USF Libraries
USF Digital Collections

Maternal interaction style, reported experiences of care, and pediatric health care utilization

MISSING IMAGE

Material Information

Title:
Maternal interaction style, reported experiences of care, and pediatric health care utilization
Physical Description:
Book
Language:
English
Creator:
Shellhorn, Wendy Lauran Struchen
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Attachment
Health
Utilization
Bonding
Experiences of care
Infant health
Immunization
Well child care
Dissertations, Academic -- Public Health -- Doctoral -- USF
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: U.S. immunization and well child-care rates are below desired levels with lower income individuals being at higher risk for receiving inadequate care. To enhance the understanding of motivating factors to health care utilization, this study explored relationships between a mother's interaction style (secure, anxious, avoidant), her reported experiences with pediatric health care and her child's utilization of pediatric health care. Participants included 126 US-born, English-speaking women with an infant 12 to18 months of age. Linear regression analyses found no bivariate associations between maternal interaction style and reported experiences of care. Poisson regression analyses measured associations of maternal interaction style, reported experiences of care, and moderating variables with health care visits and immunizations received. Main effect models found no associations between maternal interaction style and reported experiences of care. Significant associations were identified between provider ratings and sick visits. There were no associations between provider office ratings and utilization rates. When interaction style and provider/provider office ratings were included in the model, high provider ratings (P < .05) and high anxious interaction scores (p < .0001) were associated with more sick visits while higher avoidant interaction style scores (p < .01) were associated with decreased use of sick visits. Multivariate modeling identified provider rating (p < .05) and anxious interaction score (p < .01) as main effects, child's health rating as a confounder, as well as target child being mother's first, WIC/Healthy Start participation, maternal bonding and feelings about going to the doctor acting as moderators to associations between interaction style and sick/follow-up visits. Secure interaction style scores were associated with increased use of emergency department visits, controlling for the confounding effects of maternal bonding andmoderating effects of child's health status and maternal age. Findings indicate that, in some cases, maternal interaction style is associated with how and when mothers access health care for their children. The confounders and moderators identified also highlight the need for more understanding regarding what motivates individuals. Finally, there were racial and ethnic differences including higher rates of avoidant interaction styles in Black, non-Hispanic mothers. Predicting health care utilization patterns will help better target the specific needs of mothers and ultimately improve health outcomes.
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 Wendy Lauran Struchen Shellhorn.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 338 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 - 001789990
oclc - 143607781
usfldc doi - E14-SFE0001478
usfldc handle - e14.1478
System ID:
SFS0025797:00001


This item is only available as the following downloads:


Full Text

PAGE 1

iv Maternal Interaction Style, Re ported Experiences of Care, and Pediatric Health Care Utilization by Wendy Lauran Struchen Shellhorn 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 Major Professor: Kare n Kay Perrin, Ph.D. Stanley Graven, M.D. Jeffrey Kromrey, Ph.D. Charles Mahan, M.D. Lisa Simpson, M.B. Date of Approval: March 3, 2006 Key Words: attachment, health, utilization, bonding, experiences of care, infant health, immunization, well child care. Copyright 2006, Wendy Laur an Struchen Shellhorn

PAGE 2

v Dedication This dissertation is dedicated to my pa rents, Robert and Audrey Struchen, who have encouraged me to always follow my drea ms wherever they may take me. They also taught me to accept challenges, be honest, and do my best. Thank you for helping to make me who I am today.

PAGE 3

vi Acknowledgements I would like to acknowledge and thank t hose who helped me to successfully complete this learning process: Dr. Karen Kay Perrin, for your mentorsh ip as my major professor during this long doctoral training process. I also appreciate your guida nce through the bureaucracy, dodging the politics, and seeking out financia l assistance. Without your encouragement, guidance and friendship over the years, this achievement would not have been possible. Dr. Stanley Graven, for sharing you r passion for promoting the healthy development of young children, your demand for excellence, and your soft spoken commendations. Thank you for pushing me to expa nd my field of vision in the realm of maternal and child health. Dr. Jeffery Kromrey, for your ability to ma ke complex statistical concepts seem easy. You also have more patience that anyone else I know. These are special gifts that I thank you for sharing with me over the years. Dr. Charles Mahan, for shar ing your unique sense of hu mor, your wide base of knowledge, and your enthusiasm. I have learne d something new from every contact with you and hope to carry that enthusiasm forward. Dr. Lisa Simpson, for providing me encouragement and guidance. From the moment we met, you reached out to help in any way you could including helping to

