86 3.1aa Adult physical assault history -0.01 0.17 -0.03 Note: a .05; b .01 n = 350 R 2 = .07; F = 1.89, p .01

PAGE 347

328 confused with shyness, fear of public speak ing, or other common experiences that may not be considered clinical but still add to the burden of needed treatment. As noted before, individuals with high levels of in terpersonal sensitivity may have greater compliance potential as they likely rely on others to a greater extent and thus may internalize messages suggesting that what they experience is transient. A diagnosis of major depression is al so positively associated with symptom eradication and again defies the predicted direction. This may be an artifact of the data or a more complex process. The two variables th at would be expected to at least approach significance in support of major depression ar e depression symptoms and anti-depressant medication. Neither of these variables is significant though both are in the predicted direction of increased symptoms or medication use would be negatively associated with symptom eradication. One additional variable, lifetime hospitalizations, is negatively associated with symptom eradication. Larger numbers of hosp italizations usually indi cate a more severe course of illness, implying a more constant st ate of active symptoms that, in turn, reduces belief in being symptom free. However, though statistically significant, the low parameter estimate suggests that this is not an issue of critical clinical relevance. Research question 3.2: Are social factors associated with symptom eradication? The next analysis focuses on the relati onship between domain 2 social variables and symptom eradication. The model is significant (F = 3.80, p .0001) and accounts for 14% of the variance, indicating th at social factors have a greate r association with belief in symptom eradication then clinical factors. The results are summarized in Table 85. Income and education follow the same patter n as noted in discussion of the control

PAGE 348

329 variable model above (Table 82). Higher leve ls of income and educ ation are associated with decreased belief in symp tom eradication. Stigma resistance is the lone stigma scale significantly associated with eradication. The positive association demonstrates Table 85 Results of OLS Regression Testing the Im pact of Social Variables (Domain 2) on Recovery Strategy 6 (Symptom Eradication) H 1 Variable b SE t Age -0.00 0.00 -0.57 Gender -0.01 0.14 -0.09 Income -0.16 0.06 -2.71 b Education -0.18 0.07 -2.46 b Stigma 3.2a Alienation 0.11 0.18 0.66 3.2b Stereotype endorsement -0.06 0.20 -0.31 3.2c Discrimination experience -0.12 0.16 -0.73 3.2d Social withdrawal -0.14 0.21 -0.67 3.2e Stigma resistance 0.41 0.18 2.29 a Social Support and Related 3.2f Partner or best friend support 0.02 0.08 0.29 3.2g Family support -0.03 0.05 -0.56 3.2h Provider support 0.09 0.09 1.07 3.2i Friend support -0.02 0.08 -0.38 3.2j Community involvement 0.12 0.06 2.00 a 3.2k Trust -0.11 0.07 -1.63 Empowerment 3.2l Self-esteem/self-efficacy 0.23 0.18 1.25 3.2m Power/powerlessness -0.54 0.18 -3.07 b 3.2n Community activism and autonomy -0.06 0.18 -0.35 3.2o Optimism and control over the future 0.62 0.18 3.51 c 3.2p Righteous anger -0.09 0.12 -0.69 Note: a .05; b .01; c .001 n = 350; R 2 = .14; F = 3.80, p .0001

PAGE 349

330 resistance to stigmatizing messages is accompanied by an increased belief in eradication. Community involvement is also positively associated w ith eradication. Experiences of community integration may enhance belongingness and increase hope in achieving a symptom free lifestyle, though this is speculative as the mechanisms that drive this and other associ ations noted can only be hypothesized. The empowerment scales have two significant associations. Contrary to the hypothesized direction, increased endorsement of personal power is associated with decr eased sanctioning of eradication. In this case, power may not be related to symptom eradication but with increased ability to cope with symptoms. Th e final significant association is in line with the predicted direction showing optimism and control over the future to be positively associated with symptom eradication. Research question 3.3: Are service factors associated with symptom eradication? Service variables were assessed in asso ciation with symptom eradication and the model was found to have an overall significant relationship (F = 3.09, p .01), explaining 4% of the variance. However, only the control variab les of income and education were significant contributors to the model (Table 86) and were negatively associated. None of the service variables approached significance. Research question 3: Are individual, social or serv ice factors associated with symptom eradication? Main effects model Symptom eradication was regressed on the combined independent variable domains 1-3 with a signifi cant result explaining 19% of the variance (F = 2.62, p .0001) (Table 87). The variables selected for their theoretical and empirical relations to

PAGE 350

331 Table 86 Results of OLS Regression Testing the Im pact of Service Variables (Domain 3) on Recovery Strategy 6 (Symptom Eradication) Variable b SE t Age 0.00 0.01 0.40 Gender -0.18 0.14 -1.28 Income -0.13 0.06 -2.18 a Education -0.24 0.07 -3.31 c 3.3a Total number of services -0.10 0.07 -1.47 3.3b Total service hours per month -0.00 0.01 -0.27 3.3c Average satisfaction score 0.03 0.07 0.46 Note: a .05; c .001 n = 350 R 2 = .04; F = 3.09, p .01 recovery are the least associated with this dependent variable in comparison to the other five recovery strategies inve stigated. Factors associated with symptom eradication are not substantially contained in this model. The final sample size for the main effect model was 342, indicating eight observation s were deleted from the analysis due to violations of at least two of three cutpoints for the diagnostic tests us ed to evaluate the data. The full model no longer shows a significant asso ciation between educat ion and eradication. Income remains significant in association wi th the dependent variable. Likewise, major depression and interpersonal sensitivity remains positively associated with symptom eradication, opposite of the predicted directi on. Paranoia was negatively correlated with the strategy, signifying a drop in the expectation of symptom erad ication of .33 for each one point increase in paranoia.

PAGE 351

332 Table 87 Main Effects Results of OLS Regression Test ing the Impact of Domain 1-3 Variables on Recovery Strategy 6 (Symptom Eradication) H 1 Variable b SE t Age 0.00 0.01 0.22 Gender 0.12 0.16 0.75 Income -0.17 0.06 -2.90 b Education -0.14 0.08 -1.78 Domain 1 3.1a Major depression diagnosis 0.47 0.19 2.49 b 3.1b Somatization -0.03 0.11 -0.31 3.1c Obsessive-compulsive 0.03 0.14 0.25 3.1d Interpersonal sens itivity 0.40 0.14 2.92 b 3.1e Anxiety -0.03 0.14 -0.23 3.1f Depression -0.02 0.15 -0.13 3.1g Hostility 0.08 0.10 0.85 3.1h Phobic anxiety 0.01 0.11 0.09 3.1i Paranoia -0.33 0.13 -2.52 b 3.1j Psychoticism -0.05 0.13 -0.41 3.1k Substance abuse history 0.05 0.14 0.36 3.1l Hospitalization history -0.01 0.01 -1.31 3.1m Hospitalized in last year 0.12 0.15 0.79 3.1n Age of onset 0.00 0.01 0.49 3.1o Medication: Anti-depressant -0.27 0.18 -1.53 3.1p Medication: Anti-psychotic 0.03 0.15 0.20 3.1q Medication: Anti-manic 0.12 0.16 0.78 3.1r Medication: Anti-anxiety 0.03 0.14 0.19 3.1s Medication: Other psychotropic 0.05 0.15 0.33 3.1t Current employment 0.17 0.18 0.96 3.1u Years of employment -0.00 0.01 -0.69 3.1v Nuclear family mental illness -0.07 0.06 -1.28 3.1w Extended family mental illness 0.02 0.11 0.21 3.1x Child sexual abuse history -0.22 0.16 -1.37 3.1y Child physical abuse history 0.23 0.15 1.51 3.1z Adult sexual assault history -0.19 0.18 -1.02 3.1aa Adult physical assault history -0.01 0.16 -0.08 Domain 2 Stigma 3.2a Alienation 0.19 0.18 1.01 3.2b Stereotype endorsement -0.27 0.20 -1.33 3.2c Discrimination experience -0.02 0.17 -0.10 3.2d Social withdrawal -0.13 0.22 -0.61 3.2e Stigma resistance 0.29 0.19 1.47

PAGE 352

333 Table 87 (cont.) H 1 Variable b SE t Social Support and Related 3.2f Partner or best friend support 0.11 0.08 1.38 3.2g Family support -0.03 0.05 -0.68 3.2h Provider support 0.13 0.10 1.31 3.2i Friend support -0.09 0.08 -1.23 3.2j Community involvement 0.19 0.06 2.84 b 3.2k Trust -0.12 0.08 -1.56 Empowerment 3.2l Self-esteem/self-efficacy 0.23 0.20 1.17 3.2m Power/powerlessness -0.65 0.19 -3.46 c 3.2n Community activism and autonomy -0.30 0.19 -1.55 3.2o Optimism and control over the future 0.70 0.19 3.77 c 3.2p Righteous anger -0.08 0.13 -0.61 Domain 3 3.3a Total number of services -0.04 0.08 -0.52 3.3b Total service hours per month -0.01 0.01 -0.60 3.3c Average satisfaction score -0.10 0.07 -1.31 Note: b .01; c .001 n = 342 R 2 = .19; F = 2.62, p .0001 Turning attention to the domain 2 social variables, none of the stigma scales are significantly associated with eradication. Community i nvolvement is the only support variable associated with eradication. Possible sec ondary gains from community involvement may place symptoms in a different perspective or distract attention away from symptoms, lessening impact and raisi ng endorsement of eradication. The same process may be working for the positive asso ciation noted for optimism and control over the future, an empowerment scale. The st rong negative associati on highlighted in the domain 1 isolated model (Table 84) between power/powerlessness and symptom

PAGE 353

334 eradication is repeated in the full model. As noted, this relation is against the predicted direction. Lastly, none of the service ( domain 3) variables were significant. The limited model investigating significant factors from the main effects model with symptom eradication is presented next (Table 88). The model is significant (F = 11.96, p .0001) and explains slightly less varian ce (18%) compared to the main effects model (19%). The majority of the variables increased in magnitude with the exception of interpersonal sensitivity and there were no ch anges in direction of associations. Figure 9 displays the variance explained for each model where eradication was regressed on sets of independent variables. Social factors e xplain the most variance in this model. The lack of explanation from the clinical dom ain 1 variables, considering the focus on symptoms in the strategy, was unexpected. However, with only 19% of the variance explained, it is conceivable that appropriate clinical variables remain untapped. The Table 88 Reduced Model Results of OLS Regression Te sting the Impact of Domain 1-3 Variables on Recovery Strategy 6 (Symptom Eradication) Variable b SE t Income -0.17 0.05 -3.09 c Interpersonal sens itivity 0.25 0.10 2.36 b Paranoia -0.30 0.10 -3.02 c Major depression diagnosis 0.47 0.16 2.85 b Community involvement 0.17 0.05 3.24 c Power/powerlessness -0.71 0.15 -4.64 d Optimism and control over the future 0.72 0.14 5.17 d Note: b .01; c .001 n = 342 R 2 = .18; F = 11.96, p .0001

PAGE 354

same argument can be used for service vari ables. Services are, obviously, a primary avenue for symptom control. Specific type s of services may have a greater impact on symptom eradication then others, though this is beyond the scope of the current study to evaluate. Table 89 summarizes the significant findings from each model pertaining to symptom eradication. Higher levels of income are associated with less endorsement of symptom eradication in all models. More educ ation was also negativel y associated with Figure 9 Amount of Variance Explained by Each Domain for Symptom Eradication Control Variables (R2 = .04) Domain 1: Clinical and Historical Factors (R2 = .07) Main Effects Model R2 = .19 Reduced Main Effects Model R2 = .18 Symptom Eradication Domain 2: Social Factors (R 2 = .14) Domain 3: Service Factors (R 2 = .04) 335

PAGE 355

336 Table 89 Summary Table: Results of OL S Regression Testing the Impact of Domain 1-3 Variables on Recovery Strategy 6 (Symptom Eradication) Model 1 Model 2 Model 3 Model 4 Model 5 Age Sex Income Education Psychosomatic Obsessive Compulsive Interpersonal Sensitivity Anxiety Depression Hostility Phobia Paranoia Psychoticism Major Depression Diagnosis Substance History Lifetime Hospitalizations Hospitalized in Last Year Age of Onset Anti-Depressant Medication Anti-Psychotic Medication Anti-Manic Medication Anti-Anxiety Medication Other Psychotropic Medication Currently Employed Years Employed Lifetime Nuclear Family Mental Illness Extended Family Mental Illness Child Sexual Abuse Child Physical Abuse Adult Sexual Assault Adult Physical Assault Alienation Stereotype Endorsement Discrimination Occurrence Social Withdrawal Stigma Resistance Partner of Best Friend Support Family Support

PAGE 356

337 Table 89 (cont.) Provider Support Friends Support Community Involvement Trust Self-Esteem and Self-Efficacy Power and Powerlessness Community Activism and Autonomy Optimism and Control Over the Future Righteous Anger Total Number of Services Total Contact Hours Average Satisfaction M1 = Model 1: n = 350; Control variables M2 = Model 2: n = 350; Control variab les + Individual and historical factors M3 = Model 3: n = 350; Contro l variables + Social factors M4 = Model 4: n = 350; Control variables + Service factors M5 = Model 5; n = 342; Full (main effects) model (post diagnostics) eradication in four of five models, with the main effects model approaching but not achieving significance. Interpersonal se nsitivity and power/p owerlessness were significant in the two models th ey were entered but in the di rection opposite of what were predicted in both models. The diagnosi s of major depression was also positively associated with eradication, against predicti on, but could represent a combination of the desire to be depression free and the hope that this may oc cur, and the burgeoning public awareness, mostly generated by pharmaceutical company advertisement, of depression and claims for potential symptom free lifestyles.

PAGE 357

338 Community involvement and optimism were consistently and positively associated with the belief in symptom er adication and represent openings for further investigation. Variables significant in earlier models that were not significant in main effects include lifetime hospitalizations (dom ain 1) and stigma resistance (domain 2). Summary of Research Question 3, Inves tigation of Recovery Strategies All the analyses ta rgeting the six recove ry strategies were performed in response to the hypotheses embedded in research quest ion 3. The point was made earlier in the chapter that addressing each hypothesis in turn was unwieldy. Instead, the presentation of results summarized in Tables 48-89 addre ssed in more general terms the findings for each hypothesis. Table 90 reports specifica lly for each hypothesis whether it was supported in multivariate analysis. The followi ng criteria were used for deciding whether the hypothesis was supported or not: Yes, if the main effects model supports the hypothesis No, if the main effects mode l does not support the hypothesis Partial, if the hypothesis is supported in a partial (singular domain) model but not in the main effects model Reverse, if the association was significant but in the opposite direction then predicted in the main effects model The results for yes, no, or partial are genera lly consistent with the literature base available The clinical/historical domain (domain 1) had multiple significant associations with recovery strategies, t hough with no clear pattern. Out of seventeen variables that

PAGE 358

339 were significantly associated, only the symptom scale interpersonal sensitivity (significant to four recovery strategies, of which one dropped out in analysis of the reduced model) and a primary diagnosis of an anxiety disorder (significant to two strategies, though one dropped out in the reduced model) were significant to more than one strategy. Thus, clinical c ovariates are important to reco very strategies, but different ones are important to different strategies. Th is further suggests that operationalization of the strategies described in this study could be effected by specific clinical or historical factors that would need to be accounted for in development of recovery programs. Symptom levels in bivariate analysis are negatively associated with virtually every recovery strategy. This uniformity of dir ection indicates that symptom control and coping strategies, if duplicated in a future ca usal research design, could be an important prerequisite for strategy utiliz ation. In multivariate results, three recovery strategies positive future orientation, meaningfulness, personal control, and hope, and symptom eradication were significantly negatively a ssociated with anxiet y, depression, hostility or paranoia. Anxiety requires special mention. Recovery lit erature, when focused at all on specific symptoms, has concentrated the bu lk of effort on studyi ng the relationship of depression to recovery. Depr ession was significantly but weakly associated with positive future orientation only and, diagnosticall y, major depression was significant only to symptom eradication. A diagnos is of anxiety along with symp toms of anxiety including phobia are associated with three strategies. Anxiety, in this study, has a more powerful negative association then depression when considering the strength of association and number of strategies affected. Indeed, symp toms have the highest degree of association with recovery strategies of the different clin ical/historical groupings in this study. Other

PAGE 359

340 Table 90 Summary of Multivariate Results fo r Research Question 3 Hypotheses Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom eradication Hypothesis Statement Supported? Supported? Supported? Supported? Supported? Supported? Hyp3.1a Less clinically severe diagnoses for participants will be associated with greater endorsement for a recovery strategy Yes Yes Reverse Hyp3.1b Lower somatization symptoms reported by participants will be as sociated with greater endorsement for a recovery strategy Hyp3.1c Lower obsessive-compulsive symptoms reported by participants will be associated with greater endorsement for a recovery strategy Hyp3.1d Lower interpersonal sensitivity reported by participants will be as sociated with greater endorsement for a recovery strategy Reverse Reverse Reverse Reverse Hyp3.1e Lower symptoms of anxiety reported by participants will be as sociated with greater endorsement for a recovery strategy Reverse Hyp3.1f Lower symptoms of depression reported by participants will be as sociated with greater endorsement for a recovery strategy Partial Yes Partial Partial Hyp3.1g Lower hostility reported by participants will be associated with greater endorsement for a recovery strategy Yes Hyp3.1h Lower phobic anxiety reported by participants will be associated with greater endorsement for a recovery strategy Yes Partial

PAGE 360

341 Table 90 (cont.) Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom eradication Hypothesis Statement Supported? Supported? Supported? Supported? Supported? Supported? Hyp3.1i Lower paranoid ideation reported by participants will be associated with greater endorsement for a recovery strategy Yes Hyp3.1j Lower psychoticism reported by participants will be associated with greater endorsement for a recovery strategy Hyp3.1k Absence of a comorbid substance use disorder will be associated with greater endorsement for a recovery strategy for participants Reverse Hyp3.1l Lower numbers of lifetime psychiatric hospitalizations will be associated with greater endorsement for a recovery strategy for participants Yes Hyp3.1m No psychiatric hospitalization in the last year will be associated with greater endorsement for a recovery strategy for participants Yes Hyp3.1n Older age at onset of disorder will be associated with greater endorsement for a recovery strategy for participants Hyp3.1o Absence of prescribed anti-depressant medication will be associated with greater endorsement for a recovery strategy for participants Reverse Partial Hyp3.1p Absence of prescribed anti-psychotic medication will be associated with greater endorsement for a recovery strategy for participants Yes

PAGE 361

342 Table 90 (cont.) Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom eradication Hypothesis Statement Supported? Supported? Supported? Supported? Supported? Supported? Hyp3.1q Absence of prescribed anti-manic medication will be associated with greater endorsement for a recovery strategy for participants Hyp3.1r Absence of prescribed an ti-anxiety medication will be associated with greater endorsement for a recovery strategy for participants Yes Hyp3.1s Absence of any other prescribed psychotropic medication will be associated with greater endorsement for a recovery strategy for participants Hyp3.1t Being currently employed will be associated with greater endorsement for a recovery strategy for participants Hyp3.1u Greater number of years worked will be associated with greater endorsement for a recovery strategy for participants Reverse Partial Reverse Partial Hyp3.1v Absence of familial mental illness in first-degree relatives will be associated with greater endorsement for a recovery strategy for participants Reverse Hyp3.1w Absence of familial mental illness in extended family members will be associated with greater endorsement for a recovery strategy for participants Yes Hyp3.1x Absence of child sexual abuse will be associated with greater endorsement for a recovery strategy for participants Yes

PAGE 362

343 Table 90 (cont.) Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom eradication Hypothesis Statement Supported? Supported? Supported? Supported? Supported? Supported? Hyp3.1y Absence of child physical abuse will be associated with greater endorsement for a recovery strategy for participants Hyp3.1z Absence of adult sexual assault will be associated with greater endorsement for a recovery strategy for participants Hyp3.1aa Absence of adult physical assault will be associated with greater endorsement for a recovery strategy for participants Reverse Partial Hyp3.3a Lower feelings of alienation will be associated with higher recovery expectancy for participants Partial Yes Hyp3.3b Lower respondent endorsement of mental illness stereotypes will be associat ed with higher recovery expectancy for participants Yes Hyp3.3c Lower number of discrimination experiences will be associated with higher recovery expectancy for participants Hyp3.3d Lower endorsement of soci al withdrawal will be associated with higher recovery expectancy for participants Yes Hyp3.3e Greater endorsement of stigma resistance will be associated with higher recovery expectancy for participants Yes Yes Yes Hyp3.3f Greater support through intimate partner or a best friend will be associated with higher recovery expectancy for participants Yes

PAGE 363

344 Table 90 (cont.) Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom eradication Hypothesis Statement Supported? Supported? Supported? Supported? Supported? Supported? Hyp3.3g Greater support through family members will be associated with higher recovery expectancy for participants Yes Yes Hyp3.3h Greater support via mental health providers will be associated with higher recovery expectancy for participants Yes Yes Yes Hyp3.3i Greater support through friendship will be associated with higher recovery expectancy for participants Yes Hyp3.3j Greater feelings of being connected to the community will be associat ed with higher recovery expectancy for participants Yes Yes Reverse Yes Hyp3.3k Greater trust in the motivation of others will be associated with higher recovery expectancy for participants Hyp3.3l Higher ratings of self-est eem and self-efficacy will be associated with higher recovery expectancy for participants Yes Yes Partial Yes Hyp3.3m Higher ratings of personal power will be associated with higher recovery expectancy for participants Yes Yes Hyp3.3n Greater involvement in the community will be associated with higher recovery expectancy for participants Yes Hyp3.3o Greater optimism and confidence in personal control over the future will be associated with higher recovery expectancy for participants Yes Yes Yes Yes

PAGE 364

345 Table 90 (cont.) Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control and Hope Recognizing Support Help Seeking Symptom eradication Hypothesis Statement Supported? Supported? Supported? Supported? Supported? Supported? Hyp3.3p Higher ratings of righteous anger will be associated with higher recovery expectancy for participants Reverse Hyp3.3a Total number of services will be associated with higher recovery expectancies for participants Hyp3.3b The average number of contact hours per month (averaged over one year of service) will be associated with higher re covery expectancies for participants Partial Hyp3.3c Satisfaction level with services will be associated with higher recovery exp ectancies for participants Partial Partial Partial Yes Yes Note: Partial indicates that the hypothesis is supported in an isolated (singular domain) model but not in the main effects mo del Reverse indicates that the association was signifi cant but in the opposite direction then predicted in the main effe cts model

PAGE 365

346 groupings include diagnosis, medication, hospitalization, familial mental illness, employment, and abuse history. Historical factors such as child abuse, familial mental illness, employment, and hospitalization history were not strongly associ ated with recovery st rategies. Of special interest is the lack of significance for abuse and assault experiences in multivariate analysis. Though 41% of males and 65% of fe males in the study admitted to one or more experiences with child sexual abuse, child physical abuse, adult se xual assault and/or adult physical assault, th e lack of association with recove ry strategies was surprising. There are fourteen social independent variables in significant relationship with recovery strategies from domain 2, out of sixteen variables tested. Unlike domain 1 clinical variables, fully half of the variable s are significantly associated with two or more recovery strategies. Each st rategy is significant with at l east three covariates, accounting for the highest level of variance explained of the three domains. This provides direction for future research and potential reco very program planning with hope that operationalizing the social domain will affect multiple strategies, theoretically maximizing resource investment. The three sub-domains are not equally influential. Social support independent variables have th e most powerful association with recovery strategies followed closely by empowerment variab les, with all six strategies significantly associated with at least one support or empowerment variable. Stigmatization and discrimination vary in the magnitude of association with strategies 1-4 and have no significant association with stra tegies 5 (help seeking) and 6 (symptom eradication). A brief summary of results separately for each co nstruct of domain 2 is provided next due to the high degree of significant asso ciation with each recovery strategy.

PAGE 366

347 Social support. Social support variables include support from an intimate partner or best friend, family support, provider support, support from friends, community involvement and trust. Both in bivariate and multivariate analysis, tr ust was not significantly associated with any recovery strategy. Partner, family, friend, and provider support were each significantly positively associated with all recovery strate gies except symptom eradication in bivariate analysis. For symptom eradication, only provider support and community involvement were significant. None of the correlations were of exceptional magnitude ranging from low to barely moderate in strength. In multivariate analysis, it is interest ing that social support is significantly associated with each recovery strategy, though which variables are significant is not consistent across strategies. Support from family members and by a provider is the two most powerfully associated support variable s. Importance of familial support is consistent with previous research and will be elaborated in Chapter 5. Provider support has not been measured as a support covariate to recovery in previ ous research and was significant in association with positive fu ture orientation, meaningfulness, personal control, and hope, and help seek ing. What cannot be detected in this data is why provider support is so significant and wh ether providers are substitute supports for other, more mainstream, relationships. Chapter 2 deta iled the limited social networks of many individuals with mental illness, which le nds credence to the idea of provider as replacement of trusted support normally given by family and friends. Community involvement is significantly asso ciated with four recovery strategies. This supports the recovery movements emphasis on being part of and valued by

PAGE 367

348 community while at the same time valuing the community in return. In retrospect, more precise measurement of whether the responde nt felt that they were valued by their community would have measured the reciprocity aspect of this variable with greater precision. What can be stated is that belongingness to comm unity is a significant aspect of recovery that is more diffuse than the targeted personal relati onships measured by the other scales of the support constr uct and yet of e qual importance. Empowerment Empowerment variables of self-estee m/self-efficacy, power/powerlessness, community activism and autonomy, and optimism and control over the future were significantly positively associated with recovery strategies 1-5 (excluding strategy 6, symptom eradication) in bivariate analysis except for power/powerless correlated with strategy 3 (meaningfulness, personal contro l, and hope), which was surprisingly weak (r(348) = .07, ns). The magnitude of the associations betwee n empowerment, as a central concept of recovery, and recovery strategies would be expected to be strong. However, the majority of the correlations are mild to m oderate with only stra tegies 1-2 correlated above .60 with self-esteem/self-efficacy. A fu ture analysis will investigate whether the SMI or OP population samples or stratificati on by recovery expectan cy would explain the unexpectedly weak a ssociations noted. In multivariate analysis, self-esteem/self-efficacy was the most consistently significant covariate across recove ry strategies. Indeed, much like social support, one or more empowerment variables were strongly associated with each of the recovery strategies. A surprise was power/powerlessne ss being only significant in association with help seeking and symptom eradication. A sense of personal power is consistently

PAGE 368

349 mentioned in the recovery literature and this study may be m easuring the construct differently then how the qualitative and firs t person literature describes it at a more individual level. Optimism and control over the future was weakly associated with two recovery strategies (meaningfulness, personal control, and hope and recognizing support), moderately associated with effective illne ss management and strongly associated with symptom eradication. Surprisingly, optimis m and control over the future was not significantly associated with strategy 2, positive future orientation though it was positively associated with effective illness management, meaningfulness, personal control and hope, and sympto m eradication. Stigma The results outlined earlier in this chap ter shows that the bivariate correlations between stigma variables and recovery strate gies are, but for two exceptions, negative. All relationships are statistic ally significant between the stigma variables (alienation, stereotype endorsement, discrimination occu rrence, social wit hdrawal and stigma resistance) and five of six recovery strategi es. Only strategy 6, symptom eradication, is not significantly associated w ith stigma measures. This fi nding supports the results from the qualitative literature, wh ich will be fully discussed in Chapter 5. Stigma does negatively affect recovery for respondents w ith the SMI sample reporting more stigma experiences and the OP sample greater stigma resistance. The reader will also recall from the discussion of recovery expectancy above that individuals that do not believe they will recover experience more stigmati zation effects as evidenced by significantly higher scores on every stigma scale. Thus, severity of illness and non-belief in recovery are associated with stigma experience s in direct bivariate relationships.

PAGE 369

350 In multivariate analysis, stigma variables are not significantly related to help seeking or symptom eradicati on. Alienation, a cornerstone of concern for the recovery movement, is only weakly associated with one recovery strategy, recognizing support. Personal endorsement of mental illness stereotypes is significant only with meaningfulness, personal control, and hope (strategy 3). Social wit hdrawal is also only associated with one strategy, in this case posi tive future orientation. Stigma resistance is the only stigma variable signi ficant to two strategies, eff ective illness management and meaningfulness, personal control, and hope. Thus, much like the clinical variables discussed in the beginning of this summary section, stigma does appear to play a role in sanctioning of recovery strate gies, with different aspects of stigma associating with different strategies. When considering domain 3, services are a fact of life for individuals with mental illness. Not all the afflicted are served and many in need of services cannot access them. Clinical services are still the mainstay of treatment and the number of services, the amount of time spent receiving services, and satisfaction with serv ices are reasonable proxies for severity of mental illness. T hus, the SMI sample received significantly more services and total number of contact hours per month compared to the OP sample. Satisfaction with services was not significantly different. Th e results are reversed when segmenting by recovery expectancy instead of severity. There is no significant difference in number of services or time spent in services by expectancy status. However, those not expecting to recover are significantly less sa tisfied with the services they receive. Noted at several points in this document, the service domain (domain 3) was the least rigorous in analysis. Of the three variables assessed, only satisfaction with services

PAGE 370

351 was consistently significant in bivariate analys is with recovery strategies. In multivariate analysis, service satisfaction was exceptionally significant in association with strategy 4 (recognizing support) and strate gy 5 (help seeking). A point of concern is the lack of association between service related variables and effec tive illness management and symptom eradication, arguably the most clinically related recove ry strategies, as well as positive future orientation and meaningfulness, personal control, and hope, the two nonassociated strategies that reflect the recovery movement. This concludes the presentation of results targeting research question three. The following sections will address the effects of recovery expectancy (the dependent variable for research questi on 2) as a mediator between domains 1-3, for evaluating research question 4a, and in interaction with specific variables (m oderating effects) for research question 4b. In addition, the moderati ng effects of severity of mental illness will be assessed. Focus on Mediation Figure 1, page 6, Chapter 1, displays a proposed direct relationship between domain 1-3 independent variab les and recovery strategies (dependent variables). The results pertinent to these relationships were described in the s ection above addressing research question 3. Figure 1 also displays with recovery strategi es are arbitrated by recovery expectancy. Research question 4a, addressing mediation, and accompanying hypotheses are reproduced below and target ed in the following analysis. RQ4a: Does the expectation of recovery mediate the relationship between individual, social and service factors and recovery strategies?

PAGE 371

352 Hyp4a.1: Participants expectation of recovery will mediate the relationship between individual factors and recovery strategies. Hyp4a.2: Participants expectation of recovery will mediate the relationship between social factors a nd recovery strategies. Hyp4a.3: Participants expectation of recovery will mediate the relationship between service factors a nd recovery strategies. The subsequent discussion will present the results from the main effects models only. The process articulated when describing results for research que stions 2 and 3 that included reporting on each model in turn will not be duplicated in the discussion of this section. Using a progression of models a llows observation of changes when new variables are entered. However, the results of all main effe cts mediating relationships investigated were non-significant meaning only very minor partial mediating effects were noted, if present at all, and no full medi ation of effects due to recovery expectancy were uncovered. Some of the singular dom ain models had slightly more intense mediating effects. Tables summarizing the pa rtial mediating models separately for each domain and recovery strategy are included in Appendix E. The focus of this research from the st art was to describe the respondents and the information they offered with full apprecia tion of the context in which the respondents live. Main effects models capture the cont ext to the greatest degree since the lived experience of each respondent includes clinic al, social and service factors, experiences, and relationships simultaneously. However, though non-significant, it is still useful to briefly review the results of mediated main effect models and present them in summary

PAGE 372

353 tables to fully appreciate the results with and without the addition of the mediating variable. Mediating effects are det ected primarily by observing th e change in the parameter estimate and standard error. To aid in obs ervation of changes, the following tables contain only the signif icant variables of the main effect and mediating models for each of the six recovery strategies. To detect medi ation, the main effect model was run with the mediating variable added and then comparis ons between parameter estimates, full model F-statistics and changes in variance explained (R 2 ) were assessed. What is the Mediating Effect of Recovery Expectancy on Do mains 1-3 for each of the Recovery Strategies? Tables 91-96 summarizes the findings from the mediating model and compares it to the main effects model for each strategy. Discussion of strate gy 1, effective illness management, summarized in Table 91, is offere d as an example. The same points could be made in a discussion of each of the other strategies and their relationship to the mediating effect of recovery expectancy. Re ferencing Table 91, there is no change in variance explained (0.67 for both models) and the strength of the full model is only slightly reduced (F = 14.31, p .0001 versus 14.64, p .0001). Every parameter estimate is approximately the same with some minor fluctuation noted in the corresponding t-test with th e exception of interpersonal sensitivity which has a .01 change in the estimate. Thus, there is neith er full nor substantive partial mediation via recovery expectation of any independent vari able in relation to its association with effective illness management. Such minute ch anges in parameter estimates would predict

PAGE 373

354 a very small and nonsignificant parameter es timate for recovery expectancy, and this prediction is met. Table 92 summarizes the mediating model results for positive future orientation and is followed by Table 93 (mean ingfulness, personal control, and hope), Table 94 (recognizing support), Table 95 (h elp seeking), and Table 96 (symptom eradication). The lack of mediating effect for recovery expectancy demonstrates no support for research question 4a and hypothese s 4a.1 to 4a.3. Possible reasons for the lack of significance as a mediator will be discussed in Chapter 5. Table 91 Mediating Effect of Recovery Expectancy for the Association Between Domain 1-3 Variables and Recovery Strategy 1 (Effective Illness Management) Main Effects Model Mediating Model Variable b (SE) t b (SE) t Interpersonal sensitivity 0.15 (0.05) 2.82 b 0.14 (0.05) 2.74 b Anxiety diagnosis -0.20 (0.09) -2.30 a -0.20 (0.09) -2.30 a Medication: Anti-anxiety -0.16 (0.06) -2.99 b -0.16 (0.06) -2.93 b Stigma resistance -0.40 (0.08) -5.31 d -0.40 (0.07) -5.30 d Community involvement 0.08 (0.03) 3.18 b 0.08 (0.02) 3.15 b Self-esteem/self-efficacy 0.28 (0.08) 3.69 c 0.28 (0.08) 3.68 c Optimism and control over the future 0.18 (0.07) 2.57 b 0.18 (0.07) 2.58 b Recovery expectancy 0.02 (0.05) 0.41 Model Fit Statistics Model F 14.64 d 14.31 d Adjusted R 2 0.67 0.67 Note: a .05; b .01; c .001; d .0001 n = 335

PAGE 374

355 Table 92 Mediating Effect of Recovery Expectancy in the Association Between Domain 1-3 Variables and Recovery Strategy 2 (Positive Future Orientation) Main Effects Model Mediating Model Variable b t b t Education -0.05 (0.03) -2.07 a -0.05 (0.02) -2.09 b Anxiety 0.12 (0.04) 2.73 b 0.13 (0.05) 2.77 b Depression -0.11 (0.05) -2.21 a -0.10 (0.05) -2.05 a Phobic anxiety -0.12 (0.03) -3.31 c -0.11 (0.03) -3.23 c Schizophrenia diagnosis -0.19 (0.07) -2.68 b -0.19 (0.07) -2.67 b Substance abuse history 0.10 (0.05) 2.05 a 0.10 (0.05) 2.15 a Hospitalized in last year -0.10 (0.05) -2.07 a -0.10 (0.05) -2.15 a Social withdrawal -0.18 (0.07) -2.59 b -0.19 (0.07) -2.77 b Family support 0.05 (0.02) 3.11 b 0.05 (0.02) 3.03 b Provider support 0.10 (0.03) 3.25 c 0.10 (0.03) 3.31 c Community involvement 0.05 (0.02) 2.47 b 0.05 (0.02) 2.40 a Self-esteem/self-efficacy 0.67 (0.06) 10.34 d 0.66 (0.06) 10.37 d Recovery expectancy 0.07 (0.04) 1.45 Model Fit Statistics Model F 17.72 d 17.48 d Adjusted R 2 0.71 0.71 Note: a .05; b .01; c .001; d .0001 n = 335 Table 93 Mediating Effect of Recovery Expectancy in the Association Between Domain 1-3 Variables and Recovery Strategy 3 (Meaningfulness, Personal Control, and Hope) Main Effects Model Mediating Model Variable b t b t Interpersonal sensitivity 0.11 (0.04) 2.76 b .0.11 (0.04) 2.59 b Hostility -0.08 (0.03) 2.55 b -0.08 (0.03) -2.57 b Anxiety Diagnosis -0.24 (0.07) -3.43 c -0.24 (0.07) -3.42 c Medication: Anti-psychotic -0.12 (0.05) -2.58 b -0.12 (0.05) -2.49 b Child sexual abuse history 0.10 (0.05) 2.08 a 0.10 (0.05) 2.03 a Stereotype endorsement -0.20 (0.06) -3.34 c -0.19 (0.06) -3.18 c Stigma resistance -0.16 (0.06) -2.79 b -0.16 (0.06) -2.77 b Family support 0.04 (0.01) 2.46 b 0.04 (0.01) 2.36 b Provider support 0.07 (0.03) 2.55 b 0.08 (0.03) 2.62 b Friend support 0.06 (0.02) 2.46 b 0.06 (0.02) 2.48 b Community activism and autonomy 0.35 (0.06) 5.92 d 0.35 (0.06) 5.89 d Optimism and control over the future 0.11 (0.06) 1.98 a 0.11 (0.06) 2.04 a Recovery expectancy 0.07 (0.04) 1.63 Model Fit Statistics Model F 7.27 d 7.22 d Adjusted R 2 0.48 0.49 Note: a .05; b .01; c .001; d .0001 n = 334

PAGE 375

356 Table 94 Mediating Effect of Recovery Expectancy in the Association Between Domain 1-3 Variables and Recovery Strategy 4 (Recognizing Support) Main Effects Model Mediating Model Variable b t b t Other Diagnosis 0.39 (0.18) 2.14 a 0.39 (0.18) 2.14 a Nuclear family mental illness 0.07 (0.03) 2.15 a 0.07 (0.03) 2.13 a Extended family mental illness -0.11 (0.06) -1.95 a -0.12 (0.06) -1.96 a Alienation -0.22 (0.10) -2.15 a -0.22 (0.10) -2.16 a Partner or best friend support 0.22 (0.04) 5.09 d 0.22 (0.04) 5.09 d Family support 0.23 (0.03) 8.02 d 0.23 (0.03) 8.01 d Community involvement -0.08 (0.04) -2.08 a -0.07 (0.04) -2.06 a Optimism and control over the future 0.24 (0.11) 2.22 a 0.24 (0.11) 2.21 a Average satisfaction score 0.17 (0.04) 3.98 d 0.17 (0.04) 3.99 d Recovery expectancy -0.02 (0.08) -0.30 Model Fit Statistics Model F 10.98 d 10.73 d Adjusted R 2 0.60 0.60 Note: a .05; b .01; c .001; d .0001 n = 337 Table 95 Mediating Effect of Recovery Expectancy in th e Association Between Domain 1-3 Variables and Recovery Strategy 5 (Help Seeking) Main Effects Model Mediating Model Variable b t b t Age 0.01 (0.00) 2.10 a 0.01 (0.00) 2.11 a Interpersonal sensitivity 0.20 (0.09) 2.18 a 0.19 (0.09) 2.13 a Provider support 0.14 (0.06) 2.19 a 0.14 (0.06) 2.19 a Self-esteem/self-efficacy 0.36 (0.13) 2.73 b 0.36 (0.13) 2.72 b Power/powerlessness 0.25 (0.12) 2.05 a 0.25 (0.12) 2.04 a Righteous anger -0.19 (0.08) -2.20 a -0.19 (0.09) -2.20 a Average satisfaction score 0.12 (0.05) 2.52 b 0.12 (0.05) 2.49 b Recovery expectancy 0.02 (0.09) 0.25 Model Fit Statistics Model F 3.32 d 3.25 d Adjusted R 2 0.25 0.25 Note: a .05; b .01; c .001; d .0001 n = 343

PAGE 376

357 Table 96 Mediating Effect of Recovery Expectancy in th e Association Between Domain 1-3 Variables and Recovery Strategy 6 (Symptom Eradication) Main Effects Model Mediating Model Variable b t b t Income -0.17 (0.06) -2.90 b -0.18 (0.06) -2.95 b Interpersonal sensitivity 0.40 (0.14) 2.92 b 0.37 (0.14) 2.74 b Paranoia -0.33 (0.13) -2.52 b -0.33 (0.13 ) -2.56 b Major depression diagnosis 0.47 (0.19) 2.49 b 0.49 (0.19) 2.63 b Community involvement 0.19 (0.06) 2.84 b 0.18 (0.06) 2.74 b Power/powerlessness -0.65 (0.19) -3.46 c -0.66 (0.19) -3.53 c Optimism and control over the future 0.70 (0.19) 3.77 c 0.71 (0.18) 3.81 c Recovery expectancy 0.23 (0.14) 1.60 Model Fit Statistics Model F 2.62 d 2.63 d Adjusted R 2 0.19 0.20 Note: a .05; b .01; c .001; d .0001 n = 342 Focus on Moderation Referring again to Figure 1, page 6, Chapter 1, two variables will be considered for moderating, or interaction effects with select independent variables. Recovery expectancy, thus far, has not been signifi cantly associated with independent variables (see discussion of research question 2) nor ha s expectancy had a mediating effect (see previous section). Recovery expectancy will now be tested for moderating effects to satisfy research question 4b and its hypotheses. The second variable, severity of illness, will be assessed for moderating effects, addr essing research question 5 and the connected hypotheses. Moderating effects, also known as interaction effects, simultaneously test the independent effect of two variables (e.g. interpersonal sensitivity and recovery expectancy) with the addition of an intera ction term (e.g., interpersonal sensitivity x

PAGE 377

358 recovery expectancy). A moderating effect is present if the interaction term is significant. The following criteria were used to sy stematically choose variables to assess for moderating effects. The variable was significant in main eff ect analysis with at least 3 recovery strategies The variable was significantly related to the moderator (either recovery expectancy or severity of illness) in bivariate analysis The interaction terms created were cons istent with the current literature A total of six variables met the criteria. Thus, the following six in teraction terms were created: 1. Recovery expectancy x interpersonal sensitivity 2. Recovery expectancy x family support 3. Recovery expectancy x provider support 4. Recovery expectancy x community involvement 5. Recovery expectancy x self-esteem/self-efficacy 6. Recovery expectancy x optimism and control Each interaction was evaluated separately for each domain and then together with the main effects models for each recovery st rategy. Previously mentioned in chapter 3, though initially hypothesized, evaluation of these models could be considered ad hoc, and a Bonferonni correction was used (.05 / 6 interaction terms), with a p-value of .0083 required for an association to be considered significant. In ad dition, because of the addition of several new variables, multicollin earity was again assessed, this time for the

PAGE 378

359 largest potential model, main effects plus the si x interaction terms. No variables violated the .20 tolerance limit and all variab les were retained in analysis. Research Question 4b: Evaluation of Recovery Expectancy Moderating Effects Research question 4b and hypothese s 4b.1-4b.3 inquire whether recovery expectancy in interaction w ith select variables will produce a significant interaction effect. The research question a nd hypotheses are reproduced below. RQ4b: Does the expectation of recovery mode rate the relationship between individual, social and service factors and recovery strategies? Hyp4b.1: Participants expectation of recovery will moderate the relationship between individual factors and recovery strategies. Post Hoc: Post hoc analyses will target spec ific variables that are found to be statistically and clinically relevant to recovery Hyp4b.2: Participants expectation of recovery will moderate the relationship between social factors a nd recovery strategies. Post Hoc: Post hoc analyses will target spec ific variables that are found to be statistically and clinically relevant to recovery Hyp4b.3: Participants expectation of recovery will moderate the relationship between service factors a nd recovery strategies. Post Hoc: Post hoc analyses will target spec ific variables that are found to be statistically and clinically relevant to recovery Identical to mediating analysis, there were no significant moderating effects, either in domain specific or main effect models Thus, hypotheses 4b.1-4b.3 are not confirmed. Tables 97-102 summarize the intera ction statistics for each recovery strategy.

PAGE 379

360 Table 97 Moderating Effects of Recovery Expectancy in the Association of Domain 1-3 Variables on Recovery Strategy 1 (Effective Illness Management) Moderating Variables b SE t Recovery expectancy x interpersonal sensitivity -0.01 0.05 -0.27 Recovery expectancy x family support 0.02 0.04 0.63 Recovery expectancy x provider support 0.01 0.06 0.20 Recovery expectancy x community involvement 0.03 0.05 0.53 Recovery expectancy x self-esteem/self-efficacy -0.20 0.12 -1.58 Recovery expectancy x optimism and control 0.13 0.13 1.05 Note: n = 335 F = 12.99, p .0001; R 2 = .67 Table 98 Moderating Effects of Recovery Expectancy in the Association of Domain 1-3 Variables on Recovery Strategy 2 (Positive Future Orientation) Moderating Variables b SE t Recovery expectancy x interpersonal sensitivity -0.07 0.04 -1.82 Recovery expectancy x family support -0.04 0.03 -1.27 Recovery expectancy x provider support 0.03 0.05 0.67 Recovery expectancy x comm unity involvement -0.02 0.04 -0.59 Recovery expectancy x self-esteem/self-efficacy -0.05 0.10 -0.50 Recovery expectancy x optimism and control 0.15 0.10 1.46 Note: n = 335 F = 16.10, p .0001; R 2 = .72

PAGE 380

361 Table 99 Moderating Effects of Recovery Expectancy in the Association of Domain 1-3 Variables on Recovery Strategy 3 (Meaningfuln ess, Personal Control, and Hope) Moderating Variables b SE t Recovery expectancy x interpersonal sensitivity -0.04 0.04 -1.14 Recovery expectancy x family support -0.03 0.03 -0.92 Recovery expectancy x provider support -0.02 0.05 -0.44 Recovery expectancy x community involvement 0.01 0.04 0.33 Recovery expectancy x self-esteem/self-efficacy 0.02 0.10 0.25 Recovery expectancy x optimism and control 0.08 0.10 0.80 Note: n = 334 F = 6.63, p .0001; R 2 = .49 Table 100 Moderating Effects of Recovery Expectancy in the Association of Domain 1-3 Variables on Recovery Strategy 4 (Support Recognition) Moderating Variables b SE t Recovery expectancy x interpersonal sensitivity 0.03 0.07 0.38 Recovery expectancy x family support -0.03 0.06 -0.56 Recovery expectancy x provider support 0.00 0.09 0.02 Recovery expectancy x comm unity involvement -0.02 0.08 -0.35 Recovery expectancy x self-esteem/self-efficacy 0.22 0.19 1.13 Recovery expectancy x optimism and control -0.17 0.20 -0.88 Note: n = 337 F = 9.69, p .0001; R 2 = .59

PAGE 381

362 Table 101 Moderating Effects of Recovery Expectancy in the Association of Domain 1-3 Variables on Recovery Strategy 5 (Help Seeking) Moderating Variables b SE t Recovery expectancy x interpersonal sensitivity -0.14 0.08 -1.71 Recovery expectancy x family support 0.00 0.06 0.06 Recovery expectancy x provider support 0.14 0.10 1.37 Recovery expectancy x community involvement 0.03 0.09 0.34 Recovery expectancy x self-esteem/self-efficacy -0.29 0.21 -1.36 Recovery expectancy x optimism and control 0.15 0.22 0.67 Note: n = 337 F = 3.05, p .0001; R 2 = .25 Table 102 Moderating Effects of Recovery Expectancy in the Association of Domain 1-3 Variables on Recovery Strategy 6 (Symptom Eradication) Moderating Variables b SE t Recovery expectancy x interpersonal sensitivity -0.24 0.12 -1.94 Recovery expectancy x family support -0.01 0.09 -0.19 Recovery expectancy x provider support -0.20 0.15 -1.34 Recovery expectancy x comm unity involvement -0.05 0.13 -0.37 Recovery expectancy x self-esteem/self-efficacy 0.27 0.32 0.85 Recovery expectancy x optimism and control 0.35 0.33 1.06 Note: n = 342 F = 2.64, p .0001; R 2 = .21

PAGE 382

363 The tables reflect only the interaction terms since there was minimal effect on the main effects models. For all strate gies, there was a minute reduction in the full model statistic (F-statistic) with no change in significance level and a slight fl uctuation in variance explained of less then one-half of a percentage point. Research Question 5: Evaluation of Illness Severity Moderating Effects Illness severity was dichotomized by creating a variable that di stinguished between the severely mentally ill (SMI) sample and the outpatient (OP) sample. Interaction terms were created with the same six variables in th e previous moderating effect analysis (interpersona l sensitivity, family support, provider support, community involvement, self-esteem, and optimism). The Bonferroni correction was used (p .0083 to be considered significant) and multicollinearity was again assessed via the tolerance statistic. No variable breached the cutpoint and all were retained. Each interaction term was evaluated independently for each domain and then with the main effect models. Since there was little variance in the result s, the grouped model is presented for their association with each strategy. Research question 5 and accompanying hypothe ses were evaluated in this analysis and are reproduced next: RQ5: Does severity of mental illness moderate the relationship between individual, social and service factors and recovery strategies? Hyp5.1: Severity of mental illness will moderate the relationship between individual factors and recovery strategies for participants. Post Hoc: Post hoc analyses will target spec ific variables that are found to be statistically and clinically relevant to recovery

PAGE 383

364 Hyp5.2: Severity of mental illness will mode rate the relationship between social factors and recovery stra tegies for participants. Post Hoc: Post hoc analyses will target spec ific variables that are found to be statistically and clinically relevant to recovery Hyp5.3: Severity of mental illness will mode rate the relationship between service factors and recovery stra tegies for participants. Identical to the previous moderating analysis, there was no interaction effects detected at the single domain (partial models) or main effect models Only the interaction terms are presented in Tables 103-108, summarizing the in teraction effects for each recovery strategy. Tables summarizing th e moderating models separately for each domain and recovery strategy are included in appendix E. Once again, there were small, non-significant, changes in the full model statistic and adjusted R 2 This analysis targets the label of se vere mental illness. Many respondents spontaneously spoke about the label during data collection. Mental illness labels have been empirically tested and found to have le ss effect then first suspected (Link, Cullen, Struening, Shrout, Dohrenwend, 1989), though it is still significant as part of the lived context when examined in qualitative re search (Angermeyer & Matschinger, 2005; Walby, 2003b), and is considered a reason to alter behavior a nd perception. In this light, the lack of moderating effect s involving severity of ment al illness was unexpected.

PAGE 384

365 Table 103 Moderating Effects of Illness Severity in the A ssociation of Domain 1-3 Variables on Recovery Strategy 1 (Effective Illness Management) Moderating Variables b SE t Illness Severity x interpersonal sensitivity -0.06 0.05 -1.20 Illness Severity x family support 0.01 0.04 0.33 Illness Severity x provider support -0.08 0.06 -1.39 Illness Severity x community involvement 0.07 0.05 1.50 Illness Severity x self-esteem/self-efficacy -0.09 0.12 -0.74 Illness Severity x optimism and control 0.13 0.13 1.01 Note: n = 335 F = 13.18, p .0001; R 2 = 0.67 Table 104 Moderating Effects of Illness Severity in th e Association of Domain 1-3 Variables on Recovery Strategy 2 (Positive Future Orientation) Moderating Variables b SE t Illness Severity x interpersonal sensitivity 0.00 0.04 0.15 Illness Severity x family support -0.03 0.03 -0.90 Illness Severity x provider support -0.11 0.05 -2.13 Illness Severity x commun ity involvement -0.01 0.04 -0.25 Illness Severity x self-esteem/self-efficacy 0.08 0.10 -0.76 Illness Severity x optimism and control 0.30 0.11 2.66 Note: n = 335 F = 16.37, p .0001; R 2 = 72

PAGE 385

366 Table 105 Moderating Effects of Illness Severity in th e Association of Domain 1-3 Variables on Recovery Strategy 3 (Meaningfulness Personal Control, and Hope) Moderating Variables b SE t Illness Severity x interpersonal sensitivity 0.08 0.04 2.09 Illness Severity x family support 0.00 0.03 0.10 Illness Severity x provider support -0.09 0.05 -1.94 Illness Severity x commun ity involvement -0.00 0.04 -0.09 Illness Severity x self-esteem/self-efficacy -0.07 0.10 -0.73 Illness Severity x optimism and control 0.16 0.10 1.59 Note: n = 334 F = 6.73, p .0001; R 2 = 0.49 Table 106 Moderating Effects of Illness Severity in th e Association of Domain 1-3 Variables on Recovery Strategy 4 (Recognizing Support) Moderating Variables b SE t Illness Severity x interpersonal sensitivity 0.10 0.07 1.46 Illness Severity x family support -0.02 0.05 -0.41 Illness Severity x provider support 0.04 0.09 0.41 Illness Severity x commun ity involvement -0.03 0.07 -0.38 Illness Severity x self-esteem/self-efficacy 0.02 0.19 0.10 Illness Severity x optimism and control -0.00 0.20 -0.02 Note: n = 337 F = 9.92, p .0001; R 2 = 0.60

PAGE 386

367 Table 107 Moderating Effects of Illness Severity in th e Association of Domain 1-3 Variables on Recovery Strategy 5 (Help Seeking) Moderating Variables b SE t Illness Severity x interpersonal sensitivity -0.01 0.08 -0.14 Illness Severity x family support 0.05 0.06 0.90 Illness Severity x provider support -0.11 0.10 -1.16 Illness Severity x community involvement 0.04 0.08 0.45 Illness Severity x self-esteem/self-efficacy -0.50 0.22 -2.29 Illness Severity x optimism and control 0.57 0.23 2.50 Note: n = 337 F = 3.13, p .0001; R 2 = 0.25 Table 108 Moderating Effects of Illness Severity in th e Association of Domain 1-3 Variables on Recovery Strategy 6 (Symptom Eradication) Moderating Variables b SE t Illness Severity x interpersonal sensitivity 0.01 0.12 0.11 Illness Severity x family support -0.08 0.09 -0.89 Illness Severity x provider support 0.19 0.16 1.18 Illness Severity x commun ity involvement -0.12 0.13 -0.96 Illness Severity x self-esteem/self-efficacy -0.15 0.33 -0.46 Illness Severity x optimism and control 0.14 0.35 0.42 Note: n = 342 F = 2.38, p .0001; R 2 = 0.18 The absence of significant mediating a nd moderating relationships does not rule out that other factors or constr ucts untested in th is study could operate in that capacity or

PAGE 387

368 that other variables that were included might have significant interaction effects that were not tested. What can be stated from the evidence presented is that all the hypotheses attached to research questions 4a, 4b, and 5 are not supported. The direct effects captured in the discu ssion of research question 3 that assessed the associations between the three domains of independent variables and the dependent recovery strategies constitute the main results of this analysis. The strong model fit and percentage of variance captured (rangi ng from 19-71%) argues for continued investigation into direct eff ects. However, recovery is a complicated concept for an individual to grasp and utilize and is ev en more so when considering commonalities within and across samples. Thus, other potential mediators and m oderators should note be neglected for affect on recovery strategies in future research. Summary of Associations with Recovery Strategies Multiple significant associations were noted in bivariate and multivariate analyses investigating associations between recovery expectanc y, recovery strategies and independent variables partitione d into historical/clinical (d omain 1), social (domain 2), and service (domain 3) domains. Recovery expectancy was tested first in association with domains 1-3 independent variables using logistic regression. There were no significant associations detected. Recovery expectancy is, however, significantly negatively associated with each recovery strate gy in bivariate analysis suggesting that the belief in recovery may play an important role in proactive use of recovery strategies. However, significant associations were obser ved only in direct effect models testing associations between recovery strategies and domains 1-3 independent variables. Regarding recovery expectancy, there we re no significant mediating or moderating

PAGE 388

369 associations detected when investigating recovery expectancy as a mediating or moderating effect in the asso ciation between recovery st rategies and domains 1-3. Severity of mental illness was also found to be non-significant as a potential moderating variable. Analysis of direct effects suggests that for each recovery strategy investigated, domain 2 social support variables have the strongest association. Historical/clinical factors vary in the degree they associate w ith each strategy. Servi ce factors are generally unimportant in this analysis, but it should be emphasized that only a few service factors were investigated and that al l are summary measures of num ber of services, contact hours and satisfaction. An analysis of specific clinical and support services may provide a different result. Table 109 summarizes the significant associations between domains 1-3 independent variables and the six strategies of (1) effective il lness management, (2) positive future orientation, (3) meaningfulness, personal control, and hope, (4) support recognition, (5) help seeking, and (6) sympto m eradication. Information from Table 109 will be used to highlight a brief summary of the significant associations with each recovery strategy in turn. Strategy 1: Effec tive illness management. Effective illness management is surprisingly unrelated to clinical/historical factors. Interpersonal sensitivity and an ti-anxiety medication are the two clinical variables significant in the full model and reduced full model analyses. The literature on recovery would suggest a stronger relationshi p with a diagnosis of depression, symptoms of depression and experiences of child abuse or adult assault. Similarly important to the

PAGE 389

370 current study and future research is the gene ral lack of associati on between measurements of symptom levels and endorsement of effective illness management and the other recovery strategies of th is study. Though highly significan t in bivariate analysis, symptoms lose their significance in multivariate analysis. This may lend credence to the call to understand recovery as a comple x phenomenon that reduces the focus on symptoms in the recovery movement literature. Significant to this strategy, and the other fi ve strategies as well, is the importance of domain 2 social factors. Significantly associated with strategy 1 is community involvement, a social support variable, se lf-esteem/self-efficacy and optimism and control over the future, both empowerment variables. Collectively with stigma resistance, these variables explain 54% of the variance in effective illness management, with the reduced full model explaining 63%. Taken together, the combination of these variables suggests that illness management not be addr essed in isolation. Community involvement and stigma resistance are logi cally linked by their underlying dimension of relating to others. A belief in future success is also associated with illness management but the glue to these relationships strictly from an interpretive basis, is self-esteem. Selfesteem is, in part, the confiden ce to interact and be involved. It is also to believe in oneself in spite of barriers and obstacles. The lack of any significant service variables (domain 3) and the relative lack of significant association with clinical variables leads to the observation that illness management be cons idered more then adherence to treatment, but should also emphasize the reinforcement pot ential of social inte raction coupled with an assessment of stigma resistance and steps to increase self-esteem.

PAGE 390

371 Table 109 Summary of full model significant associations between recovery strategies and independent variables (domains 1-3) Total Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control, and Hope Recognizing Support Help Seeking Symptom Eradication Control Age 1 X Education 1 X Income 1 XX Domain 1 Interpersonal sensitivity 4 XX XX X* XX Anxiety 1 XX Depression 1 X Hostility 1 XX Phobia 1 XXX Paranoia 1 XX Schizophrenia 1 XX Major depression 1 XX Anxiety diagnosis 2 X* XXX Other diagnosis 1 X Anti-anxiety medication 1 XX Anti-psychotic medication 1 XX Substance use diagnosis 1 X Hospitalized in last year 1 X Nuclear family mental illness 1 X Extended family mental illness 1 X Child sexual abuse 1 X

PAGE 391

372 Table 109 (continued) Effective Illness Management Positive Future Orientation Meaningfulness, Personal Control, and Hope Recognizing Support Help Seeking Symptom Eradication Domain 2 Alienation 1 X Stereotype endorsement 1 XXX Social withdrawal 1 XX Stigma resistance 2 XXXX XX Support: Partner or best friend 1 XXXX Support: Family 3 XX XX XXXX Support: Provider 3 XXX XX X Support: Friends 1 XX Community Involvement 4 XXX XX X XX Self-esteem/efficacy 3 XXX XXXX XX Power/Powerlessness 2 X XXX Community activism and autonomy 1 XXXX Optimism and control over the future 4 XX X X XXX Righteous anger 1 XX* Domain 3 Average satisfaction 2 XXXX XX R 2 (full model) 67% 71% 48% 60% 25% 19% R 2 (reduced model) 18% 27% 63% 68% 49% 61% Note: XXXX = p .0001; XXX = p .001; XX = p .01; X = p .05 *These variables are no longer significantly associated in the reduced models

PAGE 392

373 Strategy 2: Positive future orientation. A positive future orientation is negatively associated with three of nine symptom scales, the most highly associat ed strategy with the symptom scales. General anxiety and phobic anxiety are moderately to highly associ ated with a future outlook and only mildly associated with depression. This is consistent with the literature that indicates that pervasive anxiety and depr ession limit planning and belief in the future. Only schizophrenia is associated with a limited futu re orientation, which may be a factor of the illness or a byproduct of the clinical system th at consistently presents the message of a life long debilitating illness. Psychotropic medication is not associated with a positive future orientation, nor is child abuse, adult assault or familial mental illness. However, this is the only strategy that is significantly associated with hospitalization, in this case negatively associated with hospi talization in the last year. Noted earlier, a substance abuse diagnosis is positively associated with this strategy, perhaps reflecting the emphasis on a positive future in the recovery emphasis of substan ce abuse treatment. Multivariate analysis of domain 2 variables demonstrates an association of moderate magnitude only with so cial withdrawal for stigma va riables. This suggests that being socially connected is relevant to a positive future. Being socially involved is reinforced by family and provider support, community involvement and self-esteem/selfefficacy, all in significant association with pos itive future orientat ion. Hypothetically, provider support plays an apparent positive role in supporting future growth and happiness, without forgetting that there is no evidence to support a causal relationship. It is intriguing that provider support was more strongly associated with strategy 2 then familial support and that support from friends and a best friend or partner were not

PAGE 393

374 significant. In isolation thes e factors explain 64% of the va riance. Together with domain 1 and domain 3 independent variables 68% of the variance in positive future orientation is explained in the reduced ma in effects model. Like strategy 1, there is no significant association with any of the three domain 3 covariates. Strategy 3: Meaningfulne ss, personal control, and hope Strategy 3 also has several signif icant associations with domain 1 clinical/historical variables. A diagnosis of an anxiety disorder is significant, yet there are no significant anxiety symp toms. Interpersonal sensitivity and hostility symptom scales are positively associated but these do not make sense in conjunction with the other domain 1 significant variables. Another ex ample is the significance of anti-psychotic medication, almost uniformly prescribed to individuals with schizophrenia or schizoaffective disorder, neither of whic h are significant, and the additional nonsignificance of psychoticism and paranoid sy mptom scales. Child sexual abuse is significantly associated to mean ingfulness, personal control, and hope, and logically so as there is considerable resear ch support of diminished hope for untreated survivors of abuse, representing the only significant rela tionship between any abuse/assault variable and any recovery strategy. It should be noted that in the descript ion of the development of the strategies via factor analysis in Chapter 3, that this strategy was the least conceptually coherent, which may explain why the pattern of result s from domain 1 are similarly confusing. Focusing on domain 2, this strategy has the highest number of significant covariates that, together w ith domain 1 covariates, explai n 49% of the variance in the reduced main effects model. In isolation, domain 2 social variables explain 36% of the

PAGE 394

375 variance. Two stigma variable s are found to be significantly a ssociated with strategy 3: stereotype endorsement and stigma resistance. It is sensible that a person that does not endorse stereotypes for others would not endorse them for his or herself either. Unlike strategy 1, however, social support variab les of family, provider and friends are significantly related while self-esteem/self-effica cy is not. A possible explanation is that meaning and hope are based in part on others supporting our efforts paired with resisting negative messages. Liking and believing in our abilities (self-esteem) is surprising in its absence, but might indicate that, for this samp le at least, that hope, meaning, and control comes more from without then within. This is partially s upported by a weak association with optimism and control over the future, what can reasonably be described as reflecting a more internal belief that the person has agency. This might also suggest that both internal and external factors are involved in meaning, c ontrol, and hope, with external having a greater effect. The most powerful social variable in association with strategy 3 is community activism and autonomy. Being active is logi cally connected with non-endorsement of stereotypes and stigma resistance, what can be viewed as having the courage to interact with the community. It also makes sense that activism and autonomy require a degree of support and interaction to maintain ones sens e of control and hope. However, because of the large number of clinical variables that were also significant, these reflections on social variables acting in isolation are attempts to unders tand meaningfulness, personal control, and hope in an artificial separation.

PAGE 395

376 Strategy 4: Recognizing support Recognition of support is only weakly associated with three domain 1 variables. A diagnosis of other, though m ildly significant, is generally uninterpretable. It is interesting that familial mental illness is significant to only this strategy and may represent the recognition of the inability to l ook to a mentally ill close or extended family member for support. Recognition of support is closely associat ed, as would be expected, with support variables. Domain 2 variables independently account for 52% of th e variance in support recognition. The reduced full model explains 61%, again favoring domain 2 variables for percent of variance explained. Support recognition is strongly associated with support from a partner or best friend and support from family. This is re asonable as these are generally the most intimate relationships and offer comfort and ad vice if an individual is not sure on whether someone or something is supportive. These ar e also primary sources of support in their own right. Decreased alienation is also mildly associ ated with strategy 4. This is the only recovery strategy that alienati on is significantly associated wi th both in social variable only models and full models. Alienation ha s been called the key to stigmatization (Goffman, 1963) and was predicted to be powerfully associated with all strategies. In reviewing the items encompassing the stigma scale, this is the one stigma scale that may leave some doubt in the label used to describe its items (see Table 9, Chapter 3, page 144). The scale is supposed to measure the level of disenfranchi sement experienced by the respondent, but may instead be measuring in stead of or in addition to shame, guilt and

PAGE 396

377 disappointment due to the realization of having a mental illness and the subsequent feelings of inferiority. This may not be alienation which is likely related to interpersonal relating, in other words shame, guilt and disappointment provoked by negative interactions. A more accurate descriptive labe l then alienation might be negative identity. Community involvement is mildly associated with support recognition. Interacting with the environment in which help is being assessed only makes sense. The mildly significant association with optim ism and control over the future, the only empowerment variable significant with th is strategy, could reflect security and confidence to appraise for help and, if needed, seek help, the subject of the next recovery strategy. Average satisfaction with services, th e only domain 3 variable significantly associated with any recovery strategy, is highly positively associated with recognizing support. This would suggest that provider support would be highly associated as well, yet there is no significant association with the domain 2 provider support variable. However, this may reflect a general sense of support from the c linical/support system instead of a primary provider relationship, which is feasible since over 90% of the respondents are seen by multiple providers at the partnering mental health center. Strategy 5: Help seeking Help seeking is not significantly associ ated with any domain 1 clinical/historical variables in the reduced main effect model a nd only weakly associated with interpersonal sensitivity in the full model analysis. It a ppears that help seeki ng is socially and not clinically based.

PAGE 397

378 Within the social domain help seeking is not significantly associated with any stigma scales. Provider support, a natural re lationship to seek help in, is the one significant social support variable. Empowerme nt carries the bulk of the responsibility for explaining the change in variance in help seeking. Self-esteem/self-efficacy and power/powerlessness are the two variables associ ated with help seeking. Righteous anger is also significantly related to help seeking in main effects, but drops from significance in the reduced main effects model. Individuals with low self-esteem are less likely to ask for help due to lack of co nfidence and fear of embarrass ment and a reduction in selfworth. Help seeking is a risk and individuals contending with mental illness are, for the most part, not usually risk takers, preferri ng to remain as unobtrusive as possible. Preliminary research for this study asked i ndividuals diagnosed with a severe mental illness who they went to for help and why. Not surprisingly, almost every respondent stated their provider, perhaps one other person, usually a parent, and mo st stated that they avoided asking for assistance out of concern of appearing weak or foolish or because they didnt believe they would receive the help they sought (Walby, 2003b). The power of the domain 2 covariates are marked when consid ering that the reduced main effects model explains 27% of the variance for help seek ing while the social domain tested alone explains 24%. Similar to strategy 4, but with less ma gnitude, average sati sfaction is strongly associated with help seeking. This is reas onable when considering that a mental health center is normally a place that individuals seek help from and would expect to receive said help from their providers.

PAGE 398

379 Strategy 6: Symptom eradication Symptom eradication is the single it em strategy included to gain greater understanding of a primary c ontroversy between clinical recovery and the recovery movement. This is not to imply that every clinician seeks to totally eliminate symptoms in each consumer they serve. The emphasis doe s appear to be different, with discussions of recovery focused on symp toms for many clinicians while focused away from symptoms unless they become problematic for recovery oriented services. Noted previously in this document, individuals w ith mental illness are willing to tolerate a certain level of symptoms in order to avoid overmedication and changes in their appearance and personality as a medicati on side effect (Walby, 2003c). Symptom eradication is most highly associ ated with the social domain (R 2 = .14) with the reduced main effects model explaining 18% of the variance. Howe ver, domain 1 variables of interpersonal sensitivity, ac tive symptoms of paranoia, and a diagnosis of major depression are negatively associated with e ndorsement of symptom eradication. Noted earlier in this chapter, para noia and recurrent depression ar e highly taxing symptoms that would complicate a belief in living a symptom free life. Symptom eradication is not significantly associated w ith any stigma sub-domains in main effect models. Community involve ment is the lone social support scale of significance. Increased community involveme nt could point toward less severe illness and the ability to cope and interact, leading the individual to believe that symptoms will eliminate permanently. This is certainly not out of the question fo r individuals dealing with mild anxiety or an adjustment disord er. This hypothesis is reinforced by the two empowerment variables significantly associated with strategy 6. Power/powerlessness

PAGE 399

380 and optimism and control over the future are both highly significant. Optimism and perceived personal power also de note a belief in the future, a belief that could include a symptom free life.

PAGE 400

381 Chapter 5 Synthesis of Research Findings Investigation into correlates of recovery, indeed, even de fining recovery, is still a relatively new field of research. Whereas clin ical recovery is concerned with empirically based symptom reduction delivered through clini cal services and expe rtise, the recovery movement emphasizes self directed and pers onalized approaches to recovery. Only recently have researchers begun to explore pote ntial strategies for recovery that reflect recovery choices, behaviors, and beliefs ba sed on the principles of the recovery movement. Despite some confusion and fragment ation in research efforts and defining of concepts, the recovery movement is likely here to stay as it is central to the New Freedom Commission on Mental Health (New Freedom Commission on Mental Health, 2003) and is a primary focus for the State of Florid a (J. Watts, personal communication, January 18, 2007). Thus, this chapter will begin with a di scussion of the findings regarding recovery expectancy as it pertained to research ques tions 1 and 2. A principal focus of this discussion will be recovery expectancy and severity of illness and how this contributes in theoretical terms to future research on recovery. Recovery strategies will be reviewed a nd evaluated in the context of research question 3. The explanatory contribution of each domain will follow discussion of research strategies and in turn will be followed by a discussi on of theoretical implications

PAGE 401

382 for empowerment, stigma, social comparison and social support theories. A brief discussion of additional factor s that may influence both belie f in recovery and strategy will be offered. Finally, subsequent sections will address study limitations and strengths, the relevance of this research to public health, and will conclude with suggestions for future research. Recovery Expectancy Research question 2 asked whether ther e were significant associations between the belief in recovery and specific clinical/h istorical (domain 1), social (domain 2), and service (domain 3) covariates. The question of recovery expectancy or belief that one will recover, as far as can be determined, is unique to this study. In the recovery literature, the expectation or belief in ones recovery has not been explicitly included in recovery research, though rec overy strategies (also called themes or pathways in the literature) have received a ttention (Corrigan & Ralph, 2005; Resnick, Fontana, Lehman, Rosenheck, 2005), and are the topi c of the next section. As noted in Chapter 4, bivariate analysis revealed many significant associations between domains 1-3 variables and recovery expectancy, with significance also noted in multivariate analysis after controlling for speci fic factors. Factors that are significantly related to recovery expectancy in multivaria te analysis are worthy of future research given the strength of association between rec overy expectancy and recovery strategies. However, since this is a first attempt at inve stigating recovery belie fs and the distinction between severe mental illness and less severe illness is relevant in research and clinical practice, a brief discussion of belief in recovery and severity of illness is warranted.

PAGE 402

383 Recovery expectancy and severity of mental illness If one was to assume that results from bivariate analysis told the whole tale, then a profile of someone that believes in recovery could be depicted as someone with limited active symptoms, not diagnosed with bipol ar disorder, restricted dependence on psychiatric medications, who is socially connected and supported by his or her community, family and friends, and that has in ternalized a sense of power, self-esteem, and concern for others. With this descripti on, it is easy to see why belief in recovery is more difficult for individuals labeled SMI sin ce these are often areas of greater challenge for any socially and economica lly marginalized population. Be lief in recovery may hinge on more subtle gradations of support, empowe rment, and stigma resistance as well as stable control of active symptoms and acce ss to recovery and/or empowerment based services. Belief in recovery may also relate to other factors not included in this study such as the experience of hope, the promise of a future that includes quality of life and access to necessary financial, educati onal, and employment resources. Belief in recovery may vary due to em otionally compelling tu rning points. As this study was cross-sectional, the stability of beliefs across time is unknown. However, belief in anything will change with new information and experience and there is no reason to assume that recovery beliefs are impe rvious to changes in perception of self and environment. Research with individuals with disabilities, the closest parallel the literature can offer at this time, found that meaning making coupled with social support and self-understanding, actively replacing losses with gains and accommodating to

PAGE 403

384 changes enhanced self-belief (King, et al., 2003). This is consistent with the recovery and empowerment literature with its emphasis on personal growth, planning for obstacles, being willing to change and increa sing supports. However, this comparison is based on two assumptions. First, that self-belie f is similar to belief in ones recovery and second, that belief in recovery would be processed similarly for SMI and OP sample participants. These assumptions, though logi cal, are without empi rical validation. To reiterate, this study attempted to broaden the concept of recovery to include individuals not labeled SMI, though those labeled SMI have been an almost exclusive focus of the recovery and related empowerment literature. In addition, it is important to remember that the recovery and empowerment literature addresses belief implicitly. Being admonished to believe in ones recovery is not the same as having the belief that one will recover and there is an apparent but not ye t understood impact of severity of illness on this belief, as demonstrated in this study vi a the significant results from research question 1 that found that individuals fr om the OP sample were much more likely to endorse their own recovery. Assuming that self-belief in disabled individuals is sim ilar to the belief that one will recover, the findings in this study are consistent with the literature, especially the emphasis on symptom control (reflecting self -understanding and self-care) and the powerful association with social support on re covery beliefs and strategies. The main point is that change in recove ry belief is likely to occur due to alterations in relationships, intrapsychic processes, and the lived environment, identical to areas of change considered important to individual s with disabilities.

PAGE 404

385 There is evidence from this study to support that recovery expect ancy is important for utilization of recovery strategies, though the mechanism of how this works is not clear. The lack of clarity is due, in part, to a methodologi cal issue in this study. The strategies derived and inves tigated were structured more on a is this important or relevant to you question format than a how do you implement or operationalize this format. The question of belief in recovery may very well reflect a more action oriented or behavioral aspect of recovery. Belie ving in recovery may access in respondents a cognitive process of recalling behaviors that facilitate their own r ecovery, resulting in belief or lack of belief in their own recove ry based on the evaluation of success of their actions. Operationalizing beliefs can incl ude planning and action. When considering how beliefs are operationalized, models in which recovery beliefs as well as strategies could be included would require attention to be liefs of the individual and how they act upon these beliefs to reinforce expectancy, choice of strategy and action taken in response to the strategy (Hughes, Hill, Budd, 1997; Kinderman, Setzu, Lobban, Salmon, 2006). Noted above, those labeled SMI in this study were significantly less likely to believe in their own recovery. However, it is important to not lose sight that there are individuals labeled SMI that be lieve in their recovery. Futu re research could ascertain the differences between this subgroup and t hose that do not sanction their own recovery within the SMI labeled population. In the future, efforts should be made to determine the effect of changes in recovery beliefs on utilizatio n of recovery strate gies, important in order to understand the bridge between clinical recovery and the recovery movement. However, a key component missing in this study is the cognitive impact of mental illness over time and its effect on recovery belief. Though this could certainly be considered a

PAGE 405

386 limitation to this study neurocogni tive deficit is discussed at this point because of the potential impact on recovery that is relevant more for the SMI than the OP sample. Neurocognitive deficit as impedime nt to recovery expectancy. Deficits in neurocognition were not measur ed in this study but could be exerting influence on recovery expect ancy, as well as choice of recovery strategy, perceived empowerment, susceptibility to stigma, and i ndeed nearly every f actor investigated. Deficits are important predictors of functi onal outcome in chronic mental conditions like the ones faced by all in the SMI sample and perhaps the more severe members of the OP sample. Memory deficits alter social behavior and perception of se lf and symptoms over time and could impact perception of recovery beliefs as well. Further, intact memory predicts better recovery of social and interpersonal skills (Smith, Hull, Huppert, Silverstein, 2002). The few respondents who needed extended time to complete the surveys were at highest risk of neurocognitive deficits. However, there answers were consistent across surveys and the small numbe r, four total responde nts, are unlikely to affect the results in this study. Cognitive deficits increase the risk for av ersive life experiences, which leads to dysfunctional behavior and beliefs. Defi cits are more profound and impact more domains of functioning for individuals diagnos ed with schizophrenia or schizoaffective disorder, the vast majority of which are in the SMI population. Negative beliefs of self and others, misattribution of interactions, jumping to conclusions, low expectation of success and pleasure, and withdrawal from inte raction to conserve wh at is perceived as limited intellectual capacities are common e xperiences for individuals with mental illness, with increasing impairment the longe r and more severe that symptoms are

PAGE 406

387 experienced (Beck & Rector, 2005). Together, memory and other deficits unmeasured in this study may negatively impact recovery e xpectancy. Not to screen for neuroncognitive deficits was a design choice that imposes a limit that should be considered. Cognitive deficit impact on data collecti on will be more fully explicated in the limitations section below. Recovery Strategies Recovery strategies were tested for direct associations with domains 1-3 independent variables in research question 3. Prior to that an alysis a factor analysis was completed on the combined items from the R ecovery Assessment Scale (Corrigan et al., 2005) and the Personal Vision of Recovery Questionnaire (Borowitz-Ensfield, 1998). The recovery strategies derived from factor analysis and inves tigated for this research are consistent with, but not identical, with other research on recovery. The strategies include (1) effective illness management, (2) positive future orientation, (3) meaningfulness, personal control, and hope, (4), recognizing support, (5) help seeking, and (6) symptom eradication. Corrigan, Salzer, Ralph, Sangster, Keck ( 2004), developers of the RAS, described five factors that included: (1) personal c onfidence and hope: (2) willingness to ask for help: (3) goal and success orie ntation: (4) reliance on ot hers: and (5) no domination by symptoms. The five factors of the PV RQ (Borowitz-Ensfield, 1998) included: (1) recovery through support: (2) re covery through personal challe nges; (3) recovery through professional assistance; (4) recovery through action and help-seeking; and (5) recovery through affirmation. The factors or recovery strategies for this st udy are markedly close

PAGE 407

388 to the RAS in items and description with a minor contribution from the PVRQ. Why the RAS was favored over the PVRQ in this study s factor analysis cannot be determined with confidence. Perhaps the RAS resonated to a greater degree w ith respondents or the PVRQs more direct referenc ing of recovery in each item was somehow processed by the respondents in a way that disallowed the items to mesh in analysis. A future analysis is planned that will test for differences on th e RAS and PVRQ scales as designed by their authors to compare SMI and OP responses and determine if there are significant differences between samples on the two recovery instruments. During the course of this research other instruments targeting recovery strategies or pathways were tested by other research groups. A brief review of recovery instruments was conducted to establish whether the RAS and PVRQ items remained consistent with current measurements of rec overy. The Mental Health Recovery Measure (MHRM) was designed utilizing a grounded th eory analysis of interview sessions targeting recovery (Bullock & Young, 2003). Reliability and Rasch modeling reduced the 41 original items to a 30-item instrument. Subscales and internal consistency of the MHRM are: (1) overcoming stuckness ( = .60); (2) self-empowerment ( = .82); (3) learning and self-redefinition ( = .79); (4) ba sic functioning ( = .62); (5) overall wellbeing ( = .86); (6) new potentials ( = .62); and (7) advocacy/enrichment ( = .66). A comparison of the scales reveals that the stuckness and basic functioning scales are similar to the effective illness management st rategy in this study (str ategy 1). Similarly, the learning and redefinition scale of the MHRM parallels strategy 2, positive future orientation. Strategy 3, meani ngfulness, personal control, and hope are consistent with overall well-being and new poten tial scales of the MHRM.

PAGE 408

389 The Consumer Recovery Outcomes System (CROS) was developed through focus groups and interviews with consumers and prov iders of mental health services, resulting in 38-item consumer and provider versions of the CROS (Bloom & Miller, 2005). The scales, internal consistency ( ) and test re-test (r) of the CROS consumer version are as follows: (1) hope for the future ( = .90, r = .70); (2) daily functioning ( = .83, r = .69); (3) coping with clinical symptoms ( = .86, r = .76); (4) quality of life ( = .84, r = .75); and (5) treatment satisfaction ( = .79, r = .74). Review of scale descriptions finds parallels with strategy 1 (effective illness ma nagement) and scales 2 and 3 of the CROS. Positive future orientation (strategy 2) is somewhat related to some items and the description of scale 4 of th e CROS. Strategy 3 (meaningful ness, personal control and hope) appears to be strongly relate d to scale 1 of the CROS. The Recovery Process Inventory (RPI) was designed with the assistance of 459 individuals using focus groups and prelimin ary surveys with severe mental illness resulting in a 22-item 6-factor inst rument that include: (1) anguish ( = .78); (2) connected to others ( = .71); (3) confidence and purpose ( = .77); (4) ot hers care and help ( = 56); (5), li ving situation ( = .71); and (6) hopeful and cares for self ( = 81) (Jerrell, Cousins, & Roberts, 2006). The firs t scale, anguish, appear s unique compared to this study and the instruments developed in ot her studies that were reviewed. Items in this scale are similar to the stigma scale us ed in this study, for example discriminated against, isolated and alone, and lost and hope less. This is a cons istent problem when investigating recovery, dete rmining which are components of recovery and which are covariates. Scale 2 of the RPI is similar to the help seeking scale in this study. Confidence and purpose (scale 3 of the RPI) has several nearly identical items to strategy

PAGE 409

390 2, positive future orientation. The fourth scal e of the RPI is similar to the recognizing support strategy. The final two scales of the RPI are comprised of only two items, violating a fundamental rule in development of scales via factor an alysis and are not well matched with any of the recove ry strategies in this study. The Hamilton County Recovery Initiative developed a 21 item survey that was reduced to 7 items loading only on two fact ors (Borkin, et al., 2000). Factor 1 was recovery is possible and needs faith and fact or 2 was recovery is difficult and differs among people. The description of the analysis revealed fragmented response patterns likely due to including individuals diagnosed with a mental disorder, professionals, family members, students and others (no fu rther information provided), all answering the same questions. Further, as the respondents were allowed to self-identify what response group they belonged to, there was evidence that some individuals we re not clear on what a particular category was and misidentified th emselves. However, the purpose of this survey was to provide a fast screening instru ment to determine the recovery readiness of individuals and their supports. The 7-items from the two factors have items similar to strategies 1 and 2 for this study with the ex ception of one item that states to recover requires faith, which does resemble one item from the RAS that was not incorporated in the strategies for this study. Scales referencing help seeking (scale 5) in this study are no t represented in the MHRM, CROS, RPI or Hamilton County studies, though there are some items that bear a resemblance. It appears reasonable to state that other published surveys remain consistent with the types of information regarding recovery that this research investigated. This is important for placing this effort into the larger context of recovery

PAGE 410

391 research as well as supporting generalizabil ity. An additional methodological point is that all the surveys described in this study were created with the assistance of individuals labeled severely mentally ill or severely and persistently mentally ill. Thus, none of the surveys were developed with i nput from individuals with less severity, represented in this study by the OP sample. As far as can be ascertained, input from individuals representative of the OP population has never been included in recovery measurement development. This is consistent with th e focus on the severely afflicted in recovery research and intervention. As the concept of recovery has been found to be useful to the less severely ill in this study, a future research project might consider investigating what aspects of recovery would be endorsed fo r measurement development by a less severe sample and to compare for significant differences with a SMI sample. The strategies examined in this study are also consistent with the literature developed by individuals in recovery and t hose investigating the recovery movement. This is likely due, in part, to the phras ing and emphasis on recovery. The RAS only references symptoms five of 41 questions and mentions services only once. The symptom questions are general and left to the respondent to iden tify internally what symptoms mean to them. The PVRQ me ntions symptoms and diagnosis once and medications twice. Services and hospitaliza tion are mentioned once each as well. Both instruments leaned heavily towards recove ry movement concepts and phrasing and steered relatively clear of clinical recovery concepts. A more comprehensive approach would be achieved if this st udy were replicated to include more measures of clinical recovery and to evaluate the impact of th e three domains of independent variables on clinical recovery measures as dependent va riables. Considering the high number of

PAGE 411

392 hospitalizations for many of the SMI group a nd the near equal measures of symptom levels for SMI and OP samples, clinical recove ry may have relevance not detected in this study. Explanatory Contribution of I ndependent Variable Domains Next offered is a brief discussion of each independent variable domain and, when available, comparisons to results from relevant studies. In general, the social domain has the highest degree of association with utilizati on of recovery strategi es and with recovery expectancy. The clinical do main has a less consistent relationship with recovery strategies or expectancy but cannot be discounted and might act ually play a larger role in recovery than some advocates of the recove ry movement may desire to sanction. The service domain was not especially pertinen t in this analysis though there was an intriguing pattern that will be noted. Domain 1: Clinical/historical factors. The clinical/historical domain has obvious relevance to recove ry with increased symptoms and recent hospitalizations decreasing endorsement of recovery expectancy and recovery strategies. Symptoms may inte rfere with planning or executing recovery actions, implicating cognitive, emotional and behavioral effects on recovery. Active symptoms demand attention and could be inte rpreted as a failure to recover. The recovery literature reduces the importance of symptoms by suggesting that recovery can occur in the face of some symptoms (Chadw ick, 1997; Marshall, Crowe, Oakes, Deane, Kavanagh, 2007). However, when confront ed with overwhelming symptoms there appears to be agreement with the clinical literature that suggests directi ng all necessary

PAGE 412

393 resources toward symptom reduction (Fava, Ruini, Beliase, 2007; McCay, et al., 2006; Wilson, et al, 2006). Findings from the litera ture are consistent with this study where elevated symptoms were negatively asso ciated with recovery belief and strategy endorsement in nearly every case of bivariate analysis and in some cases in multivariate analysis as well. Focusing next on diagnosis, the SMI and OP samples were, for the most part, categorized as expected. Schi zophrenia and schizoaffective di sorder were far more likely to be diagnosed in the SMI population, wh ile anxiety disorders, dysthymia and adjustment disorders were dia gnosed more in the OP sample The one surprising finding was the diagnosis of bipolar disorder, whic h was nearly identical in number for both samples. This may reflect an increase in diagnosis of less severe cases. However, a dialogue with medical and administrative prof essionals at the community mental health partner agency did not clarify the unexpected level of inclusion of bipolar diagnosis in the OP sample. However, since bipolar disorder was the most prevalent disorder for the total sample, it may be a contributor to some of the unexpected similarities between the SMI and OP samples, especially symptom levels. It should be reemphasi zed that psychiatric diagnosis is not especially accurate in th e community mental health setting where psychiatrists and nurse practit ioners are often given only 45-minutes to complete an initial psychiatric history and diagnostic eval uation without the assist ance of any formal diagnostic instruments. Other studies of r ecovery have included mu ch higher numbers of individuals diagnosed with schi zophrenia and schizo affective disorder than found in this sample. For instance, nearly 50% of the re spondents were diagnosed with schizophrenia or schizoaffective disorder in the study validating the Recovery Process Inventory

PAGE 413

394 (Jerrell, Cousins, & Roberts, 2006), while the percentages were from 15% to 20% higher for the study developing the Health Rec overy Measure (MHRM) (Bullock & Young, 2003), and the Consumer Recovery Outcomes System (CROS) (Bloom & Miller, 2005), compared to only 22% in this study. As hypothesized, the SMI sample was pres cribed considerably more medications both in types, number of distinct medicati ons, and dosages. These differences were verified in chart abstractions as a check ag ainst reporting bias for medication usage. To review, in bivariate associati on with recovery strategies, an ti-depressant, anti-anxiety and anti-manic medications were significan t only with strategy 1 (effective illness management), and strategy 2 (positive future orientation), with higher average scores on the strategy associated with less use of th e medication. Only anti-anxiety medication remained significant in multivariate analysis and only for effective illness management. Noted earlier, anti-psychotic medication was significantly associated with strategy 3 (meaningfulness, personal control, and hope) in multivariate analysis and not significant with any strategy in bivariate analysis. Individuals from the OP sample may have qualitatively different symptoms or have more effective non-prescription based coping strategies compared to the SMI sample. Me dications are prescribed for symptom control in the majority of cases and it was noted earlier that the symptom levels were nearly identical for the OP and SMI samples, leadi ng to the expectation of approximately equal levels of medication use. The prevalence of assault by type and gende r in this study is below what has been reported in the literature. For instance, the findings of Goodman, et al. (2001) in an investigation of victimization of individuals with mental i llness found that lifetime sexual

PAGE 414

395 assault for females was 68% and males 40%, whereas for females in this study it was 49% and males 12%. Listing results of th e Goodman, et al. (2001) study by type of assault and gender group, with this studies re sults in parentheses, we find child sexual abuse for females at 49% (54%) and males 29 % (15%), child physical assault for females at 54% (44%) and males 58% (27%), adult se xual assault for females at 57% (44%) and males 25% (10%), and, lastly, adult physical assault for females at 75% (47%) and males 79% (19%). Only for child sexual abuse is the prevalence in this study greater than the Goodman et al., study. Similar discrepanc ies are noted with other studies of victimization and the mentally ill (Muese r, et al., 1998; Shack, Averill, Kopecky, Krajewski, Gummattira, 2004). One plausible expl anation is that this sample is a mix of severe and less severe respondents whereas all of the studies cited we re exclusively made up of individuals with a severe mental illn ess. Another possibl e explanation is the method used to collect the data. Four ques tions at the end of data collection asked whether the participant had even been sexually abused or physically abused as a child and sexually or physically assaulted as an adult. If affirmative, a follow-up question asked whether the abuse/assault had occurred more than once. The goal was to limit stress on the participant during data collection in or der to buffer against symptom exacerbation, which, though successful, might have allowe d respondents to respond incorrectly. As described in Chapter 3, a follow-up ch art abstraction was completed to obtain additional information relevant to this analysis as well as to provide a double-check on participant responses. The chart abstr actions did not iden tify additional abuse experiences for any participant. However, it should be noted that the number of respondents admitting to abuse or assault was 34% higher than the abuse recorded in the

PAGE 415

396 charts reviewed. This suggests an informati on collection bias by clinical staff regarding abuse or at least a lack of recording this da ta in the clinical records. Regardless, the literature supports abuse experiences as salien t to the onset and course of mental illness and an impediment to recovery (Alexander, Meunxenmaier, Dumont, Auslander, 2005). The lack of association of abus e variables with endorsement of recovery strategies in this study is counter to the literature and requires further exploration. In summary, the clinical and historical factors survey ed in this study were more comprehensive than in other st udies investigating recovery. Efforts were made to target factors that had either theoretical or empirical support for a significant association with recovery. This is the first study to include familial mental illness, hospitalization history, or prescribed psychotropic medications in analysis. The majority of findings were consistent with the empirical literature. Some results were inconsistent and this may be due, in part, to differences in diagnosti c composition, trauma experiences, and other factors. Clinical and historical factors have a potentially powerful role in affecting belief in recovery and use of recovery strategies. Domain 2: Social factors. Domain 2 social factors have the la rgest number and magnitude of significant associations of the three predictor domains w ith the six dependent recovery strategies. This provides empirical support for the rec overy movements focus on social support, community cohesion, fairness and availability of employment, empowerment, and social justice (Brown, 2001; Fisher & Chamberlin, 2007; Masterson & Owen, 2006). This also provides indirect support for recovery oriented services provided in clinical settings

PAGE 416

397 (Marshall, Crowe, Oades, Deane, Kava nagh, 2007; Ratzlaff, McDiarmid, Marty, Rapp, 2006). Social support was powerfully associated with all six recove ry strategies, though from where the support was received was diffe rent for most strategies. Support from partners/best friends or from friends in gene ral was surprisingly associated with only recognizing support (strategy 4) and meaningfulness, personal control, and hope (strategy 3). This differs from what was voiced in preliminary qualitative re search for this study (Walby, 2003a). In that study, 90% of the individuals stated that support from friends was their greatest support or at least tied with familial support. However, the pilot study sample was recruited from a psychiatric da y treatment program where socialization and building of peer support was emphasi zed, possibly biasing results. Support from family and providers were the most highly associated support variables in this study with bot h recovery expectancy and four of six strategies. This could partially reflect a population bias si nce for many individuals with mental illness these are the key supports left them as friends, partners and even extended family distance themselves after onset of illness a nd possible changes in behavior interacting with stigmatizing beliefs and expectations (MacDonald, et al, 2004). The majority of consumer social network support is provided by families, limited numbers of friendships, informal groups and, in some cases, provider support. Drop-in cen ters, clubhouses and support groups offer more formal and consistent support when available, but the supply of such organizations is stil l far below the demand (Cowell, et al., 2003; Macias, Propst, Rodican, Boyd, 2001).