PAGE 4

vii expand my limited knowledge of health policy, rearranging your sche dule for my needs, and providing me office space. Dr. Mario Hernandez, for your endless s upport during this dissertation process. Your flexibility as an employer and mentor ha s been endless. When I asked for anything work, time off from work, advice, or co-a uthorship opportunities you gave it to me without hesitation. Thank you for being you. Dr. David Shern, Dr. Bob Friedman, and th e Louis de la Parte Florida Mental Health Institute, for the support given to me the past 18 years. Thank you for flexible employment opportunities, moral support, in corporation of my dissertation in your ACHA contract so that I could recruit for my study, and for waiving fees for data analysis regarding my dissertation. My family, for always being there when I need them, and for not being there during those times when I needed to be left alone to write. You ar e an inspiration and I will do my best to make you proud. Finally, my best friend and husband, Alan Shellhorn, for always being there to encourage and support me through this challe nging time in my life. Without you by my side, this dissertation woul d not have been possible. Supported in part by the Pediatric Clinic al Research Center of All Childrens Hospital and the University of South Flor ida, and the Maternal and Child Health Bureau, R60 MC 00003-01, Department of Health and Human Services, Health Resources and Services Administration.

PAGE 5

i Table of Contents List of Tables ................................................................................................................ ..ix List of Figures ............................................................................................................... .xiv Abstract .................................................................................................................xv Chapter One: Introduction................................................................................................1 Importance of Preventive Health Care..................................................................2 Factors that Influence Behavior............................................................................3 Genetics and Behavior..............................................................................3 Attachment Theory, Interaction Styles, and Behavior..............................4 Adult Attachments....................................................................................5 Reported Experiences of Care..................................................................7 Approaches to Enhancing Services...............................................8 Other Factors Influencing Health Care Utilization Behaviors..................9 Statement of the Problem....................................................................................10 Research Questions.............................................................................................11 Research Hypotheses..........................................................................................12 Significance of the Study....................................................................................13 Definition of Terms.............................................................................................13 Chapter Two: Literature Review....................................................................................16 Introduction.........................................................................................................16 Health Care Utilization Patte rns in the United States.........................................18 Well Child Care and Immunization Guidelines..................................................23 Health Care Utilization Barriers.........................................................................26 Approaches to Enhancing Services.........................................................29 Health Care Disparities.......................................................................................29 Understanding the Culture..................................................................................32 Effective Interventions........................................................................................36

PAGE 6

ii Factors that Influence Behavior..........................................................................38 Genetics and Behavior........................................................................................38 Attachment Theory and H ealth Care Utilization................................................39 Attachment Bond....................................................................................45 Adult Attachments..................................................................................46 Patient-Provider Attachment a nd Health Care Utilization......................46 Interaction Styles....................................................................................49 The Role of Cognition and Context........................................................51 Other Behavioral Motivators..............................................................................53 Maternal Depression...............................................................................54 Experiences of Care Reported by Patients..........................................................55 Consumer Assessment of Health Plans Survey..................................................57 Identifying a Target Population..........................................................................61 Advantages and Disadvantages of Face to Face Interviews...............................62 Summary of Research.........................................................................................64 Chapter Three: Methods.................................................................................................65 Purpose of the Study...........................................................................................65 Funding and Other Resources.............................................................................65 Design .................................................................................................................66 Study Setting...........................................................................................68 HIPAA and the Protectio n of Human Subjects.......................................69 Study Population.....................................................................................69 Sampling Framework..............................................................................71 Recruitment.............................................................................................71 Letter to Mothers.........................................................................72 Other Recruitment Approaches...................................................72 Targeted Recruitment..................................................................73 Data Collection Instruments...............................................................................73 Screening Form.......................................................................................74 Demographic Questionnaire...................................................................74 Reported Experience of Care Questionnaire...........................................75 Relationship Scales Questionnaire..........................................................77