PAGE 417

398 What is unique in this study is the ev aluation of separate networks of support whereas other studies utilized general measures of overall support. In order to compare with other studies, the importan ce of social support in this study must be evaluated in the context of the recovery strategi es identified. This makes it di fficult to contrast directly with other studies that generate d slightly different recovery strategies or conceptualized social support differently, but some gene ral comparisons can be made. A study by Skarsater, et al. (2003) emphasized how importa nt social support was for recovery from major depression. This study was limited to women with major depression and thus reduces comparability, but the qualitative re search described strongly emphasized the need for social support in nearly all aspect s of recovery. A quant itative study targeting individuals with bipolar disorder found similar results (Johnson, Lundstrom, AbergWistedt, & Mathe, 2003). In this study, low social support was positively associated with relapse and partial versus full remission. Soci al support measured in this effort appears consistent with the findings of other studies and provides additional reinforcement for the role of support in th e recovery literature. Empowerment is a core feature of rec overy oriented services offered in community mental health settings. Supported living, supported employment and supported education as well as strengths ba sed case management share features in common with the recovery movement and are often described as empowering. The level of empowerment depends on how the services are delivered with some supported services considered more empowering through providing choices to the individual while others are viewed as coercive with choices dictated. To be sure, while community-based programs in clinical settings discuss empowerment of consumers, there is seldom a focus

PAGE 418

399 on different aspects of empowerment, for inst ance in strengthening personal power and choice, raising self-esteem and promoting community involvement. Likewise, while consumers are encouraged to increase their social networks, it is rare that they are provided financial support to do so. An important aspect of giving back to the community is community involvement and activism. Together, these were significan t for five of six recovery strategies and were highly significant in biva riate analysis with recovery expectancy. In a study by Corrigan (2002), the benefit of empowe ring community opport unities was found to increase recovery though anti-stigma effect s and integration opport unities, especially when coupled with empowering and collabo rative services that reinforce positive community messages. This study reflects the majority of the litera ture that emphasizes having the community accommodate to the a fflicted individual and less with the reciprocity of giving back to the community. A separate study that investigated micro, meso and macro level of empowerment a nd their affect on recovery found that community integration, choice a nd control (similar to the power/powerle ssness factor in this study) were significantly related to positive recovery (Nelson, Lord, & Ochocka, 2001). This study emphasized the benefits fo r the consumer of giving back to the community and how this facilitates true inte gration versus a feeling of belonging that maintains the separateness of the individual as a sick person. This is more in line with the concept of community involvement investig ated in this study and provides support for the results of this study. Self-esteem is central to empowerment a nd was the most consistent and powerful covariate with recovery strate gies in this research, a findi ng duplicated in the recovery

PAGE 419

400 literature (Corrigan, et al., 1999; Hutchi nson, Anthony, Massaro, & Rogers, 2007). In both studies cited, self-esteem was related to confidence, educati onal achievement, and perceived success in recovery based activities or training a nd were investigated in an empowerment framework. However, self-est eem is more often investigated with recovery as a separate construct, not em bedded in an empowerment model. Though not measured in this study, self-esteem in the cont ext of its influences on recovery has been empirically linked with quality of life (G uerje, Harvey, & Herman, 2004), a possible construct for future research discussed be low. Supportive services have also been recognized for their benefits to recovery, mediated by the effects of higher versus lower self-esteem (Ratzlaff, McDiamid, Marty, & Rapp, 2006). What this study and the literature in the field lack is evaluating for causal direction of self-esteem and recovery. Does self-esteem lead to recovery or is r ecovery enhancing for self-esteem? Or is the relationship bidirectional? Stigma is a powerful reality in the lives of many indivi duals with mental illness. It is a barrier recognized by the recovery movement and research, but is often not specifically addressed in clini cal settings. The preliminary qualitative research for this study included questions regard ing stigmatization and discri mination experiences. Of over thirty in-depth interviews, all reported discriminatory and stigmatizing experiences and yet all stated that they had never discussed these experi ences with their provider nor had the provider broached the subject (Walby, 2003a; Walby 2003b). The same respondents spoke eloquently abou t their need for recovery an d their fear in discussing recovery with anyone outside of the mental h ealth field or with other consumers for fear

PAGE 420

401 of ridicule and rejection. This lived experience of cons umers is congruent with the results derived from both bivariate and mu ltivariate analysis in this study. Despite expectations, stigma was not as pow erful a construct as anticipated in this study. Alienation and social withdrawal were the two scales that we re expected to be strongly negatively associated with recovery expectancy and strate gies. Though bivariate analysis was consistent with this predicti on, multivariate analysis found both scales only mildly to moderately associated with one recovery strategy each. Similarly, stigma resistance was only associated with two strategies in multivariate analysis. Noted in a previous chapter, a methodologic decision was made to concentr ate on internalized stigma versus recollection of discriminatory experiences. This may have affected the results as individuals would be expected to generate grea ter emotional resistance to facing internalized stigma then recounting di screte discriminatory experiences. Whereas internalized stigma could prompt feelings of shame, guilt, and lowered self-esteem, discriminatory experiences could be expr essed with anger and righteous indignation. Though lacking intensity and breadth of a ssociation, the results of stigma in analysis with recovery expect ancy and strategies was in the predicted direction. Stigma has been found to be negatively associated with different aspects of recovery in multiple studies. Stigma and employment was research ed in the context of recovery and found that a key for success in employment for seve rely mentally ill individuals was in part building successful small enterprises with fl exible scheduling in order to maximize the return for the employer and limit the opportuni ties for discrimination and stigmatizing interactions (Warner & Mandiberg, 2004). Th is in turn led to increased recovery expectations and experiences. Longitudi nal survey data fr om a sample of 610

PAGE 421

402 participants with severe mental illness dem onstrated that stigma, symptoms, self-concept and life satisfaction (as aspect s of empowerment) were all pa rt of the recovery process and have a reciprocal effect on each othe r (Markowitz, 2001). During the course of analysis for this study, there was a positive significant association found in bivariate correlational analysis between every symptom s cale and each stigma scale with mild to moderate magnitudes (see Appendix F, Tabl e 161). Stigma and empowerment were found to be negatively associated and signifi cant as well (Appendix F, Table 179). These findings are consistent with the Warner and Mandiberg (2004) and Markowitz (2001) studies. It can be stat ed that the results of the stigma an alysis in relation to recovery and in relation to other independent variables is consistent with the overall pattern in the literature. Domain 3: Service factors. The only consistent finding in bivariate and multivariate analysis was the significant association between satisfaction with se rvices and endorsement of recovery expectancy and strategies. The research on re covery and services has taken two tracks. First, research has evaluated the degree th at traditional servic es have served an empowerment or recovery purpose. Second, re covery oriented services were developed and tested to ascertain the degree that the services met the recove ry expectations of choice, collaboration, consumer led, and empowerment orientation. However, research into the effectiveness of traditional therapies has consistently emphasized the quality of the relationship and satisfaction with the service offered as the key to therapeutic change, regardless of the treatment modality or technique used (Hill, 1989).

PAGE 422

403 The concept of service satisfaction guided the operationalization of the service domain. This concept resonate s in the recovery and medical literature with remarkably similar results to this study. The recove ry movement emphasizes recovery principle based services for all aspects of care, not just psychiatric. Satisfaction with primary care treatment was found to be related to the qua lity of the consulta tion, the influence of stereotypes on the behavior of the physic ian, giving of hope to the consumer by the doctor, and the user friendliness of the pr imary care organization (Lester, Tritter, & England, 2003). The American Association of Community Psychiatrists Guidelines for Recovery Oriented Services in Community Mental Health Systems recognizes recovery based services as a major cultural shift th at replaces paternalistic services with collaborative and satisfying services (Sowers, 2005). Service satisfaction based on recovery principles is central to the guidel ines and the expectation is that community mental health centers nationally will adopt and act within these guidelines. These studies stress the importance of choice in medical and psychiatric serv ice utilization. However, an important difference in the Sowers review of the guidelines and this study is that service satisfaction was measured in general for all services, whether ostensibly recovery oriented (e.g., supported housi ng) or clinical in nature (e.g., a medication review appointment). The Sowers article addresses recovery based services only. However, service satisfaction in clinical services is key to success (H ill, 1989) and this apparently translates to recovery based services as well.

PAGE 423

404 Implications for Underlying Theories Us ed to Generate and Assess Study Results Multiple theoretical perspectives were evaluated when developing the background, methods, and analytic pl an of this effort. The findings of this research have been, in general, consistent with theories of social support, empowerment, and stigma. Attribution theory and labeling theory were also considered and will be discussed under stigma theory. Brief discussion targeting each theoretical area is provided next to more firmly embed the results into the theoretical pe rspectives used to generate the study. Social support theory. Social support theory emphasizes the health benefits of being connected to others and the receipt of different forms of support, including emotional, instrumental, appraisal and informational support. Soci al support theory also emphasizes the reciprocal nature of support and its relevance for establishing and maintaining positive and mutual relationships. The benefits to recovery from mental illness from social support have been documented (Corrigan, et al., 1999; Corrigan & Phelan, 2004; Frese, Stanley, Kress, & Vogel-Scibilia, 2001). Comparison of social support reporting between SMI and OP samples found the SMI sample with slightly higher scores for support for all scales except support from friends. Though not significant, it is interesting that the sample expected to be the most socially isolated reported that they are the least. The same comparison based on recovery expectancy is in the predicted direction, with individuals sanctioning recovery with higher scores on all six social suppor t scales and significant for partner or best friend, family, and friend s upport as well as community involvement.

PAGE 424

405 Recovery and social support have been linked conceptually from the beginning of the recovery movement and recovery re search (Anthony, 1993; Leavy, 1983). Social support was first investigated in the context of recovery from substance abuse and then in recovery for those with co-occurring subs tance abuse and mental health diagnoses (Laudet, Magura, Vogel, & Knight, 2000). Securing and maintaining a positive and flexible social network is a cornerstone of peer supported recovery. Social support means more than acquaintanceship or alleviation of boredom in recovery. It is a network of friends, family, and professionals that assist the consumer w ith emotional stability, choice making, practical needs, to ch allenge them when needed a nd, important to the mutual aspect of recovery, to expect the consumer to give back to the level they are able when reasonable. Reciprocation is considered vital to persona l growth and the lack of reciprocation that is inherent in the clini cal relationship is a stumbling block to the joining of the clinical and peer s upported recovery initiatives. Provider support was measured in pa rt to assess whether such support is significant to recovery strategies as a small st ep toward unifying clinical recovery and the recovery movement. In this study, provider support was not signifi cantly different in association with recovery expectancy between the SMI and OP samples. However, in multivariate analysis provider support is signi ficantly associated with three strategies, positive future orientation, meaningfulness, pe rsonal control and hope, and help seeking. It should be noted that provide r support in significant associat ion with recovery strategies may be in part due to all respondents having one or more providers. It is unknown if the same significance would be found for individuals that were no longer in a provider-

PAGE 425

406 consumer relationship and were asked to re call whether provide support was significant to their recovery efforts. The significance of provider support compared to other forms of support is relevant to a new theoretical process in assessing correlates of recovery. Support from a provider is a one-way expectati on. Indeed, the provider that receives substantial support in return will be viewed as having crosse d professional boundaries. Genuine affection does often develop between a provider and consumer, but once the consumer has received from the provider sufficient resour ces where the provider is no longer needed, then the relationship is expected to end. This is counter to the recovery movement that values and expects relationshi ps to continue until and unless they become unhealthy, and only then are they terminated. Social suppor t is intended to be h ealthy and positive, but may not always be welcome and could be percei ved as negative. This type of negative support is known as dissupport. The social support, social dissupport cont inuum was described in a brief article by Malone (1988). In this arti cle, the author describes the negative consequences of clingy, angry, selfish and demanding aspects of suppor t that force one individual onto another. This was found to be anxiety and stress produc ing to the recipients of this unwanted support. Though little more was written about dissupport, it conceptually fits with another concept that began to gain popularity around the same time, expressed emotion, which will be described later in this ch apter (Anderson, Hogarty, Bayer, Needleman, 1984). Dissupport can be organized by type of support as social support is. Emotional support remains relatively constant across time and age and when not desired is experienced as intrusive, confusing, and manipulative. Instrumental (practical) supports

PAGE 426

407 tend to decline with age and when offered when not wanted or desired can be perceived by the recipient or others as enabling. This is particularly relevant to individuals with mental illness who may adapt a victim role and be gin to lose benefit fr om social support. Dissupport in general, which can also be con ceptualized as relational strain, tends to decline with age as well (Due, Holste in, Rikke, Modvig, & Avlund, 1999). Though not investigated in this study, a secondary analysis by age, gender and provider of support would better evaluate whether so cial support from this study is consistent with current theoretical expectations. Two additional theoretical sub-constructs or pathways to investigate in social support include the effect of social distance and the degree that social support acts as a buffer between stress and symptomatic illness and the subsequent effect on recovery belief and active us e of recovery strategies. Level of desired social distance may me diate the level of social support received by an individual with mental illness. Soci al distance is the degree that one individual would agree to be in a relationship with an other that in turn may reduce stigma and increase social support through contact (Alexander & Link, 2003). Preferences and levels of contact reduce fears of violen ce and unexplainable behavior, increasing the comfort level of contact and providing in creased support and decreased perception of stereotypes (Angermeyer, Beck & Matschinger, 2003). Social distance may be how stigma operationalizes its impact on social suppo rt and thus affecting the course of illness and sequela faced by the afflicted. Though th e relationship between social distance and perception of dangerous and fear is partic ularly strong (Corriga n, Green, Lundin, Kubiak, Penn, 2001), the mediating relationship of soci al distance on stigma and social support

PAGE 427

408 and that relationship on recove ry has not been evaluated. This would be an additional level of theory to investigate when consider ing the effects of social support on health. Social support has been investigated as both moderator and mediator of the impact of stress on illness fo r at least thirty years (Cobb, 1976; Lin, Ensel, Simeone, & Kuo, 1979). Social support buffers stress in vulnerable groups in cluding those with mental illness (Strous, Ratner, Gibel, Poniz ovsky, Ristner, 2005; Thoits, 1982). Common stressors for mentally ill populations are cons istent with other marginalized groups and include financial problems, unsafe neighbor hoods, unstable housi ng, discrimination and impaired access to resources. The vulnerabil ity and damage caused by these stressors are magnified with the addition of a mental illn ess that may exacerbate further due to these same stressors. In effect, a cyclical process may occur. Recovery is meant to target the clinical and social aspects of mental illness, but often in the same environment of these more structural stress agents. What is not known is the effect of stressors like financial problems and unsafe neighborhoods on recovery belief and use of recovery strategies while targeting recovery base d services to the consumer and whether social support mediates this relationship. The direct effect of aspects of social support on recovery strategies was supported in this study and thus it would be a reasonable assumption to test features of social support theory for mediating and/or moderating effects. Empowerment theory When assessing the impact of empowerm ent variables in this study, the majority of correlations are in the expected direction for bivariate and multivariate analyses. The association of empowerment variables with re covery strategies was not as powerful or uniform as predicted, though there were several instances of strong association.

PAGE 428

409 Increased perceived empowerment is positively associated with increased endorsement of recovery strategies. Personal empowerment, cons idering the rigidity of the mental health system and funding/reimbursement streams, is inextricably linked with organizational and community empowerment. Empowerment is contextually dependent and alters over time due to changes in resource allo cation, community programming, provider philosophies, and personal successes (F oster-Fishman, Salem, Chibnall, Legler, &Yapachai, 1998; Zimmerman, Is rael & Schulz, 1992). Described in chapter 2, empowerment is a multi-level construct that includes the individual, organizations, community, and soci ety, though the majority of research has focused on individual empowerment. Defici ent organizational empowerment can be considered a possible explanation for th e depressed magnitude of empowerment associations with recovery strategies. The survey acces sing self-reported empowerment was administered in the community ment al health center building. Preliminary qualitative research for this project reveal ed a general satisfaction with the services offered by the center but also a pervasive feel ing of disempowerment, lack of choices, and concern that the consumer needed to fit th e setting and that the setting was inflexible to consumer needs (Walby, 2005). In other words, organizational empowerment, defined as a system of services that, from the top down, are responsive to and consider the empowerment of each staff member and consum er a high priority, was considered below expectation by the consumers. Interviews w ith providers and admini strators reinforced this perception and were frustrating for a st aff that can accurately be described as compassionate and wanting to provide em powering services, but are constrained by systemic and financial limitations. Thus, reflecting on personal empowerment while

PAGE 429

410 seated in a setting of orga nizational disempowerment coul d have biased the scores downward. Empowerment across disciplines shares the concepts of mobilizing, liberating, and transforming (Hur, 2006). A broader and more ecologically valid understanding of empowerment and recovery strategies should in clude these concepts. Recovery strategies at their core are organizing c oncepts of personal growth that strive for resource access, building supports through mobilization of personal assets, liberation of the constraints of mental illness and discrimination, and transfor ming or improving a sense of self that is less restricted by mental illness. These processes are nurtured and reinforced in peer run recovery organizations and need to be enhanced in clinical organizations that strive for an empowerment orientation. Along with being multi-level, empowerment is multidimensional as well as interactive since power and em powering are inevitably acted out in relationships and are based on resource acquisition. Similar to the discussion on social support, the disempowering attributes of unsafe neighbor hoods, unsafe housing and financial distress embody the multidimensionality of empowerment and include psychological, political and sociological components. A cognitiv e component of empowerment should be incorporated when considering the relations hip of the many dimensions and levels of empowerment and how these in turn are re lated to recovery. Though cognition is a individual level experience, this discussion does not suggest that empowerment in the context of recovery be relegated to individual level empowerment. On the contrary, recovery, like empowerment, is based on multi-levels and requires the cooperation of communities, organizations, families and individuals. Recovery oriented services,

PAGE 430

411 research, and even publishing ar e subject, in part, to politic al and power preferences and beliefs. Returning to the cognitive, the impact on the complex thought process of mental illness has not been considered when investig ating the relationship of empowerment and recovery. Expanding the theory of empowerment to account for the changes in cognition and how altered perceptions accommodate empowerment principles and the desire and ability to recover is a logical next step. Further, the topi c of active symptom distortion should only be part of the discussion, and a sm all part at that. Active symptoms for the vast majority of individuals with a mental illness are short term and transient compared to times of lucidity and focus. There is, however a long term, subtle and pervasive effect of mental illness that affects perceptions. The slight distortions in long term processing should be considered when investigating empow erment, recovery and the mentally ill. Whether these changes are structural, neurolog ical, psychological, or stress induced is irrelevant. What is hypothesi zed here is that the experien ce of empowerment in current theory may be different for individuals w ith mental illness compared to empowering nonmentally ill and non-sy mptomatic individuals. Stigma theory. There are many different models of stigma and most do not meet the requirements to be considered a theory. However, there are several theories that appear related to stigma that can be considered. For this discussion, attribution theory and modified labeling theory will be considered in relation to the findings of this study. In addition, self-stigmatization theory will be evaluated since internalized stigma was measured instead of discrimination experiences.

PAGE 431

412 In multivariate analysis, the association between stigma variables and recovery strategies is not as uniform as noted in bi variate analysis, but se veral associations are strong. Stigma resistance emerged as the mo st relevant aspect of stigmatization to recovery measured in this study. A possi ble interpretation of this relationship incorporates the theory of self-stigmatization described by Goffman (1961). Managing the illness requires admitting th e illness and coping with external stigmatization. Even more threatening because of its insidious nature is internalized stigma based in part on self-disappointment over perceived weakness as well as activation of attributions of blame, shame and disgust that were learned before the illness manifested and projected outward toward anyone deemed to be abnor mal (Corrigan, 2000; Corrigan, Markowitz, Watson, Rowan, & Kubiak, 2003). The person must come to terms with an illness that was formally derided in others. Self-stigma involves focusing on the self what was once transmitted to others in the form of negativ e and discriminatory thoughts, feelings, and actions. Resisting stigma includes resisti ng self-stigma and caring enough for the self to take care of the illness to maximize qualit y of life. Self-stigma would benefit from additional research, especially for better defining the role of stigma resistance, identifying relevant correlates, and how to operationalize the concept for practical usage. A useful first step might be to identify how the OP sample generates and demonstrates stigma resistance as they are significantly more likely to endorse resistance and then use the information obtained as a basis for improving stigma resistance for individuals labeled SMI. Structural violence including impris onment, economic hardship, lack of opportunities to gain and maintain resour ces, systematic exclusion, homelessness,

PAGE 432

413 neighborhood violence and denial of rights are shaped by stig ma when conceptualized as an ecologic construct operati ng at individual, organizationa l, community and societal levels (Kelly, 2005). Aspects of structural violence are then internal ized as self-stigma. It should be noted that this is similar to the processes described above when considering stressors affecting social support and impediments to empowerment. Together, this reinforces the ecological perspective noted in Chapter 2 as relevant to the understanding and support for recovery. Recovery occurs da ily in the face of structural violence and empowerment barriers. Returning to and adding the concept of cognition, how an individual internalizes these experiences may reinforce internalized stigma or, if cognitive processes allow, protect the individual as well. Th us, internalization of stigma is a product of early learning, interna lized societal messages regarding the unpredictability and undesireability of mentally ill individuals, reinforced by structural violence and maintained via paternalistic provision of services to the afflicted. A complete understanding of recovery requires the ecological framework discussed here. Internalization of stigma can also be cons idered via attributio n and labeling theory. Attribution theory postulate s that human beings take into account the possible reasons for a persons behavior or illness and that when th e illness is considered selfinduced, e.g., alcoholism, then the person is to blame and will be considered via emotions of anger and fear, resulting in coercive and discriminatory behavior (Corrigan, et al, 2003). Unfortunately, many hold the opinion that mental illness is self-induced through inappropriate lifestyle choices a nd behaviors. Thus, attributio n theory is an interpersonal not intrapersonal theor y. But must it be so?

PAGE 433

414 There was no direct measure of attribu tions in this study. However, what was measured was internalized stigma based on, in all likelihood, the negative internalizations of non-mentally ill individuals beliefs projected on the respondents. In addition, however, it is possible that the respondents themselves have internalized negative attributes of mental illness and of themselves as mentally ill individuals in the form of self-stigma. Thus, internalized negative attributes of the self may be projected back onto the self, increasing self-stigma and the sequela that follow. This theoretically links selfstigma and attribution theory in such a way that is testable in the future. The final theoretical construct that is also considered with stigma theory is modified labeling theory. Link (1989) f ound that psychiatric patients as well as community respondents felt that psychiatric pa tients would be rejected by society and that secrecy and withdrawal are necessary to cope with the societal response. However, subsequent research found that the label is re jected if perceived to include an unrealistic or negative stereotype that is experienced as foreign to the individual (Camp, Finlay, & Lyons, 2002). Thus, it can be argued that those that do not rej ect the label have internalized the attributes proj ected by others in the form of self-stigma. The theme of self-stigma noted throughout this chapter is not an accident. The rela tive lack of current research on self-stigma processe s and their relevance to reco very is considered a large gap in the current understanding of recovery based services.

PAGE 434

415 Additional Factors or Constructs That Might Influence Recovery Expectancy and Recovery Strategies During the course of the analysis and in terpretation of the da ta, three additional concepts with apparent relevance to recovery were noted. These were not investigated in this study; however the following discussion wi ll explain each concept and place it in the context of recovery. Expressed emotion (EE) is a measure of the amount of emotion displayed to an afflicted individual, usually by immediate family members or other cl ose relationships in extended family and friendships. Negative as pects of EE have received, by far, the most attention in the literature. Parents of i ndividuals diagnosed with schizophrenia have received the most attention in the EE literature. There are three agreed upon components of EE that include hostility, emotional over-involvement, and critical comments (Kreisman & Blementhal, 1995). High levels of EE are communicated as hostility, blame, non-tolerance and critical appraisal of the individual. Indi viduals with high-EE tend to view the illness as under the persons control and were more likely to criticize behavioral deficits (Weisma n, Nuechterlein, Goldstein, & Snyd er, 1998). This is similar to attribution theory and se lf-stigma noted above (Weisma n, Nuechterlein, Goldstein, & Snyder, 1998). For individuals with schizophrenia, life satisfacti on and quality of life were improved when their parents had lo w-EE and engaged in mutually supportive interactions (Greenberg, K nudsen, Aschbrenner, 2006). Recovery may be influenced by the degree to which an individual is exposed to EE. Expressed emotion has been linked to in creases in active symp toms and this study established that active symptoms reduce belief in recovery and endorsement of recovery

PAGE 435

416 strategies (Pourmand, Kavanagh, & Vaugha n, 2005). High EE families express less emotional support though there may be little difference in instrumental support, though when offered the support is often given gr udgingly, maximizing shame. Reinforcement of self-stigma is also theoretically feas ible with high EE families though there is no empirical support for this assertion. Indee d, the simultaneous investigation of social support, stigma, and empowerment with EE has received surprisingly little attention. Thus, expressed emotion is discussed here as a potentially useful avenue of future research to identify how social support aff ects recovery and how dissupport, enacted as EE, can inhibit recovery. Insight is the second additional construc t to be briefly considered. Previous research has demonstrated that active illness management is strongly associated with level of insight (Watson, et al., 2006), espe cially medication adhe rence. It can be hypothesized that sufficient insight into the presence of ment al illness and the accompanying sequela would lead to increase d belief in the expectation of recovery, leading to active use of rec overy strategies. However, depending on the quality and strength of internalized illne ss attributions, the o pposite effect is possible, with decreased belief in recovery and less strategy use. Insight in relation to mental illness is de fined as illness awareness, belief in the necessity of intervention (pro fessional, peer and/or persona l), and the attribution of symptoms to an illness process that is res ponsive to intervention. However, insight is only weakly responsive to psyc hosocial interventions (Startup, Jackson, & Startup, 2006), and there is no evidence that be lief in recovery can be taugh t with any more success.

PAGE 436

417 Nevertheless, it appears reasonable that recovery strategies might be the core concepts for recovery based programs that operationalizes and teaches the how to of recovery. Quality of life is the final construct that may influence recovery and that was not included in this study. Measures of quality of life and recovery pathways are positively associated (Corrigan, et al., 1999) Returning to a stable leve l of baseline functioning and then improving quality of life are basic tenets of the recovery move ment and indicates a conceptual link between quality of life and recovery (Young & Ensing, 1999). Quality of life has also been linked to self-esteem, em powerment, and social support, all sub-groups of the most highly associated domain, the soci al domain, in this st udy (Guraje, Harvey, & Herman, 2004; Rogers, Anthony, & Lyass, 2004; Young & Ensing, 1999). A belief in ones ability to recover may be, in part, a re flection of current quality of life as an indicator of how life is and thus how life may remain. Quality of life measurement has been measured frequently in the investigation of severe mental illness. However, it has onl y received limited atten tion in the recovery literature. Low self-esteem was found to be a risk factor for impaired quality of life for individuals that had clinica lly recovered from psychosis (Gureji, Harvey, & Herman, 2004). Clinical recovery was defined as complete amelioration of positive psychotic symptoms and the continued presence of negative symptoms. Factors of social support, stigma, insight and other aspects of empowerme nt were not included in this study. This limits interpretability from the perspective of the recovery movement and is representative of the more limited and target ed studies that address recovery and quality of life simultaneously.

PAGE 437

418 To summarize, this section was included to draw attenti on once again to the complexity of recovery and, despite all attempts to develop a comprehensive study, the potential covariates and construc ts not included in this study. This is a form of limitation of this study, with other limitations the subject of the next section. Study Limitations There are several key limitations to this study that affects both internal validity and generalizability. When appropriate, ot her limitations were noted in previous sections. First is the cro ss-sectional design of the study that limits understanding of the directionality of effects. Next, the measurement of recovery expectancy lacked rigor, which may have contributed to the lack of significant results in multivariate analysis. Specific threats to internal va lidity are limitations of the data collection instruments, potential misclassification of respondents to the severely ment ally ill (SMI) and outpatient (OP) samples, and the effect of symptom levels on atte ntional behavior and cognitive processing of respondents duri ng data collection. Limitations to generalizability (external vali dity) include the representativeness of the study sample and the limited geographic area where the study was located. The direction of effects cannot be directly observed in a cross-section design. Though in some cases a reasonable argument can be made that, for instance, symptoms were present before a type of medication was prescribed, in most cases the direction of association can be determined, but not the que stion of causality. Other study designs are required to test these results for di rection and longevity of effects.

PAGE 438

419 Limitations to measurement of recovery expectancy. Recovery expectancy was introduced as a new concept in understanding the concept of recovery. When conceptualized, it was expected to perhaps have a limited relationship with endorsement of recovery strategies. Similarly, the concept was not expected to be strongly related to the thr ee independent variables domains. However, the results outlined in Chapter 4 indicate that th e concept is strongly related in bivariate analysis with recovery strategies and indepe ndent variables. However, the dichotomous measure of recovery expectancy does not pr ovide sufficient variance of what appears to be a complex construct. This could expl ain, in association with the independent variables, the lack of significant results in multivariate analysis. Thus, recovery expectancy results should be carefully consider ed in light of the somewhat crude measure of the construct. There is sufficient evid ence to consider the construct for further investigation into its definition and conceptualization. Limitations to instrument choice and availability. The instrument measuring recovery e xpectancy was developed specifically for this study. It has the benefit of being simp le and straightforward, but did not measure explicitly what each participant believed recovery means to them. Thus, what a respondent was considering when indicating a non-endorsement of personal recovery may be due to unrealistic expect ations of oneself, a belief th at mental illness is never cured or controlled, inability to attain wealth or edu cation, or many other possible explanations. Understanding why individuals do not believe they will recover and using this information to develop categories of rec overy of social, clinical, service related or other groupings would allow for targeted program development and unity between

PAGE 439

420 professional and peer run recovery programs, increasing the chance of recovery program success at a population level. Recovery expectancy was measured as a dichotomous variable that lacks variance. Expectancy is likely a complex cons truct that the simple measure used in this study does not sufficiently represent. Thus the results from the analyses involving recovery expectancy should be considered with caution. The Personal Vision of Recovery Ques tionnaire (PVRQ) has not been used outside of the study that developed the instru ment. It was chosen because it was one of few measures available at the design stage of the study and that it was developed in consensus with individuals recove ring from mental illness, an a priori critical point when seeking instruments for this study. The lack of additional exposure and evidence of validity should be acknowledged as a limitation to measuring the primary criterion of the study. Another reason the instrument was chos en was that each question in the PVRQ directly references the rec overy of the individual by stat ing that the content of the question is relevant to my recovery. This was different from the question format used in the Recovery Assessment Scale (RAS) wh ere each question references the individual in an I statement but does not mention reco very specifically in any question. The goal was to offer both personalized and more ge neral question formats to better capture the breadth of recovery belief and experien ce. The RAS has had more exposure in subsequent research and was included in a recent list of recovery measurement instruments with proven validity where the PVRQ was not (Campbell-Orde, Chamberlin, Carpenter, & Stephen, 2005; Corrigan, et al. 2005). The RAS has also been used to validate a more recently developed recovery instrument (Jerrell, Cousins, & Roberts,

PAGE 440

421 2006) and is considered to be a valid and genera lizable instrument in the field of recovery research and measurement. Like the PVRQ it has the benefit of being developed in cooperation with individuals rec overing from mental illness. The Support and Community Connectedness Survey (SCCS) was developed specifically for this study and has not been independently validated. Though other social support surveys do exist, they were not viewed as specific enough for tapping into unique social support, trust, and community connecti on needs and experiences of the mentally ill population. The development of the instru ment did allow for a thorough reliability analysis using both internal consistency and te st re-test correlation. However, there was not sufficient opportunity for a thorough validity analysis, limiting the confidence in the results of the analysis. The Internalized Stigma of Mental Il lness (ISMI) scale was also not validated beyond the initial study that developed the scale. This, again, is an important limitation to consider when interpreting study findings. Other scales were available but tended to measure discrimination and stigmatizing experiences not specific to mental illness or, in some cases, specific to mental illness, for instance the Stigma Discrimination Survey (SDS) (Wahl, 1999), but concentrated on how many discrimination and stigmatization experiences each respondent recalled. Knowing how ofte n or to what degree an individual has experienced discrimination does not indicate how such experiences are internalized and how they aff ect the recovery beliefs and st rategies of the respondent, a focus considered of more importance in this study. Thus, the ISMI was selected as it was designed to measure the internalized response to stigma and discrimination.

PAGE 441

422 The final measurement concern pertains to domain 3 (the service domain). The service domain was analyzed with less rigor co mpared to domains 1 and 2. In measuring for service factors, there was considerable care taken in determining which services, how long the individual was involved with each serv ice and the number of contact hours per month for each service. The respondents info rmation was validated in chart reviews and found to be approximately 99% accurate. Howe ver, the choice was ma de to utilize only the summary measures of number of servi ces, average satisfaction, and average contact hours per month in order to control the numbe r of independent variables to maximize efficiency of statistical power. This choice may account for the lack of significant associations noted with the excep tion of service satisfaction. Limitations due to resp ondent characteristics. Measuring by self-report is difficult unde r ideal conditions due to recall bias and other confounding factors. Adding the potenti ally distracting and limiting experience of psychiatric symptoms compounds the concer n. Additionally, many of the respondents were prescribed one or more psychotropic me dications, often at high dosages. Several respondents talked freely about active sy mptoms they were experiencing and life stressors that they were managing. Needi ng to redirect respondents back to survey response was reported consistently by the stud ent assistants and e xperienced by the lead researcher as well. No formal data were kept to quantify this potential issue but the anecdotal information indicates that this might have been an issue beyond what is normally experienced in survey research. This concern was li kely offset to a degree by the empathic and non-pressured atmosphere f acilitated by the resear ch team but it is important to acknowledge that the study desi gn did not account for perceptual distortions

PAGE 442

423 experienced by any respondent. This is a co mmon dilemma for research with individuals diagnosed with a mental illness and all effort was made to provide empathic communication for rapport building and ad equate time required for a non-stressful experience (Dworkin, 1992). Limitations to generalizability. One issue of generalizability is the representativeness of the study sample to the population it is tasked to repr esent. This sample represen ts individuals with mental illness that are in some phase of recovery but that are also receiving clinical and/or supportive services in a community mental h ealth center setting. It is likely that individuals with mental illness that are work ing for recovery with no assistance from or belonging to a clinical setting might have si gnificant differences from those that are receiving services. Further, nearly all data were collected in the community mental health center setting and this could have aff ected the responses of the respondents. It is reasonable to assume that the responses that they might have given in their home or some other setting might have been di fferent from the responses provided in the clin ical setting. The study location was a semi-rural county in the southeastern United States that has limited minority representation, mainly midd le level incomes, and is proximate to a large metropolitan area but without readily available tran sportation between the County and the city. Further research will be requi red to see if the resu lts of this study are replicated in other geog raphic locations with a mo re diverse population and socioeconomics. Finally, and this is both a strength and a limitation, th e inclusion of individuals that were not labeled severely mentally is unique to this study and represents an

PAGE 443

424 expansion of the recovery concept. However, the similarity of the OP and SMI samples on key clinical domain 1 variables, for inst ance bipolar disorder in the diagnostic variable, the majority of the symptom scales, familial mental illness and others, indicates a parallel between the samples that was unexpect ed. Because of this it cannot be stated with certainty whether these samples are actual ly two samples or whether this degree of likeness is consistent in other clinical settings. This shoul d not be overstated as a major concern, considering the number of other variables in which there were significant differences, but caution is still suggested. In addition, the majority of results are reported based on belief in recovery and degree of e ndorsement of recovery strategies, and thus crossover the two samples and are unaffected by the similarity between OP and SMI. At worst, this does limit interpretation of results until the question of one or two samples is fully clarified. Study Strengths A primary strength of this study was the re search design that virtually eliminated missing data. Missing data was a design i ssue of paramount importance during the formative stage of the study. Minimal missi ng data increases confidence in internal validity and was a key factor in training of the research team (see Appendix D). Each research assistant role-played the data co llection process from greeting to incentive multiple times with emphasis on double checking each instrument as it was completed to quickly locate missing values and to ask the respondent to reconsider a response. In addition, if a value was found to be missing and the deletion was not intentional, the staff was trained to help the respondent review the questions from that point forward to ensure

PAGE 444

425 that an incorrect value wasnt populating th e rest of the questions. Because of the exploratory nature of this study, it was emphasized how missing data hinders the ability to explain and understand the phenomena under investigation. Data may be missing due to the study respondents, the study design, or the interaction of the study design and the respondents (McKnight, McKnight, Sidani, & Fi gueredo, 2007). Any of these can affect sample size, reducing statistical power and casting doubt on results, and potentially result in poor representation of the population studied. An additional strength of this study is th e capture of multiple domains of possible covariates that are contextually representa tive of the recovery literature and related concepts, i.e. reintegration literature. As far as can be determined, this is the first study to explore multiple clinical, social and service variables for effects on recovery strategies simultaneously. Other studies have included a limited number of clinical variables, usually represented by a short symptom surv ey (Corrigan, et al., 2005). Such a limited perspective was not deemed adequate for understanding the complexity of recovery. The results of this study indicates that multiple significant direct effects were found for clinical and service factors, with the domain 2 social fa ctors (stigma, social support, empowerment) being the domain with the st rongest associations. However, that all domains were significantly re presented serves to highlight the complexity of mental illness recovery while suggesting that soci al factors receive increased attention. What may be another first was the intr oduction of the belief in personal recovery, what was titled recovery expectancy throughout the study, into relevance as an additional concept influencing recovery. Although the fa ctors that influence recovery expectancy were not significantly detected, the associa tion between expectancy and each strategy is

PAGE 445

426 intriguing and warrants further theoretical and empirical in vestigation. Belief in an ultimate positive outcome motivates individuals to accomplish many life tasks other than recovery from mental illness, or other illnes ses for that matter. Education, employment, a good relationship and other goals are based, in part, on believing that the goal is worthwhile and the energy and resources wo rth investing. Preliminary qualitative research that helped direct the development of this study indicated that individuals with mental illness seek to find meaning and direction in their lives and work to limit the effect of the illness while at the same time growing outside of the illness (Walby, 2003b). These are all facets of what the recovery m ovement has described as necessary for an individual to be walking a recovery path. In fact, a consistent complaint of the respondents in the qualitative studies was the focus on symptom reduction and illness management that clinical serv ices and professional relationships tended toward. Thus, expectancy has been touched on in other qualitative research (C hadwick, 1997; Corrigan, Calabrese, Diwan, Keogh, Keck, & Mussey, 20 02), but has not been investigated systematically. A final strength of this study was the expa nsion of the recovery concept to include individuals with less severe mental illness. Noted in the limitations section above, a significant difference between the OP and SMI samples was not obvious within the clinical domain. However, there were signifi cant differences in the social and service domains. Although the majority of those who report they wi ll recover were in the OP sample, it is important to note that there were also individuals in the SMI sample that believed they would recover while some individuals in the OP sample believed that they would never recover. Expanding the recovery co ncept to include more variance in illness

PAGE 446

427 would increase resources available for the focus on recovery while also reducing the isolation of individuals with a severe ment al illness and allowing mental illness to be thought of as a continuum base d not only on clinical impair ment but also on level of support, degree of empowerment, and other factors shown to be relevant in this and other studies. Implications for Public Health Mental illness remains a highly prevalent category of illness in the United States, contributing to substantial morbidity and increased mortality (Miller, Paschall, & Swendsen, 2006; Wasylenki, 1994). Due to th e high prevalence of individuals with mental illness, as well as multilevel stigma ra nging from individual acts of stigmatization to societal stigma, and lack of parity in insurance coverage, many individuals often go without treatment. This increases the burden of disease in the population and, via cooccurring disorders, negatively affects medi cal conditions in need of treatment and prevention efforts relevant to public health as a whole. There ha s been considerable effort to bring worthwhile programs to indi viduals in need. Clinical and research professionals have worked in unison to de velop and test evidence-based programs that are effective for many types of mental illness (Drake, Merrens, & Lynde, 2005). However, evidence-based programs are often poorly disseminated and/or implemented with little fidelity. The ga ps in evidence-based program access as well as the need for community based supports are being targeted by the consumer recovery and empowerment movement. The recovery movement has begun to demonstrate empirical support for peer run programs, supportive services (e.g., supported employment and

PAGE 447

428 housing), and self help organizations. The artificial divide between professional and peer run services is being challenge d in some locations and with greater consistency (Nelson, Lord, Ochocka, 2001). The research detailed in this document adds to the recovery literature and consumer movement by contributi ng to the literature targ eting the health of a large subset of the population dealing with increased morbidity and risk of mortality due to mental illness. The social domain in this research effort is the most highly associated domain for virtually all the recovery strategies wh en taking into account both the number of independent variables in signifi cant association with the strategies and the strength of the relationships. This implication of non-cl inical factors endorsed by individuals in recovery both supports the consumer perspec tive as valid in conjunction with clinical services and is consistent with the emphasi s on social factors when considering public healths contribution to population health. Public health research into social factor s often targets social capital and/or social support. Disagreement continues on whether bonding social capital is associated with individuals and their social ne tworks or whether social capit al is a collective resource associated with communities (Poortinga, 2006). However, there is agreement that social support and social capital are related and that both lead to better health outcomes (Berkman & Clark, 2003; Kawachi & Berkman, 2001). Indeed, a comparison of prevention efforts that target social support and community cohesion or connectedness fair well against more indi vidualized health behavior approaches when targeting depression or heart disease (Leskela, Rytsal a, Komulainen, Melartin, Sokero, LestelaMielonen, 2006; Lomas, 1998). Poor social ties and support ar e a consistent risk factor

PAGE 448

429 for morbidity related to depression and hear t disease (Everson-Rose & Lewis, 2005). In this study, community connectedness was found to be significantly associated with effective illness management, positive future orientation, recognizing support and symptom eradication. This research also found that support from family members, providers, intimate partners and friendships are strongly associated with sanctioning of recovery strategies. It is a reasonable a ssumption that such suppor t is ameliorative for active disease processes as well as preventativ e for relapse or onset of disorder. Thus, this research is consistent with public healths focus on community intervention and provision of social support. Public healths focus on social justice parallels the empowerment focus of the recovery movement, with social justice c onsidered a macro-level organizing construct and empowerment focused more on individual or organizational development (Fondacaro & Weinberg, 2002). The recovery movement views empowerment as a central concept and is dedicated to operationalizing the concept and making it more accessible, and teachable, to individuals with mental illness. This reflects the previously mentioned relationship between population level social capital and individual or familial level social support. Whereas social support is easier to target programmatically compared to social capital, empowerment is easier to target then social justice. One aspect of empowerment that is strongly associated with recovery st rategies in this research is self-esteem/selfefficacy. Optimism and personal power are also significantly asso ciated. Together, empowerment factors are strongly correlated with all six of the recovery strategies in this study, resonating with public health s social justice priority.

PAGE 449

430 Public health consistently attempts to champion health and resource access for all individuals in need. This goa l is in part what stimulated the decision to investigate the relevance of recovery for individuals not labele d severely mentally ill. Recovery services increase social contact, provide emotiona l and, when possible, instrumental support, encourages self-direction and empowerment, and brings together diverse responses for responding to symptoms, suicidal behavior and other potential aspects of living with mental illness. Though individuals in this study from the OP sample have been hospitalized less in their lifetime, have greate r levels of employment, and are less likely to have been abused as children, they are as symptomatic as the SMI sample, are not significantly different in educat ion or income, and are as likely to have been hospitalized for psychiatric reasons in the last year. Thus there is impaired quali ty of life for both the SMI and the OP samples. However, individual s that have reduced quality of life but are not labeled SMI usually do not have access to a recovery network or program. In fact, the mental health center that supported this research has a recovery program in development that will focus on empowerment and assistance to consumers, but only SMI consumers will be eligible. The results of th is study supports that the concept of recovery is relevant to individuals not labeled SMI. More importantly, th e distinction between belief and non-belief in personal recovery crosses the SMI and OP sample distinction reinforcing the spectrum concep tualization of mental illne ss. What is not known is whether exposure to recovery related services would assist the less severe individual in overcoming the debilitating aspects of mental illness, a potentially productive focus of continued research.

PAGE 450

431 A counter-argument is that the SMI label is itself a negative label that may add to the need for specialized recove ry services. Further, indi viduals that do not meet the criteria for the SMI label may resent the implication of receiving the same services as those with the label. Anecdot ally, individuals from the OP sample often verbalized that they enjoyed helping the mental health cente r and others dealing w ith symptoms whereas the SMI sample often expressed hope that part icipation in the research would be helpful to themselves and help to provide more re sources for their needs. Discussion during team meetings of the research team revealed that this was consistent in the experience of each team member. So, while individuals in the OP sample share some characteristics with the SMI sample, there is no indication th at they would agree that recovery based services would be relevant to them. A final implication for public health is that this study adds to the knowledge needed for developing a more comprehensive recovery program. Recovery program is defined here as a program that focuses on clin ical, social, and service factors as well as the independent peer service process more t ypical of the recovery movement. What is needed is a comprehensive, logically and sc ientifically developed program that embraces the best of both sides of what is the same co in, a melding of clinical and consumer based recovery programming. Noted above, there are researchers that are challenging the division between clinical/professional and pe er run services, considering both to be correct and complimentary if given the opportunity. There are cl early clinical issues that must be taken into consideration, with th e consumer having the freedom to choose a course of treatment and even to reject prof essional assistance if they choose a complete peer assisted or even a lone pa th. The option to reassess and e ngage in services as part of

PAGE 451

432 a team with professionally trained service providers must remain open, however. This study will hopefully make a professional aware of the important social support, stigmatization, and empowerment facets of recovery as well and promote a consumer as expert model that increases team work betw een the social and professional network of each consumer. Suggestions for Future Research Suggestions for research include both a broader research agenda targeting recovery and associated constr ucts and next steps for the current study. In addition, throughout this chapter several brief comment s were made suggesting a new course of research or an additional analysis planned with the data collected. Future research should utilize multiple res earch designs that target the recovery construct as well as covariates that ha ve evidence supporting a relationship with recovery. Though not used to the extent it used to be, th e use of case study methodology could be useful for understanding the subtle ties of recovery as well as for tracking changes in recovery expectancy and use of recovery strategies over the course of different disorders, all while collecting and unde rstanding the contextual data that appears vital for a complete understa nding of recovery. For in stance, though a particular diagnosis was seldom significant in the current study, no attempts were made to validate the diagnosis, review diagnostic history, or assess the symptom scales in relation to diagnostic criteria. Thus, diagnosis might play a more powerful role than detected in this study. Nested case studies tend to smaller sample sizes than survey re search, but a larger sample could be used to investigate diagnosis and other factors with greater depth. Case

PAGE 452

433 studies could be aggregated and supported with qualitative research targeting recovery. Though there is some qualitative research publishede, there is certainly room for addressing recovery qualitativel y and learning from the indivi duals engaged in their own recovery. In addition, policy an alysis from the federal leve l on down to local would be useful to more fully understand how policy is being developed in support of recovery as well as how it is implemented. Further, evalua tion of current recovery programs with a mixed focus on consumer internalization and change within the recovery model as well as how service providers are embracing and implementing the model would add additional information for a thor ough understanding of recovery. Information derived from qualitative and case study methodology, policy analysis and evaluations of current program efforts could be used to identify correlates of recovery as well as aid in the design of evidence-based recovery programs. Further, larger studies with greater pow er in order to identify poten tial mediators and moderators could elaborate on the strategic information obtained in other studies and analyses. Finally, more sophisticated research designs targeting individual, or ganizational (clinical, consumer run, social welfare or support), co mmunity and policy va riables of interest utilizing multilevel modeling such as HLM or si milar analyses that utilizes nested data would capture both the individua l and contextual information. An immediate improvement to the current study would be to repeat the study in a different sample population using a longitu dinal framework. Currently, there are no published studies that track recovery strategi es or pathways overtime. As noted, this might also be the first study to measure re covery expectancy. Assessing expectancy longitudinally would be of interest to detect the clinical, social and service factors that

PAGE 453

434 associate significantly with expectancy and to understand if thes e factors change over time. This is especially important consideri ng none of the independe nt variables in this study were significantly asso ciated with expectancy. Prior to the longitudinal study discus sed above, an in-depth, qualitative investigation of the operati onalization of recovery belie fs and how strategies are behaviorally enacted would offer opportunitie s to improve the choice of independent variables to be addressed as well as possibly suggest other, more contextually specific, recovery strategies. In addition, effort s hould be made to categorize the reasons that individuals do not believe they will recover. A stratif ied sample utilizing a semistructured interview schedule of the current study, stratified by expectancy (yes/no) and degree of strategy endorsemen t (high/medium/low), would be one way to structure the sampling for the qualitative study. The use of mixed methods either simultaneously or sequentially would, as noted in the discussion above, include important contextual factors and capitalize on the strengths of both qualita tive and quantitative research paradigms.

PAGE 454

435 References Abrahamson, D. J., Steele, A. P., & Abrahamson, L. S. (2003). Practice, policy, and parity: The politics of persis tence. Professional Psychology: Research and Practice 34 (5), 535-539. Ackerson, B. J., & Harrison, W. D. (2000). Practitioners' perceptions of empowerment. Families in Society 81(3), 238-244. Aghevli, M. A., Blanchard, J. J., & Horan, W. P. (2003). The expression and experience of emotion in schizophrenia: a study of social interactions. Psychiatry Research 119(3), 261-270. Agresti, A., & Finlay, B. (1997). Statistical Methods for the Social Sciences Upper Saddle River, New Jersey: Prentice Hall, Inc. Ahrens, A. H., & Alloy, L. B. (1997). Social comparison processes in depression. In B. P. Buunk & F. X. Gibbons (Eds.), Health, coping, and well-being : Perspectives from social comparison theory ( pp. 389-410). London: Lawrence Erlbaum Associates, Publishers. Alaszewski, A., & Manthorpe, J. (1995). Dur kheim, social integration and suicide rates. Nursing Times 91(25), 34-35. Alexander, L. A., & Link, B. G. (2003). The impact of contact on stigmatizing attitudes toward people with mental illness. Journal of Mental Health 12(3), 271-289. Alexander, M. J., Muenzenmaier, K., Dumont J., & Auslander, M. (2005). Daring to pick up the pieces in the puzzle: A c onsumer-survivor model of healing from childhood sexual abuse. In R. O. Ralph & P. Corrigan (Eds.), Recovery in mental illness: Broadening our Understanding of Wellness (pp. 207-232). Washington, DC: American Psychological Association. American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (4th ed.) Washington, DC: Amer ican Psychiatric Association. Anderson, C. M., Hogarty, G. E., Bayer, T., & Needleman, R. (1984). Expressed emotion and social networks of parent s of schizophrenic patients. British Journal of Psychiatry, 144, 247-255.

PAGE 455

436 Andresen, R., Oades, L., & Caputi, P. (2003). The experience of recovery from schizophrenia: towards an empirically validated stage model. Australian and New Zealand Journal of Psychiatry 37 (5), 586-594. Angermeyer, M. C., Beck, M., & Matschinger, H. (2003). Determinants of the public's preference for social distance fr om people with schizophrenia. Canadian Journal of Psychiatry 48(10), 663-668. Angermeyer, M. C., & Matschinger, H. (2003). The stigma of mental illness: Effects of labelling on public attitudes toward s people with mental disorder. Acta Psychiatrica Scandinavica 108, 304-309. Angermeyer, M. C., Matschinger, H., & Corriga n, P. W. (2004). Familiarity with mental illness and social distance from people w ith schizophrenia and major depression: Testing a model using data from a representative population survey. Schizophrenia Research 69(2-3), 175-182. Anthony, W. A. (1993). Recovery from mental illness: The gu iding vision of the mental health service system in the 1990s. Psychosocial Rehabilitation Journal 16(4), 11-23. Auerbach, E. S. (2001). The individual placement and support model vs. the menu approach to supported employment: Wh ere does occupational therapy fit in? Occupational Therapy in Mental Health 17(2), 1-19. Auquier, P., Simeoni, M. C., Sapin, C., Reine, G., Aghababian, V., Cramer, J., et al. (2003). Development and validation of a pa tient-based health-re lated quality of life questionnaire in schizophrenia: the S-QoL. Schizophrenia Research 63(1-2), 137-149. Azrin, S. T., & Goldman, H. H. (2005). Eviden ce-based practice emerges. In R. Drake, M. R. Merrens & D. Lynde (Eds.), Evidence-based mental health practice: A textbook (pp. 67-94). New York, NY: W.W. Norton & Company. Baron, R. M., & Kenny, D. A. (1986). The mode rator-mediator variab le distinction in social psychological research: Conc eptual, strategic, and statistical considerations. Journal of Personality and Social Psychology 51(6), 1173-1182. Barton, R. (1999). Psychosocial rehabilitation services in community support systems: A review of outcomes and policy recommendations. Psychiatric Services 50 525534. Bassman, R. (2001). Whose reality is it anyw ay? Consumers/survi vors/ex-patients can speak for themselves. Journal of Humanistic Psychology, 41(4), 11-35.

PAGE 456

437 Baxter, E. A., & Diehl, A. (1998). Emotiona l changes: Consumer and family members recovering from the trauma of mental illness. Psychiatric Rehabilitation Journal 21, 349-355. Beck, A. T., & Rector, N. A. (2005). Cognitive approaches to schizophrenia: Theory and Therapy. Annual Review of Clinical Psychology 1(1), 577-606. Becker, D. R., & Drake, R. E. (2003a). Conceptual, historical, and ideological underpinnings of supported employment. In A working life for people with severe mental illness Becker, D. R., & Drake, R. E. (2003b). A working life for people with severe mental illness New York, NY: Oxford University Press. Bell, M. D. (2001). Object-relati ons and reality-testing deficits in schizophrenia. In P. W. Corrigan & D. L. Penn (Eds.), Social cognition and schizophrenia (pp. 285-314). Washington, D.C.: American Psychological Association. Bellamy, C. D., & Mowbray, C. T. (1998). Supported employment as an empowerment intervention for people with mental illness. Journal of Community Psychology 26(5), 401-413. Bender, A. W. (1995). The relationship between health lo cus of control, perceived selfefficacy, hardiness, and recovery in schizophrenia Unpublished Quantitative, Correlational, Catholic University of America, Washington, DC. Bentley, K. J. (2000). Empowering our own: Peer leadership training for a drop-in center. Psychiatric Rehabilitation Journal 24(2), 174-178. Berghofer, G., Schmidl, F., Rudas, S., Steine r, E., & Schmitz, M. (2002). Predictors of treatment discontinuity in outpatient mental health care. Social Psychiatry and Psychiatric Epidemiology 37(6), 276-282. Berkman, L. F., & Clark, C. (2003). Neighbor hoods and networks: The construction of safe places and bridges. In I. Kawachi & L. F. Berkman (Eds.), Neighborhoods and Health (pp. 288-302). New York, NY: Oxford University Press. Berkman, L. F., & Glass, T. (2000). Social in tegration, social netw orks, social support, and health. In L. F. Berkman & I. Kawachi (Eds.), Social Epidemiology (pp. 137173). New York: Oxford University Press. Bjorkman, T., Hansson, L., & Sandlund, M. (2002). Outcome of case management based on the strengths model compared to standa rd care. A randomised controlled trial. Social Psychiatry and Psychiatric Epidemiology 37(4), 147-152.

PAGE 457

438 Blankertz, L. (2001). Cognitive components of self esteem for individuals with severe mental illness. American Journal of Orthopsychiatry, 71 (4), 457-465. Blascovich, J., Mendes, W. B ., Hunter, S. B., & Lickel, B. (2000). Stigma, threat, and social interactions. In T. F. Heatherton, R. E. Kleck, M. R. Hebl & J. G. Hull (Eds.), The social psychology of stigma (pp. 307-333). New York: The Guilford Press. Bledsoe, C. (2001). Unique eyes and diffe rent windows of opportunity: The consumer provider perspective. In C. E. Brown (Ed.), Recovery and wellness: Models of hope and empowerment for people with mental illness (pp. 23-42). New York: Hawthorn Press, Inc. Bloom, B. L., & Muller, A. (2005). The Consumer Recovery Outcomes System (CROS 3.0): Assessing Clinical Status and Progress in Person with Severe and Persistent Mental Illness Unpublished manuscript. Boardman, J., Grove, B., Perkins, R., & She pherd, G. (2003). Work and employment for people with psychiatric disabilities. British Journal of Psychiatry 182 (6), 467468. Bola, J. R., & Mosher, L. R. (2003). Treatment of acute psychosis without neuroleptics: Two-year outcomes from the Soteria Project. Journal of Nervous and Mental Disease, 191 (4), 219-229. Borg, M., & Kristiansen, K. (2004). R ecovery-oriented professionals: Helping relationships in mental health services. Journal of Mental Health 13(5), 493-505. Borkin, J. R., Steffen, J. J., Ensfield, L. B ., Krzton, K., Wishnick, H., Wilder, K., et al. (2000). Recovery Attitudes Questionnaire: Development and Evaluation. Psychiatric Rehabilitation Journal 24(2), 95-102. Borowtiz-Ensfield, L. (1998). The Personal Vision of Recovery Questionnaire (PVRQ): The development of a consumer-derived scale Unpublished Instrument, Quantitative, University of Cincinnati, Cincinnati, Ohio. Brent, E. E., Mirielli, E. J., & Thompson, A. (1993). Ex-Sample: An expert system for determining sample size. Brodsky, B. S., Qquendo, M. A., Ellis, S. P., Haas, G. L., Malone, K. M., & Mann, J. J. (2001). The relationship of childhood abuse to impulsivity and suicidal behavior in adults with major depression. American Journal of Psychiatry 158(11), 18711877. Brown, C. E. (Ed.). (2001). Recovery and wellness: Models of hope and empowerment for people with mental illness New York: Hawthorn Press, Inc.

PAGE 458

439 Bruce, S. E., Yonkers, K. A., Otto, M. W., Eisen, J. L., Weisberg, R. B., Pagano, M. E., et al. (2005). Influence of Psychiatric Comorbidity on Recovery and Recurrence in Generalized Anxiety Diso rder, Social Phobia, and Panic Disorder: A 12-Year Prospective Study. American Journal of Psychiatry 162(6), 1179-1187. Brunt, D., & Hansson, L. (2002). The social networks of persons with severe mental illness in in-patient settings and supported community settings. Journal of Mental Health 11(6), 611-621. Buchanan, J. (1995). Social support and schizo phrenia: A review of the literature. Archives of Psychiatric Nursing, 9(2), 68-78. Bullock, W. A., & Young, S. L. (2003, August, 2003). The Mental Health Recovery Measure (MHRM). Paper presented at the Annual Conference of the American Psychological Association, Toronto, Canada. Byrne, B. C., Voorberg, N., Schofield, R., Br owne, G., Gafni, A., Schuster, M., et al. (1999). Health education or empowerment education with individuals with a serious persistent ps ychiatric disability. Psychiatric Rehabilitation Journal 22(4), 368-380. Byrne, C. (1999). Facilitating empowerment groups: Dismantling professional boundaries. Issues in Mental Health Nursing 20(1), 55-71. Bystritsky, A., Liberman, R. P., Hwang, S. S., & Wallace, C. J. (2001). Social and independent living and quality of life in obsessive-compulsive disorder and schizophrenia. Depression and Anxiety 14, 214-218. Calabrese, J. D., & Corrigan, P. (2005). Be yond dementia praecox: Findings from longterm follow-up studies of schizophrenia. In R. O. Ralph & P. Corrigan (Eds.), Recovery in mental illness: Broadening our understanding of wellness (pp. 6384). Washington, DC: American Psychological Association. Camp, D. L., Finlay, W. M. L., & Lyons, E. (2002). Is low self-esteem an inevitable consequence of stigma? An example from women with chronic mental health problems. Social Science and Medicine 55, 823-834. Carling, P. J. (1990). Major mental illne ss, housing, and supports: the promise of community integration. American Psychologist 45, 969-975. Carlson, G. A., Bromet, E. J., Driessens, C ., Mojtabai, R., & Schwartz, J. E. (2002). Age of onset, childhood psychopathology, and 2year outcome in psychotic bipolar disorder. American Journal of Psychiatry 159 (2), 307-309.

PAGE 459

440 Carlson, G. A., Bromet, E. J., & Sievers, S. (2000). Phenomenology and outcome of subjects with earlyand a dult-onset psychotic mania. American Journal of Psychiatry, 157(2), 213-219. Carpenter, K. M., & Hittner, J. B. (1995). Di mensional characteristics of the SCL-90-R: Evaluation of gender differences in dually diagnosed inpatients. Journal of Clinical Psychology, 51(3), 383-390. Carpenter, W. T., Gold, J. M., Lahti, A. C., Queern, C. A., Conley, R. R., Bartko, J. J., et al. (2000). Decisional capacity for informed consent in schizophrenia research. Archives of General Psychiatry 57(6), 533-538. Cassano, G. B., Rucci, P., Frank, E., Fagiolini, A., Dell'Osso, L., Shear, M. K., et al. (2004). The mood spectrum in unipolar and bipolar disorder: Arguments for a unitary approach. American Journal of Psychiatry 161(7), 1264-1269. Centers for Disease Control and Prevention. (2005). HIV Basic Statistics. Retrieved January 15, 2005, from http://www.cdc.gov/hiv/stats.htm#hivest Chadwick, P. K. (1997). Recovery from psychosis: Learning more from patients. Journal of Mental Health 6(6), 577-588. Cheasty, M., Clare, A. W., & Collins, C. (1998). Relation between sexual abuse in childhood and adult depression: Case-control study. British Medical Journal 316(7126), 198-201. Chernomas, W. M., Clarke, D. E., & Chishol m, F. A. (2000). Perspectives of women living with schizophrenia. Psychiatric Services 51 (12), 1517-1521. Chinman, M. J., Rosenheck, R., Lam, J. A ., & Davidson, L. (2000). Comparing consumer and nonconsumer provided case management services for homeless persons with serious mental illness. Journal of Nervous and Mental Disease 188(7), 446-453. Clarke, G. N., Herinckx, H. A., Kinney, R. F., Paulson, R. I., Cutler, D. L., Lewis, K., et al. (2000). Psychiatric hospitalization, a rrests, emergency room visits, and homelessness of clients with serious and persisten mental illness: Findings from a randomized trial of two ACT programs vs. usual care. Mental Health Services Research 2(3), 155-164. Clinton, M., Lunney, P., Edwards, J., Weir, D., & Barr, J. (1998). Perceived social support and community adaptation in schizophrenia. Journal of Advanced Nursing, 27, 955-965. Cobb, S. (1976). Social support as a moderator of stress. Psychosomatic Medicine 38 300-314.

PAGE 460

441 Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum, Publishers. Cohen, S., & Syme, S. L. (1985). Social support and health Orlando, FL: Academic Press. Collins, M. E., Mowbray, C. T., & Bybee, D. (2000). Characteristics predicting successful outcomes of participants w ith severe mental illness in supported education. Psychiatric Services 51(6), 774-780. Copeland, M. E. (2001). Wellness recovery action plan: A system for monitoring, reducing and eliminating uncomfortable or dangerous physical symptoms and emotional feelings. In C. E. Brown (Ed.), Recovery and wellness: Models of hope and empowerment for people with mental illness (pp. 127-150). New York: Hawthorn Press, Inc. Corrigan, P., & Calabrese, J. D. (2005). Stra tegies for assessing and diminishing selfstigma. In P. Corrigan (Ed.), On the stigma of mental illn ess: Practical strategies for research and social change (pp. 239-256). Washington, DC: American Psychological Association. Corrigan, P., & Ralph, R. O. (2005). Introducti on: Recovery as consumer vision and research paradigm. In R. O. Ralph & P. Corrigan (Eds.), Recovery in mental illness: Broadening our understanding of wellness (pp. 3-18). Washington, DC: American Psychological Association. Corrigan, P., Thompson, V., Lambert, D., Sangster, Y., Noel, J. G., & Campbell, J. (2003). Perceptions of Discrimination Among Persons With Serious Mental Illness. Psychiatric Services 54(8), 1105-1110. Corrigan, P. W. (1998). The impact of stigma on severe mental illness. Cognitive and Behavioral Practice 5, 201-222. Corrigan, P. W. (2000). Mental health stigma as social attribution: Implications for research methods and attitude change. Clinical Psychology: Science and Practice 7(1), 48-67. Corrigan, P. W. (2002). Empowerment and seriou s mental illness: Treatment partnerships and community opportunities. Psychiatric Quarterly 73(3), 217-228. Corrigan, P. W., Calabrese, J. D., Diwan, S. E., Keogh, C. B., Keck, L., & Mussey, C. (2002). Some recovery processes in mutu al-help groups for persons with mental illness; I: Qualitative analysis of program materials and testimonies. Community Mental Health Journal 38(4), 287-301.

PAGE 461

442 Corrigan, P. W., Faber, D., Rashid, F., & L eary, M. (1999). The construct validity of empowerment among consumers of mental health services. Schizophrenia Research 38 (1), 77-84. Corrigan, P. W., & Garman, A. N. (1997). C onsiderations for research on consumer empowerment and psychosocial interventions. Psychiatric Services, 48 (3), 347352. Corrigan, P. W., Giffort, D., Rashid, F., L eary, M., & Okeke, I. (1999). Recovery as a psychological construct. Community Mental Health Journal, 35(3), 231-239. Corrigan, P. W., Markowitz, F. E., Watson, A. C., Rowan, D., & Kubiak, M. A. (2003). An attribution model of public discriminati on toward persons with mental illness. Journal of Health and Social Behavior 44(June), 162-169. Corrigan, P. W., & Phelan, S. M. (2004). So cial support and recovery in people with serious mental illnesses. Community Mental Health Journal, 40(6), 513-523. Corrigan, P. W., River, L. P., Lundin, R. K., Uphoff-Wasowski, K., Campion, J., Mathisen, J., et al. (2000). Stigmatizing attributions about mental illness. Journal of Community Psychology, 28(1), 91-102. Corrigan, P. W., Rowan, D., Gr een, A., Lundin, R. K., River, L. P., Uphoff-Wasowski, K., et al. (2002). Challenging two mental illness stigmas: Personal responsibility and dangerousness. Schizophrenia Bulletin 28(2), 293-309. Corrigan, P. W., Salzer, M. S., Ralph, R. O ., Sangster, Y., & Keck, L. (2005). Examining the factor structure of the Recovery Assessment Scale. Schizophrenia Bulletin 30(4), 1035-1041. Corrigan, P. W., Watson, A. C., Heyrman, M. L., Warpinski, A., Gracia, G., Slopen, N., et al. (2005). Structural S tigma in State Legislation. Psychiatric Services 56 (5), 557-563. Cowell, A. J., Pollio, D. E., North, C. S., St ewart, A. M., McCabe, M. M., & Anderson, D. W. (2003). Deriving se rvice costs for a clubhouse psychosocial rehabilitation program. Administration & Policy in Mental Health 30(4), 323-340. Coyne, J. (1990). Interpersonal processes in depression. In G. Keitner (Ed.), Depression and Families: Impact and Treatment (pp. 36-52). Washington, DC: APPI Press. Crawford, M. J., de Jonge, E., Freema n, G. K., & Weaver, T. (2004). Providing continuity of care for people with seve re mental illness: A narrative review. Social Psychiatry and Psychiatric Epidemiology 39(4), 265-272.

PAGE 462

443 Cresswell, C. M., Kuipers, L., & Power, M. J. (1992). Social networks and support in long-term psychiatric patients. Psychological Medicine 22, 1019-1026. Crocker, J. (1999). Social stigma and self-est eem: Situational constr uction of self-worth. Journal of Experime ntal Social Psychology 35, 89-107. Crocker, J., & Quinn, D. M. (2000). Social stig ma and the self: Meanings, situations, and self-esteem. In T. F. Heatherton, R. E. Kleck, M. R. Hebl & J. G. Hull (Eds.), The social psychology of stigma (pp. 153-183). New York: The Guilford Press. Cunningham, K., Wolbert, R., Graziano, A ., & Slocum, J. (2005). Acceptance and Change: The Dialectic of Recovery. Psychiatric Rehabilitation Journal 29(2), 146-148. Cutler, C. G. (2001). Self-care agency and sy mptom management in patients treated for mood disorder. Archives of Psychiatric Nursing 15(1), 24-31. Cutting, L. P., & Docherty, N. M. (2000). Sc hizophrenia Outpatients' Perceptions of Their Parents': Is Expressed Emotion a Factor? Journal of Abnormal Psychology 109(2), 266-272. Dalgin, R. S., & Gilbride, D. (2002). Perspectiv es of people with psychiatric disabilities on employment disclosure. Psychiatric Rehabilitation Journal 26(3), 306-310. Darves-Bornoz, J.-M., Lemperiere, T., Degi ovanni, A., & Gaillard, P. (1995). Sexual victimization in women with schi zophrenia and bipolar disorder. Social Psychiatry & Psychiatric Epidemiology 30(2), 78-84. Das, S. K., Malhotra, S., & Basu, D. ( 1999). Family study of acute and transient psychotic disorders: Comparison with schizophrenia. Social Psychiatry and Psychiatric Epidemiology 34(6), 328-332. Davidson, L. (2003). Living outside ment al illness: Qualitative studies of recovery in schizophrenia New York, NY: New York University Press. Davidson, L., Sells, D. J., Sangster, S., & O'Conn ell, M. J. (2005). Qualitative studies of recovery: What can we learn from the person? In R. O. Ralph & P. Corrigan (Eds.), Recovery in mental illness: Br oadening our understanding of wellness (pp. 147-170). Washington, DC: American Psychological Association. Deegan, P. (2005). Recovery as a journey of the heart. In L. Davidson, C. Harding & L. Spaniol (Eds.), Recovery from severe mental i llnesses: Research evidence and implications for practice (Vol. 1, pp. 57-68). Boston: Center for Psychiatric Rehabilitation.

PAGE 463

444 Deegan, P. E. (1998). Recovery: The lived experience of rehabilitation. Rehabilitation Psychology, 43(1), 11-19. Deegan, P. E. (2001). Recovery as a self-dir ected process of healing and transformation. In C. E. Brown (Ed.), Recovery and wellness: Models of hope and empowerment for people with mental illness (pp. 5-22). New York: Hawthorn Press, Inc. Dell'Osso, L., Pini, S., Casano, G. B., Mastrocinq ue, C., Seckinger, R. A., Saettoni, M., et al. (2002). Insight into illness in patient s with mania, mixed mania, bipolar depression and major depressi on with psychotic features. Bipolar Disorders, 4, 315-322. Dempsey, I., & Foreman, P. (1997). Toward a clarification of empowerment as an outcome of disability service provision. International Journal of Disability, Development and Education 44(4), 287-303. Derogatis, L. (1994). SCL-90-R: Symptom Checklist 90-Revised. Administration, scoring and procedures manual (3rd ed.) Minneapolis, MN: National Computer Systems Inc. Dickerson, F. B. (1998). Strategi es that foster empowerment. Cognitive and Behavioral Practice 5(2), 255-275. Dickinson, D., Green, G., Hayes, C., Gilheany, B., & Whittaker, A. (2002). Social network and social support characteristics amongst individuals recently discharged from acute psychiatric units. J ournal of Psychiatric and Mental Health Nursing, 9, 183-189. Dixon, L., Goldberg, R., Lehman, A., & McNary, S. (2001). The impact of health status on work, symptoms, and functional outco mes in severe mental illness. Journal of Nervous and Mental Disease, 189(1), 17-23. Dovidio, J. F., Major, B., & Crocker, J. (2000) Stigma: Introduction and overview. In T. F. Heatherton, R. E. Kleck, M. R. Hebl & J. G. Hull (Eds.), The social psychology of stigma (pp. 1-30). New York: Guildford Publishers, Inc. Drake, R. E. (2005). The principles of evid ence-based mental health treatment. In R. Drake, M. R. Merrens & D. Lynde (Eds.), Evidence-based mental health practice: A textbook (pp. 45-66). New York, NY: W.W. Norton & Company. Due, P., Holstein, B., Rikke, L., Modvig, J., & Avlund, K. (1999). Social relations: Network, support and relational strain. Social Science and Medicine 48 (5), 661673.

PAGE 464

445 Duggan, C., Sham, P., Minne, C., Lee, A., & Murray, R. (1998). Family history as a predictor of poor long-term outcome in depression. British Journal of Psychiatry, 173, 527-530. Dworkin, R. J. (1992). Researching persons with mental illness Thousand Oaks, CA: Sage Publications. Eaton, W. W., Badawi, M., & Melton, B. (1995). Prodromes and precursors: Epidemiologic data for primary preven tion of disorders with slow onset. American Journal of Psychiatry 152 (7), 967-972. Elliott, D. M., Mok, D. S., & Briere, J. (2004). Adult sexual assault: Prevalence, symptomatology, and sex differences in the general population. Journal of Traumatic Stress 17 (3), 203-211. Essock, S. M., Mueser, K. T., Drake, R. E., Covell, N. H., McHugo, G. J., Frisman, L. K., et al. (2006). Comparison of ACT and Standard Case Management for Delivering Integrated Treatment for Co-occurring Disorders. Psychiatric Services 57 (2), 185-196. Evans, J., & Repper, J. (2000). Employment, social inclusion and mental health. Journal of Psychiatric and Mental Health Nursing 7, 15-24. Evenson, R. C., Holland, R. A., Mehta, S., & Yasin, F. (1980). Factor analysis of the Symptom Checklist-90. Psychological Reports 46, 695-699. Everson-Rose, S. A., & Lewis, T. T. ( 2005). Psychosocial factors and cardiovascular diseases. Annual Review of Public Health 26 (1), 469-500. Evert, H., Harvey, C., Trauer, T., & Herrma n, H. (2003). The relati onship between social networks and occupational and self-care functioning in people with psychosis. Social Psychiatry and Psychiatric Epidemiology 38(4), 180-188. Falloon, I. R. H., Heid, T., Roncone, R., Coverd ale, J. H., & Laidlaw, T. (1998). Optimal treatment strategies to enhan ce recovery from schizophrenia. Australian & New Zealand Journal of Psychiatry 32 (1), 43-49. Fallot, R. D., & Harris, M. (2002). The Trauma Recovery and Empowerment Model (TREM): Conceptual and Practical Issues in a Group Intervention for Women. Community Mental Health Journal, 38(6), 475-485. Fassino, S., Ferrero, A., Piero, A ., Mongelli, E., Caviglia, M. L., Delsedime, N., et al. (2004). Rehabilitative residential treatm ent in patients with severe mental disorders: Personality features a ssociated with short-term outcome. American Journal of Orthopsychiatry 74(1), 33-42.

PAGE 465

446 Fava, G. A., Ruini, C., & Belaise, C. (2007). The concept of recovery in major depression. Psychological Medicine 37(3), 307-317. Fawcett, S. B., Paine-Andrews, A., Francisco, V. T., Schultz, J. A., Richter, K. P., Lewis, R. K., et al. (1996). Empowering community health initiatives through evaluation. In D. M. Fetterman, S. Kaftarian & A. Wandersman (Eds.), Empowerment evaluation: Knowledge and tools for self-assessment and accountability (pp. 161187). Thousand Oaks, CA: Sage Publications. Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117149. Finlay, W. M. L., Dinos, S., & Lyons, E. ( 2001). Stigma and multiple social comparisons in people with schizophrenia. European Journal of Social Psychology 31, 579592. Fisher, D. (2003). People are more important than pills in recovery from mental disorder. Journal of Humanistic Psychology 43 (2), 65-68. Fisher, D. B., & Chamberlin, J. (2007). Consumer-Directed Tr ansformation to a Recovery-Based Mental Health System National Empowerment Center. Fitzsimons, R. F. (2002). Empowerment and its implications for clinical practice in mental health: A review. Journal of Mental Health 11(5), 481-499. Fogelson, D. L., Neuchterlein, K. H., Asarnow R. F., Payne, D. L., & Subotnik, K. L. (2004). Validity of the family histor y method for diagnosing schizophrenia, schizophrenia-related psychoses, and schizophrenia-spectrum personality disorders in first-degree relativ es of schizophrenia probands. Schizophrenia Research 68 (2-3), 309-317. Fondacaro, M. R., & Weinberg, D. (2002). Con cepts of Social Ju stice in Community Psychology: Toward a Social Ecological Epistemology. American Journal of Community Psychology, 30(4), 473-492. Foster-Fishman, P. G., Salem, D. A., Chi bnall, S., Legler, R., & Yapchai, C. (1998). Empirical support for the critical assump tions of empowerment theory. American Journal of Community Psychology 26 (4), 507-536. Frank, E., Prien, R. F., Jarrett, R. B., Keller, M. B., Kupfer, D. J., & Lavori, P. W. (1991). Conceptualization and rationale for cons ensus definitions of terms in major depressive disorder: Remission, re covery, relapse, and recurrence. Archives of General Psychiatry 48, 851-855.

PAGE 466

447 Frese, F. J., 3rd. (1998). Advocacy, recover y, and the challenges of consumerism for schizophrenia. Psychiatric Clinics of North America 21(1), 233-249. Frese, F. J., Stanley, J., Kress, K., & Voge l-Scibilia, S. (2001). Integrating evidencebased practices and the recovery model. Psychiatric Services 52, 1462-1468. Friedman, S., Smith, L., Fogel, D., Paradis, C., Viswanathan, R., Ackerman, R., et al. (2002). The incidence and influe nce of early traumatic life events in patients with panic disorder: A comparison with other psychiatric outpatients. Journal of Anxiety Disorders, 16(3), 259-272. Friedrich, R. M., Hollingsworth, B., Hradek, E., Friedrich, H. B., & Culp, K. R. (1999). Family and client perspectives on alternat ive residential settings for persons with severe mental illness. Psychiatric Services, 50 (4), 509-514. Gandhi, D. H., Bogrov, M. U., Osher, F. C., & Myers, C. P. (2003). A Comparison of the Patterns of Drug Use among Patients with and without Severe Mental Illness. American Journal on Addictions, 12(5), 424-431. Gates, L. B., Akabas, S. H., & Zwelling, E. (2001). Have I got a worker for you: Creating employment opportunities for peopl e with psychiatric disability. Administration & Policy in Mental Health 28(4), 319-325. Gearon, J. S., Bellack, A. S., Rachbeisel, J., & Dixon, L. (2001). Dr ug-use behavior and correlates in people with schizophrenia. Addictive Behaviors 26(1), 51-61. Gerlach, J., & Larsen, E. B. (1999). Subject ive expereince and ment al side-effects of antipsychotic medication. Acta Psychiatrica Scandinavia 395 113-117. Giron, M., & Gomez-Beneyto, M. (2004). Rela tionship between family attitudes and social functioning in schizophrenia: A nine-month follow-up prospective study in Spain. Journal of Nervous and Mental Disease 192 (6), 414-420. Gitlin, M. J., Nuechterlein, K. H., Subotnik, K. L., Ventura, J., Mintz, J., Fogelson, D. L., et al. (2001). Clinical outcome following neuroleptic discontinuation in patients with remitted recent-onset schizophrenia. American Journal of Psychiatry 158(11), 1835-1842. Glick, I. D., Murray, S. R., Vasudevan, P., Ma rden, S. R., & Hu, R. J. (2001). Treatment with atypical antipsychotics: New indications and new populations. Journal of Psychiatric Research 35 (3), 187-191. Goering, P., Durbin, J., Foster, J., Boyles, S., Babiak, T., & Lancee, B. (1992). Social networks of resident s in supportive housing. Community Mental Health Journal 28(3), 199-214.

PAGE 467

448 Goffman, I. (1963). Stigma: Notes on the manage ment of spoiled identity New York: Simon & Schuster. Goodman, L. A., Rosenberg, S. D., Mueser, K. T., & Drake, R. E. (1997). Physical and sexual assault history in women with serious mental illness: Prevalence, correlates, treatment, and future research directions. Schizophrenia Bulletin 23(4), 685-696. Goodman, L. A., Salyers, M. P., Mueser, K. T., Rosenberg, S. D., Swartz, M., Essock, S. M., et al. (2001). Recent victimization in women and men with severe mental illness: Prevalence and correlates. Journal of Traumatic Stress 14(4), 615-632. Gouvier, W. D., Sytsma-Jordan, S., & Mayville S. (2003). Patterns of discrimination in hiring job applicants with disabilities: The role of disa bility type, job complexity, and public contact. Rehabilitation Psychology, 48 (3), 175-181. Green, G., Hayes, C., Dickinson, D., Whittaker, A., & Gilheany, b. (2002). The role and impact of social relation ships upon well-being reported by mental health services users: A qualitative study. Journal of Mental Health, 11 (5), 565-579. Green, M. F. (2001). Schizophrenia revealed: From neurons to social interactions New York, NY: W.W. Norton Publishers. Greenberg, J. S., Knudsen, K. J., & Aschbr enner, K. A. (2006). Prosocial Family Processes and the Quality of Life of Persons with Schizophrenia. Psychiatric Services 57 (12), 1771-1777. Grunebaum, M., Aquila, R., Portera, L., Leon, A. C., & Weiden, P. (1999). Predictors of management problems in supported housing: a pilot study. Community Mental Health Journal 35(2), 127-133. Grusky, O., Tierney, K., Manderscheid, R. W., & Grusky, D. B. (1985). Social bonding and community adjustment of ch ronically mentally ill adults. Journal of Health and Social Behavior 26(March), 49-63. Gureje, O., Harvey, C., & Herrman, H. (2004). Self-esteem in patients who have recovered from psychosis: profile and relationship to quality of life. Australian & New Zealand Journal of Psychiatry 38(5), 334-338. Hafkenscheid, A. (1993). Psychometic evaluation of the symptom checklist (SCL-90) in pyschiatric inpatients. Personality and Individual Differences 14(6), 751-756. Handler, J., Doel, K., Henry, A., & Lucca, A. (2003). Implementing supported employment services in a real-world setting. Psychiatric Services 54(7), 960-962.

PAGE 468

449 Hartman, L. L. (1996). The social reintegration of formerly institutionalized mental patients into a community setting Unpublished Dissertation; Empirical Study, University of Texas, Austin. Hatcher, L. (1994). A step-by-step approach to using th e SAS system for factor analysis and structural equation modeling. Cary, NC: SAS Institute, Inc. Hatcher, L., & Stepanski, E. J. (1994). A step-by-step approach to using the SAS system for univariate and multivariate statistics Cary, NC: SAS Publishing. Hayden, E. P., & Klein, D. N. (2001). Outcome of dysthymic disorder at 5-year followup: The effect of familial psyc hopathology, early adversity, personality, comorbidity, and chronic stress. American Journal of Psychiatry 158 (11), 18641870. Health, N. F. C. o. M. (2003). New Freedom Commission on Mental Health, Achieving the Promise: Transforming Mental He alth Care in America, Final Report (NonPeer Reviewed No. SMA-03-3832). Rockv ille, MD: Department of Health and Human Services. Herman, N. J. (1993). Return to sender: Rein tegrative stigma-management strategies of ex-psychiatric patients. Journal of Contemporary Ethnography 22(3), 295-330. Hess, R. E., & Mercer, W. M. (2001). Empowe rment effects of teaching leadership skills to adults with a severe mental illness and their families. Psychiatric Rehabilitation Journal 24(3), 257-265. Hinshaw, S. P., & Cicchetti, D. (2002). Stig ma and mental disorder: Conceptions of illness, public attitudes, personal disclosure, and social policy. Development and Psychopathology, 12, 555-598. Hogan, M. F. (2003). The President's New Freedom Commission: Recommendations to transform mental health care in America. Psychiatric Services, 54(11), 14671474. Hollon, S. D., Thase, M. E., & Markowitz, J. C. (2003). Treatment and prevention of depression. Psychological Science in the Public Interest 3(2), 39-77. Holmbeck, G. N. (1997). Toward Terminological Conceptual, and Statistical Clarity in the Study of Mediators and Moderators: Examples From the Child-Clinical and Pediatric Psychology Literatures. Journal of Consulting and Clinical Psychology 65(4), 599-610.

PAGE 469

450 Holzinger, A., Kilian, R., Lindenbach, I., Pets cheleit, A., & Angermeyer, M. C. (2003). Patients' and their relatives' causa l explanations of schizophrenia. Social Psychiatry and Psychiatric Epidemiology 38, 155-162. Holzinger, A., Loffler, W., Muller, P., Priebe, S., & Angermeyer, M. C. (2002). Subjective illness theory and antipsyc hotic medication compliance by patients with schizophrenia. Journal of Nervous and Mental Disease 190(9), 597-603. Honey, A. (2004). Benefits and drawbacks of employment: Perspectives of people with mental illness. Qualitative Health Research 14(3), 381-395. Hooley, J. M., & Gotlib, I. H. (2000). A diat hesis-stress conceptua lization of expressed emotion and clinical outcome. Applied and Preventive Psychology 9(3), 135-151. House, J. S. (1988). Work stress and social support Reading, MA: Addison-Wesley, Press. Hughes, I., Hill, B., & Budd, R. (1997). Complia nce with antipsychotic medication: From theory to practice. Journal of Mental Health 6(5), 473-490. Humberstone, V. (2002). The experiences of people with schizophrenia living in supported accommodation: A qualitative study using grounded theory methodology. Australian and New Zealand Journal of Psychiatry, 36, 367-372. Hur, M. H. (2006). Empowerment in terms of theoretical persp ectives: Exploring a typology of the process and co mponents across disciplines. Journal of Community Psychology, 34(5), 523-540. Hutchinson, D. S., Anthony, W., Massaro, J., & Rogers, E. S. (2007). Evaluation of a Combined Supported Computer Education and Employment Training Program for Persons with Psychiatric Disabilities. Psychiatric Rehabilitation Journal 30(3), 189-197. Isaac, S., & Michael, W. B. (1995). Handbook in research and evaluation (3rd ed.). San Diego, CA: EDITS, Educational a nd Industrial Testing Services. Jerrell, J. M., Cousins, V. C ., & Roberts, K. M. (2006). Psychometrics of the Recovery Process Inventory. The Journal of Behavioral Health Services and Research 33(4), 464-473. Johnson, L., Lundstrom, O., Aberg-Wistedt, A., & Mathe, A. A. (2003). Social support in bipolar disorder: Its releva nce to remission and relapse. Bipolar Disorders, 5(2), 129-137.