PAGE 7

iii Health Care Utilization Data Abstraction Form.....................................78 Instrument Validity and Re liability Procedures..................................................79 Assessment of Validity...........................................................................79 Panel of Experts......................................................................................79 Pilot Study...............................................................................................81 Assessment of Reliability.......................................................................83 Data Analysis......................................................................................................83 Phase One: Data Reduction....................................................................86 Study Variables...........................................................................86 Component One: Descriptive Statistics......................................87 Component Two: Factor Analysis..............................................88 Component Three: Bi variate Statistics.......................................90 Phase Two: Interaction Styl e and Experiences of Care..........................91 Phase Three: Interaction Style, Expe riences of Care, and Utilization ...92 Phase Four: Moderating Factors.............................................................94 Sample Size Calculation.....................................................................................96 Chapter Four: Results.....................................................................................................97 Recruitment of Study Sample.............................................................................97 Data Analysis Phase One: Data Reduction and Transformation........................99 Component One: Descriptive Statistics................................................100 Maternal Demographics............................................................100 Target Child Health Issues........................................................106 Interaction Style........................................................................109 Reported Experiences of Care Questions.................................111 Health Care Utilization.............................................................113 Component Two: Factor Analysis........................................................114 Component Three: Biva riate Associations...........................................116 Data Analysis Phase Two: Interac tion Style and Reporte d Experiences of Care.............................................................................................................120 Research Question One.........................................................................120 Research Question Two........................................................................121

PAGE 8

iv Phase Three: Interaction Style, Expe riences of Care and Health Care Utilization...................................................................................................122 Research Question Three......................................................................122 Phase Four ......................................................................................................126 Research Question Four........................................................................126 Chapter Five: Synthesis of the Research......................................................................151 Utilization of Health Care Services..................................................................151 Utilization of Well Child Care Visits Finding/Implication...................152 Utilization of Sick and Follow-up Visits Finding/Implication.............152 Utilization of Emergency Department Visits Finding/Implication.......155 Utilization of Immunizations Finding/Implication...............................156 Experiences of Care Ratings.............................................................................158 Provider Ratings of Experiences of Care Finding/Implication.............158 Provider Office Ratings of Experien ces of Care Finding/Implication..159 Maternal Interaction St yle Finding/Implication................................................159 High Anxious Interaction Scores..........................................................160 High Avoidant Interaction Scores.........................................................160 Impact of Differing Interaction Scores.................................................161 Interaction Style Re liability Issues...................................................................163 Additional Study Findings................................................................................164 Limitations of the Study....................................................................................165 Implications for Action.....................................................................................165 Medicaid Changes.................................................................................166 Health Care Providers...........................................................................168 Other Programs.....................................................................................170 Recommendations for Future Research............................................................171 Replication of Study Us ing Other Populations.....................................171 Maternal Bonding.................................................................................171 Measurement of Feelings, Attitudes, and Perceptions..........................172 Incorporation of providers in Research Activities................................172 Family Planning and Baby Spacing......................................................173 Electronic Medical Record...................................................................174

PAGE 9

v References..................................................................................................................... 175 Appendices....................................................................................................................1 93 Appendix A. Well-Child Visits................................................................194 Appendix B. Immunization Schedule......................................................195 Appendix C. Catch-Up Immunization Schedule......................................196 Appendix D. Funding Announcement......................................................197 Appendix E. University of South Florida IRB Approval.........................198 Appendix F. Florida Department of Health IRB Approval.....................200 Appendix G. Tampa Ge neral IRB Approval............................................202 Appendix H. Interview Screening Questions ..........................................203 Appendix I. Recruitment Flier ...............................................................204 Appendix J. Recruitment Letter .............................................................205 Appendix K. Demographic Survey Instrument........................................207 Appendix L. CAHPS Core Questions .....................................................209 Appendix M. Relationship Scales Questionnaire......................................211 Appendix N. Medical Record Abstraction Questions..............................213 Appendix O. Recruitment of Study Sample.............................................215 Appendix P. Maternal Health History.....................................................220 Appendix Q. Pediatric Health Care Issues...............................................223 Appendix R. Maternal Interacti on Style: Relationship Scales Questionnaire................................................................226 Appendix S. Reported Experiences of Care............................................228 Appendix T. Calculation of an Odds Ratio..............................................238 Appendix U. Hypothesis 4 Po isson Regression Models..........................241 Appendix U.1 Well Child Care Vi sits, Interaction Style, and provider Ratings Race/Ethnicity (Adjusted) (N=126) .......242 Appendix U.2 Well Child Care Vi sits, Interaction Style, and provider Ratings by Birth Or der (Adjusted)(N=126) ........243 Appendix U.3 Well Child Care Vi sits, Interaction Style, and Provider Ratings by Childs Health Rating (Adjusted) (N=126) ..............................................................................244 Appendix U.4 Well Child Care Vi sits, Interaction Style, and Provider Ratings by Mothers Health Rating (Adjusted)(N=126) .............................................................245