PAGE 470

451 Jorgensen, P., Nordentoft, M., Abel, M. B ., Gouliaev, G., Jeppesen, P., & Kassow, P. (2000). Early detection and assertive co mmunity treatment of young psychotics: The OPUS study: Rationale and design of the trial. Social Psychiatry and Psychiatric Epidemiology 35(7), 283-287. Judd, L. L., & Akiskal, H. S. (2003). The prev alence and disability of bipolar spectrum disorders in the US population: Re-ana lysis of the ECA database taking into account subthreshold cases. Journal of Affective Disorders 73 123-131. Kapur, S. (2003). Psychosis as a state of ab errant salience: A framework linking biology, phenomenology, and pharmacology in schizophrenia. American Journal of Psychiatry, 160(1), 13-23. Kawachi, I., & Berkman, L. F. (2001). Social ties and mental health. Journal of Urban Health: Bulletin of the Ne w York Academy of Medicine 78 (3), 458-467. Keitner, G., Miller, I., & Epstein, N. (1990). The functioning of families with major depression. International Journal of Family Psychiatry 7, 11-16. Kelly, B. D. (2005). Structural violence and schizophrenia. Social Science and Medicine 61(3), 721-730. Kennedy, N., Abbot, R., & Paykel, E. S. (2004). Longitudinal syndromal and subsyndromal symptoms after severe de pression: 10-year follow-up study. British Journal of Psychiatry, 184 (4), 330-336. Kessler, R. C. (2000). Psychiatric epidemiology: Selected recent advances and future directions. Bulletin of the World Health Organization 78(4), 464-474. Kessler, R. C., Davis, C. G., & Kendler, K. S. (1997). Childhood adversity and adult psychiatric disorder in the U. S. National Comorbidity Survey. Psychological Medicine, 27 1101-1119. Kessler, R. C., Leaf, P. J., Bruce, M. L., Fr iedman, R. M., Grosser, R. C., Kennedy, C., et al. (1996). The 12-month prevalence and correlate s of serious mental illness (SMI) Rockville, Maryland: U.S. Departme nt of Health and Human Services. Kessler, R. C., Merikangas, K. R., Berglund, P., Eaton, W. W., Koretz, D. S., & Walters, E. E. (2003). Mild disorders should not be eliminated from the DSM-IV. Archives of General Psychiatry 60(Nov), 1117-1122. Kessler, R. C., Mickelson, K. D., & Williams, D. R. (1999). The prevalence, distribution, and mental health correlates of percei ved discrimination in the United States. Journal of Health and Social Behavior 40(September), 208-230.

PAGE 471

452 Kessler, R. C., & Walters, E. (2002). The natio nal comorbidity survey. In M. T. Tsuang & M. Tohen (Eds.), Textbook in psychiatric epidemiology (2nd ed., pp. 343-362). New York: Wiley-Liss, Inc. Kessler, R. C., Walters, E. E., & Forthofer, M. S. (1998). The social consequences of psychiatric disorders, III: Pr obability of marital stability. American Journal of Psychiatry, 155(8), 1092-1096. Kessler, R. C., & Zhao, S. (1999). Overview of descriptive epid emiology of mental disorders. In C. S. Aneshensel & J. C. Phelan (Eds.), Handbook of the sociology of mental health (pp. 127-150). New York: Kluwer Academic / Plenum Publishers. Kessler, R. C., Zhao, S., Katz, S. J., Kouzis A. C., Frank, R. G., Edlund, M. J., et al. (1999). Past-year use of outpatient servi ces for psychiatric problems in the National Comorbidity Survey. American Journal of Psychiatry 156, 115-123. Kinderman, P., Setzu, E., Lobban, F., & Salmon, P. (2006). Illness beliefs in schizophrenia. Social Science & Medicine 63 (7), 1900-1911. King, G., Cathers, T., Brown, E., Specht, J. A., Willoughby, C., Polgar, J. M., et al. (2003). Turing points and protective processe s in the lives of people with chronic disabilities. Qualitative Health Research 13(2), 184-206. Klein, D. N., Schwartz, J. E., Rose, S., & Leader, J. B. (2000). Five-year course and outcome of dysthymnic disorder: A pr ospective, naturalis tic follow-up study. American Journal of Psychiatry 157 (6), 931-939. Klein, D. N., Shankman, S. A., & Rose, S. (2006). Ten-Year Prospective Follow-Up Study of the Naturalistic Course of Dysthymic Disorder and Double Depression. American Journal of Psychiatry 163 (5), 872-880. Knight, M. T. D., Wykes, T., & Haywar d, P. (2003). 'People don't understand": An investigation of stigma in schizophrenia using Interpretative Phenomenological Analysis (IPA). Journal of Mental Health 12 (3), 209-222. Kraepelin, E. (1913). Lectures on clinical psychiatry (3rd ed.). New York, NY: William Wood. Kraepelin, E. (1919). Dementia praecox and paraphrenia (R. M. Barclay, Trans.). Chicago: Chicago Medical Books. Kramer, R. C. (2000). Poverty, in equality, and youth violence. Annals of the American Academy of Political and Social Science 567, 123-139.

PAGE 472

453 Kreisman, D. D., & Blumenthal, R. (1995). Emotional overinvolvement: A review and examination of its role in expressed emotion. In J. R. Greenley (Ed.), Research in community and mental health (pp. 3-40). Greenwich: JAI Press, Inc. Kruger, A. (2000a). Empowerment in social work practice with the psychiatrically disabled: Model and method Smith College Studies in Social Work, 70(3), 427439. Kruger, A. (2000b). Schizophrenia: Recovery and hope. Psychiatric Rehabilitation Journal 24, 29-37. Laudet, A. B., Magura, S., Vogel, H. S., & Knight, E. (2000). Support, Mutual Aid and Recovery from Dual Diagnosis. Community Mental Health Journal 36(5), 457476. Laudet, A. B., Magura, S., Vogel, H. S., & Kni ght, E. L. (2002). Interest in and obstacles to pursuing work among unemploye d dually diagnosed individuals. Substance Use & Misuse 37(2), 145-170. Leavy, R. L. (1983). Social support and psychological disorder: A review. Journal of Community Psychology, 11, 3-21. Lencz, T., Smith, C. W., Auther, A., Correll, C. U., & Cornblatt, B. (2004). Nonspecific and attenuated negative symptoms in patients at clinical high-risk for schizophrenia. Schizophrenia Research 68(1), 37-48. Leskel, U., RytsL, H., Komulainen, E., Melartin, T., Sokero, P., Lestel-Mielonen, P., et al. (2006). The influence of adve rsity and perceived social support on the outcome of major depressive disorder in subjects with different levels of depressive symptoms. Psychological Medicine, 36 (6), 779-788. Lester, L., Tritter, J. Q., & England, E. (2003). Satisfaction with primary care: The perspectives of people with schizophrenia. Family Practice 20 (5), 508-513. Letvak, S. (2002). The importance of soci al support for rural mental health. Issues in Mental Health Nursing, 23, 249-261. Levitan, R. D., Rector, N. A., Sheldon, T., & Goering, P. (2003). Childhood adversities associated with major depression and/or anxiety disorders in a community sample of Ontario: Issues of co -morbidity and specificity. Depression and Anxiety 17(1), 34-42. Levy, M. I. (2005). Law and Psychiatry: Stigma and ment al illness in the United States Retrieved February 28, 2005

PAGE 473

454 Liberman, R. P., & Kopelowicz, A. (2005). R ecovery from schizophrenia: A criterionbased definition. In R. O. Ralph & P. Corrigan (Eds.), Recovery in mental illness: Broadening our understanding of wellness (pp. 101-130). Washington, DC: American Psychological Association. Lin, N., Ensel, W. M., Simeone, R. S., & K uo, W. (1979). Social support, stressful life events, and illness: A m odel and an empirical test. Journal of Health and Social Behavior, 20 (2), 108-119. Linhorst, D. M., Hamilton, G., Young, E., & Eckert, A. (2002). Opportunities and barriers to empowering people with severe mental illness through participation in treatment planning. Social Work 47 (4), 425-434. Link, B. G. (1987). Understanding labeling effect s in the area of mental disorders: An assessment of the effects of expectations of rejection. American Sociological Review 52(1), 96-112. Link, B. G. (2001). Stigma: Many mechanisms require multifaceted responses. Epidemiologia e Psichiatria Sociale 10(1), 8-11. Link, B. G., Cullen, F. T., Frank, J., & Woznia k, J. F. (1987). The social rejection of former mental patients: Understanding why labels matter. American Journal of Sociology 92 (6), 1461-1500. Link, B. G., Cullen, F. T., Struening, E. L., Shrout, P. E., & Dohrenwend, B. P. (1989). A modified labeling theory approach to ment al disorders: An empirical assessment. American Sociological Review 54(3), 400-423. Link, B. G., & Phelan, J. C. (2001). Conceptualizing stigma. Annual Review of Sociology 27, 363-385. Link, B. G., Phelan, J. C., Bresnahan, M ., Stueve, A., & Pescosolido, B. A. (1999). Public conceptions of mental illness: Labels, causes, dangerousness, and social distance. American Journal of Public Health 89 (9), 1328-1333. Link, B. G., Struening, E. L., Rahav, M., Ph elan, J. C., & Nuttbrock, L. (1997). On stigma and its consequences: Evidence from a longitudinal study of men with dual diagnoses of mental illness and substance abuse. Journal of Health and Social Behavior 38, 177-190. Lipton, F. R., Cohen, C. I., Fisher, E., & Ka tz, S. E. (1981). Schizophrenia: A network crisis. Schizophrenia Bulletin 7, 144-151.

PAGE 474

455 Lipton, F. R., Siegel, C., Hannigan, A., Samu els, J., & Baker, S. (2000). Tenure in supportive housing for homeless persons with severe mental illness. Psychiatric Services 51, 479-486. Littrell, K. H., & Littrell, S. H. (1998). I ssues of reintegration and rehabilitation in schizophrenia. Psychiatric Annals, 28 (7), 371-377. Lomas, J. (1998). Social capital and hea lth: Implications for public health and epidemiology. Social Science and Medicine 47 (9), 1181-1888. Lysaker, P. H., Buck, K. D., Hammoud, K., Taylor, A. C., & Roe, D. (2006). Associations of symptoms, psychosocial f unction and hope with qualities of selfexperience in schizophrenia: Comparisons of objective and subjective indicators of health. Schizophrenia Research 82 (2-3), 241-249. Lysaker, P. H., Campbell, K., & Johannesen, J. K. (2005). Hope, Awareness of Illness, and Coping in Schizophrenia Spectrum Di sorders: Evidence of an Interaction. Journal of Nervous & Mental Disease 193(5), 287-292. Macchi, M., Peretti, L., & Barabino, C. (1996). Experience of a day center for young psychotics. International Journal of Mental Health 25(3), 68-72. Macdonald, E. M., Hayes, R. L., & Baglioni A. J. (2000). The quantity and quality of social networks of young people with early psychosis compared with closely matched controls. Schizophrenia Research 46, 25-30. Macias, C., Propst, R., Rodican, C., & Boyd, J. (2001). Strategic planning for ICCD clubhouse implementation: Development of the clubhouse research and evaluation screening survey (CRESS). Mental Health Services Research 3(3), 155-167. Magura, S., Laudet, A. B., Mahmood, D., Rose nblum, A., Vogel, H. S., & Knight, E. L. (2003). Role of self-help processe s in achieving abstinence among dually diagnosed persons. Addictive Behaviors 28(3), 399-413. Mahoney-Holst, C. (2005). From psychosis a nd alienation to recove ry. In L. Davidson, C. Harding & L. Spaniol (Eds.), Recovery from severe mental illnesses: Research evidence and implica tions for practice (Vol. 1, pp. 479-484). Boston: Center for Psychiatric Rehabilitation. Malm, U., Ivarsson, B., Allebeck, P., & Fa lloon, I. R. (2003). Integrated care in schizophrenia: a 2-year randomized cont rolled study of two community-based treatment programs. Acta Psychiatrica Scandinavica 107(6), 415-423.

PAGE 475

456 Malone, J. (1988). The social suppor t, social dissupport continuum. Journal of Psychosocial Nursing 26 (12), 18-22. Mancini, M. A., Hardiman, E. R., & Laws on, H. A. (2005). Making Sense of It All: Consumer Providers' Theories About Fact ors Facilitating and Impeding Recovery from Psychiatric Disabilities. Psychiatric Rehabilitation Journal 29(1), 48-55. Markowitz, F. E. (2001). Modeling processes in recovery from mental illness: Relationships between symptoms, life satisfaction, and self-concept. Journal of Health and Social Behavior 42(March), 64-79. Marrone, J., & Golowka, E. (2005). If work ma kes people with mental illness sick, what do unemployment, poverty, and social is olation cause? In L. Davidson, C. Harding & L. Spaniol (Eds.), Recovery from severe ment al illnesses: Research evidence and implica tions for practice (Vol. 1, pp. 451-463). Boston: Center for Psychiatric Rehabilitation. Marshall, S. L., Crowe, T. P., Oades, L. G., Deane, F. F., & Kavanagh, D. J. (2007). A Review of Consumer Involvement in Evaluations of Case Management: Consistency With a Recovery Paradigm. Psychiatric Services 58(3), 396-401. Marwaha, S., & Johnson, S. (2004). Schi zophrenia and employment: A review. Social Psychiatry and Psychiatric Epidemiology 39(5), 337-349. Masterson, S., & Owen, S. (2006). Mental health service user 's social and individual empowerment: Using theories of power to elucidate far-reaching strategies. Journal of Mental Health 15(1), 19-34. McCay, E., Beanlands, H., Leszcz, M., Goerin g, P., Seeman, M. V., Ryan, K., et al. (2006). A Group Intervention to Promot e Healthy Self-Concepts and Guide Recovery in First Episode Schizophrenia: A Pilot Study. Psychiatric Rehabilitation Journal 30(2), 105-111. McFarlane, A. C., Bookless, C., & Air, T. (2001). Posttraumatic stress disorder in a general psychiatric inpatient population. Journal of Traumatic Stress 14 (4), 633645. McGorry, P. D. (1992). The concept of recove ry and secondary prev ention in psychotic disorders. Australian & New Zealand Journal of Psychiatry, 26, 3-17. McGruder, J. H. (2001). Life experience is not a disease or why medicalizing madness is counterproductive to recover y. In C. E. Brown (Ed.), Recovery and wellness: Models of hope and empowerment for people with mental illness (pp. 59-80). New York: Hawthorn Press, Inc.

PAGE 476

457 McGurk, S. R., & Mueser, K. T. (2004). Cognitive functioning and employment in severe mental illness. Journal of Nervous and Mental Disease 191 (12), 789-798. McGurk, S. R., Mueser, K. T., Harvey, P. D., LaPuglia, R., & Marder, J. (2003). Cognitive and symptom predictors of work outcomes for clients with schizophrenia in supported employment. Psychiatric Services 54(8), 1129-1135. McHugo, G. J., Drake, R. E., Teague, G. B ., & Xie, H. (1999). Fidelity to Assertive Community Treatment and Client Outcomes in the New Hampshire Dual Disorders Study. Psychiatric Services 50(6), 818-824. McKnight, P. E., McKnight, K. M., Si dani, S., & Figueredo, A. J. (2007). Missing Data: A Gentle Introduction New York: The Guilford Press. Meddings, S., & Perkins, R. (2002). What 'ge tting better' means to staff and users of a rehabilitation service: An exploratory study. Journal of Mental Health 11 (3), 319-325. Meeks, S., & Murrell, S. A. (1994). Service pr oviders in the social networks of clients with severe mental illness. Schizophrenia Bulletin, 20(2), 399-406. Mettee, D. R., & Smith, G. R. (1977). Social comparison and interpersonal attration: The case for dissimilarity. In J. M. Suls & R. L. Miller (Eds.), Social comparison processes: Theoretica l and empirical perspectives (pp. 69-102). Washington, DC: Hemisphere. Miller, B. J., Paschall, C. B., III, & Sve ndsen, D. P. (2006). Mortality and Medical Comorbidity Among Patients With Serious Mental Illness. Psychiatric Services 57(10), 1482-1487. Miller, T. J., McGlashan, T. H., Rosen, J. L ., Somjee, L., Markovich, P. J., Stein, K., et al. (2002). Prospective diagnosis of the initial prodrome for schizophrenia based on the structured intervie w for the prodromal syndromes: Preliminary evidence of interrater reliability and predictive validity. American Journal of Psychiatry 159(5), 863-865. Milliken, P. J., & Northcott, H. C. (2003) Redefining parental identity: caregiving and schizophrenia. Qualitative Health Research 13(1), 100-113. Mintz, J., Mintz, L. I., Arruda, M. J., & Hwang, S. S. (1992). Treatments of depression and the functional capacity to work. Archives of General Psychiatry 49, 761-788. Morin, R. C., & Seidman, E. (1986). A social network approach and the revolving door patient. Schizophrenia Bulletin 12(2), 262-273.

PAGE 477

458 Mowbray, C. T., Gutierrez, L. M., Bellamy, C. D., Sziilvagyi, S., & Strauss, S. (2003). Replication of a psychosocial rehabilita tion program: A case study analysis of supported education. Journal of Community Psychology, 31 (5), 437-457. Mueser, K. T., Becker, D. R., & Wolf e, R. (2001). Supported employment, job preferences, job tenu re and satisfaction. Journal of Mental Health 10 (4), 411417. Mueser, K. T., Corrigan, P. W., Hilton, D. W., Tanzman, B., Schaub, A., Gingerich, S. L., et al. (2002). Illness management a nd recovery: A review of the research. Psychiatric Services 53, 1272-1284. Mueser, K. T., Goodman, L. B., Trumbetta, S. L., Rosenberg, S. D., Osher, F. C., Vidaver, R., et al. (1998). Trauma and posttraumatic stress disorder in severe mental illness. Journal of Consulting & Clinical Psychology 66(3), 493-499. Mueser, K. T., Torrey, W. C., Lynde, D ., Singer, P., & Drake, R. E. (2003). Implementing evidence-based practices for people with severe mental illness. Behavior Modification 27(3), 387-411. Nagle, S., Cook, J. V., & Polatajko, H. J. (2002). I'm doing as much as I can: Occupational choices of persons with a severe and persistent mental illness. Journal of Occupational Science 9 (2), 72-81. Nelson, G., Lord, J., & Ochocka, J. (2001a). Empowerment and mental health in community: Narratives of psyc hiatric consumer/survivors. Journal of Community and Applied Social Psychology 11(2), 125-142. Nelson, G., Lord, J., & Ochocka, J. (2001b). Shifting the paradigm in community mental health Toronto, Canada: Universi ty of Toronto Press. New Freedom Commission on Me ntal Health. (2003). New Freedom Commission on Mental Health, Achieving the Promise: Transforming Mental Health Care in America, Final Report. (No. SMA-03-3832). Rockville, MD: Department of Health and Human Services. Newman, S. J. (2001). Housing attributes and serious mental illness: implications for research and practice. Psychiatric Services, 52 (10), 1309-1317. Nicolson, R., Brookner, F. B., Lenane, M., Goch man, P., Ingraham, L. J., Egan, M. F., et al. (2003). Parental schizophrenia spec trum disorders in childhood-onset and adult-onset schizophrenia. American Journal of Psychiatry 160 (3), 490-495. Nieves, E. J. (2002). The effectiveness of the assertive community treatment model. Administration & Policy in Mental Health 29 (6), 461-480.

PAGE 478

459 Nikelly, A. G. (2001). The role of envir onment in mental health: Individual empowerment through so cial restructuring. Journal of Applied Behavioral Science 37 (3), 305-323. Noble, L. M., Douglas, B. C., & Newman, S. P. (2001). What do patients expect of psychiatric services? A systematic and critical review of empirical studies. Social Science and Medicine 52 (7), 985-998. Nolan, P., & Clancy, A. (1995). Survey of th e social networks of people with severe mental health problems. Journal of Psychiatric and Mental Health Nursing 2, 131-142. O'Brien, M. P., Gordon, J. L., Bearden, C. E., Lopez, S. R., Kopelowicz, A., & Cannon, T. D. (2006). Positive family environment predicts improvement in symptoms and social functioning among adolescents at imminent risk for onset of psychosis. Schizophrenia Research 81(2-3), 269-275. Ochocka, J., Nelson, G., & Janzen, R. (2005) Moving forward: Negotiating self and external circumstances in recovery. Psychiatric Rehabilitation Journal 28(4), 315-322. Olstad, R., Sexton, H., & Sogaard, A. J. (1999). The Finnmark study. Social support, social network and mental distre ss in a prospective population study. Social Psychiatry and Psychiatric Epidemiology 34, 519-525. Olstad, R., Sexton, H., & Sogaard, A. J. (2001). The Finnmark Study. A prospective population study of the social support buffe r hypothesis, specific stressors and mental distress. Social Psychiatry and Psychiatric Epidemiology 36, 582-589. Ormel, J., Oldehinkel, A. J., & Brilman, E. I. (2001). The interplay and etiological continuity of neuroticism, difficulties, and life events in the etiology of major and subsyndromal, first and recurrent de pressive episodes in later life. American Journal of Psychiatry, 158 (6), 885-891. Paulson, R., Herinckx, H., Demmler, J., Clarke, G., Cutler, D., & Birecree, E. (1999). Comparing practice patterns of consumer and non-consumer mental health service providers. Community Mental Health Journal 35 (3), 251-269. Paulson, R. I., Post, R. L., Herinckx, H. A ., & Risser, P. (2002). Beyond components: Using fidelity scales to measure and a ssure choice in program implementation and quality assurance. Community Mental Health Journal 38(2), 119-128.

PAGE 479

460 Penn, D. L., Kolmaier, J. R., & Corrigan, P. W. (2000). Interpersonal factors contributing to the stigma of schizophrenia: Social skills, perceived attractiveness, and symptoms. Schizophrenia Research 45, 37-45. Perese, E. F. (1997). Unmet needs of persons with chronic mental il lnesses: relationship to their adaptation to community living. Issues in Mental Health Nursing 18(1), 19-34. Perlick, D. A., Rosenheck, R. A., Clarkin, J. F., Sirey, J. A., Salahi, J., Struening, E. L., et al. (2001). Adverse effects of perceived stigma on social adaptation of persons diagnosed with bipola r affective disorder. Psychiatric Services, 52(12), 16271632. Pettie, D., & Triolo, A. M. (1999). The search for identity and meaning in the recovery process. Psychiatric Rehabilitation Journal 22, 255-262. Pevalin, D. J., & Goldberg, D. P. (2003). Soci al precursors to onset and recovery from episodes of common mental illness. Psychological Medicine 33(2), 299-306. Phelan, J. C. (2002). Genetic base of mental illness -a cure for stigma? Trends in Neurosciences, 25(8), 430-431. Pittman, D. C., Parsons, R., & Peterson, R. W. (1990). Easing the way: A multifaceted approach to day treatment. Journal of Psychosocial Nursing 28(11), 7-11. Poortinga, W. (2006). Social capital: An in dividual or collective resource for health? Social Science & Medicine 62(2), 292-302. Pourmand, D., Kavanagh, D. J., & Vaughan, K. (2005). Expressed emotion as predictor of relapse in patients with comorbid psychoses and substance use disorder. Australian & New Zealand Journal of Psychiatry 39(6), 473-478. Provencher, H. L., Gregg, R., Mead, S., & Mu eser, K. (2002). The role of work in the recovery of persons with psychiatric disabilities. Psychiatric Rehabilitation Journal 26(2), 132-144. Pyne, J. M., Bean, D., & Sullivan, G. (2001). Characteristics of patients with schizophrenia who do not believe they are mentally ill. Journal of Nervous and Mental Disease 189(3), 146-153. Rachbeisel, J., Scott, J., & Dixon, L. (1999) Co-occurring severe mental illness and substance use disorders: A review of recent literature. Psychiatric Services, 50, 1427-1434. Ralph, R. O. (2000). Recovery. Psychiatric Rehabilitation Journal 4(3), 480-517.

PAGE 480

461 Ralph, R. O., & Corrigan, P. W. (Eds.). (2005). Recovery in mental i llness: Broadening our understanding of wellness Washington, DC: American Psychological Association. Rappaport, J. (1987). Terms of empowerment/exe mplars of prevention: Toward a theory for community psychology. American Journal of Community Psychology, 15(2), 121-148. Rappaport, J. (1995). Empowerment meets narra tive: Listening to stories and creating settings. American Journal of Community Psychology 23(5), 795-807. Ratzlaff, S., McDiarmid, D., Marty, D., & Rapp, C. (2006). The Kansas consumer as provider program: Measuring the effects of a supported educ ation initiative. Psychiatric Rehabilitation Journal 29(3), 174-182. Rauter, U. K., Leonard, C. E., & Swett, C. P. (1996). SCL-90-R factor structure in an acute involuntary adult psyc hiatric inpatient sample. Journal of Clinical Psychology, 52(6), 625-629. Rempfer, M. V., & Knott, J. (2001). Participatory action research: A model for establishing partnerships between mental health res earchers and persons with psychiatric disabilities. In C. E. Brown (Ed.), Recovery and wellness: Models of hope and empowerment for people with mental illness (pp. 151-166). New York: Hawthorn Press, Inc. Repper, J., & Brooker, C. (1998). Serious mental health problems in the community: The significance of policy, practi ce and research. In C. Brooker & J. Repper (Eds.), Serious mental health problems in the community: Policy, practice and research (pp. 3-13). London: Balliere-Tindall. Repper, J., & Perkins, R. (2003). Social inclusion and recovery: A model for mental health practice London, Britain: Ba illiere Tindall. Resnick, S. G., Armstrong, M., Sperrazza, M., Harkness, L., & Rosenheck, R. A. (2004). A model of consumer-provide r partnership: Vet-to-vet. Psychiatric Rehabilitation Journal 28(2), 185-187. Resnick, S. G., Fontana, A., Lehman, A. F., & Rosenheck, R. A. (2005). An empirical conceptualization of th e recovery orientation. Schizophrenia Research 75 (1), 119-128. Resnick, S. G., Rosenheck, R. A., & Lehman, A. F. (2004). An exploratory analysis of correlates of recovery. Psychiatric Services 55(5), 540-547.

PAGE 481

462 Riecher-Rossler, A., & Hafner, H. (2000). Gender aspects in schizophr enia: Bridging the border between social a nd biological psychiatry. Acta Psychiatrica Scandinavica 102(Suppl 407), 58-62. Ritsher, J. B., Otilingam, P. G., & Grajales M. (2003). Internalized stigma of mental illness: Psychometric properties of a new measure. Psychiatry Research 121(1), 31-49. Robinson, D. G., Woerner, M. G., McMenima n, M., Mendelowitz, A., & Bilder, R. M. (2004). Symptomatic and functional r ecovery from a first episode of schizophrenia or schizoaffective disorder. American Journal of Psychiatry 161(3), 473-479. Rogers, E. S., Anthony, W., & Lyass, A. (2004) The Nature and Dimensions of Social Support Among Individuals with Severe Mental Illnesses. Community Mental Health Journal 40(5), 437-450. Rogers, S. E., Chamberlin, J., Ellison, M. L., & Crean, T. (1997). A consumerconstructed scale to measure empowerme nt among users of mental health services. Psychiatric Services, 48 (8), 1042-1047. Rosenfield, S. (1997). Labeling mental illness: The effects of received services and perceived stigma on life satisfaction. American Sociological Review 62, 660-672. Roth, D., & Crane-Ross, D. (2002). Impact of services, met needs, and service empowerment on consumer outcomes. Mental Health Services Research 4(1), 43-56. Rothbard, A. B., Metraux, S., & Blank, M. B. (2003). Cost of care for Medicaid recipients with serious mental illness and HIV infection or AIDS. Psychiatric Services 54 (12), 1240-1246. Rungreangkulkij, S., & Chesla, C. (2001). Smooth a heart with water: Thai mothers care for a child with schizophrenia. Archives of Psychiatric Nursing 15(3), 120-127. Rusch, N., Lieb, K., Bohus, M., & Corrigan, P. W. (2006). Brief Reports: Self-Stigma, Empowerment, and Perceived Legitim acy of Discrimination Among Women With Mental Illness. Psychiatric Services 57 (3), 399-402. Sacco, W. P., & Beck, A. T. (1995). Cognitive theory and therapy. In E. E. Beckham & W. R. Leber (Eds.), Handbook of Depression (2nd ed., pp. 329-351). New York, NY: Guildford Press.

PAGE 482

463 Salkever, D., Domino, M. E., Burns, B. J., Sa ntos, A. B., Deci, P. A., Dias, J., et al. (1999). Assertive community treatment for people with severe mental illness: the effect on hospital use and costs. Health Services Research 34 (2), 577-601. Sayce, L. (1998). Stigma, discrimination and social exclusion: What's in a word? Journal of Mental Health 7 (4), 331-343. Scheff, T. (1974). The labeling theory of mental illness. American Sociological Review 39, 444-452. Scheid, T. L., & Anderson, C. (1995). Living w ith chronic mental illness: understanding the role of work. Community Mental Health Journal 31(2), 163-176. Schenkel, L. S., Spaulding, W. D., DiLillo, D., & Silverstei n, S. M. (2005). Histories of childhood maltreatment in schizophrenia: Relationships with premorbid functioning, symptomatology, and cognitive deficits. Schizophrenia Research 76(2-3), 273-286. Schwoeri, L., & Sholevar, G. P. (1994). The fa mily transmission of depression. In G. P. Sholevar (Ed.), The transmission of depressi on in families and children (pp. 101122). Northvale, New Jersey: Jason Aronson, Inc. Secker, J., Gulliver, P., Peck, E., Robinson, J ., Bell, R., & Hughes, J. (2002). Evaluation of community mental health services: co mparison of a primary care mental health team and an extended day hospital service. Health and Social Care in the Community 9(6), 495-503. Segal, S. P., Silverman, C., & Temkin, T. (1995). Measuring empowerment in client-run self-help agencies. Community Mental Health Journal 31(3), 215-227. Segal, Z. V., Pearson, J. L., & Thase, M. E. (2003). Challenges in preventing relapse in major depression: Report of a National Institute of Mental Health Workshop on state of the science of relaps e prevention in major depression. Journal of Affective Disorders, 77 (2), 97-108. Sells, D. J., Stayner, D., & Davidson, L. (2004). Recovering the self in schizophrenia: An integrative review of qualitative studies. Psychiatric Quarterly, 75(1), 87-97. Services, U. S. D. o. H. a. H. (1999). Mental health: A report of the Surgeon General. Rockville, MD: U.S. Department of H ealth and Human Services, Substance Abuse and Mental Health Services Admi nsitration, Center for Mental Health Services, National Institutes of Health, National Institute of Mental Health.

PAGE 483

464 Shack, A. V., Averill, P. M., Kopecky, C., Kr ajewski, K., & Gummattira, P. (2004). Prior History of Physical and Sexual A buse Among the Psychiatric Inpatient Population: A Comparison of Males and Females. Psychiatric Quarterly 75(4), 343-359. Shahar, G., Chinman, M., Sells, D., & Davidson, L. (2003). An action model of socially disruptive behaviors committed by person with severe mental illness: The role of self-reported childhood abuse and suspiciousness-hostility. Psychiatry 66(1), 4252. Shahar, G., & Davidson, L. (2003). Depressive symptoms erode self -esteem in severe mental illness: A three-wave, cross-lagged study. Journal of Consulting and Clinical Psychology, 71(5), 890-900. Sheppard, L. C., & Teasdale, J. D. (2004) How does dysfunctional thinking decrease during recovery from major depression? Journal of Abnormal Psychology 113(1), 64-71. Shern, D. L., & Evans, M. E. (2005). The importa nce of research in mental health service delivery. In R. Drake, M. R. Merrens & D. Lynde (Eds.), Evidence-based mental health practice: A textbook (pp. 123-140). New York, NY: W.W. Norton & Company. Shumway, M., Saunders, T., Shern, D. L., Pines, E., Downs, A., Burbine, T., et al. (2003). Preferences for schizophrenia treatment outcomes among public policy makers, consumers, families, and providers. Psychiatric Services 54 (8), 11241128. Simon, N. M., Otto, M. W., Wisniewski, S. R., Fossey, M., Sagduyu, K., Frank, E., et al. (2004). Anxiety Disorder Comorbidity in Bipolar Disorder Pa tients: Data From the First 500 Participants in the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD). American Journal of Psychiatry 161(12), 22222229. Skarsater, I., Agren, H., & Dencker, K. ( 2001). Subjective lack of social support and presence of dependent stressful life ev ents characterize patients suffering from major depression compared with healthy volunteers. Journal of Psychiatric and Mental Health Nursing, 8(2), 107-114. Skarsater, I., Dencker, K., Bergbom, I., Haggstrom, L., & Fridlund, B. (2003). Women's conceptions of coping with major depr ession in daily life: A qualitative salutogenic approach. Issues in Mental Health Nursing 24, 419-439.

PAGE 484

465 Smith, T. A., Bellack, A. S., & Liberman, R. P. (1996). Social skills training for schizophrenia: Review and future directions. Clinical Psychology Review 16(7), 599-617. Smith, T. E., Hull, J. W., Huppert, J. D., & Silverstein, S. M. (2002). Recovery from psychosis in schizophrenia and schi zoaffective disorder: Symptoms and neurocognitive rate-limiters for the development of social behavior skills. Schizophrenia Research 55(3), 229-237. Smith, T. E., Hull, J. W., MacKain, S. J., Wa llace, C. J., Rattenni, L. A., Goodman, M., et al. (1996). Training hospitalized patient s with schizophrenia in community reintegration skills. Psychiatric Services 47(10), 1099-1103. Solomon, A., Haaga, d. A., & Arnow, B. A. (20 01). Is clinical depre ssion distinct from subthreshold depressive symptoms? A review of the continuity issue in depression research. Journal of Nervous and Mental Disease 189(8), 498-506. Solomon, D. A., Keller, M. B., Leon, A. C., Meu ller, T. I., Lavori, P. W., Shea, M. T., et al. (2000). Multiple recurrences of major depressive disorder. American Journal of Psychiatry 157(2), 229-233. Sorgaard, K. W., Hansson, L., Heikkila, J ., Vinding, H. R., Bjarnason, O., BegntssonTops, A., et al. (2001). Predictors of social relations in persons with schizophrenia living in the community: A Nordic multicentre study. Social Psychiatry and Psychiatric Epidemiology 36, 13-19. Sourander, A., Pihlakoski, L., Aromaa, M., Rautava, P., Helenius, H., & Sillanpaa, M. (2006). Early predictors of parentand self-reported perceived global psychological difficulties among adolescents A prospective cohort study from age 3 to age 15. Social Psychiatry and Psychiatric Epidemiology 41(3), 173-182. Sowers, W. E. (2005). Transforming System s of Care: The American Association of Community Psychiatrists Guidelines for Recovery Oriented Services Community Mental Health Systems. Community Mental Health Journal, 41(6), 757-774. Spaniol, L., Wewiorski, N., Gagne, C., & Anthony, W. A. (2002). The process of recovery from schizophrenia. International Review of Psychiatry, 14, 327-336. Spaniol, L., Wewlorski, N. J., Gagne, C., & Anthony, W. A. (2005). The process of recovery from schizophrenia. In L. Davidson, C. Harding & L. Spaniol (Eds.), Recovery from severe mental illnesses: Research evidence and implications for practice (Vol. 1, pp. 82-99). Boston: Center for Psychiatric Rehabilitation.

PAGE 485

466 Spindel, P., & Nugent, J. (2000). Pola r opposites: Empowerment philosophy and Assertive Community Treatment (ACT). Ethical Human Sciences and Services 2(2), 93-100. Srebnik, D., Livingston, J., Gordon, L., & King, D. (1995). Housing choice and community success for individuals with se rious and persistent mental illness. Community Mental Health Journal, 31(2), 139-152. Stangor, C., & Crandall, C. S. (2000). Threat and the social construction of stigma. In T. F. Heatherton, R. E. Kleck, M. R. Hebl & J. G. Hull (Eds.), The social psychology of stigma (pp. 62-87). New York: The Guilford Press. Stephens, J. H., Richard, P., & McHugh, P. R. (1997). Long-term follow-up of patients hospitalized for schi zophrenia, 1913-1940. Journal of Nervous and Mental Disease, 185 (12), 715-721. Stroul, B. (1989). Community support systems fo r persons with long-term mental illness: A conceptual framework. Psychosocial Rehabilitation Journal 12(3), 9-26. Strous, R. D., Ratner, Y., Gibel, A., Poniz ovsky, A., & Ritsner, M. (2005). Longitudinal assessment of coping abilities at exacerba tion and stabilizati on in schizophrenia. Comprehensive Psychiatry 46(3), 167-175. Stuart, H., & Arboleda-Florez, J. (2001). Community atti tudes toward people with schizophrenia. Canadian Journal of Psychiatry 46, 245-252. Svedberg, B., Backenroth-Ohsako, G., & Lut zen, K. (2003). On the path to recovery: Patients' experiences of treatment with long-acting injections of antipsychotic medication. International Journal of Mental Health Nursing 12(2), 110-118. Tait, L., Birchwood, M., & Trower, P. (2003). Predicting engagement with services for psychosis: Insight, symptoms and recovery style. British Journal of Psychiatry 182(2), 123-129. Thoits, P. A. (1982). Life stress, social support, and psychologi cal vulnerability: Epidemiological considerations. Journal of Community Psychology, 10, 341-362. Thoits, P. A. (1995). Stress, coping, and social support processes: Where are we? What next? Journal of Health and Social Behavior 35(Special issue, 53-79). Thompson, V. L., Noel, J. G., & Campbell, J. (2004). Stigmatization, discrimination, and mental health: The impact of multiple identity status. American Journal of Orthopsychiatry, 74(4), 529-544.

PAGE 486

467 Torgalsboen, A.-K. (2005). What is recovery in schizophrenia? In L. Davidson, C. Harding & L. Spaniol (Eds.), Recovery from severe ment al illnesses: Research evidence and implica tions for practice (Vol. 1, pp. 302-315). Boston: Center for Psychiatric Rehabilitation. Ustun, T. B. (1999). The global bu rden of mental disorders. American Journal of Public Health 89(9), 1315-1318. Ustun, T. B., Ayuso-Mateos, J. L., Chatterji, S., Mathers, C., & Murray, C. J. L. (2004). Global burden of depressive disorders in the year 2000. British Journal of Psychiatry, 184(5), 386-392. Valiakalayil, A., Paulson, L. A., & Tibbo, P. (2004). Burden in adolescent children of parents with schizophrenia: Th e Edmonton High-Risk Project. Social Psychiatry and Psychiatric Epidemiology, 39(7), 528-535. van Os, J., & Sham, P. (2003). Gene-envir onment correlation a nd interaction in schizophrenia. In R. M. Murray, P. B. Jones, E. Susser, J. van Os & M. Cannon (Eds.), The epidemiology of schizophrenia (pp. 235-253). Cambridge: Cambridge University Press. van Os, J., Verdoux, H., Maurice-Tison, S., Gay, B., Liraud, F., Salamon, R., et al. (1999). Self-reported psychos is-like symptoms and the continuum of psychosis. Social Psychiatry and Psychiatric Epidemiology 34, 459-463. Vassend, O., & Skrondal, A. (1999). The probl em of structural indeterminacy in multidimensional symptom report instruments. The case of SCL-90-R. Behaviour Research and Therapy 37, 685-701. Vaughan, G., & Hansen, C. (2004). 'Like minds, like mine': A New Zealand project to counter the stigma and discrimination associated with mental illness. Australasian Psychiatry, 12(2), 113-117. Velligan, D. I., & Miller, A. L. (1999). C ognitive dysfunction in schizophrenia and its importance to outcome. The place of atypical antipsychotics in treatment. Journal of Clinical Psychiatry, 60, 25-29. Verdoux, H., & van Os, J. (2002). Psychotic sy mptoms in non-clinical populations and the continuum of psychosis. Schizophrenia Research 54, 59-65. Vourdas, A., Pipe, R., Corrigall, R., & Frangou, S. (2003). Increased developmental deviance and premorbid dysfunction in early onset schizophrenia. Schizophrenia Research 62 (1-2), 13-22.

PAGE 487

468 Wahl, O. F. (1999). Telling is risky business: Mental health consume rs confront stigma New Brunswick, NJ: Rutgers University Press. Walby, G. (2003a). A qualitative inquiry into the subjective experience of the mentally ill Tampa, FL: Department of Community a nd Family Health, College of Public Health, University of South Florida. Walby, G. (2003b). Reflexivity, reintegration and the se verely mentally ill: A qualitative analysis. Tampa, FL: Department of Community and Family Health, College of Public Health, University of South Florida. Walby, G. (2005). Report of qualitative findings from phase I of a multi-phase evaluation of mental health services provided to individuals with seve re mental illness New Port Richey, FL. Walby, G., & Mae, C. (2003). A qualitative inquiry of mental illness: Social networks, employment and role stress Paper presented at the American Public Health Association Annual Mee ting, San Francisco, CA. Walsh, J. (1994). The social networks of seri ously mentally ill persons receiving case management services. Journal of Case Management 3(1), 27-35. Warner, R., & Mandiberg, J. (2004). Changing the environment of sc hizophrenia at the community level. Australasian Psychiatry 11, s58-s64. Wasylenki, D. A. (1994). The cost of schizophrenia. Canadian Journal of Psychiatry 39(9 Suppl 2), S65-69. Watson, P. W. B., Garety, P. A., Weinman, J., Dunn, G., Bebbington, P. E., Fowler, D., et al. (2006). Emotional dysfunction in schi zophrenia spectrum psychosis: the role of illness perceptions. Psychological Medicine 36 (6), 761-770. Watts, D. J. (1999). Networks, dynamics, and the small-world phenomenon. American Journal of Sociology 105 (2), 493-527. Weisman, A. G., Nuechterlein, K. H., Gold stein, M. J., & Snyder, K. S. (1998). Expressed Emotion, Attributions, and Schizophrenia Symptom Dimensions. Journal of Abnormal Psychology 107 (2), 355-359. Wexler, B. E., Lyons, L., Lyons, H., & Mazure C. M. (1997). Physical and sexual abuse during childhood and development of psychiatric illnesses during adulthood. Journal of Nervous & Mental Disease 185(8), 522-524.

PAGE 488

469 White, V. E. (2002). Developing counseling objectives and empowering clients: A strength-based intervention. Journal of Mental Health Counseling 24 (3), 270279. White, W., Boyle, M. G., & Loveland, D. (2005). Recovery from addiction and from mental illness: Shared and contrasting lessons. In R. O. Ralph & P. Corrigan (Eds.), Recovery in mental illness: Br oadening our understanding of wellness (pp. 233-258). Washington, DC: American Psychological Association. Wilder-Willis, K. E., Shear, P. K., Steffen, J. J., & Borkin, J. (2002). The relationship between cognitive dysfunction and coping abilities in schizophrenia. Schizophrenia Research 55, 259-267. Williams, C. C., & Collins, A. A. (2002). Th e social construction of disability in schizophrenia. Qualitative Health Research 12(3), 297-309. Williams, P., Barclay, L., & Schmied, V. (2004) Defining social support in context: A necessary step in improving resear ch, intervention, and practice. Qualitative Health Research 14(7), 942-960. Wilson, S., Washington, L. A., Engel, J. M., Ciol, M. A., & Jensen, M. P. (2006). Perceived Social Support, Psychological Adjustment, and Functional Ability in Youths With Physical Disabilities. Rehabilitation Psychology 51(4), 322-330. Wirshing, D. A., Wirshing, W. C., Marder, S. R., Liberman, R. P., & Mintz, J. (1998). Informed consent: Assessment and comprehension. American Journal of Psychiatry, 155(11), 1508-1511. Wood, J. V., Heimpel, S. A., Newby-Clark, I. R., & Ross, M. (2005). Snatching Defeat From the Jaws of Victory: Self-Est eem Differences in the Experience and Anticipation of Success. Journal of Personality and Social Psychology 89 (5), 764-780. Wowra, S. A., & McCarter, R. (1999). Valid ation of the empowerm ent scale with an outpatient mental health population. Psychiatric Services 50 (7), 959-961. Yanos, P. T., Primavera, L. H., & Knight, E. L. (2001). Consumer-Run Service Participation, Recovery of Social Functioning, and the Mediating Role of Psychological Factors. Psychiatric Services 52(4), 493-500. Yanos, P. T., Rosenfield, S., & Horwitz, A. V. (2001). Negative and supportive social interactions and quality of life among persons diagnosed with severe mental illness. Community Mental Health Journal 37(5), 405-419.