PAGE 10

vi Appendix U.5 Well Child Care Vi sits, Interaction Style, and Provider Ratings by WIC/Healthy Start Participation (Adjusted)(N=126) .............................................................246 Appendix U.6 Well Child Care Vi sits, Interaction Style, and Provider Ratings by Mothe rs Employment Status (Adjusted)(N=126) .............................................................247 Appendix U.7 Well Child Care Vi sits, Interaction Style, and Provider Ratings by Mothers Feelings About Providers (Adjusted)(N=126) ............................................248 Appendix U.8 Well Child Care Vi sits, Interaction Style, and Provider Ratings by Mothers Age (Adjusted)(N=126) .............................................................249 Appendix U.9 Well Child Care Vi sits, Interaction Style, and Provider Ratings by Mothers Bonding Score (Adjusted)(N=126) .............................................................250 Appendix U.10 Well Child Care Vi sits, Interaction Style, and Provider Office Ratings by Race/Ethnicity (Adjusted)(N=126) .............................................................251 Appendix U.11 Well Child Care Vi sits, Interaction Style, and Provider Office Rati ngs by Birth Order (Adjusted)(N=126) .............................................................252 Appendix U.12 Well Child Care Vi sits, Interaction Style, and Provider Office Ratings by Childs Health Rating (Adjusted)(N=126) .............................................................253 Appendix U.13 Well Child Care Vi sits, Interaction Style, and Provider Office Ratings by Mothers Health Rating (Adjusted)(N=126) .............................................................254 Appendix U.14 Well Child Care Vi sits, Interaction Style, and Provider Office Ratings by WIC/Healthy Start Participation (Adjusted)(N=126) .......................................255 Appendix U.15 Well Child Care Vi sits, Interaction Style, and Provider Office Ratings by Mothers Employment Status (Adjusted)(N=126) ..................................................256 Appendix U.16 Well Child Care Vi sits, Interaction Style, and Provider Office Ratings by Mo thers Feelings About Provider Offices (Adjusted)(N=126) .................................257 Appendix U.17 Well Child Care Vi sits, Interaction Style, and Provider Office Ratings by Mothers Age (Adjusted)(N=126) .............................................................258

PAGE 11

vii Appendix U.18 Well Child Care Vi sits, Interaction Style, and Provider Office Ratings by Mothers Bonding Score (Adjusted)(N=126) .............................................................259 Appendix U.19 Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Race/Ethnicity (Adjusted)(N=126) ..................260 Appendix U.20 Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Birth Order (Adjusted)(N=126) .......................261 Appendix U.21 Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Childs Health Ra ting (Adjusted)(N=126) ......262 Appendix U.22 Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Mothers Health Rating (Adjusted)(N=126) .............................................................263 Appendix U.23 Sick/Follow-up Visits, Interaction Style, and Provider Ratings by WIC/Healthy Start Participation (Adjusted)(N=126) .............................................................264 Appendix U.24 Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Mothers Employment Status (Adjusted)(N=126) .............................................................265 Appendix U.25 Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Mothers Fee lings About Providers (Adjusted)(N=126) .............................................................266 Appendix U.26 Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Mothers Age (Adjusted)(N=126) ...................267 Appendix U.27 Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Mothers Bonding Score (Adjusted)(N=126) .............................................................268 Appendix U.28 Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Race/Ethnicity (Adjusted)(N=126) ......269 Appendix U.29 Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Birth Orde r (Adjusted)(N=126) ............270 Appendix U.30 Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Childs Health Rating (Adjusted)(N=126) .............................................................271 Appendix U.31 Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Mothers Health Rating (Adjusted)(N=126) .............................................................272 Appendix U.32 Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by WIC/Hea lthy Start Participation (Adjusted)(N=126) .............................................................273

PAGE 12

viii Appendix U.33 Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Mothers Employment Status (Adjusted)(N=126) .............................................................274 Appendix U.34 Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Mothers Feelings About Provider Offices (Adjusted)(N=126) ................................................275 Appendix U.35 Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Mothers Age (Adjusted)(N=126) ........276 Appendix U.36 Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Mothers Bonding Score (Adjusted)(N=126) .............................................................277 Appendix U.37 Predicted Sick and Follow-Up Visits by Avoidant Interaction Style and Whether Target Child was Mother's First Controlling for Provider Office and Interaction Styles (Secur e, Anxious, Avoidant).................278 Appendix U.38 Predicted Sick a nd Follow-Up Visits by Anxious Interaction Style and WIC/H ealthy Start Participation Controlling for Provider Offi ce and Interaction Styles (Secure, Anxious, Avoidant)...............................................279 Appendix U.39 Predicted Sick a nd Follow-Up Visits by Anxious Interaction Style and Fee lings About Going to the Doctor Controlling for Provider Office and Interaction Styles (Secur e, Anxious, Avoidant).................280 Appendix U.40 Predicted Sick a nd Follow-Up Visits by Anxious Interaction Style and Bondi ng Issues Controlling for Provider Office and Interaction Styles (Secure, Anxious, Avoidant).............................................................281 Appendix U.41 Emergency Department Visits, Interaction Style, and Provider Ratings by Race/Ethnicity (Adjusted)(N=126) .............................................................282 Appendix U.42 Emergency Department Visits, Interaction Style, and Provider Ratings by Birth Or der (Adjusted)(N=126) .......283 Appendix U.43 Emergency Department Visits, Interaction Style, and Provider Ratings by Childs Health Rating (Adjusted)(N=126) .............................................................284 Appendix U.44 Emergency Department Visits, Interaction Style, and Provider Ratings by Mothers Health Rating (Adjusted)(N=126) .............................................................285

PAGE 13

ix Appendix U.45 Emergency Department Visits, Interaction Style, and Provider Ratings by WIC/Healthy Start Participation (Adjusted)(N=126) .............................................................286 Appendix U.46 Emergency Department Visits, Interaction Style, and Provider Ratings by Mothe rs Employment Status (Adjusted)(N=126) .............................................................287 Appendix U.47 Emergency Department Visits, Interaction Style, and Provider Ratings by Mothers Feelings About Providers (Adjusted) (N=126) ...........................................288 Appendix U.48 Emergency Department Visits, Interaction Style, and Provider Ratings by Mothers Age (Adjusted)(N=126) .............................................................289 Appendix U.49 Emergency Department Visits, Interaction Style, and Provider Ratings by Mothers Bonding Score (Adjusted)(N=126) .............................................................290 Appendix U.50 Emergency Department Visits, Interaction Style, and Provider Office Ratings by Race/Ethnicity (Adjusted)(N=126) .............................................................291 Appendix U.51 Emergency Department Visits, Interaction Style, and Provider Office Rati ngs by Birth Order (Adjusted)(N=126) .............................................................292 Appendix U.52 Emergency Department Visits, Interaction Style, and Provider Office Ratings by Childs Health Rating (Adjusted)(N=126) .............................................................293 Appendix U.53 Emergency Department Visits, Interaction Style, and Provider Office Ratings by Mothers Health Rating (Adjusted)(N=126) .............................................................294 Appendix U.54 Emergency Department Visits, Interaction Style, and Provider Office Ratings by WIC/Healthy Start Participation (Adjusted)(N=126) .......................................295 Appendix U.55 Emergency Department Visits, Interaction Style, and Provider Office Ratings by Mothers Employment Status (Adjusted)(N=126) ..................................................296 Appendix U.56 Emergency Department Visits, Interaction Style, and Provider Office Ratings by Mo thers Feelings About Provider Offices (Adjusted)(N=126) .................................297 Appendix U.57 Emergency Department Visits, Interaction Style, and Provider Office Ratings by Mothers Age (Adjusted)(N=126) .............................................................298