PAGE 489

470 Young, A. S., Chinman, M., Forquer, S. L., Kni ght, E. L., Vogel, H., Miller, A., et al. (2005). Use of a Consumer-Led Interven tion to Improve Provider Competencies. Psychiatric Services 56 (8), 967-975. Young, A. S., Forquer, S. L., Tran, A., Starz ynski, M., & Shatkin, J. (2000). Identifying clinical competencies that support rehabilitation and em powerment in individuals with severe mental illness. Journal of Behavioral Health Services and Research, 27(3), 321-333. Young, S. L., & Ensing, D. S. (1999). Exploring recovery from the perspective of people with psychiatric disabilities. Psychiatric Rehabilitation Journal 22(3), 219-231. Yurkovich, E., Smyer, T., & Dean, L. (1999) Maintaining health: Proactive clientoriented community day treatment centr es for the chronic mentally ill. Journal of Psychiatric and Mental Health Nursing 6(1), 61-69. Zar, J. H. (1999). Biostatistical Analysis (4th ed.). Upper Saddle River, New Jersey: Simon & Schuster. Ziguras, S. J., Klimidis, S., Lambert, T. J., & Jackson, A. C. (2001). Determinants of antipsychotic medication compliance in a multicultural population. Community Mental Health Journal 37(3), 273-283. Zimmerman, M. A. (2000). Empowerment theo ry: Psychological, organizational, and community levels of analysis. In J. Rappaport & E. Seidman (Eds.), Handbook of community psychology (pp. 43-64). New York: Kluwer Academic / Plenum Publishers. Zimmerman, M. A., Israel, B. A., & Schulz, A. J. (1992). Further explorations in empowerment theory: An empirical an alysis of psychological empowerment. American Journal of Community Psychology 20 (6), 707-727. Zimmerman, M. A., & Warschausky, S. (1998). Empowerment theory for rehabilitation research: Conceptual an d methodological issues. Rehabilitation Psychology, 43(1), 3-16. Zygmunt, A., Olfson, M., Boyer, C. A., & Mech anic, D. (2002). Interventions to improve medication adherence in schizophrenia. American Journal of Psychiatry 159(10), 1653-1664.

PAGE 490

471 Appendices

PAGE 491

472 Appendix A Measurement Tools Recovery Expectancy Checklist (REC) Recovery Assessment Scale (RAS) Personal Vision of Recovery Questionnaire (PVRQ) Symptom Checklist 90-Revised (SCL-90-R) (Copyrighted, unable to duplicate) Support and Community Connection Survey (SCCS) Empowerment Survey (ES) Internalized Stigma of Me ntal Illness Scale (ISMI) Service Satisfaction Questionnaire (SSQ) History and Demographic Form (HDF) Chart Abstraction Tool

PAGE 492

473 REC Age in Years ____________________ Gender _________________________ Test Date ____________________________ MM/DD/YY Please answer the following four (4) questions by indicating the res ponse that best fits your personal beliefs and choices on the s ubject of recovery from mental illness 1. Do you think that recovery from me ntal illness will ever be possible for you? YES NO 2. To what degree do you th ink you will eventually recover? __ 1 _____________ 2 ______________ 3 _______________ 4 _______________ 5 ____ None A Little Somewhat A Lot Completely 3. What will recovery mean to you? Please select all that apply a. I will be without mental illness a. ________ b. I will be symptom free b. ________ c. I will be able to function, even with some symptoms c. ________ d. I will be able to stay out of the hospital/crisis unit d. ________ e. I will be off psychiatric medications e. ________ f. I will be able to work full time f. ________ g. I will be able to work part time g. ________ 3. h. In a few brief sentences, please describe in your own words what recovery means to you _____________________________________________________________________ _____________________________________________________________________ _____________________________________________________________________ _____________________________________________________________________ _____________________________________________________________________

PAGE 493

474 4. Which of the following are important for your recovery? Please check all that apply. a. Contact with my primary provider/case manager/therapist a. ________ b. Contact with my psychiatri st or nurse practitioner b. ________ c. Contact with other individuals with mental illness c. ________ d. Psychiatric medication d. ________ e. Parents e. ________ f. Siblings (brothers and sisters) f. ________ g. Other family g. ________ h. Spouse or significant other h. ________ i. Friends i. _________ j. School/Attending class/Education j. _________ k. Learning about or study ing mental disorders k. ________ l. Work (paid) l. ________ m. Work (volunteer) m. ________ n. Art or other creative expression n. ________ o. Reading o. ________ p. Quiet time/time alone p. ________ q. Exercise or following an exercise program q. ________ r. Eating a healthy diet r. ________ s. Spirituality (NOT attendi ng services, religious functions) s. ________ t. Religious affiliation (e.g. attending services) t. ________ u. Meditation, relaxation or other similar experience u. ________ v. Travel v. ________ w. Being hopeful w. ________ x. Having choices in different aspects of life x. ________ y1. Other y1. __________________________________________________ y2. Other y2. __________________________________________________ y3. Other y3. __________________________________________________

PAGE 494

475 RAS Age in Years ____________________ Gender _________________________ Test Date ____________________________ MM/DD/YY I am going to read a list of statements that describe how people sometimes feel about themselves and their lives. Pl ease listen carefully to each one and indicate the response that best describes the exte nt to which you agree or di sagree with the statement. For each of these statements, please indicate whether you strongly disagree (1), disagree (2), not sure (3), agree (4), or strongly agree (5) with these statements. Strongly Disagree Disagree Not Sure Agree Strongly Agree 1 I have a desire to succeed 1 2 3 4 5 2 I have my own plan for how to stay or become well 1 2 3 4 5 3 I have goals in life that I want to reach 1 2 3 4 5 4 I believe I can meet my current personal goals 1 2 3 4 5 5 I have a purpose in life 1 2 3 4 5 6 Even when I dont care about myself, other people do 1 2 3 4 5 7 I understand how to control the symptoms of my mental illness 1 2 3 4 5 8 I can handle it if I get sick again 1 2 3 4 5 9 I can identify what triggers the symptoms of my mental illness 1 2 3 4 5 10 I can help myself become better 1 2 3 4 5 11 Fear doesnt stop me from living the way I want to 1 2 3 4 5

PAGE 495

476 Strongly Disagree Disagree Not Sure Agree Strongly Agree 12 I know that there are mental health services that do help me 1 2 3 4 5 13 There are things that I can do that help me deal with unwanted symptoms 1 2 3 4 5 14 I can handle what happens in my life 1 2 3 4 5 15 I like myself 1 2 3 4 5 16 If people really knew me, they would like me 1 2 3 4 5 17 I am a better person than before my experience with mental illness 1 2 3 4 5 18 Although my symptoms may get worse, I know I can handle it 1 2 3 4 5 19 If I keep trying, I will continue to get better 1 2 3 4 5 20 I have an idea of who I want to become 1 2 3 4 5 21 Things happen for a reason 1 2 3 4 5 22 Something good will eventually happen 1 2 3 4 5 23 I am the person most responsible for my own improvement 1 2 3 4 5 24 Im hopeful about my future 1 2 3 4 5 25 I continue to have new interests 1 2 3 4 5 26 It is important to have fun 1 2 3 4 5 27 Coping with my mental illness is no longer the main focus of my life 1 2 3 4 5

PAGE 496

477 Strongly Disagree Disagree Not Sure Agree Strongly Agree 28 My symptoms interfere less and less with my life 1 2 3 4 5 29 My symptoms seem to be a problem for shorter periods of time each time they occur 1 2 3 4 5 30 I know when to ask for help 1 2 3 4 5 31 I am willing to ask for help 1 2 3 4 5 32 I ask for help when I need it 1 2 3 4 5 33 Being able to work is important to me 1 2 3 4 5 34 I know what helps me get better 1 2 3 4 5 35 I can learn from my mistakes 1 2 3 4 5 36 I can handle stress 1 2 3 4 5 37 I have people I can count on 1 2 3 4 5 38 I can identify the early warning signs of becoming sick 1 2 3 4 5 39 Even when I dont believe in myself, other people do 1 2 3 4 5 40 It is important to have a variety of friends 1 2 3 4 5 41 It is important to have healthy habits 1 2 3 4 5

PAGE 497

478PVRQ Age in Years ____________________ Gender _________________________ Test Date ____________________________ MM/DD/YY We are interested in your beliefs about your own recovery from mental illness. By recovery we mean the way you have learned to cope with your mental illness and go forward with your life. Please answ er all the questions, whether or not you consider yourself to be in recovery right now. Please read each of the following statements. Circle the ratin g that most closely matches your opinion: S trongly Agree (5); Agree (4): Neither Agree nor Disagree (3); Disagree (2); or Strongly Disagree (1) 1. Spirituality is a part of my recovery 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 2. I am responsible for my own recovery 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 3. People who expect very little of me interfere w ith my recovery 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 4. Recovery means becomi ng satisfied with my life 5 4 3 2 1 Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree

PAGE 498

4795. Hope is important for my recovery 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 6. Being diagnosed correctly is necessary for my recovery 5 4 3 2 1 Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 7. Family support is im portant for m y recovery 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 8. Sticking up for clients rights is a part of my recovery 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 9. Having something meaningful to do is important for my recovery 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 10. Helping others is part of my recovery 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 11. Asking for help is a part of my recovery 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree

PAGE 499

48012. I know people who are recoveri ng from problems similar to mine 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 13. Recovery means my symptoms will be easier to control 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 14. Recovery means I will be free of symptoms 5 4 3 2 1 Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 15. Self-help groups are im portant to my recovery 5 4 3 2 1 Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 16. Recovery means getting more control of my life 5 4 3 2 1 Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 17. The cause of mental illness is not important for my recovery 5 4 3 2 1 Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 18. At times, treatment against my wi shes is necessary for my recovery. For example, involuntary hospitalization, forced medication, or community probate 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree

PAGE 500

48119. Support from a special person, such as a spouse or partner, is important for my recovery 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 20. Recovery means I will not be mentally ill anymore 5 4 3 2 1 Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 21. I am convinced that medication can help me to recover 5 4 3 2 1 Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 22. Side effects from my medication make it harder for me to recover 5 4 3 2 1 Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 23. Recovery involves finding new meaning in my life 5 4 3 2 1 Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree 24. Support from mental health profe ssionals is important for my recovery 5 4 3 2 1____ Strongly Agree Ne ither Agree Disagree Strongly Agree nor Disagree Disagree

PAGE 501

482 SCCS Age in Years ____________________ Gender _________________________ Test Date ____________________________ MM/DD/YY Instructions : This survey is designed to assess how a person feels they fit in their community and among their family and friends Please read each question carefully, and circle the number that best describes how much that statem ent describes you. Please take your time and answer all questions. Strongly Disagree Disagree Mildly Disagree Mildly Agree Agree Strongly Agree 1 My partner (or best friend) and I support each other equally 1 2 3 4 5 6 2 My friends help me feel good about myself 1 2 3 4 5 6 3 My provider helps me with my emotional stability 1 2 3 4 5 6 4 I have a close relationship with my family 1 2 3 4 5 6 5 If I make a mistake, my friends point it out so I will correct it 1 2 3 4 5 6 6 There are only a few people I completely trust 1 2 3 4 5 6 7 My family meets many of my needs 1 2 3 4 5 6 8 Most of my friends help me in whatever way I need 1 2 3 4 5 6 9 People are trustworthy, but society is not 1 2 3 4 5 6 10 If I am making a mistake, my provider will always point it out 1 2 3 4 5 6 11 My intimate partner (or best friend) helps me in many ways 1 2 3 4 5 6 12 I can always turn to my partner (or best friend) for advice if I am confused 1 2 3 4 5 6 13 My partner (or best friend) praises me and cheers for me when I accomplish something 1 2 3 4 5 6

PAGE 502

483 Strongly Disagree Disagree Mildly Disagree Mildly Agree Agree Strongly Agree 14 Most people will do the wrong thing if they know they will not be caught 1 2 3 4 5 6 15 When I am sad or feeling blue, I can always turn to my family 1 2 3 4 5 6 16 I wish my partner (or best friend) would try harder not to hurt my feelings 1 2 3 4 5 6 17 When I do something wrong, my partner (or best friend) points it out to me 1 2 3 4 5 6 18 Helping me feel good is what my partner or best friend does best 1 2 3 4 5 6 19 I volunteer my time to organizations when I can 1 2 3 4 5 6 20 Sometimes I feel that my provider does not take the time to hear me 1 2 3 4 5 6 21 When my back is in a corner, I can count on my partner (or best friend) to support me 1 2 3 4 5 6 22 My friends spend time with my intimate partner (or best friend) and I 1 2 3 4 5 6 23 My closest relationships usually last for two years or more 1 2 3 4 5 6 24 If I need money or help with a bill, my family almost always gives it to me 1 2 3 4 5 6 25 Most of the time people are looking out for themselves, not trying to be helpful to others 1 2 3 4 5 6 26 The more I give to the community, the more I want to 1 2 3 4 5 6 27 My friends know who my provider is 1 2 3 4 5 6 28 I keep myself informed on community issues 1 2 3 4 5 6 29 If I need to know something, I ask my provider 1 2 3 4 5 6 30 If I need help fixing or making something, my family helps me 1 2 3 4 5 6 31 I am one of the first to contribute to community projects or concerns 1 2 3 4 5 6 32 I feel I give back to the community for what I take 1 2 3 4 5 6 33 If I need to know something my family usually has the answer 1 2 3 4 5 6 34 My partner (or best friend) gives me things or helps me do things to make my life easier 1 2 3 4 5 6

PAGE 503

484 ES Age in Years ____________________ Gender _________________________ Test Date ____________________________ MM/DD/YY Please respond to each item. Read each que stion carefully, and circ le the response that best describes your level of agreement with the statement. 1 I generally accomplish what I set out to do Strongly Agree Agree Disagree Strongly Disagree 2 I feel powerless most of the time Strongly Agree Agree Disagree Strongly Disagree 3 People have a right to make their own decisions, even if they are bad ones Strongly Agree Agree Disagree Strongly Disagree 4 People are limited only by what they think possible Strongly Agree Agree Disagree Strongly Disagree 5 Getting angry about something is often the first step toward changing it Strongly Agree Agree Disagree Strongly Disagree 6 I have a positive attitude about myself Strongly Agree Agree Disagree Strongly Disagree 7 Making waves never gets you anywhere Strongly Agree Agree Disagree Strongly Disagree 8 People should try to live their lives the way they want to Strongly Agree Agree Disagree Strongly Disagree 9 I can pretty much determine what will happen in my life Strongly Agree Agree Disagree Strongly Disagree 10 People have no right to get angry just because they dont like something Strongly Agree Agree Disagree Strongly Disagree 11 When I make plans, I am almost certain to make them work Strongly Agree Agree Disagree Strongly Disagree 12 You cant fight city hall Strongly Agree Agree Disagree Strongly Disagree

PAGE 504

485 13 People working together can have an effect on their community Strongly Agree Agree Disagree Strongly Disagree 14 I am generally optimistic about the future Strongly Agree Agree Disagree Strongly Disagree 15 Getting angry about something never helps Strongly Agree Agree Disagree Strongly Disagree 16 I am usually confident about the decisions I make Strongly Agree Agree Disagree Strongly Disagree 17 When I am unsure about something, I usually go along with the group Strongly Agree Agree Disagree Strongly Disagree 18 People have more power if they join together as a group Strongly Agree Agree Disagree Strongly Disagree 19 Very often a problem can be solved by taking action Strongly Agree Agree Disagree Strongly Disagree 20 I am often able to overcome barriers Strongly Agree Agree Disagree Strongly Disagree 21 Experts are in the best position to decide what people should do or learn Strongly Agree Agree Disagree Strongly Disagree 22 Working with others in my community can help to change things for the better Strongly Agree Agree Disagree Strongly Disagree 23 I feel I am a person of worth, at least on an equal basis with others Strongly Agree Agree Disagree Strongly Disagree 24 Most of the misfortunes in my life were due to bad luck Strongly Agree Agree Disagree Strongly Disagree 25 I see myself as a capable person Strongly Agree Agree Disagree Strongly Disagree 26 Usually, I feel alone Strongly Agree Agree Disagree Strongly Disagree 27 I am able to do things as well as most other people Strongly Agree Agree Disagree Strongly Disagree 28 I feel I have a number of good qualities Strongly Agree Agree Disagree Strongly Disagree

PAGE 505

486 ISMI Age in Years ____________________ Gender _________________________ Test Date ____________________________ MM/DD/YY Please respond to each item. We are going to us e the term mental illn ess in the rest of this questionnaire, but please th ink of it as whatever you feel is the best term for it. For each question, please mark wh ether you strongly disagree ( 1) disagree ( 2) agree ( 3) or strongly agree ( 4) 1 In general, I am able to live life the way I want to Strongly Disagree Disagree Agree Strongly Agree 2 I feel out of place in the world because I have a mental illness Strongly Disagree Disagree Agree Strongly Agree 3 Stereotypes about the mentally ill apply to me Strongly Disagree Disagree Agree Strongly Agree 4 People discriminate against me because I have a mental illness Strongly Disagree Disagree Agree Strongly Agree 5 I dont talk about myself much because I dont want to burden others with my mental illness Strongly Disagree Disagree Agree Strongly Agree 6 Having a mental illness has spoiled my life Strongly Disagree Disagree Agree Strongly Agree 7 People can tell that I have a mental illness by the way I look Strongly Disagree Disagree Agree Strongly Agree 8 Others think that I cant achieve much in life because I have a mental illness Strongly Disagree Disagree Agree Strongly Agree 9 I dont socialize as much as I used to because my mental illness might make me look or behave weird Strongly Disagree Disagree Agree Strongly Agree 10 I feel comfortable being seen in public with an obviously mentally ill person Strongly Disagree Disagree Agree Strongly Agree 11 People without mental illness could not possibly understand me Strongly Disagree Disagree Agree Strongly Agree

PAGE 506

487 12 Mentally ill people tend to be violent Strongly Disagree Disagree Agree Strongly Agree 13 People ignore me or take me less seriously just because I have a mental illness Strongly Disagree Disagree Agree Strongly Agree 14 Negative stereotypes about mental illness keep me isolated from the normal world Strongly Disagree Disagree Agree Strongly Agree 15 I can have a good, fulfilling life, despite my mental illness Strongly Disagree Disagree Agree Strongly Agree 16 I am embarrassed or ashamed that I have a mental illness Strongly Disagree Disagree Agree Strongly Agree 17 Because I have a mental illness, I need others to make most decisions for me Strongly Disagree Disagree Agree Strongly Agree 18 People often patronize me, or treat me like a child, just because I have a mental illness Strongly Disagree Disagree Agree Strongly Agree 19 I stay away from social situations in order to protect my family or friends from embarrassment Strongly Disagree Disagree Agree Strongly Agree 20 People with mental illness make important contributions to society Strongly Disagree Disagree Agree Strongly Agree 21 I am disappointed in myself for having a mental illness Strongly Disagree Disagree Agree Strongly Agree 22 People with mental illness cannot live a good, rewarding life Strongly Disagree Disagree Agree Strongly Agree 23 Nobody would be interested in getting close to me because I have a mental illness Strongly Disagree Disagree Agree Strongly Agree 24 Being around people who dont have a mental illness makes me feel out of place or inadequate Strongly Disagree Disagree Agree Strongly Agree 25 Living with mental illness has made me a tough survivor Strongly Disagree Disagree Agree Strongly Agree 26 I feel inferior to others who dont have a mental illness Strongly Disagree Disagree Agree Strongly Agree

PAGE 507

488 27 Mentally ill people shouldnt get married Strongly Disagree Disagree Agree Strongly Agree 28 I avoid getting close to people who dont have a mental illness to avoid rejection Strongly Disagree Disagree Agree Strongly Agree 29 I cant contribute anything to society because I have a mental illness Strongly Disagree Disagree Agree Strongly Agree

PAGE 508

489 SSQ Age in Years ____________________ Gender _________________________ Test Date ____________________________ MM/DD/YY Note: For Level of Satisfaction : 6 = Very Satisfied; 5= Satisfied; 4= Somewhat Satisfied; 3 = Somewhat Dissatisfied; 2=Dissatisfied; 1=Very Dissatisfied Service Type Receiving ? Hours/ Month Level of Satisfaction Contributes to Recovery? Outpatient Medical/Medication Y N 6 5 4 3 2 1 Y N Outpatient therapy (individual) Y N 6 5 4 3 2 1 Y N Outpatient therapy (relationship/couple) Y N 6 5 4 3 2 1 Y N Case management (general) Y N 6 5 4 3 2 1 Y N Case management (intensive) Y N 6 5 4 3 2 1 Y N Recovery Program Y N 6 5 4 3 2 1 Y N Supported living (financial only) Y N 6 5 4 3 2 1 Y N Supported living (agency/low supervision) Y N 6 5 4 3 2 1 Y N Supported living (agency/supervised) Y N 6 5 4 3 2 1 Y N Supported employment Y N 6 5 4 3 2 1 Y N Dialectical Behavior Therapy (DBT) Y N 6 5 4 3 2 1 Y N Forensic Services Y N 6 5 4 3 2 1 Y N PREP day program Y N 6 5 4 3 2 1 Y N Drop-In Center Y N 6 5 4 3 2 1 Y N Totals/Averages

PAGE 509

490 HDF Age in Years ____________________ Gender _________________________ Test Date ____________________________ MM/DD/YY 1. Education : Please check which res ponse most applies to you 1. Less than High School _______ 2. High School Graduate _______ 3. 1-2 years of College or Associates Degree _______ 4. More than 2 years of college or Bachelors Degree _______ 5. Graduate School _______ If a degree(s) was/were earned, what degree was obtained and what was the degree in? _______________________________________________________________ _______________________________________________________________ 2.A Income: Please indicate your approximate personal annual income $0.0 $10,000 _______ $10,001 $15,000 _______ $15,001 $20,000 _______ $20,001 $25,000 _______ $25,001 $30,000 _______ $30,001 $40,000 _______ $40,001 $50,000 _______ $50,001 $75,000 _______ $75,001 more _______ 2.B Income : Please indicate your approximate combined annual income if different than personal income (if more than yourself contributes financially to your household). (circle if N/A) $0.0 $10,000 _______ $10,001 $15,000 _______ $15,001 $20,000 _______ $20,001 $25,000 _______

PAGE 510

491 $25,001 $30,000 _______ $30,001 $40,000 _______ $40,001 $50,000 _______ $50,001 $75,000 _______ $75,001 more _______ 3. Employment: 3.A How many years were you employed full time in your life? _______________ 3.B How many years were you employed part time in your life? _______________ 3.C Are you currently working? (YES / NO) _____________ (if no, skip to question 4) 3.D If you are working how would you de scribe your employment (please check one)? 1. Fulltime paid employment _______ (# of hours/week _______) 2. Part time paid employment _______ (# of hours/week _______) 3. Volunteer/non-paid employment _______ (# of hours/week _______) 3.E What is your job? ____________________________________________________ 4. Diagnosis What is your psychiatric diagnosis? _________________________________________ 5. Hospitalization 5.A How many times have you been in a hospi tal or crisis unit because of mental health reasons over your life time? ___________________________ 5.B When was the last time you were in th e hospital or crisis unit (indicate in number of years or months back)? ________________________________ 5.C How old were you the first time you we re in a hospital or crisis unit for mental health reasons? ______________________________

PAGE 511

492 6. Age of Onset 6.A How old were you when you were first officially diagnosed? ____________ 6.B How old were you when you first real ized that you had a mental health problem? _________ 7. Family History of Mental Illness 7.A Do you have any family members th at you know were officially diagnosed with a mental health problem? For each of the following, indicate yes or no and the diagnosis if know. 1. Father YES NO Diagnosis ___________________ 2. Mother YES NO Diagnosis ___________________ 3. Brother YES NO Diagnosis ___________________ 4. Sister YES NO Diagnosis ___________________ 5. Aunt or Uncle YES NO Diagnosis ___________________ 6. Grandparent YES NO Diagnosis ___________________ 7. Other (e.g. cousin) YES NO Diagnosis ___________________ 7.B Do you have any family members th at you suspect had a mental illness but were not officially diagnosed as far as you know? For each of the following, indicate yes or no and what they were showing that made you believe this. 1. Father YE S NO Symptom ____________________ 2. Mother YES NO Symptom ____________________ 3. Brother YES NO Symptom ____________________ 4. Sister YES NO Symptom ____________________ 5. Aunt or Uncle YES NO Symptom ____________________ 6. Grandparent YES NO Symptom ____________________ 7. Other (e.g. cousin) YES NO Symptom ____________________ 8. Medication Please indicate if you are currently pres cribed and taking any of the following (check all that apply) 1. Anti-Depressant Medication YES NO 2. Anti-Psychotic Medication YES NO 3. Anti-Manic Medication YES NO 4. Anti-Anxiety Medication YES NO 5. Other Psychiatric Medication YES NO

PAGE 512

493 9. Substance Use History Please indicate if you have ever been diagnosed with an alcohol or substance dependence diagnosis (e.g. alcoholism, alcohol dependence, cocaine addiction, etc.) YES NO If yes, what was the addictive substance? ____________________________________________________________ 10. History of Abuse It is recognized that this is an area of sensitivity and privacy. Please do not answer these questions if you feel that it is too stressful for you. If you do answer the questions, please understand that we are not seeking deta ils but simply an acknowledgement of whether you have e xperienced any of the following. Please note for the purposes of this study, child sexual abuse is defined as touching with or without penetrati on an individual younger than age 18, by someone at least 5 years older than th e individual, for the purposes of sexual gratification. 10.A Child Sexual Abuse YES NO More than one time? YES NO 10.B Child Physical Abuse YES NO More than one time? YES NO 10.C Adult Sexual Assault YES NO More than one time? YES NO 10.D Adult Physical Assault YES NO More than one time? YES NO

PAGE 513

494 Chart Abstraction Tool Participant # ____________ Age ________________ Gender ________ Date __________ Education : List greatest level of education comple ted noted in record (e.g. high school graduate, college graduate (BA or BS), associates degree, etc.): ___________________________________________________ Income : Indicate annual income in thousands. If monthly income is given multiply x12. ____________________________ Employment Is the participant currently employed? YES NO Fulltime (hours/week) _____________ Part time (hours/week) _____________ Volunteer (hours/week) _____________ Total number of years employed (if available) _____________________ Age of Onset Indicate how old respondent was at first diagnosis (if available) ___________________ Indicate how old respondent was when fi rst symptoms are noted (if available) ___________________ Diagnostic History Axis 1 Diagnoses (listed in order) ( psychiatric and substance related ) go back two years (start at most recent) Diagnosis 1: Name ______________________ DSM-IV# _______________ Diagnosis 2: Name ______________________ DSM-IV# _______________ Diagnosis 3: Name ______________________ DSM-IV# _______________ Diagnosis 4: Name ______________________ DSM-IV# _______________

PAGE 514

495 Axis 2 Diagnoses (listed in order) (N umber and Name) go back two years (start as most recent) Diagnosis 1: Name ______________________ DSM-IV# _______________ Diagnosis 2: Name ______________________ DSM-IV# _______________ Diagnosis 3: Name ______________________ DSM-IV# _______________ Axis 3 Diagnoses (listed in order) (Name) go back two years (start at most recent) Medical Diagnosis/Problem 1: _________________________________________ Medical Diagnosis/Problem 2: _________________________________________ Medical Diagnosis/Problem 3: _________________________________________ Axis 5: Most recent Global Assessment of Functioning (GAF) score ___________________ Substance Use If a substance use problem is indicated but not listed as a diagnosis indicate here YES NO What substance or subs tances were involved? _________________________________________________________________ Hospitalization Dates of hospitalizations or crisis unit adm issions in the last two (2) years start with most recent. Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU Date Admitted ___________ Length of Stay __________ Hospital CSU

PAGE 515

496 Medication Indicate medications prescribed for the last two years start with most recent. Type of Medication Date Name Psychosis De press Manic Anxiety Other Child Abuse Evidence of child abuse (child sexual abuse (CSA), child physical abuse (PSA) Single Incident CSA YES NO Multiple Incident CSA YES NO Single Incident PSA YES NO Multiple Incident PSA YES NO Adult Assault/Abuse Evidence of adult sexual assault (S A) or adult physical assault (PA) Single Incident Adult SA YES NO Multiple Incident Adult SA YES NO Single Incident Adult PA YES NO Multiple Incident Adult PA YES NO

PAGE 516

497 Familial Mental Illness Indication of DIAGNOSE D mental illness for 1. Father YES NO Diagnosis _____________________ 2. Mother YES NO Diagnosis _____________________ 3. Brother YES NO Diagnosis _____________________ 4. Sister YES NO Diagnosis _____________________ 5. Aunt or Uncle YES NO Diagnosis _____________________ 6. Grandparent YES NO Diagnosis _____________________ 7. Other (e.g. cousin) YES NO Diagnosis _____________________ Indication of SYMPTOMATIC individual without official diagnosis 1. Father YE S NO Symptom _____________________ 2. Mother YES NO Symptom _____________________ 3. Brother YE S NO Symptom _____________________ 4. Sister YES NO Symptom _____________________ 5. Aunt or Uncle YES NO Symptom _____________________ 6. Grandparent YES NO Symptom _____________________ 7. Other (e.g. cousin) YES NO Symptom _____________________ Services Indicate if the respondent is curren tly utilizing the following services Outpatient Medical (psychiatrist/nurse practitioner) YES NO Outpatient Therapy (Individual) YES NO Outpatient Therapy (Couple/Family) YES NO Case Management (Standard) YES NO Case Management (Intensive) YES NO Recovery Program YES NO Supported Living (Financial Only) YES NO Supported Living (Agency/Low Supervision) YES NO Supported Living (Agency/Supervised) YES NO Supported Employment YES NO Dialectical Behavior Therapy YES NO Forensic Services YES NO PREP Day Program YES NO Drop-In Center YES NO

PAGE 517

498 Appendix B Development of the Support and Community Connectedness Survey Background and Theoretical Constructs For many with a severe mental illness (S MI), the chronic course of the disease process leads to esca lating levels of social exclusion. The 1998 white paper Modernizing mental health highlighted how th e prejudice attached to mental ill health and the failure to understand its causes, leads to discrimi nation and social exclusion (DoH 1998, p. 1.11). There have been severa l studies that have addressed social networks in the SMI populat ion (Becker, Leese, Clarks on, Taylor, et al, 1998; Brugha, Wing, Brewin, MacCarthy, & Lesage, 1993; Brunt & Hansson, 2002; Homes-Ebor & Riger, 1992). Evidence suggests that the SM I population would have better treatment compliance and success in remaining out of psychiatric facilitie s if their social network/social support struct ure was both larger and more flexible (Fried, Johnsen, Starrett, Calloway & Morrissey, 1998; Goering, Durbin, Foster, Boyles, Babiak, & Lancee, 1992; MacDonald, Hayes & Baglioni, 2000). There have been fewer studies directed to perceived social support for those with a SMI (Letvak, 2002). Very few studies have looked at the st ructural implications of soci al networks in relation to interactional aspects of so cial support (Hall & Nelson, 1996). There are currently no studies that address empirically the SMI population and their percep tions of community support (perceived social cohesion, the individua l or micro-level of social capital) or studies that link these constructs with treatment compliance, maintenance in the community or social inclusion. Thus, there is no evidence for or ag ainst the idea that

PAGE 518

499 perceived community level integration is beneficial to client stability or that it enhances social inclusion. In an attempt to extend current social theory, the construct of Perceived Social Integration (PSI) was develope d to understand the relationshi ps between the constructs that appear to make up PSI. These constructs include social network satisfaction (SSN), perceived social support (PSS) and perceived social co hesion (PSC). Together, information related to the three sub-constr ucts noted were evaluated in a survey developed to capture the constr ucts individually and to le nd support to the hypothesis that these constructs together partially mediat e between life as a psychiatric patient and marginalization from society. The instrument consists of Likert st yle questions that can be completed quickly and can be administered by a clients provider. The purpose of this additional appended information is to describe the development of the instrument that gathers data simultaneously on each of the th ree sub-constructs and to explicate their relationship to the construct of PSI. Procedure Two rounds of content validity were co mpleted. The first was focused on the 112 questions generated to represent the three c onstructs SNS, PSS, and PSC. Each question was developed with a construct and sub-constr uct in mind, targeting the questions to the cluster x sub-construct models in Table 110 and Table 111. The methods and tables are fully described in the next section. Be fore item construction was completed, local experts were contacted and asked if they w ould participate in the content validity phase of a new instrument. Two groups were queried. The first was five professors at USF who were experts in the social constructs being investigated. Four agreed to participate.

PAGE 519

500 A second group of reviewers consisted of experts in the SMI population. Population experts had to have a minimum of a Masters degree in psychology or a related field and at least three years working with the SMI population. All five contacted agreed to participate. All participants were sent a packet of information that included an introduction section, description of the cons tructs and sub-constructs, deta iled instructions on how to proceed, a separate page with the sub-constr ucts listed and numbered to make scoring easier, and a large, multi-page matrix of the questions. For each question, the respondent was asked to decide which construct the question pertained to, which sub-construct within each construct, how appropriate th ey viewed the question and how clear the phrasing was (1=clear, 2=needs work, 3=poor for appropriateness and phrasing). A comment section was included for each question in case the participant had an alternative choice to wording or some other comment. Finally, a brief description of the question development and a list of references were incl uded in each packet. It was suggested that each participant read these onl y after completing the task. Once the information was returned, a la rge table was created with each question and response included to ascertain the accuracy of responses compared to the authors beliefs of which construct each item fit. D ecision rules were generated to determine which questions advanced to round two of th e content validity assessment. As noted there were a total of four prof essors and five clinicians who participated by answering all or nearly all the questions. Each professo r had their answer multiplied by two (2.0) because the knowledge of the constructs wa s considered of higher importance then knowledge of the population. Clinician respons es were not weighte d. A perfect score

PAGE 520

501 was 13 ((4 professors x 2) + (5 clinicians x 1) = 13). If a question received a score of 10 or greater, it was included w ith no or minor modifications. For instance, all questions with partner or intimate partner in them were altered by adding or best friend in response to one professors suggestion that many of the SMI may not have intimate partners but may have a best fr iend. This level of alteration is not expected to affect the respondents answer, but will allow more respondents to answer. Questions with a final score of 7-9 were considered feasible with modification. Fe w of these were considered modifiable after review and most were disc arded. Any question scoring a six or below was automatically discarded. The original concept was to have the participants review all the questions a second time. This was not followed because se veral participants complained of the time needed to complete the first pass. It was decided that the second pass would involve only those questions with sufficient substantive ch ange that a more careful check was needed (total of nine) and any additional questions generated (total of four). Thus, the second round of content validity assessment was on thir teen questions. The information sent to the participants for the first review of items is included at the end of this appendix and is titled Content Validity First Pass. Four additional questions were generated to fit gaps in the theoretical matrices (tables 1 and 2) where original questions were all eliminated. These questions and the second review for content valid ity are also included at the end of this appendix and is titled Content Validity Second Pass. Item Development All items for this survey were designed following an in-depth examination of the literature. Social networks and social support were conceptualized using both clusters

PAGE 521

502 and sub-constructs. For both constructs the clusters were family, friends, neighbors/acquaintances, intimate partners and mental health providers. Not all subconstructs were considered of equal importance for each cluster. Regarding social networks, the direction and clustering subconstructs were considered of primary importance, with magnitude (size) of sec ondary importance (Table 110). This deductive process was used for each cluster, generating items for evaluation. Table 110 Social networks: Cluster x sub-construct Cluster Dimensions Magnitude Directionality Durability Homogeneity Clustering Family + + Friends + + + + Neighbors + Intimates + + + Providers + + Note: + = primary importance; = seconda ry importance, empty = no importance, not addressed Social support was evaluated along the same clusters but different dimensions. The importance of the sub-construct for each construct is demonstrated in Table 111. Social cohesion did not have literature support for addressing the concepts of community involvement, reciprocity and trus t in relation to the clusters noted above. Because of this, items were constructed independent of clusters, thus more generally. However, the concepts of trust and community involvement were considered of slightly higher importance than reciprocity, generating a few mo re items in each than in reciprocity. The belief is that a sense of community involve ment, caring and trust will lead to higher

PAGE 522

503 levels of reciprocity. Howeve r, reciprocity might require higher levels of proactive behavior and access to resources such as tr ansportation, money and time. This is Table 111 Social support: Cluster x sub-construct Cluster Dimensions Emotional Instrumental Informational Appraisal Dissonance Family + + + + Friends + + + Neighbors Intimates + + + + Providers + + + Note: + = primary importance; = seconda ry importance, empty = no importance, not addressed especially true for active involvement, i.e. volunteering, that would require investment of scare resources, but the more passive aspects of community involvement would not. The lack of resources available to the SMI population is, unfortunately, a daily aspect of their lives. Thus, reciprocity was considered of less importance for both pragmatic and theoretical reasons. Content Validity Results First Pass Of the 112 original questions, 76 scored a 10 or higher. Not all of these were included in the pilot instrument and four additional questions were designed for the second round. Table 112, in the column la beled CV1 Results, summarizes which questions were retained, eliminated, and upda ted. If a question was updated, the altered question is included below that question in boldface. These questions formed the

PAGE 523

504 Table 112 Question and Results Matrix Q # Question Construct Sub-Construct CV1 Score CV1 Results CV2 Results 1 Most people will do the wrong thing if they know they wont be caught Cohesion Trust 11 Keep 2 When I do something wrong, my partner or best friend points it out to me Support Appraisal 12 Keep 3 I would like to have different friends than I do now Support Emotional 4 Drop 4 When I am sad or feeling blue, I can always turn to my family Support Emotional 11 Keep 5 I feel comfortable openly expressing my feelings Support Emotional 11 Drop 6 My family spends time with my friends and I Network Clustering 12 Keep 7 I am not lonely because I know my neighbors Support Emotional 4 Drop 8 I wish my partner or best friend would try harder not to hurt my feelings Support Dissonance 13 Keep 9 I can count on my counselors to be honest about how they see me Support Appraisal 4 Drop 10 People are trustworthy, but society is not Cohesion Trust 12 Keep 11 I belong to a group that is not a church group Cohesion Community 8 Updated I belong to a church or other community group Keep 12 It is okay that I dont work because I am mentally ill Cohesion Community 7 Drop 13 My friends are often angry with me Support Dissonance 10 Keep 14 I have a close relationship with my family Support Emotional 12 Keep 15 I volunteer my time to organizations when I can Cohesion Community 13 Keep 16 Sometimes I feel that my counselor doesnt take the time to hear me Support Dissonance 11 Keep 17 People without mental illness have more people to help them then I do Support General 3 Drop 18 Overall, I feel that enough people care about me and support me Support General 7 Drop 19 Most of my friends help me feel good, but help me in many other ways too Support Instrumental 12 Keep 20 I change friends quickly Network Durability 12 Keep 21 My intimate partner has the same types of problems as I do Network Homogeneity 4 Drop

PAGE 524

505 Table 112 (cont.) Q # Question Construct Sub-Construct CV1 Score CV1 Results CV2 Results 22 Sometimes I help my providers with their problems Network Directionality 4 Drop 23 I take more than I give in relationships Support Directionality 8 Drop 24 I feel criticized when my family says theyre trying to help me Support Dissonance 12 Keep 25 I generally trust the federal government to do what is right Cohesion Trust 13 Keep 26 I am close to my mental health providers Support Emotional 9 Updated My counselor helps me with my emotional stability Keep 27 As far as society is concerned, people without a mental illness are better than people with a mental illness Cohesion Reciprocity 8 Updated As far as society is concerned, people without a mental illness do more for the community then people with a mental illness Keep 28 I feel safe to walk in my neighborhood Cohesion Trust 10 Keep 29 I often visit with my neighbors at my home or theirs Network Directionality 8 Drop 30 My symptoms get in the way of my having a good relationship Support General 9 Drop 31 When my back is in a corner, I can count on my partner or best friend to support me Support Instrumental 12 Keep 32 When I need to know something, my friends usually have the answer Support Information 13 Keep 33 I provide as much for my partner or best friend as my he or she does for me Network Directionality 5 Drop 34 My friends spend time with my intimate partner or best friend and I Network Clustering 12 Keep 35 My family helps me Network Directionality 2 Drop 36 My partner or best friend doesnt feel I can do anything right Support Dissonance 13 Keep 37 I make a point to vote during elections Cohesion Community 13 Keep 38 My closest relationships usually last for two years or more Network Durability 10 Keep 39 I have had some of the same friends for many years Network Durability 12 Keep 40 Sometimes when I need advice or information, friends or family make sure I cant get it Support General 11 Keep

PAGE 525

506 Table 112 (cont.) Q # Question Construct Sub-Construct CV1 Score CV1 Results CV2 Results 41 My counselor knows or has met my intimate partner or best friend Network Clustering 11 Keep 42 If I need money or help with a bill, my family almost always gives it to me Support Instrumental 12 Keep 43 I can laugh with my neighbors and feel good about myself with them Support Appraisal 8 Drop 44 When I need a shoulder to cry on, I have at least one good friend to turn to Support Emotional 9 Updated I have at least one good friend to turn to if I am feeling sad or afraid Keep 45 I cant disagree with others because I am mentally ill Support General 4 Drop 46 My intimate partner or best friend helps me in many ways Support Emotional 13 Keep 47 My friends wont leave me alone when I ask them to Support Dissonance 12 Keep 48 I would say that I have a large family Network Magnitude 12 Keep 49 I need to be careful because my community wont help me Cohesion Community 13 Keep 50 My family meets several of my needs Support Instrumental 13 Keep 51 Most of the time people are looking out for themselves, not trying to be helpful to others Cohesion Reciprocity 12 Keep 52 If I make a mistake, my friends point it out so I will correct it Support Appraisal 13 Keep 53 You can trust family but no one else Cohesion Trust 7 Drop 54 I feel that I belong and Im at home in the community Cohesion Community 11 Keep 55 Helping me feel good is what my partner or best friend does best Support Emotional 13 Keep 56 Other people with a mental illness have more fun in life than I do Support General 5 Drop 57 My family has had contact with my counselor(s) Network Clustering 11 Keep 58 I have lacked friendships most of my life Network Magnitude 10 Keep 59 My family isnt honest with me about what I am doing Support Appraisal 11 Keep 60 If I am making a mistake, my counselor will always point it out Support Appraisal 10 Keep 61 I know most of my neighbors by name Network Magnitude 8 Drop

PAGE 526

507 Table 112 (cont.) Q # Question Construct Sub-Construct CV1 Score CV1 Results CV2 Results 62 In too many of my relationships, people take advantage of me Support General 7 Drop 63 I feel comfortable going to a neighbor and borrowing something Support Instrumental 5 Drop 64 I generally trust my local government to do what is right Cohesion Trust 12 Keep 65 The more I give to the community, the more I want to Cohesion Reciprocity 12 Keep 66 My friends know who my counselors are Network Clustering 12 Keep 67 In many ways, most of my neighbors are a lot like me Network Homogeneity 10 Keep 68 I have friendships with people who are not mental health clients Network Homogeneity 12 Keep 69 I keep myself informed on community issues Cohesion Community 12 Keep 70 If I need to know something, I ask my counselor(s) Support Information 12 Keep 71 I have many friendships with mental health clients Network Homogeneity 9 Updated Most of my friends are also mental health clients Keep 72 My family spends time with my intimate partner (or best friend) and I Network Clustering 11 Keep 73 I choose intimate partners partially by how much we have in common Network Homogeneity 9 Updated I choose my friends by how much we have in common Keep 74 If I need help fixing or making something, my family helps me Support Instrumental 12 Keep 75 My counselor tells me too much what to do without fully talking it over with me Support Dissonance 12 Keep 76 If I am in trouble, I can go to a neighbor for help Cohesion Trust 5 Drop 77 I care that the community does well Cohesion Community 9 Updated I care for my community and try to help it when I can Keep 78 There are only a few people I completely trust Cohesion Trust 10 Keep 79 I want to have my family living closer to me Support Emotional 3 Drop 80 I know that many of my neighbors are mental health clients Network Homogeneity 10 Keep

PAGE 527

508 Table 112 (cont.) Q # Question Construct Sub-Construct CV1 Score CV1 Results CV2 Results 81 I am able to help my family when they need or ask Network Directionality 2 Drop 82 I cannot be in a relationship because I am mentally ill Support General 6 Drop 83 I feel I give back to the community for what I take Cohesion Reciprocity 13 Keep 84 I know whom to call if I have a concern about my local government Cohesion Community 10 Keep 85 My partner gives me things or helps me do things to make my life easier Support Instrumental 12 Keep 86 There must be a lot wrong for me to end an intimate relationship Network Durability 5 Drop 87 Sometimes I feel that my family is too close to me Support Dissonance 11 Keep 88 I find I get a new counselor too often Network Durability 12 Keep 89 If I need food, money, or other help, my friends are always there Support Instrumental 11 Keep 90 I generally trust my state government to do what is right Cohesion Trust 13 Keep 91 My counselor(s) lift my spirits Support Emotional 11 Drop 92 My partner or best friend is always telling me what to do Support Dissonance 13 Keep 93 Too many people block me from expressing what I feel Support General 13 Drop 94 In general, I have enough friends Network Magnitude 12 Keep 95 My counselor talks about his or her life with me Network Directionality 13 Keep 96 Only people who are also mentally ill are honest with me Support General 0 Drop 97 My friends say nice things to me and help me feel good about myself Support Appraisal 12 Keep 98 I am a trusting person Network Trust 7 Drop 99 My partner or best friend praises me and cheers for me when I accomplish something Support Appraisal 13 Keep 100 I worry about bad things happening in my community Cohesion Trust 8 Drop 101 I help my friends as much as they help me Network Directionality 6 Drop 102 My family comes over or offers advice when I dont want them to Support Dissonance 13 Keep 103 I can always turn to my best friend or partner for advice if I am confused Support Information 11 Keep

PAGE 528

509 Table 112 (cont.) Q # Question Construct Sub-Construct CV1 Score CV1 Results CV2 Results 104 If I were NOT mentally ill, I would care more about things in the community Cohesion Community 8 Updated Mental illness stops a person from caring about their community Keep 105 My community doesnt care about me, so I stay to myself Cohesion Reciprocity 11 Keep 106 I believe that most people are basically good Cohesion Trust 9 Updated I believe that in a pinch, I could trust a stranger to help me Keep 107 If I need to know something my family usually has the answer Support Information 13 Keep 108 My counselor helps me in more ways then talking and counseling Support Emotional 9 Drop 109 Because of my mental illness, I cant give back to the community Cohesion Reciprocity 10 Drop 110 I believe that you cant be too careful in life Cohesion Trust 7 Drop 111 I feel I have a voice in local government Cohesion Community 13 Keep 112 I belong to a church Cohesion Community 8 Drop Additional Questions Round Two Content Validity 113 I try to provide for my family as much as they provide for me Network Directionality Keep 114 My friends give to me to the same degree I give to them Network Directionality Keep 115 My best friend or partner and I support each other equally Network Directionality Keep 116 I am one of the first to contribute to community projects or concerns Cohesion Reciprocity Keep Note: Questions in bold face represent questions altered or added for round two of content validity assessment majority of round two of content validity assessment (see below). Table 113 breaks down by construct the questions retained and eliminated. A retention rate of 68% appears inordinately high for a first pass. Discussion with the participants indicated that they had a high degree of confidence in their responses. Also, several iterations and a

PAGE 529

510 Table 113 Question retention first round content validity assessment Social Network Social Support Social Cohesion Total Original 28 53 31 112 Kept 19 34 23 76 Discarded 9 19 8 36 Total 28 53 31 112 great deal of time were invest ed in initial questio n construction along with the theoretical basis supporting each question. The weighting of professor over clinician was also considered as a possible reason for the high retention rate. As noted, this groups knowledge of the constructs was considered more important, and also was pred icted to be more accurate. Comparison of the average number of correct responses be tween the two groups was also completed. Table 114 displays the results by participant for number attempted of 112, number correct, percent right and percent wrong. The prediction of the construct group having a higher percentage of agreement with the auth or due to knowledge of the constructs was relatively accurate. Only one respondent in the construct group (#3) scored lower than any of the population group. Conversely, and against expectation, one member of the population group (#1) actually scored the highest of all participants. Assessing the correct item res ponses, three patterns were located in the response data that should be mentioned. First, only one of the questions that targeted the subconstruct of directionality in the construct of social netw orks was retained. It is hypothesized that this sub-construct, which is similar to the soci al support concept of reciprocity, was confusing to the respondents. Because the directi onality of social

PAGE 530

511 Table 114 Accuracy of content validity respondents first pass Construct Group* Population Group** Respondent 1 2 3 4 1 2 3 4 5 Attempted 106 112 110 91 107 112 112 112 112 Correct 90 93 84 79 95 65 86 61 74 % Incorrect 15.89 16.96 24.32 14.13 12.04 41.96 23.21 45.54 33.93 % Correct 84.11 83.04 75.68 85.87 87.96 58.04 76.79 54.46 66.07 Note: *Construct group = university professors **Population group = clinicians from community agencies support within social networks is important, thre e new questions were developed and targeted to this sub-construct. These questi ons are included in boldface at the bottom of Table 112. One additional question was also created to reflect the reciprocity subconstruct of social cohesion due to a slight ly higher elimination rate for that subconstruct. Second, for both social networks and social support, the neighbor cluster had nearly all questions eliminated. Upon scru tiny, it was noted that many of the neighbor questions pertained directly to those with a mental illness and were eliminated. This is related with the third finding, all questions pertaining direct ly to mental illness were eliminated through the process of question verification. The unsuccessful attempt to make items specific to the mentally ill was discussed with a few of the respondents. Uniformly, they agreed that items such as thes e would form a separate instrument, if they were legitimate at all. Measuring the constr ucts across samples with varying degrees of mental illness requires a uniform approach. Those without a mental illness would have a difficult time answering these que stions. A few questions more specific to the mentally

PAGE 531

512 ill do remain, in less direct form, and will form a nucleus for contrasting the samples after future data collection and analysis. Content Validity Results Second Pass. The second pass of the content valid ity was completed only with the construct sample. The number of questions and the time involved for clinicians who have considerably less flexibility in their tim e drove this choice. The decision rule for passing any question was the requirement of three out of four respondents finding no problem with each question. All questions passed this level of scrutiny and were included in the final survey. The decision to k eep questions is noted in the last column of Table 112. Reliability Analysis The final pilot survey contained 80 item s and was administered first to 125 individuals afflicted with me ntal illness who were asked to comment on item content, language and format. The 80 item version of the SCCS is included at the end of this appendix. There were few changes sugge sted and any changes required some consistency in those reviewing the surve y. Once the changes we re incorporated, 350 individuals agreed to pilot the survey. One hundred of these respondents were administered the survey a second time from 68 weeks later. Together, both test-retest and internal consistency were measured. A f actor analysis of test items revealed a sixfactor structure that encompassed 34 items. Details on factors and reliability estimates are provided in Ch apter 3 (pp 135-140).

PAGE 532

513 Appendix B Content Validity First Pass Introduction Thank you so much for agreeing to be a part of this process. You are being asked to participate in this task because you ar e either an expert in the population being investigated and/or the constructs. This is an effort to design an instrument that will measure simultaneously perceptions of social networks, social support, and social capital in a severely mentally ill (SMI) population. You are helpi ng with the content validation of the instrument, a fundamental and esse ntial aspect of inst rument development necessary for acceptance of the information obtained by the instrument. Also, in order to truly understand the perceptions of the SMI, it is important to be able to compare their perceptions to those without a diagnosed mental illness. Thus, this instrument will be given to other populations as well. The design, then, will need to be simple and comprehensive, a task Im counting on your input to assist me with. For the vast majority of people with a SMI, their symptoms come and go. They are not perpetually in a state of active psychosis or depressi on. Yet, the effect of the illness appears to continue unimpeded. Little research has investigated this aspect of the mental illness, and this instrument will hopefully move that line of research forward in some small way. It is hypothesized that a combination of per ception of self and perception of self in the community maintains the negative aspects of mental illness during non-symptomatic phases in addition to the fear of relapse. Through measuring these three constructs, it is hoped that the results will cont ribute to understanding how the mentally ill see themselves in the world, their sens e of belonging and their perception of

PAGE 533

514 the cohesiveness of society. This information will be useful in developing appropriate interventions. The style of the finished instrument is sti ll in development. Currently, it will be a set of forced choice (yes/no) questions and Likert style questions (six point strongly agree, agree, mildly agree, mildly disagree, disagree, strongly agree) The forced choice questions will help to guide the respondent through the questions, allo wing them to skip sections that do not pertain to them. Six points was chosen over four to allow more variance to be captured and because more choice s has a positive effect on reliability. The exact number of questions has not been decide d, and this process will help to establish that answer as well. The rationale for how questions were chosen is included in a separate appendix. I ask each participant to complete the scoring process before reviewing (or skippi ng!) the appendix. I am hoping to be able to solidify this inst rument with one pass at establishing the content validity with your assistance. Howe ver, it may require one additional iteration if substantial changes are suggested. Construct Definitions The following provides brief descriptions of each construct. Since the constructs are complex, some effort was made to cap ture the construct through addressing subdimensions found in the literature. Because of the complexity of the constructs, please note that a thorough investigation of each sub-dimension is not being attempted. It is hoped that by touching on the sub-dimensions, the greater construct will be adequately measured. However, having a brief descripti on of the sub-dimensions may assist you in identifying the larger construct that the question is aimed at.

PAGE 534

515 Social Networks Mitchell has defined a social netw ork as a specific set of linkages among a defined set of persons, with the addi tional property that the characteristics of these linkages as a whole be used to interpret the social behavior of the person involved (Mitchell, 1969 p. 4). Thus, SNs refer to the structural aspects of social relationships. They are the channels through which prag matic help as well as emotional and psychological support can be exchanged between individuals (Achat, Kawachi, Levine, Berkey, Coakley & Colditz, 1998). The overall SN represents the potential for support, whereas the perceived SN is composed of actual supports by those whom an individual knows by name, sees at least monthly, and identif ies as an important support in his or her life (Walsh, 1994). Perceived SN is the concept of greatest importance to this investigation. The questions pertaining to SNs are not expected to be comprehensive. The study being considered will also in clude a separate social network analysis. These questions are designed to provide inform ation that will be used to compare the satisfaction of current SNs to the perceived SN. The fo llowing sub-constructs were considered important and included in question development: Magnitude size of network cluster. Directionality unidirectional versus reciprocal support that passes be tween the network participants Durability length of time in relationships. Homogeneity the degree to which network member s share similar social or personal attributes.

PAGE 535

516 Cross Clustering to what degree different members of a social network are in contact with each other. Social Support Social support is the provision of actual assistance or feelings of attachment to individuals by other persons who are perceived as caring (Hobfoll, Freedy, Lane, & Geller, 1990). Social Support is considered a multi-dimensional construct. The concept of social support is operationalized in many different ways: on the basis of who is providing the support; quantity and quality of support; availability of support; and satisfaction with support (Letvak, 2002). An important aspect of support that is fre quently overlooked in th e literature is the concept of dissupport (Malone, 1988; Avison, 1996). The general assumption underlying findings is that social support has a positive value. The problem is that it can also be viewed as a source of unwelcome demands. Indeed, Wing (1978) has suggested that social withdrawal may actually be pr otective for someone with schizophrenia who may have damaged social skills. A theoretica l link exists between negative aspects of the affective dimension of SN and social dissupport. Dissupport, as a source of stress, can be received through critical, bossy or intrusive linkages. Both support and dissupport will be measured in the proposed instrument. Sub-dimensions of social support to be measured include: General/Descriptive aspects of the self that are im portant to giving or receiving social support Emotional affection, esteem, concern, respect Instrumental aid in labor, time, or money Informational suggestion, information, advice

PAGE 536

517 Appraisal affirmation, feedback, insight into the self from the other Dissonance Negative aspects of support, i.e. criticism, insults, lack of respect Social Capital Social capital has received many defi nitions. Portes (1998) suggest that social capital has come to mean the ability to secure benefits through membership in networks and other social structures. Soci al capital can be investigated at the macro (population) level and at the micro (indivi dual) level. The development of this instrument will address the individual percep tion of community level support as well as the interaction between the self and the co mmunity. Social capital is frequently considered a population level asset, making it difficult to capture in small samples. However, I believe that perceived social capital, even if the person may not think of it in that term, can be captured in small samp les and compared between these samples. Both intra-community ties (integration) and extra-community ties (linkage) will be addressed (Hawe & Schiell, 2000). Three areas commonly examined in SC are: Volunteering/community involvement This aspect of social capital will be measured through questions regarding current perceive d involvement in the community. An SMI individual who is a volunteer may volunteer for reasons different from the non-SMI. For many with an SMI, volunteering is a safer and less stressful decision than competitive employment and has nothing to do with altruism as measured by social capital. Additionally, perceived worth to the community and how well one fits in may affect how much one engages or perceives oneself to be involved in a reciprocal relationship with society. It is hoped that data from this inst rument will help to elaborate this hypothesized connection.

PAGE 537

518 Reciprocity looks at how a person believes that they are a contributor as well as a consumer of community assets, as noted above Questions will directly investigate how much give and take with society each indi vidual perceives they are involved in. Reciprocity may have a positive affect on individual level self-esteem, self-worth and, important for this study, psychiatric stability. Qualitative data from a pilot study just completed appears to reinforce this hypothesis. Trust in community safety, availability of support outside immediate SNs and in providers is a poorly researched area rega rding the SMI population. Additionally, trust that people will behave in a certain way, trust in the government, and trust in ones neighbors has been under researched in the SMI population. Instructions 1. Please make sure you have reviewed the constructs above and are comfortable with the definitions. You may not comple tely agree with how the constructs are defined. If you have a serious reserva tion, please feel free to comment on this separately. Do keep in mind that none of these constructs have an absolute consensus of definition in the literature, so I do not expect consensus at this time either. 2. Once you have reviewed the constructs, youll begin to address the questions themselves in the table provided. Please note that there are five tasks for each question: a. In the first column indicate whethe r you think this question addresses social networks (N), social support (S) or social capital (C) by placing the appropriate letter in the column. Ea ch question was developed in relation

PAGE 538

519 to a single construct. However, with the level of shared variance and explanatory ability between these constr ucts it is difficult to completely exclude two while addressing one. Co mparing your responses will allow me to see how accurate I was in development of the question. b. This next column can be skipped if you wish. However, I am including it in case you want to take a stab at wh ich sub-dimension I am attempting to capture. Below, youll find a list of numbers. You can include the number for the sub-dimension that matches the construct in the SubDimension column of the table. c. The last two columns address appropri ateness (A) and phrasing (P) of the question. Using a three level scale, pl ease indicate if you feel the question is (1) clear, (2) needs work, or (3) yo u havent the faintest idea what Im talking about. i. Appropriateness indicates the level that you feel the question fits the construct you chose for that question. ii. Phrasing evaluates the words chosen and word order. In your opinion, is the question cl ear and easily understood d. Under each question is a blank area. Please feel free (if you answered 2 or 3 to the A, C or P columns) to sugges t a different phrasing or to make any other comments. For part 2.b. above, please use the follo wing numbers if you choose to address the optional column Ive included this on a separate page in case you would like to print this page for ease of use.

PAGE 539

520 Social Networks 1. Magnitude 2. Directionality 3. Durability 4. Homogeneity 5. Cross Clustering Social Support 6. Emotional 7. Instrumental 8. Informational 9. Appraisal 10. Dissonance Social Capital 11. Volunteering / Community involvement 12. Reciprocity 13. Trust

PAGE 540

521 Q# Question Construct Sub-Dimension Appropriateness Phrasing 1 Most people will do the wrong thing if they know they wont be caught Comment: 2 When I do something wrong, my partner points it out for me Comment: 3 I would like to have different friends than I do now Comment: 4 When I am sad or feeling blue, I can always turn to my family Comment: 5 I feel comfortable openly expressing my feelings Comment: 6 My family spends time with my friends Comment: 7 I am not lonely because I know my neighbors Comment: 8 I wish my partner would try harder not to hurt my feelings Comment: 9 I can count on my counselor s to be honest about how they see me Comment:

PAGE 541

522 Q# Question Construct Sub-Dimension Appropriateness Phrasing 10 People are trustworthy, but society is not Comment: 11 I belong to a group that is not a church group Comment: 12 It is okay that I dont work because I am mentally ill Comment: 13 My friends are often angry with me Comment: 14 I have a close relationship with my family Comment: 15 I volunteer my time to organizations when I can Comment: 16 Sometimes I feel that my provider doesnt take the time to hear me Comment: 17 People without mental illness have more people to help them then I do Comment: 18 Overall, I feel that enough people care about me and support me Comment:

PAGE 542

523 Q# Question Construct Sub-Dimension Appropriateness Phrasing 19 Most of my friends help me feel good, but help me in many other ways too Comment: 20 My friendships tend to change quickly Comment: 21 My intimate partner has the same types of problems as I do Comment: 22 Sometimes I help my providers with their problems Comment: 23 I take more than I give in relationships Comment: 24 I often feel criticized wh en my family says theyre trying to help me Comment: 25 I generally trust the federa l government to do what is right Comment: 26 I am close to my mental health providers Comment: 27 As far as society is concer ned, people without a mental illness are better than people with a mental illness Comment:

PAGE 543

524 Q# Question Construct Sub-Dimension Appropriateness Phrasing 28 I feel safe to walk in my neighborhood Comment:: 29 I often visit with my neighbors at my home or theirs Comment: 30 My symptoms get in the way of my having a good relationship Comment: 31 When my backs in a corner, I can count on my partner to support me Comment: 32 When I need to know something, my friends usually have the answer Comment: 33 I provide as much for my pa rtner as my partner does for me Comment: 34 My friends spend time with my intimate partner and me Comment: 35 My family helps me Comment: 36 My partner doesnt feel I can do anything right Comment:

PAGE 544

525 Q# Question Construct Sub-Dimension Appropriateness Phrasing 37 I make a point to vote during elections Comment: 38 My intimate relationships usually last for a year or more Comment: 39 I have had some of the same friends for many years Comment: 40 Sometimes when I need advice or information, some people make sure I cant get it Comment: 41 My mental health providers know or have met my intimate partner Comment: 42 If I need money or help with a bill, my family almost always gives it to me Comment: 43 I can laugh with my neighbors and feel good about myself with them Comment: 44 When I need a shoulder to cry on, I have at least one good friend to turn to Comment: 45 I cant disagree with others because I am mentally ill Comment:

PAGE 545

526 Q# Question Construct Sub-Dimension Appropriateness Phrasing 46 My intimate partner helps me in many ways Comment: 47 My friends wont leave me alone when I want them to Comment: 48 I would say that I have a large family Comment: 49 I need to be careful because my community wont help me Comment: 50 My family meets several of my needs Comment: 51 Most of the time people are looking out for themselves, not trying to be helpful Comment: 52 If I make a mistake, my friends point it out so I will correct it Comment: 53 You can trust family but no one else Comment: 54 I feel that I belong and Im at home in the community Comment:

PAGE 546

527 Q# Question Construct Sub-Dimension Appropriateness Phrasing 55 Helping me feel good is what my partner does best Comment: 56 Other people with a mental illness have more fun in life than I do Comment: 57 My family has had contact with my mental health providers Comment: 58 I have lacked friendships most of my life Comment: 59 My family isnt honest with me about what I am doing Comment: 60 If I am making a mistake, my counselor will always point it out Comment: 61 I know most of my neighbors by name Comment: 62 In too many of my relationships, people take advantage of me Comment: 63 I feel comfortable going to a neighbor and borrowing something Comment:

PAGE 547

528 Q# Question Construct Sub-Dimension Appropriateness Phrasing 64 I generally trust my local governme nt to do what is right Comment: 65 The more I give to the community, the more I want to Comment: 66 My friends know who my mental health providers are Comment: 67 In many ways, most of my nei ghbors are a lot like me Comment: 68 I have friendships with people who are not mental health clients Comment: 69 I keep myself informed on community issues Comment: 70 If I need to know something, I ask my counselor(s) Comment: 71 I have many friendships with mental health clients Comment: 72 My family spends time with my intimate partner Comment:

PAGE 548

529 Q# Question Construct Sub-Dimension Appropriateness Phrasing 73 I choose intimate partners partially by how much we have in common Comment: 74 If I need help fixing or making something, my family helps me Comment: 75 My counselor tells me too much what to do without fully talking it over with me Comment: 76 If I am in trouble, I can go to a neighbor for help Comment: 77 I care that the community does well Comment: 78 There are only a few people I completely trust Comment: 79 I want to have my family living closer to me Comment: 80 I know that many of my ne ighbors are mental health clients Comment: 81 I am able to help my family when they need or ask Comment:

PAGE 549

530 Q# Question Construct Sub-Dimension Appropriateness Phrasing 82 I cannot be in a relationship because I am mentally ill Comment: 83 I feel I give back to the community for what I take Comment: 84 I know whom to call if I ha ve a concern about my local government Comment: 85 My partner gives me things or helps me do things to make my life easier Comment: 86 There must be a lot wrong for me to end an intimate relationship Comment: 87 Sometimes I feel that my family is too close to me Comment: 88 I find I get new mental health providers too often Comment: 89 If I need food, money, or other help, my friends are always there Comment: 90 I generally trust my state governme nt to do what is right Comment:

PAGE 550

531 Q# Question Construct Sub-Dimension Appropriateness Phrasing 91 My counselor(s) lift my spirits Comment: 92 My partner is always telling me what to do Comment: 93 Too many people block me from expressing what I feel Comment: 94 In general, I have enough friends Comment: 95 My mental health providers talk about their lives with me Comment: 96 Only people who are also me ntally ill are honest with me Comment: 97 My friends say nice things to me and help me feel good about myself Comment: 98 I am a trusting person Comment:

PAGE 551

532 Q# Question Construct Sub-Dimension Appropriateness Phrasing 99 My partner praises me and cheers for me when I accomplish something Comment: 100 I worry about bad things happening in my community Comment: 101 I help my friends as much as they help me Comment: 102 My family comes over or offers advice when I dont want them to Comment: 103 I can always turn to my partner for advice if I am confused Comment: 104 If I were NOT mentally ill, I would care more about things in the community Comment: 105 My community doesnt care about me, so I stay to myself Comment: 106 I believe that most people are basically good Comment: 107 If I need to know somethi ng my family usually has the answer Comment:

PAGE 552

533 Q# Question Construct Sub-Dimension Appropriateness Phrasing 108 My counselor helps me in more ways then talking and counseling Comment: 109 Because of my mental illne ss, I cant give back to the community Comment: 110 I believe that you cant be too careful in life Comment: 111 I feel I have a voice in local government Comment: 112 I belong to a church Comment:

PAGE 553

534 Appendix B Content Validity Second Pass Results from first pass content validity Retained Retained and Updated Discarded Social Networks Satisfaction 22 2 9 Sub-Constructs Magnitude 3 Directionality 4 Durability 4 Homogeneity 5 Clustering 6 Perceived Social Support 34 2 19 Sub-Constructs Emotional 7 Instrumental 7 Informational 5 Appraisal 5 Dissonance 10 Perceived Social Cohesion 24 5 8 Sub-Constructs Community Involvement 10 Reciprocity 6 Trust 8 The original total number of questions was 112. Of these 76 were sufficient for inclusion in the instrument. This was determ ined by the following: There were a total of four professors and five clinicians who part icipated through answering all or nearly all the questions. Each professor had their an swer multiplied by two (2.0) because the knowledge of the constructs was considered of higher importance then knowledge of the population. Clinician responses were not weighted. A perfect score was 13 ((4 professors x 2) + (5 clinicians x 1) = 13). If a question rece ived a score of 10 or greater, it was included with no or minor modifications. For instance, all questions with partner

PAGE 554

535 or intimate partner in them were altered by adding or best friend in response to one professors suggestion that many of the SMI may not have intimate partners but may have a best friend. Questions w ith a final score of 7-9 were considered feasible with modification. Few of these were considered modifiable after review and most were discarded. Any question sc oring a six or below was automatically discarded. The next task: Included are a few ques tions with modifications that I hope you will have time for review. Also, four ne w questions were developed. Three were developed to target the social network sub-dimension of dire ctionality. I believe that directionality in social networ ks is the same as reciprocity in social support theory. For networks, this merely measures the dire ction of support or aid from the respondent to those in his or her environment, the e nvironment to the respondent, or both. These questions target family, friends and intimate partners. One additional question was added to the social cohesion cons truct under the reciprocity sub-dimension to bring the questions for that area closer to the other two sub-dimensions. If you would simply review the questions Im including that received the most alterations (both old and new questions are in cluded) and also the four new questions. There are nine questions (table above, column labeled retained and updated) that were overhauled. Simply indicate to what leve l you agree with the questions and if you have an alternative wording toss it in! Ive identified the c onstructs and sub-dimensions already.

PAGE 555

536 Reworded and New Questions Q # Question Wording Construct SubDimension 1 Old I belong to a group that is not a church group Cohesion Community New I belong to a church or other community group Comments 2 Old I am close to my mental health providers Support Emotional New Me mental health provider helps me with my emotional stability Comments 3 Old As far as society is concerned, people without a mental illn ess are better than people with a mental illness Cohesion Reciprocity New As far as society is concerned, people without a mental illness do more for the community then people with a mental illness Comments 4 Old When I need a shoulder to cry on, I have at least one good friend to turn to Support Emotional New I have at least one good friend to turn to if I am feeling sad or afraid Comments 5 Old I have many friendships with mental health clients Network Homogeneity New Most of my friendships are with mental health clients Comments 6 Old I choose intimate partners partially by how much we have in common Network Homogeneity New I choose my friends by how much we have in common Comments 7 Old I care that the community does well Cohesion Community New I care for my community and try to help it when I can

PAGE 556

537 Comments 8 Old If I were NOT mentally ill, I would care more about things in the community Cohesion Community New Mental illness stops a person from caring about their community Comments 9 Old I believe that most people are basically good Cohesion Trust New I believe that most people are basically trustworthy Comments NEW QUESTIONS 1 New I try to provide for my family as much as they provide for me Network Directionality Comments 2 New My friends give to me to the same degree I give to them Network Directionality Comments 3 New My best friend or partner and I support each other equally Network Directionality Comments 4 New Im one of the first to involve myself with community projects or concerns Cohesion Community Comments

PAGE 557

538 Survey Measuring Percei ved Social Integration Pilot Draft Name (Optional) _____________________________ Birth Date _______________________ Test Date _____________________ MM/DD/YY MM/DD/YY Age in Years _____________________ Gender _______________________ Instructions : This survey is designed to assess how a person feels they fit in society and among their family and friends. Please read each question carefully, and circle the number that best describes how much that statement describes you. Please take your time and answer all questions. If you feel that a question doe s not pertain to you, answer N/A (circle 99) instead of leaving it blank If you have any questions please ask them now. N/A Strongly Disagree Disagree Mildly Disagree Mildly Agree Agree Strongly Agree 1 My best friend (or partner) and I support each other equally 99 0 1 2 3 4 5 2 My friends say nice things to me and help me feel good about myself 99 0 1 2 3 4 5 3 I generally trust my local government to do what is right 99 0 1 2 3 4 5 4 My counselor helps me with my emotional stability 99 0 1 2 3 4 5 5 I have a close relationship with my family 99 0 1 2 3 4 5 6 My friends are often angry with me 99 0 1 2 3 4 5 7 If I make a mistake, my friends point it out so I will correct it 99 0 1 2 3 4 5 8 I have friendships with people who are not mental health clients 99 0 1 2 3 4 5 9 I care for my community and try to help it when I can 99 0 1 2 3 4 5 10 There are only a few people I completely trust 99 0 1 2 3 4 5

PAGE 558

539 N/A Strongly Disagree Disagree Mildly Disagree Mildly Agree Agree Strongly Agree 11 My partner (or best friend) is always telling me what to do 99 0 1 2 3 4 5 12 My friends give to me to the same degree I give to them 99 0 1 2 3 4 5 13 I find I get a new counselor too often 99 0 1 2 3 4 5 14 I have lacked friendships most of my life 99 0 1 2 3 4 5 15 My family meets many of my needs 99 0 1 2 3 4 5 16 My partner (or best friend) doesnt feel I can do anything right 99 0 1 2 3 4 5 17 Most of my friends help me feel good, but help me in many other ways too 99 0 1 2 3 4 5 18 My family spends time with me and my friends 99 0 1 2 3 4 5 19 People are trustworthy, but society is not 99 0 1 2 3 4 5 20 When I need to know something, my friends usually have the answer 99 0 1 2 3 4 5 21 If I am making a mistake, my counselor will always point it out 99 0 1 2 3 4 5 22 My intimate partner (or best friend) helps me in many ways 99 0 1 2 3 4 5 23 Sometimes I feel that my family is too close to me 99 0 1 2 3 4 5 24 I can always turn to my best friend (or partner) for advice if I am confused 99 0 1 2 3 4 5 25 Mental illness stops a person from caring about their community 99 0 1 2 3 4 5 26 I know whom to call if I have a concern about my local government 99 0 1 2 3 4 5 27 I believe that in a pinch, I could trust a stranger to help me 99 0 1 2 3 4 5 28 My counselor talks about his or her life with me 99 0 1 2 3 4 5 29 My family spends time with me and my intimate partner (or best friend) 99 0 1 2 3 4 5

PAGE 559

540 N/A Strongly Disagree Disagree Mildly Disagree Mildly Agree Agree Strongly Agree 30 My partner (or best friend) praises me and cheers for me when I accomplish something 99 0 1 2 3 4 5 31 In general, I have enough friends 99 0 1 2 3 4 5 32 Most people will do the wrong thing if they know they wont be caught 99 0 1 2 3 4 5 33 When I am sad or feeling blue, I can always turn to my family 99 0 1 2 3 4 5 34 I belong to a church or other community group 99 0 1 2 3 4 5 35 I wish my partner (or best friend) would try harder not to hurt my feelings 99 0 1 2 3 4 5 36 I change friends quickly 99 0 1 2 3 4 5 37 I feel criticized when my family says theyre trying to help me 38 I generally trust the federal government to do what is right 99 0 1 2 3 4 5 39 As far as society is concerned, people without a mental illness can do more for the community then people with a mental illness 99 0 1 2 3 4 5 40 I have at least one good friend to turn to if I am f eeling sad or afraid 99 0 1 2 3 4 5 41 My family has had contact with my counselor(s) 99 0 1 2 3 4 5 42 Helping me feel good is what my partner or best friend does best 99 0 1 2 3 4 5 43 Most of my friends are also mental health clients 99 0 1 2 3 4 5 44 My community doesnt care about me, so I stay to myself 99 0 1 2 3 4 5 45 I feel I have a voice in local government 99 0 1 2 3 4 5 46 I have had some of the same friends for many years 99 0 1 2 3 4 5 47 When I do something wrong, my partner or best friend points it out to me 99 0 1 2 3 4 5

PAGE 560

541 N/A Strongly Disagree Disagree Mildly Disagree Mildly Agree Agree Strongly Agree 48 I volunteer my time to organizations when I can 99 0 1 2 3 4 5 49 Sometimes I feel that my counselor doesnt take the time to hear me 99 0 1 2 3 4 5 50 I feel safe to walk in my neighborhood 99 0 1 2 3 4 5 51 When my back is in a corner, I can count on my partner or best friend to support me 99 0 1 2 3 4 5 52 My friends spend time with my intimate partner (or best friend) and I 99 0 1 2 3 4 5 53 I make a point to vote during elections 99 0 1 2 3 4 5 54 My closest relations hips usually last for two years or more 99 0 1 2 3 4 5 55 Sometimes when I need advice or information, friends or family make sure I cant get it 99 0 1 2 3 4 5 56 My counselor knows or has met my intimate partner or best friend 99 0 1 2 3 4 5 57 If I need money or help with a bill, my family almost always gives it to me 99 0 1 2 3 4 5 58 My friends wont leave me alone when I ask them to 99 0 1 2 3 4 5 59 I would say that I have a large family 99 0 1 2 3 4 5 60 I need to be careful because my community wont help me 99 0 1 2 3 4 5 61 Most of the time people are looking out for themselves, not trying to be helpful to others 99 0 1 2 3 4 5 62 I feel that I belong and Im at home in the community 99 0 1 2 3 4 5 63 My family isnt honest with me about what I am doing 99 0 1 2 3 4 5 64 The more I give to the community, the more I want to 99 0 1 2 3 4 5 65 My friends know who my counselors are 99 0 1 2 3 4 5

PAGE 561

542 N/A Strongly Disagree Disagree Mildly Disagree Mildly Agree Agree Strongly Agree 66 In many ways, most of my neighbors are a lot like me 99 0 1 2 3 4 5 67 I keep myself informed on community issues 99 0 1 2 3 4 5 68 If I need to know something, I ask my counselor(s) 99 0 1 2 3 4 5 69 I choose my friends by how much we have in common 99 0 1 2 3 4 5 70 If I need help fixing or making something, my family helps me 99 0 1 2 3 4 5 71 I try to provide for my family as much as they provide for me 72 I am one of the first to contribute to community projects or concerns 99 0 1 2 3 4 5 73 I feel I give back to the community for what I take 99 0 1 2 3 4 5 74 If I need to know something my family usually has the answer 99 0 1 2 3 4 5 75 My counselor tells me too much what to do without fully talking it over with me 99 0 1 2 3 4 5 76 My family comes over or offers advice when I dont want them to 99 0 1 2 3 4 5 77 I know that many of my neighbors are mental health clients 99 0 1 2 3 4 5 78 I generally trust my state government to do what is right 99 0 1 2 3 4 5 79 If I need food, money, or other help, my friends are always there 99 0 1 2 3 4 5 80 My partner gives me things or helps me do things to make my life easier 99 0 1 2 3 4 5

PAGE 562

543 DO NOT WRITE BELOW THIS POINT SNS Score = ____________ MAG Sub-Score = __________ HOM Sub-Score = __________ DIR Sub-Score = __________ CLU Sub-Score = __________ DUR Sub-Score = __________ PSS Score = ____________ EMO Sub-Score = __________ APP Sub-Score = __________ INS Sub-Score = __________ DIS Sub-Score = __________ INF Sub-Score = __________ PSC Score = ____________ COM Sub-Score = __________ TRU Sub-Score = __________ REC Sub-Score = __________ PSI Score = ____________

PAGE 563

544 Appendix C Forms Participant Contact Letter Participant Follow-Up Contact Letter Informed Consent Informed Consent Checklist Compensation Reception Form

PAGE 564

545 Participant Contact Letter August 18, 2005 Dear Sir or Madam: This letter is an invitation to participat e in a research projec t that the Harbor and the University of South Florid a are working on in cooperation. The study is titled Factors Related to Recovery from Mental Illness. Please take a moment to read this letter and to decide if you would like to participate. You are 1 of 350 people being asked to participate. In order for you to participate, it is important for you to unders tand what the study is about and what your part in it will be. First, th e study is looking at how factors related to the individual (e.g., symptoms, diagnosis, hospitalization, experience s of stigma, social support, etc.) and the services they receive (e.g., how satisfied you are, etc.) are related to your recovery. Second, we are interested in what helps you in your recovery process and what you think is important in order to recover. So what do you have to do? If you are interested, you need to contact us through the lead researcher (Gary Walby) at the following number: 727858-3335. Please leave a message if he does not answer. Your primary provider (e.g. therapist, case manager) will know about this project as well, so you can talk to them if you want to. Next, an appointment will be made to meet at the Harbor offices nearest you. You will only have to attend one appointment During this appointment you will have the study explained in more detail, you will review an informed consent so you absolutely understand what is being asked of you, and then you will answer a series of surveys. The total time involved will range from 60-90 minutes. For your time, each individual will be paid $10.00 as a thank you for your effort. At this appointment you will be guided through the process by one of our team members. They include John Adams, Amy Br add, Jessica Burns, Christina Rickus or Gary Walby. We very much hope to hear fr om you and will look forward to our meeting. Sincerely, Gary W. Walby, M.S., M.S.P.H.

PAGE 565

546 Follow-Up Contact Letter Dear Thank you for your interest in the research study titled Factors Related to Recovery from Mental Illness. We have either spoken to you on the phone, or you returned a postcard to us th at your provider provided, or you let your primary provider know that you were interested and they contacted us via the telephone. This letter is to confirm the appointment time and date that was agreed upon in our prior conversation. Appointment Location: ______________________________________________ Appointment Date and Time __________________________________________ As a reminder, you will only have to attend one appointment During this appointment you will have the study explained in more detail, you will review an informed consent so you absolutely understand what is being asked of you and then you will answer a series of surveys. Th e total time involved will range from 60-90 minutes. For your time, each individual wi ll be given $10.00 as a thank you for your effort. If the time, date, and or location of the a ppointment noted above is either incorrect or something has happened that you must chan ge the appointment time, please contact us as soon as possible. First attempt to contac t the person listed below. If they are not available or do not return your call within 48 hours then please contact the lead researcher (Gary Wa lby) at 727-858-3335. Researcher Name ___________________________________ Phone Number _____________________________________ Thank you very much for your interest a nd assistance in this research effort. Sincerely, Gary W. Walby, M.S., M.S.P.H.

PAGE 566

Informed Consent for an Adult Social and Behavioral Sciences University of South Florida Morton Plant Mease BayCare Information for People Who Take Part in Research Studies Researchers at the University of South Fl orida (USF) study many topics. For instance, information on how people cope or recover from mental illness, what factors are important for recovery, and how people feel abou t services for mental illness. To do this, we need the help of people who agree to take part in a research study. Title of research study: Factors Related to Recovery From Mental Illness Person in charge of study: Gary Walby, M.S., M.S.P.H. Study staff who can act on behalf of the person in charge: Jessica Burns, Christina Rickus, John Adams, and Amy Bradd Where the study will be done: The Harbor Behavioral Health Care Institute Should you take part in this study? This form tells you about this research study. You can decide if you want to take part in it. You do not have to take part. Reading this form can help you decide. Before you decide: Read this form. Talk about this study with the person in charge of the study or the person explaining the study. You can have someone with you when you talk about the study. Find out what the study is about. You can ask questions: You may have questions this form does not answer. If you do, ask the person in charge of the study or study staff as you go along. You dont have to guess at things you dont understand. Ask the people doing the study to explain things in a way you can understand. After you read this form, you can: Take your time to think about it. Have a friend or family member read it. Talk it over with someone you trust. 547

PAGE 567

548 Its up to you. If you choose to be in the study, then you can sign the form. If you do not want to take part in this study, do not sign the form. Why is this research being done? The purpose of this study is to find out if individuals believe they will recover from mental illness, what ways they use to recover, and what factors are related to beliefs in recovery and choice of ways to recover. Why are you being asked to take part? We are asking you to take part in this st udy because you are a consumer/client at the Harbor and have been diagnosed with an illness. How long will you be asked to stay in the study? You will be asked to spend part of one da y in the study answering questions. You will have the option to pause answering questions if you choose and start again on a different day. If this happens then you will spend a maximum of two days in the study. How often will you need to come for study visits? A study visit is one you have with the person in charge of the study or study staff. You will need to come for one study visit, or tw o if you decide to pause and resume data collection. After agreeing to be in the study by understanding and signing this form, you will be asked a series of questions from several surveys. Most study visits, and your total time in th is study, will take about 45-90 minutes. At each visit, the person in charge of th e study or staff will assist you in answering several surveys that cover recovery belie fs, symptoms, personal and family history, experiences of stigma or empo werment and related factors. How do you get started? If you decide to take part in this study, you will need to sign this consent form. Also, to make sure that you understand what you are agreeing to, a quick assessment will also be completed to make sure that you know what is being asked of you. This is called an independent capacity assessment Once the consent form is signed, the researcher will begin to ask you questions and provide you with the surveys. It is important that each individual knows th at this form also gives permission for the researcher to look at the Har bor clinical records of the participant in order to get background information of importance (e.g. di agnostic history, hospitalization history, services provided, etc.) Will you be paid for taking part in this study? We will pay you for the time you volunteer in this study. You will be paid $10.00 for your time in answering the survey questions. There is no other compensation available. What will it cost you to take part in this study? There is no cost to you for this research.

PAGE 568

549 What are the potential benefits if you take part in this study? We dont know if you will get any benefits by ta king part in this study. It is hoped that this research will assist in program development to increase the success in recovery, but this may or may not benefit you directly. What are the risks if you take part in this study? There are no known risks to those who take part in this study. What if you get sick or hurt while you are in the study? If you need emergency care: USF does not provide emergency care. Morton Plant Mease Statement: You may be exposed to risk of injury from taking part in this study. In the event of an injury, which you believe is related to this clinical trial, medical care (including hospi talization) is available. Such treatment will be at your expense, the expense of the sponsor, or your medical insurance. Morton Plant Mease Health Care and its hospitals do not have funds available to provide payment for such injuries. You are not giving up any legal rights by signing this document, nor are you releasing Morton Plant Mease Health Care, its hospitals, or doctors, from responsibility for negligence unrelated to the nature and risk of the treatment. Further information on the above, as well as information regarding this research may be obtained from Morton Plant Mease Health Care Department of Clinical Research at (727) 461-8311. Call the person in charge of this study as soon as you can. They will need to know that you are hurt or ill. Call Gary Walby at 727-858-3335 or Emergency Services at 727-841-4455. What will we do to keep your study records private? Federal law requires us to keep your study r ecords private. All study records will have only an identification code other than this consent. All records will be stored in separate locked cabinets in a locked office. All computer records will be password protected. However, certain people may need to see your study records. By law, anyone who looks at your records must keep them confidential. The only people who will be allowed to see these records are: The study staff. People who make sure that we are doing the study in the right way. They also make sure that we protect your rights and safety: o The USF Institutional Review Board (IRB), its staff and other individuals acting on the behalf of USF o Morton Plant Mease Institutional Review Board o The United States Department of Health and Human Services (DHHS) We may publish what we find out from this st udy. Also, we have agreed to inform the Harbor management and clinical staff of th e results of the study to assist them in

PAGE 569

550 developing better recovery progr ams. Whether for publication or the Harbor, we will not use your name or anything else that would let people know who you are. What happens if you decide not to take part in this study? You should only take part in this st udy if you want to take part. If you decide not to take part: You wont be in trouble or lose any rights you normally have. You will still receive the same services at the Harbor you would normally have. What if you join the study and then later decide you want to stop? If you choose to stop while filling out t he surveys, simply tell the researcher and the process will end. You can get the answers to your questions. If you have any questions about this study, call Gary Walby at 727-858-3335. If you have questions about your rights as a person who is taking part in a study, call USF Research Compliance at (813) 974-5638.

PAGE 570

551 Consent to Take Part in this Research Study Its up to you. You can decide if you want to take part in this study. I freely give my consent to take part in this study. I understand that this is research. I have received a copy of this consent form. ________________________ ________________________ ___________ Signature Printed Name Date of Person taking part in study of Person taking part in study Statement of Person Obtaining Informed Consent I have carefully explained to the person taki ng part in the study what he or she can expect. The person who is giving consen t to take part in this study Understands the langu age that is used. Reads well enough to understand this form. Or is able to hear and understand when the form is read to him or her. Does not have any problems that could make it hard to understand what it means to take part in this study. Is not taking drugs that make it hard to understand what is being explained. To the best of my knowledge, when this person signs this form, he or she understands: What the study is about. What needs to be done. What the potential benefits might be. What the known risks might be. That taking part in the study is voluntary. ________________________ ________________________ ___________ Signature of Investigator Printe d Name of Investigator Date or authorized research investigator designated by the Principal Investigator

PAGE 571

552 Informed Consent Checklist Respondent # _______________ Date ______________ Attempt # ______________ Directions : The participant must answer all questi ons correctly and in order to be in the study. If an incorrect answer is given, re-explain the sec tion that was missed and begin the checklist again with a new form. Keep and attach ALL checklists used to the signed informed consent. Do you have the right to say no to this study? Correct _____ Incorrect _____ What is this study about? Correct _____ Incorrect _____ How long will you be asked to be in the study? Correct _____ Incorrect _____ How many times will you need to see someone in the study? Correct _____ Incorrect _____ Are there any potential risks to this study? Correct _____ Incorrect _____ What can happen if you decide not to join the study? Correct _____ Incorrect _____

PAGE 572

553 What happens if you decide to stop being in the study? Correct _____ Incorrect _____ Is there a payment for being in the study and what is it? Correct _____ Incorrect ____ Signature of Researcher__________________________________________

PAGE 573

554 Compensation Receipt Form Participant # ____________________ Gender _________________________ Test Date ____________________________ MM/DD/YY By signing this, I am stating that I have rece ived the compensation for taking part in this study. This compensation is exactly $10.00 for my time in answering the survey questions with the researcher. ___________________________________________ ________________ Participant Name Date and Time ___________________________________________ ________________ Researcher Witness Date and Time

PAGE 574

555 Appendix D Training Material for Research Assistants Training Schedule Training Week 2 Information Training Week 3 Information Psychotropic Medication Handout Medication Tables Scripts: Opening Script Consent Process Script Opening Steps Data Collection Instruments

PAGE 575

556 College of Public Health Recovery Research Team Weekly Training Schedule June 22 July 27 All meeting times are 2:00 p.m. to 5:00 p.m. Week One June 22 Introduction to study Confidentiality and human subjects protection Consent form and procedures Introduction to study forms Dissemination plan Q & A Week Two June 29 Introduction to the study population Diagnoses Study forms in-depth Possible stressful scenarios Adverse events What to do in a crisis Week Three July 6 Medications Services Data collection protocol the nitty gritty No missing data Week Four July 13 Data collection practice No missing data Week Five July 20 Data collection practice No missing data Week Six July 27 Data collection practice No missing data

PAGE 576

557 Recovery Research Training Week 2 June 29, 2005 Introduction to the study population The study population will be comprised of individuals on a continuum of mental illness. The artificial differences between the SMI (severely mentally ill) and OP (outpatient therapy clients) samples may or may not accurate until confirmed via analysis. Thus, we cannot assume that the OP sample will be high f unctioning or that the SMI sample will be low functioning. However, there are differences between individuals with higher levels of severity and those with le ss severity of mental illness. To reiterate, these differences may not accurately distinguish the two samples. These differences will be emphasized via looking briefly at clini cal, social, history, o ccupation and education, and health factors. Clinical Diagnosis : Diagnostically, we can expect some key differences due to severity of mental illness. Individuals in the SMI sample will be required to have a diagnosis of: (1) schizophrenia or a schizophre nia spectrum disorder (e.g. paranoid disorder, delusional disorder, or schizoaffective disorder), (2 ) bipolar disorder (bipolar I or bipolar II, ma nic or depressive dominant, slow or rapid cycling), and (3) major depression, most likely recurrent, with or without psyc hotic features. A fourth category (other) will be used to captu re individuals with features related to the above AND who have been designated severely mentally ill by the partnering agency. Thinking outside the sample inclus ion criteria, it is possible that individuals in the OP sample will also have one of these diagnoses but with the additional designation of in remission or in partial remission. Those with less severe illnesses will be labeled with many possible diagnoses. For instance, dysthymia and post-traumatic stress diso rder are likely, as are other anxiety disorders, somatization disorders, behavior al disorders and dissociative disorders. Codes/diagnoses likely in less severe cases and not in severe cases are adjustment disorders and V-codes. Adjustment disorders are limited, reactive symptom clusters related to some discernible life event. For instance, an adult who is processing through a divorce and becomes c linically depressed at a mild-moderate level may be diagnosed with adjustment disorder with depression. There are also adjustment disorders with anxiety or conduct problems, and several others. V-Codes are definitive of environmental problems, events, or relationship issues that are the focus of treatment. This does not preclude heightened clinical symptoms but the focus will be on the pr esenting issue. For example, V-codes include marital problem, family issues, parent-child problem, etc. Axis 2 diagnoses of personality disorders are likely to be somewhat common. For some less severe individuals these may be the primary diagnosis of clinical interest with a less severe axis one diagnosis as comorbid. For individuals with more serious illnesse s personality disorders ma y be secondary comorbid conditions. Keep in mind that the National Comorbidity Survey found that individuals labeled as SMI have a mini mum of three psychiatric diagnoses in

PAGE 577

558 approximately 93% of cases. This may or may not be reflected in the clinical records but should be kept in mind. Substance abuse diagnoses are allo wed for the study providing that it is not the Axis 1 primary diagnosis. A co morbid abuse or dependence diagnosis does not disqualify an individual. Ho wever, the burgeoning literature on cooccurring disorders strongly suggest that said individua ls have more intractable illnesses and may be experiencing, in general, more psychiatric burden from moment to moment. Impact on Data Collection : Current diagnosis an d diagnostic history will impact data collection indi rectly. Individuals with more serious diagnoses will be at risk for cognitive and emotional processing deficits that may interfere with the pace of data co llection, their ability to maintain attention, and require more assistan ce in understanding the questions. Further, they may become more easily frustrated and may also choose to discontinue the study. The study is focused on recovery from mental illness and having a bias introduced via more severe cases quitting is a potential, and worrisome, occurrence. Ex tra care to establish and maintain rapport is needed with more severe case s. Individuals with diagnoses of borderline, narcissistic, or antisocial personality disorders are more apt to be difficult respondents and these dia gnoses, if offered by the respondent, should heighten the researchers atten tion. The same is true if a cooccurring addictive process or subs tance abuse history are indicated. A very KEY POINT to keep in mind is to never lose sight of the human being behind the diagnosis. Thus terms such as a borderline or a schizophrenic are derogatory a nd will be viewed as insensitive by consumers and providers alike. Symptoms Current : Current symptom levels wi ll be detected by the Symptom Checklist-90R. This instrument will be ad ministered first for three reasons: (1) it asks questions that many of the consumers will be familiar with, easing anxiety. (2) it is the longest instrument and the consumer will feel that they have made immediate headway and will be less resist ant for other instruments and less likely to stop/quit. (3) there are several qu estions that are cues to potential behaviors/problems that can interrupt the collection process and the researcher will be able to mentally mark these and adjust his or her behavior/demeanor accordingly. These questions include: 11. Feeling easily annoyed or irritated 15. Thoughts of ending your life 24. Temper outbursts that you cant control 38. Having to do things very slowly to insure correctness 43. Feeling that you are being watche d or talked about by others\ 45. Having to check or double-check what you do 46. Difficult making decisions 54. Feeling hopeless about the future 55. Trouble concentrating 59. Having thoughts of death or dying

PAGE 578

559 67. Having urges to break or smash things 74. Getting into frequent arguments 78. Feeling so restless you couldnt sit still 81. Shouting or throwing things 84. Having thoughts about sex that bother you a lot Be alert to patterns within these ques tions and be aware that, though the SCL-90R instructs the respondent to indicate thei r experience with these in the last seven days, he or she may be experiencing thes e at the moment of data collection. Impact on Data Collection : Like diagnosis, an elevated symptom picture will influence data collection in many of the same potential ways. Symptoms may interfere with cognitiv e and emotional processing. It is important that if the individual is revealing symptoms or discernibly symptomatic that you stay focused on the task of data collection: maintaining rapport/empathy, pacing the data collection toward its endpoint, staying organized and ensuri ng as neat and complete a process as possible. Always remember that assistance is just a few steps away. Symptoms Prodromal: The developmental period that is non-symptomatic to first symptoms and then age of full onset varies by the individual and can have a great impact on current functioning. Sta tistically, the following nine individual symptoms are considered predictive of the onset of serious mental illness: 1. Reduced attention and concentration 2. Reduced energy, motivation and anergia 3. Depressive mood 4. Sleeping disturbances 5. Anxiety 6. Social withdrawal 7. Mistrust 8. Social dysfunctioning 9. Social constructs (self-esteem, self-efficacy) Other areas that could impact ability to focus on the data collection process include impact on cognitive functioning, focus and atten tion, and the capacity to work independently. The earlier the onset of serious illness the more likely that the brain has received inalte rable structural and neuroche mical changes. In most cases these are subtle and it is difficult to predict the direct impact on data collection. Attention defic it problems are predictive of potential later serious mental illness and may continue into adu lthood. Further, serous illness may deenergize an individual and or impact on their self-efficacy (belief in ability to accomplish something), increasing the time needed to complete tasks. Impact on Data Collection : The above prodromal-to-current developmental factors may slow the da ta collection process. Extra care should be taken to assist the individual in maintaining focus, offering breaks when needed, assisting the respondent in understanding what specific words mean (assisting with vocabulary, not with interpreting the question) and consistently reminding th em that they are in the position of helping us, that we need them, and that they are in control.

PAGE 579

560 Social Social Networks : A social network is the number of social contacts or relationships that an indivi dual has. This includes th e network density (how many individuals the person is in contact with) and composition (who makes up each cluster). Individuals with a more severe illness history will have smaller numbers (less density) and fewer clusters. Increased severity also means that these individuals are much more likely to re ly on family members, other individuals who are mentally ill, and families of origin for social contacts. Individuals Social and Familial Roles : The more severe the illness the greater the probability that the individual will be the identified pati ent (the sick one) in any or all social interactions. This role will in time beco me part of the self-structure. This will have a negative effect on self-esteem and confidence. Family members may lower expectations of the individual when a mental illness arises. It is not uncommon for the family to assume th at all abilities and functioning are compromised. Internalization of lower familial expectations is often followed with lowered expectations of self as well. Further, the emotional expression of family members can range from far too empathic and enmeshed to cold and distant. Internalization and hypersensitiv ity to expressed emotion is relatively common for individuals with severe mental illness. Social Skills, Social Comfort, and Social Resources : The more severe the illness the higher the probability that the individua l will have social skill deficits. This may also extend to being uncomfortable in situations like the re search situation. Keep in mind that many of the individuals with severe mental illness have very limited, often non-existent, money for social events. This level of isolation contributes to their marginalized existence and often increases eagerness for social contact. Hygiene : Individuals with mental illn ess may, usually during periods of decompensation, but not always, have poor hygiene. Individuals who are psychiatrically fragile will be screened from entering the study, decreasing this problem. However, because it may occur, I will be dispensing room fresheners to each member as part of their data collection equipment. Impact on Data Collection : Limited social networks and, possibly, reduced social participat ion may channel the more severe individual into viewing the researcher as a clinician. This in turn could trigger the individual to stray into ar eas of discussion that are c linical in nature. It is imperative at these moments for the re searcher to restate their role and purpose for the respondent. This should be done empathically but directly followed with an immediate return to data collection. Addressing role functions, individuals with mental i llness may adapt the sick role and expect assistance when none is need ed. However, the point of this research is data collection. If the res earcher is asked to read the questions and provide more direct assistance for data collec tion then please do so. Do not become embroiled in a cont rol struggle if you sense that the individual is more capable then he or she lets on. That assumes that they are aware of their capacity and ma y be deliberately downplaying their

PAGE 580

561 abilities. It is equally likely that th ey are responding in habitual patterns or role scripts and data collection is not the point to challenge these. If hygiene is a problem it is important to to lerate this issue, if it occurs, and not to comment on it, as this is a clin ical issue and not a research issue. History Individuals with mental illness have expe rienced varying levels of loneliness and isolation, difficult familial relations and abnormal developmental experiences. Development is influenced by biology and e nvironmental factors. In fact, it is a common belief that for most severe mental illnesses that biology determines onset and environment determines course and (to a great extent) recovery. Two areas of historic concern are abuse and trauma, a nd substance history/self-medication. Abuse and trauma are common experien ces for individuals with mental illness. These experiences are linked to onset, course, symptoms, and recovery. It is very likely that many of the resp ondents will have sexual or physical abuse or assault in their chil dhood and adulthood. Individuals with this history are more likely to have been abused, further comp licating the clinical picture. It is important to remember this potential hist ory from the moment of first contact, even if the actual questions are not until nearly the end of data collection. Finally, not only do many of these in dividuals live ne ar or in poverty, many of them grew up in poverty. Poverty is one of many structural, macro level risk factors for mental illness. Indeed, th e lower SES is a risk for earlier onset and poor course of illness. Impact on Data Collection : These varying historical factors will have many of the same effects as previous ly noted. Together, they indicate both vulnerability and an uncommon resilience. It cannot be assumed individuals with mental illness are un iformly fragile. It also cannot be assumed that each individual is well along on their recovery journey or that they even believe in the concep t of recovery. Thus, each research team member will need to understand their beliefs about recovery before beginning any data collection and re main as objective, empathic and pleasant as possible du ring data collection. Occupation and Education More severely ill individuals (currently and/or historically) are less likely to be employed or to have a successful history of competitive employment. Similarly, the more severely ill are less likely to ha ve college experience and are more likely to not have graduated from high school. Implications for Data Collection : With a potential poor history of work and education, it is imperative to connect with the humanity of the respondent and to actively work against be ing viewed as superior or better. This can be difficult as consumers are used to adopting a general subservience around individuals with considerable education. It is important to engage them in a collabor ative process where the consumer is helping the researcher. There is evidence in the literature that this

PAGE 581

562 improves effort and thus accuracy of research data and, more importantly, it is compatible with the public health values of equity and social justice. Health Mental illness has a profound affect on perceived health, years lived with morbidity and disability, a nd reduction in life expectan cy. General health is associated with quality of life and both are depressed due to mental illness. The more severe the history of mental illness the more chronic health conditions potentially afflict the individual. Individuals with severe mental illness are at a much higher risk for coronary conditions, ne oplasm, diabetes and other illnesses. Further, for some illnesses (e.g. schizophrenia) there is an increased risk of medical conditions due to the psychiat ric medications (e.g. diabetes due to atypical antipsychotics). Individuals with severe mental illness are also at a higher risk for blood-borne illnesses (e.g. hepatitis B and C, and HIV) compared to the non-mentally ill population. Fina lly, mental illness is often accompanied by poor health habits (e.g. smoking and dr inking) as well as poor nutrition and inactive lifestyles. Part of the bad habits is learned a nd part stem from lack of resources. Interestingly, nicotine has de monstrated an anti-hallucinatory affect, which may account in part for the number of individuals with schizophrenia who are nicotine addicted Implications for Data Collection : Mental illness takes a toll not only mentally but also physically. Many i ndividuals interviewed might be in poor condition and suffering from physical illnesses. This may slow response time, necessitate additional breaks, and disrupt concentration. The individual may be thinking about these problems to a degree and some questions may initiate discussion of these problems. It is important to be polite, listen, and steer the conv ersation and focus back to the data collection task. Possible stressful/frustrating scenarios The following are a sample of potentially frustrating experiences or habits/behaviors of the respondent. If ot her situations arise, please contact me directly and immediately. A list will be kept to update all team members. 1. Argumentative respondent : Some respondents may argue for the sake of arguing or have a disorder that is characterized with a difficult interpersonal style (e.g. borderline or antisocial personality disorders). For these individuals the researcher is asked to remain calm, do not respond negatively, defuse through low tones and possibly a positive use of humor, and stay focused on the data collection task. 2. Frequent breaks/poor attention : This is more likely for individuals with severe illnesses who have a legitimate reason for needing to break or who may be enjoying the attention and are se eking to prolong the encounter. Less severe individuals may have other items on their agenda and will complete the process with few to no interruptions. 3. Slow responders : Some individuals may spend more time than others over some or all questions. If this is a function of cognitive deficit then no amount

PAGE 582

563 of encouragement will speed them up. However, if the individual appears to be spending too much time trying to find the right answer, then the researcher can encourage them by pointing out that their first is nearly always the one that fits them best and that th ere is no right answ er to any of these questions. 4. Consistent misinterpret ation or misunderstanding of question content : For some individuals this is a consequen ce of their illness and the researcher should handle this with pa tience and compassion. The researcher needs to be careful to alter the meaning of the ques tion when explaining it or to map his or her own experience onto the question, as this will likely be confusing to the respondent. As much as possible, gi ve meaning to misunderstood words only. 5. Wanting to bring a friend or fam ily member to the data collection appointment : This should be discouraged as much as possible. Simply stating that this is not a possibility will de ter most attempts without losing the respondent. However, if the responden t insists and there is a danger of nonparticipation due to this, then one person may accompany with the following rules: They do not participate or offer any responses. If they do so they must be warned once and asked to leave on th e second infraction. If the respondent asks the person that accompanies them for an answer or an opinion, the researcher will interrupt this and will use the same guidelines as just noted (one warning then asking the accompanier to leave). 6. Respondent who makes personal inquiries : Many individuals with mental illness are lonely and isolated. They are curious about your life and may form a relatively quick attachment to you as we ll. Inquiries will range from jobs and hobbies to relationships and children. This may include your status as a prelude to asking you out or inquiring if they could see you again some time. At all times the researcher will main tain their respect for the individual, remembering that the researcher is of ten in a far more stable and productive situation. It is recommended that the re searcher reveal as little about him or herself as possible as the goal is a one-time data collection and not a social interaction. Confirming that the researcher is in a relationship in a matter-offact tone with an immediate return to data collection will often defuse other questions. 7. Flirting or sexually suggestive behavior : This could range from mildly suggestive and even humorous to quite ex plicit and uncomfortab le. The latter will, in all likelihood, neve r occur. However, if the researcher becomes uncomfortable it is suggested that they directly state this to the participant and ask them to return to task. If an unc omfortable situation remains please feel free to end the data collection, pay the incentive, and ask the person to leave. 8. Do not keep appointments or frequent reschedules : Unfortunately, not keeping appointments is common, especi ally for those with a SMI. Each participant will be able to reschedule one time only. If they do not attend the rescheduled appointment th an this should be communicated to me, and he or she will be dropped from the study. 9. Wanting to quit early and still receive the incentive : This will be seldom, if ever, I hope. If there is a legitimate reason (e.g. forgot another appointment)

PAGE 583

564 then ask them to reschedule and state that the incentive is part of the contract (the consent) that they will receive after putting forth his or her agreed upon effort. If they are obviously not co ming back or interested then do not become involved in a confrontation. If this does occur, pay the incentive and let them leave. Adverse events/What to do in a crisis For the following adverse events (more se rious than stressful or frustrating situations) please follow the pr ocedures as outlined below: If the participant becomes agitated or belligerent 1. Tell them that they appear frus trated and a break is in order 2. Leave the area 3. Locate or call the pr ovider contact person 4. Have them come into the session with you and inform the participant that the session is over 5. Do NOT reschedule 6. Pay the incentive only if they pause to ask. Do not interrupt their leaving to offer the incentive. It is important for you to separate yourself from them. If they inquire later with their primary provider about the incentive then it will be paid through the provider. If the participant beco mes sad/weepy/distraught 1. Ask them if they want a break 2. If a break appears in order then explicitly state a break will be taken 3. Ask them if they wish to continue with data collection 4. If they continue to be distraught then find the provider contact person 5. Inform the contact person what is happening and ask them to come to the session 6. Reschedule to finish data collection if possible If the participant states that they are feeling harm ful to self or others 1. Tell them that you are concerned an d that you need to speak to the provider contact pers on at your location 2. If the participant attempts to leave do NOT block their path 3. Immediately contact the provider an d inform them of EXACTLY what the participant stated 4. Follow the directions of the provider Study forms in-depth Word by word Question by question Line by line

PAGE 584

565 Recovery Research Training Week 3 June 6, 2005 1. Finish critique/review of instruments 2. Medications See handouts 3. Harbor Services See handout 4. Data collection step by step Introduction Greetings/making the re spondent comfortable General description of the study See document titled Opening Script Consent Process Hand copy of consent to the respondent Describe the consent procedure Read through the consent with the respondent Administer the Independ ent Capacity Assessment Have respondent sign TWO (2) copies of the consent Witness/sign both copies Retain one copy and give the other copy to the respondent See document titled Consent Process Data Collection Process Explain the process of using surveys Briefly overview what the information generally covers Ask if they are comfortable or need a break before beginning the collection process See document titled Data Co llection Opening Steps Data Collection Procedure Follow the order of the instruments in the same order they are distributed Introduce each survey before using it never move to a survey without introduction Read the instructions from the top of each survey See document titled Data Co llection Instruments Incentive Procedure After the final survey is completed introduce the Incentive Receipt Form Have the respondent sign it and then witness it Pay the incentive Say good bye (nicely)! See document titled Incentive Procedure

PAGE 585

566 Psychotropic Medications Overview and Cl assification of Four Types for Data Collection Symptom Relief, Not Cure Just as aspirin can reduce a fever without clearing up the infection that causes it, psychotherapeutic medications act by controlling symptoms. Like most drugs used in medicine, they correct or compensa te for some malfunction in the body. Psychotherapeutic medications do not cure ment al illness, but they do lessen its burden. In many cases, these medications can help a person get on with life despite some continuing mental pain and difficulty copi ng with problems. For example, drugs like chlorpromazine can turn off the "voices" he ard by some people with schizophrenia and help them to perceive reality more accurately. And antidepressants can lift the dark, heavy moods of depression. The degree of response ranging from little relief of symptoms to complete remission depends on a va riety of factors related to the individual and the particular diso rder being treated. How long someone must take a psychotherap eutic medication depends on the disorder. Many depressed and anxious people may need medication for a single period perhaps for several months and then never have to ta ke it again. For some conditions, such as schizophrenia or manic-depressive illness, me dication may have to be taken indefinitely or, perhaps, intermittently. Like any medication, psychotherapeutic medi cations do not produce the same effect in everyone. Some people may respond better to one medication than another. Some may need larger dosages than others do. Some e xperience annoying side effects, while others do not. Age, sex, body size, body chemistry, physical illnesses and their treatments, diet, and habits such as smoking, are some of th e factors that can influence a medication's effect. Antianxiety Medications Everyone experiences anxiety at one time or another "butterf lies in the stomach" before giving a speech or sweaty palms during a job interview are co mmon symptoms. Other symptoms of anxiety include irritability, uneas iness, jumpiness, feelings of apprehension, rapid or irregular heartbeat, stomachache, nausea, faintness, and breathing problems. Anxiety is often manageable and mild. But sometimes it can present serious problems. A high level or prolonged state of anxiety can be very incapacitating, making the activities of daily life difficult or impossible. Besides generalized an xiety, other anxiety disorders are panic, phobia, obsessive-compulsive disorder (OCD), and posttraumatic stress disorder. Phobias, which are persistent, irrational f ears and are characterized by avoidance of certain objects, places, and things, sometime s accompany anxiety. A panic attack is a severe form of anxiety that may occur s uddenly and is marked with symptoms of

PAGE 586

567 nervousness, breathlessness, pounding heart, and sweating. Sometimes the fear that one may die is present. Antianxiety medications help to calm and relax the anxious person and remove the troubling symptoms. There are a number of an tianxiety medications currently available. The preferred medications for most anxiety disorders are th e benzodiazepines. In addition to the benzodiazepines, a non-benzodiazepine, buspirone (BuSpar), is used for generalized anxiety disorders. Antidepressants are also effective for panic attacks and some phobias and are often prescribed for these conditions. They are also sometimes used for more generalized forms of anxiety, especi ally when it is accompanied by depression. The medications approved by the FDA for use in OCD are all antidepressants clomipramine, fluoxetine, and fluvoxamine. The most commonly used benzodiazepines are alprazolam (Xanax) and diazepam (Valium), followed by chlordiazepoxide (Librium, Librax, Libritabs). Benzodiazepines are relatively fast-acting medications; in contra st, buspirone must be taken daily for 2 or 3 weeks prior to exerting its antianxiety eff ect. Most benzodiazepines will begin to take effect within hours, some in even less tim e. Benzodiazepines differ in duration of action in different individuals; they may be taken two or three times a day, or sometimes only once a day. Dosage is generally started at a low level and gradually raised until symptoms are diminished or removed. The dosage will vary a great deal depending on the symptoms and the individual's body chemistry. Benzodiazepines have few side effects. Drow siness and loss of c oordination are most common; fatigue and mental slowing or confus ion can also occur. These effects make it dangerous to drive or operate some machiner y when taking benzodiazepines especially when the patient is just beginning trea tment. Other side effects are rare. Benzodiazepines combined with other medicatio ns can present a problem, notably when taken together with commonly used substances su ch as alcohol. It is wise to abstain from alcohol when taking benzodiazepines, as the interaction between benzodiazepines and alcohol can lead to serious and possibly lif e-threatening complications. Following the doctor's instructions is important. The doctor should be informed of all other medications the patient is taking, including over-the-counter preparations Benzodiazepines increase central nervous system depression when combined with alcohol, anesthetics, antihistamines, sedatives, muscle relaxant s, and some prescription pain medications. Particular benzodiazepines may influence th e action of some anticonvulsant and cardiac medications. Benzodiazepines have also been associated with abnormalities in babies born to mothers who were taking these medications during pregnancy. With benzodiazepines, there is a potential for the development of tolerance and dependence as well as the po ssibility of abuse and wit hdrawal reactions. For these reasons, the medications are generally prescrib ed for brief periods of time days or weeks and sometimes intermittently, for stressful situ ations or anxiety attacks. For the same reason, ongoing or continuous treatment with benzodiazepines is not recommended for most people. Some patients may, howe ver, need long-term treatment.

PAGE 587

568 A withdrawal reaction may occur if the tr eatment is abruptly stopped. Symptoms may include anxiety, shakiness, headache, dizzine ss, sleeplessness, loss of appetite, and, in more severe cases, fever, seizures, and psychosis. A withdrawal reaction may be mistaken for a return of the anxiety, since ma ny of the symptoms are similar. Thus, after benzodiazepines are taken for an extended pe riod, the dosage is gradually tapered off before being completely stopped. Although benzodiazepines, buspirone, tricyclic antidepressants, or SSRIs are the preferred medications for most anxiety disorders, occasionally, for specific reasons, one of the following medications may be pr escribed: antipsychotic medications; antihistamines (such as Atarax, Vistaril, a nd others); barbiturates such as phenobarbital; and beta-blockers such as propranolol (Inde ral, Inderide). Propanediols such as meprobamate (Equanil) were commonly pres cribed prior to the introduction of the benzodiazepines, but toda y rarely are used. Antidepressant Medications The kind of depression that will most likely benefit from treatment with medications is more than just "the blues." It's a condition that's prolonged, lasting 2 weeks or more, and interferes with a person's ability to carry on daily tasks and to enjoy activities that previously brought pleasure. The depressed person will seem sad, or "down," or may show a lack of interest in his surroundings. He may have trouble eating and lose weight (although some people eat more and gain weight when depressed). He may sleep too much or too little, have difficulty going to sleep, sleep restlessly, or awaken very early in the morning. He may speak of feeling guilty, worthless, or hopele ss. He may complain that his thinking is slowed down. He may lack energy, feeling "e verything's too much," or he might be agitated and jumpy. A person who is depresse d may cry. He may think and talk about killing him or herself and may even make a suicide attempt. Some people who are depressed have psychotic symptoms, such as delusions (false ideas) that are related to their depression. For instance, a psychotically depressed person might imagine that he is already dead, or "in hell," being punished. Not everyone who is depressed has all these symptoms, but everyone who is depressed has at least some of them. A depression can range in intensity from mild to severe. Antidepressants are used most widely for se rious depressions, but they can also be helpful for some milder depressions. Antidep ressants, although they are not "uppers" or stimulants, take away or reduce the sympto ms of depression and help the depressed person feel the way he did before he became depressed. Antidepressants are also used for disorders characterized principally by anxiety. They can block the symptoms of panic, including rapid heartbeat, terror, dizziness, chest pains, nausea, and breathing problems. They can also be used to treat some phobias.

PAGE 588

569 The physician chooses the particular antidepress ant to prescribe based on the individual patient's symptoms. When someone begins taking an antidepressant, improvement generally will not begin to show immediatel y. With most of these medications, it will take from 1 to 3 weeks before changes begin to occur. Some symptoms diminish early in treatment; others, later. For instance, a person' s energy level or sleeping or eating patterns may improve before his depressed mood lifts. If there is little or no change in symptoms after 5 to 6 weeks, a different medication may be tried. Some people will respond better to one than another. Since there is no certain way of determining beforehand which medication will be effective, the doctor may have to prescribe first one, then another, until an effective one is found. Treatment is continued for a minimum of several months and may last up to a year or more. While some people have one episode of depr ession and then never have another, or remain symptom-free for years, others have more frequent episodes or very long-lasting depressions that may go on for years. Some people find that their depressions become more frequent and severe as they get olde r. For these people, continuing (maintenance) treatment with antidepressants can be an effective way of reducing the frequency and severity of depressions. Those that are commonly used have no known long-term side effects and may be continued indefinitely. Th e prescribed dosage of the medication may be lowered if side effects become troublesome. Lithium can also be used for maintenance treatment of repeated depressions whether or not there is evidence of a manic or maniclike episode in the past. Dosage of antidepressants varies, depending on the type of drug, the person's body chemistry, age, and, sometimes, body weight. Dosages are generally started low and raised gradually over time until the desired effect is reached wit hout the appearance of troublesome side effects. There are a number of antidepressant medicatio ns available. They differ in their side effects and, to some extent, in their level of effectiveness. Tricyclic antidepressants (named for their chemical structure) are more commonly used for treatment of major depressions than are monoamine oxidase inhibitors (MAOIs); but MAOIs are often helpful in so-called "atypical" depressions in which there are symptoms like oversleeping, anxiety, pani c attacks, and phobias. The last few years have seen the introduc tion of a number of new antidepressants. Several of them are called "selective sero tonin reuptake inhibitors" (SSRIs). Those available at the present time in the United States are fluoxetine (Prozac), fluvoxamine (Luvox), paroxetine (Paxil), and sertraline (Zoloft). (Luvox has been approved for obsessive-compulsive disorder and Paxil ha s been approved for panic disorder.) Though structurally different from each other, all the SSRIs' antidepressant effects are due to their action on one specific neurotransmitter, serotonin. The FDA has also approved two other antidepressants that affect two neurotransmitters: serotonin and norepinephrine. They are venlafaxine (Effexor) and nefazodone (Serzone). All of these newer antidepressants seem to have less bothersome side effects than the older tricyclic antidepressants.

PAGE 589

570 The tricyclic antidepressant clomipramine (A nafranil) affects serotonin but is not as selective as the SSRIs. It was the first me dication specifically approved for use in the treatment of obsessivecompulsive disord er (OCD). Prozac and Luvox have now been approved for use with OCD. Another of the newer antidepressants, bupropion (Wellbutrin), is chemically unrelated to the other antidepressants. It has more e ffect on norepinephrine and dopamine than on serotonin. Wellbutrin has not been associated with weight gain or sexual dysfunction. It is contraindicated for individuals with, or at ri sk for, a seizure disorder or who have been diagnosed with bulimia or anorexia nervosa. Side Effects of Antidepressant Medications 1. Tricyclic Antidepressants There are a number of possibl e side effects with tricycli c antidepressants that vary, depending on the medication. For example, am itriptyline (Elavil) ma y make people feel drowsy, while protriptyline (Vivactil) hardly does this at all and, in some people, may have an opposite effect, produci ng feelings of anxiety and re stlessness. Because of this kind of variation in side effects, one antidepressant might be highly desirable for one person and not recommended for another. Tric yclics on occasion may complicate specific heart problems, and for this reason the physicia n should be aware of all such difficulties. Other side effects with tricyclics may in clude blurred vision, dry mouth, constipation, weight gain, dizziness when changing position, increased sweating, difficulty urinating, changes in sexual desire, decrease in se xual ability, muscle tw itches, fatigue, and weakness. Not all these medications produce all side effects, and not everybody gets them. Some will disappear quickly, while others may remain for the length of treatment. Some side effects are similar to symptoms of depression (for instance, fatigue and constipation). For this reason, the patient or family should discuss all symptoms with the doctor, who may change the medication or dosage. Tricyclics also may interact with thyroid hormone, antihype rtensive medications, oral contraceptives, some blood coagulants, so me sleeping medications, antipsychotic medications, diuretics, antihistamines, aspiri n, bicarbonate of soda, vitamin C, alcohol, and tobacco. An overdose of antidepressants is serious a nd potentially lethal. It requires immediate medical attention. Symptoms of an overdos e of tricyclic antidepressant medication develop within an hour and may start with rapid heartbeat, dilated pupils, flushed face, and agitation, and progress to confusion, loss of consciousness, seiz ures, irregular heart beats, cardiorespiratory collapse, and death. 2. The Newer Antidepressants

PAGE 590

571 The most common side effects of these antide pressants are gastroin testinal problems and headache. Others are insomnia, anxiety, and agitation. Because of potentially serious interaction between these medications and m onoamine oxidase inhib itors, it is advisable to stop taking one medication from 2 to 4 or 5 weeks before starting the other, depending on the specific medications involved. In addi tion, some SSRIs have been found to affect metabolism of certain other medications in the liver, creating possible drug interactions. 3. Monoamine Oxidase Inhibitors (MAOIs) MAOIs may cause some side effects similar to those of the other antidepressants. Dizziness when changing position and rapid heartbeat are common. MAOIs also react with certain foods and alcoholic beverage s (such as aged cheeses, foods containing monosodium glutamate (MSG), Chianti and other red wines), a nd other medications (such as over-the-counter cold and allergy preparations, loca l anesthetics, amphetamines, insulin, some narcotics, and antiparkinsonian medications). These reactions often do not appear for several hours. Signs may include severe high blood pressure, headache, nausea, vomiting, rapid heartbeat, possible confusion, psychotic symptoms, seizures, stroke, and coma. For this reason, people taki ng MAOIs must stay aw ay from restricted foods, drinks, and medications. They should be sure that they are furnished by their doctor or pharmacist, a list of all foods, beve rages, and other medica tions that should be avoided. Precautions to be Observed When Taking Antidepressants When taking antidepressants, it is important to tell all doctors (a nd dentists) being seen not just the one who is treati ng the depression about all medications being used, including over-the-counter preparations and alcohol. Antidepressants should be taken only in the amount prescribed and should be kept in a secure place away from children. When used with proper care, following doctors' instructi ons, antidepressants are extremely useful medications that can reverse the misery of a depression and help a person feel like him or herself again. Antimanic Medications Bipolar disorder (manic-depressive illness) is characterized by cycling mood changes: severe highs (mania) and lows (depression) Cycles may be predominantly manic or depressive with normal mood between cycles Mood swings may follow each other very closely, within hours or days, or may be sepa rated by months to years. These "highs" and "lows" may vary in intensity and severity. When someone is in a manic "high," he may be overactive, over-talkative, and have a great deal of energy. He will switch quickly from one topic to another, as if he cannot get his thoughts out fast enough; his attention span is often s hort, and he can easily be distracted. Sometimes, the "high" person is ir ritable or angry and ha s false or inflated ideas about his position or impor tance in the world. He may be very elated, full of grand

PAGE 591

572 schemes that might range from business deal s to romantic sprees. Often, he shows poor judgment in these ventures. Mania, untrea ted, may worsen to a psychotic state. Depression will show in a "low" mood, lack of energy, changes in eating and sleeping patterns, feelings of hopelessn ess, helplessness, sadness, wo rthlessness, and guilt, and sometimes thoughts of suicide. Lithium The medication used most often to combat a manic "high" is lithium. It is unusual to find mania without a subsequent or preceding pe riod of depression. Lithium evens out mood swings in both directions, so th at it is used not just for acute manic attacks or flare-ups of the illness, but also as an ongoing treatment of bipolar disorder. Lithium will diminish severe manic symptoms in about 5 to 14 days, but it may be anywhere from days to several months until the condition is fully controlled. Antipsychotic medications are sometimes used in the first several days of treatment to control manic symptoms until the lithium begins to take effect. Likewise, antidepressants may be needed in addition to lithium during the depressive phase of bipolar disorder. Someone may have one episode of bipolar disorder and never ha ve another, or be free of illness for several years. However, for those who have more than one episode, continuing (maintenance) treatment on lithium is us ually given serious consideration. Some people respond well to maintenance treat ment and have no further episodes, while others may have moderate mood swings that lessen as treatment continues. Some people may continue to have episodes that are diminished in frequency and severity. Unfortunately, some manic-depressive patients may not be helped at all. Response to treatment with lithium varies and it cannot be determined beforehand who will or will not respond to treatment. Regular blood tests are an important part of treatment with lithium. A lithium level must be checked periodically to measure the amount of the drug in the body. If too little is taken, lithium will not be effective. If too much is taken, a variety of side effects may occur. The range between an effective dose and a toxic one is small. A lithium level is routinely checked at the beginning of treatm ent to determine the best lithium dosage for the patient. Once a person is stable and on ma intenance dosage, a lithium level should be checked every few months. How much lithium a person needs to take may vary over time, depending on how ill he is, his body chemistry, and his physical condition. Anything that lowers the level of sodium (t able salt is sodium chloride) in the body may cause a lithium buildup and lead to toxicity. Reduced salt intake, heavy sweating, fever, vomiting, or diarrhea may do this. An unusual amount of exercise or a switch to a lowsalt diet is examples. It's important to be aware of conditions that lower sodium and to share this information with the doctor. The lithium dosage may have to be adjusted.

PAGE 592

573 When a person first takes lithium, he may experience side effects, such as drowsiness, weakness, nausea, vomiting, fatigue, hand tremor, or increased thirst and urination. These usually disappear or subside quickly, although hand tremor may persis t. Weight gain may also occur. Dieting will help, but crash diets should be avoi ded because they may affect the lithium level. Drinking low-calorie or no-calorie beverages will help keep weight down. Kidney changes, accompanied by increa sed thirst and urination, may develop during treatment. These conditions that ma y occur are generally manageable and are reduced by lowering the dosage. Because lithiu m may cause the thyroid gland to become underactive (hypothyroidism) or sometimes enlarged (goiter), thyroid function monitoring is a part of the therapy. To restore normal thyroid function, thyroid hormone is given along with lithium. Because of possible complications, lithium ma y either not be recommended or may be given with caution when a person has exis ting thyroid, kidney, or heart disorders, epilepsy, or brain damage. Women of child-b earing age should be aware that lithium increases the risk of congenital malformations in babies born to women taking lithium. Special caution should be taken during the first 3 months of pregnancy. Lithium, when combined with certain othe r medications, can have unwanted effects. Some diuretics substances that remove wate r from the body increase the level of lithium and can cause toxicity. Other diuretics, like co ffee and tea, can lower the level of lithium. Signs of lithium toxicity may include naus ea, vomiting, drowsiness, mental dullness, slurred speech, confusion, dizziness, muscle twitching, irregular he artbeat, and blurred vision. A serious lithium overdose can be lifethreatening. Someone w ho is taking lithium should tell all the doctors incl uding dentistshe sees about all other medications he is taking. With regular monitoring, lithium is a safe and effective drug that enables many people, who otherwise would suffer from incapacitati ng mood swings, to lead normal lives. Anticonvulsants Not all patients with symptoms of mania benefit from lithium. Some have been found to respond to another type of me dication, the anticonvulsant me dications that are usually used to treat epilepsy. Carbamazepine (Tegretol) is the anticonvulsant that has been most widely used. Manic-depressive patients who cycle rapidly that is, they change from mania to depression and back again over the course of hours or days, rather than months seem to respond particularly well to carbamazepine. Early side effects of carbamazepine, a lthough generally mild, include drowsiness, dizziness, confusion, disturbed vision, perceptual distortio ns, memory impairment, and nausea. They are usually transient and often respond to temporary dosage reduction. Another common but generally mild adverse eff ect is the lowering of the white blood cell count which requires periodic blood tests to moni tor against the rare possibility of more serious, even life-threatening, bone marrow depr ession. Also serious are the skin rashes

PAGE 593

574 that can occur in 15 to 20 percent of patient s. These rashes are sometimes severe enough to require discontinuatio n of the medication. In 1995, the anticonvulsant divalproex sodium (Depakote) was approved by the Food and Drug Administration for manic-depressive illne ss. Clinical trials have shown it to have effectiveness in controlling manic symptoms e quivalent to that of lithium; it is effective in both rapid-cycling and non-rapid-cycling bipolar. Though divalproex can cause gastrointestinal side effects, the incidence is low. Other adverse effects occasionally reported are headache, double vi sion, dizziness, anxiety, or confusion. Because in some cases divalproex has caused liver disfunction, liver function tests should be performed prior to therapy and at frequent intervals thereafter, particularly during the first six months of therapy. Antipsychotic Medications A person who is psychotic is out of touch w ith reality. He may "hear voices" or have strange and untrue ideas (for example, thinki ng that others can hear his thoughts, or are trying to harm him, or that he is the President of the United States or some other famous person).* He may get excited or angry for no a pparent reason, or spend a lot of time off by himself, or in bed, sleeping during the day and staying awake at night. He may neglect his appearance, not bathing or changing clothes, and may become difficult to communicate with saying things that make no sense, or barely talking at all. These kinds of behaviors are symptoms of ps ychotic illness, the principal form of which is schizophrenia. All of the symptoms may not be present when someone is psychotic, but some of them always are. Antipsychotic medications, as their name suggests, act against these symptoms. These medications cannot "cur e" the illness, but they can take away many of the symptoms or make them milder. In some cases, they can shorten the course of the illness as well. There are a number of antipsychotic (neuroleptic) medications available. They all work; the main differences are in the potency that is, the dosage (amount) pr escribed to produce therapeutic effects and the side effects. Some people might think that the higher the dose of medication, the more seri ous the illness, but this is not always true. A doctor will consider several factors when prescribing an antipsychotic medication, besides how "ill" someone is. These include the patient's age, body weight, and type of medication. Past history is im portant, too. If a person took a particular medication before and it worked, the doctor is likely to pres cribe the same one again. Some less potent drugs, like chlorpromazine (Thorazine), are prescribed in higher numbers of milligrams than others of high potency, like haloperidol (Haldol). If a person has to take a large amount of a "high-dose" antipsychotic medication, such as chlorpromazine, to get the same effect as a small amount of a "low-dose" medication, such as haloperidol, why doesn't the doctor just prescribe "low-dose" medications? The

PAGE 594

575 main reason is the difference in their side eff ects (actions of the medication other than the one intended for the illness). These medicati ons vary in their side effects, and some people have more trouble with certain side effects than others. A side effect may sometimes be desirable. For instance, th e sedative effect of some antipsychotic medications is useful for pa tients who have trouble sleepi ng or who become agitated during the day. Unlike some prescription drugs, which must be taken several times during the day, antipsychotic medications can usually be ta ken just once a day. Thus, patients can reduce daytime side effects by taking the medicati ons once, before bed. Some antipsychotic medications are available in forms that can be injected once or twice a month, thus assuring that the medicine is being taken reliably. Most side effects of antipsychotic medications are mild. Many common ones disappear after the first few weeks of treatment. Thes e include drowsiness, rapid heartbeat, and dizziness when changing position. Some people gain weight while taking antipsy chotic medications and may have to change their diet to control their weight. Other side effects th at may be caused by some antipsychotic medications include decrease in sexual ability or interest, problems with menstrual periods, sunburn, or skin rashes. If a side effect is especially troublesome, it should be discussed with the doctor who may prescribe a diff erent medication, change the dosage level or schedule, or prescribe an additional medication to control the side effects. Movement difficulties may occur with the use of antipsychotic medications, although most of them can be controlled with an anticholinergic medication. These movement problems include muscle spasms of the neck, eye, back, or other muscles; restlessness and pacing; a general slowi ng-down of movement and spee ch; and a shuffling walk. Some of these side effects may look like psyc hotic or neurologic (Parkinson's disease) symptoms, but aren't. If they are severe, or persist with continued treatment with an antipsychotic, it is important to notify the doc tor, who might either change the medication or prescribe an additional one to control the side effects. Just as people vary in their responses to anti psychotic medications, they also vary in their speed of improvement. Some symptoms dimini sh in days, while others take weeks or months. For many patients, substantial impr ovement is seen by the sixth week of treatment, although this is not true in ev ery case. If someone does not seem to be improving, a different type of medication may be tried. Even if a person is feeling better or complete ly well, he should not just stop taking the medication. Continuing to see the doctor wh ile tapering off medi cation is important. Some people may need to take medication for an extended period of time, or even indefinitely. These people usually have ch ronic (long-term, con tinuous) schizophrenic disorders, or have a history of repeated sc hizophrenic episodes, a nd are likely to become ill again. Also, in some cases a person who ha s experienced one or two severe episodes may need medication indefinitely. In these ca ses, medication may be continued in as low

PAGE 595

576 a dosage as possible to maintain control of symptoms. This approach, called maintenance treatment, prevents relapse in many people a nd removes or reduces symptoms for others. While maintenance treatment is helpful for many people, a drawback for some is the possibility of developing long-term side e ffects, particularly a condition called tardive dyskinesia. Involuntary movements charac terize this condition. These abnormal movements most often occur around the mouth, but are sometimes seen in other muscle areas such as the trunk, pelvis, or diaphrag m. The disorder may range from mild to severe. For some people, it cannot be reve rsed, while others recover partially or completely. Tardive dyskinesia is seen mo st often after long-term treatment with antipsychotic medications. There is a higher incidence in women, w ith the risk rising with age. There is no way to determine whet her someone will develop this condition, and if it develops, whether the patient will recover. At present, there is no effective treatment for tardive dyskinesia. The possible risks of long-term treatment with antipsychotic medications must be weighed against the be nefits in each individual case by patient, family, and doctor. Antipsychotic medications can produce unwanted effects when taken in combination with other medications. Therefore, the doctor shou ld be told about all medicine being taken, including over-the-counter preparations, and the extent of the use of alcohol. Some antipsychotic medications interfere with the action of antihypertensive medications (taken for high blood pressure), anticonvulsant s (taken for epilepsy), and medications used for Parkinson's disease. Some antipsy chotic medications add to the effects of alcohol and other central ne rvous system depressants, such as antihistamines, antidepressants, barbiturates, some sleepi ng and pain medications, and narcotics. Atypical neuroleptics In 1990, clozapine (Clozaril), an "atypical neuroleptic," was introduced in the United States. In clinical trials, this medication was found to be mo re effective than traditional antipsychotic medications in individuals with treatment-resistant schizophrenia, and the risk of tardive dyskinesia is lower. However, because of the potential side effect of a serious blood disorder, agranulocytosis, patients who are on clozapine must have a blood test each week. The expense involved in this monitoring, together with the cost of the medication, has made maintenance on cl ozapine difficult for many persons with schizophrenia. However, 5 years after its introduction in the United States, approximately 58,000 persons were being treated with clozapine. Since clozapine's approval in the United St ates, other atypical ne uroleptics (also called atypical antipsychotics) have been introduce d. Risperidone (Risperd al) was released in 1994, olanzapine (Zyprexa) in 1996, and queti apine (Seroquel) in 1997. Several other atypical neuroleptics are in development. While they have some side effects, these newer medications are generally better tolerated than either clozapine or the the traditional antipsychotics, and they do not cause agranuloc ytosis. Like clozapine, they have shown little tendency to give rise to tard