PAGE 14

x Appendix U.58 Emergency Department Visits, Interaction Style, and Provider Office Ratings by Mothers Bonding Score (Adjusted)(N=126) .............................................................299 Appendix U.59 Predicted Emergency Department Visits by Secure Interaction Style and Moth er's Age Controlling for Provider Office and Interaction Styles................................300 Appendix U.60 Predicted Emergency Department Visits by Secure Interaction Style and Child Health Rating Controlling for Provider Office and Inte raction Styles (Secure, Anxious, Avoidant).............................................................301 Appendix U.61 Immunizations, In teraction Style, and Provider Ratings by Race/Ethnicity (Adjusted)(N=126) ..................302 Appendix U.62 Immunizations, Inte raction Style, and Provider Ratings by Birth Order (Adjusted)(N=126) .......................303 Appendix U.63 Immunizations, In teraction Style, and Provider Ratings by Childs Health Ra ting (Adjusted)(N=126) ......304 Appendix U.64 Immunizations, In teraction Style, and Provider Ratings by Mothers Health Rating (Adjusted)(N=126) .............................................................305 Appendix U.65 Immunizations, In teraction Style, and Provider Ratings by WIC/Healthy Start Participation (Adjusted)(N=126) .............................................................306 Appendix U.66 Immunizations, In teraction Style, and Provider Ratings by Mothers Employment Status (Adjusted)(N=126) .............................................................307 Appendix U.67 Immunizations, In teraction Style, and Provider Ratings by Mothers Fee lings About Providers (Adjusted)(N=126) .............................................................308 Appendix U.68 Immunizations, In teraction Style, and Provider Ratings by Mothers Age (Adjusted)(N=126) ...................309 Appendix U.69 Immunizations, In teraction Style, and Provider Ratings by Mothers Bonding Score (Adjusted)(N=126) .............................................................310 Appendix U.70 Immunizations, Intera ction Style, and Provider Office Ratings by Race/Ethnicity (Adjusted)(N=126) ..................311 Appendix U.71 Immunizations, Intera ction Style, and Provider Office Ratings by Birth Order (Adjusted)(N=126) .......................312 Appendix U.72 Immunizations, Intera ction Style, and Provider Office Ratings by Childs Health Ra ting (Adjusted)(N=126) ......313

PAGE 15

xi Appendix U.73 Immunizations, Intera ction Style, and Provider Office Ratings by Mothers Health Rating (Adjusted) (N=126) ..............................................................................314 Appendix U.74 Immunizations, Intera ction Style, and Provider Office Ratings by WIC/ Healt hy Start Participation (Adjusted)(N=126) .............................................................315 Appendix U.75 Immunizations, Intera ction Style, and Provider Office Ratings by Mothers Employment Status (Adjusted)(N=126) .............................................................316 Appendix U.76 Immunizations, Intera ction Style, and Provider Office Ratings by Mothers Feeli ngs About Provider Offices (Adjusted)(N=126) .............................................................317 Appendix U.77 Immunizations, Intera ction Style, and Provider Office Ratings by Mothers Age (Adjusted)(N=126) ...................318 Appendix U.78 Immunizations, Inter action Style, and Provider Office Ratings by Mothers Bonding Score (Adjusted)(N=126) .............................................................319 Appendix V Well Child Care Data Recording Issues.............................320 Appendix W Measurement of Mate rnal Feelings About Doctors and Maternal Bonding.........................................................321 Appendix X Issues regarding the Measurement of Immunizations........323 Appendix Y National Comparisons of Ra tings of Experiences of Care.324 Appendix Z Additional Study Findings, Limitations, and Strengths......327 About the Author................................................................................................End Page

PAGE 16

xii List of Tables Table 1. Final St udy Population (N=139)................................................................98 Table 2. Comparison of Study Partic ipants with Complete Data Versus Those Excluded (N=139)...........................................................................99 Table 3. County of Residenc e by Race/Ethnicity (N=126)..................................100 Table 4. Marital Status (N=126)............................................................................101 Table 5. Maternal Education (N=126)...................................................................102 Table 6. Maternal Age (N=126)............................................................................103 Table 7. Current Maternal Health Status (N=126).................................................104 Table 8. Planning of Pregna ncy for Target Child (N=126)...................................105 Table 9. Pregnanc y History (N=126).....................................................................106 Table 10. Child Health Issues (N=126)...................................................................106 Table 11. Employment and Daycare (N=126).........................................................108 Table 12 Other Child-Serving Programs (N=126)..................................................109 Table 13. Maternal Inte raction Style (N=126).........................................................110 Table 14. Item Total Correlations (N=126).............................................................110 Table 15. Consumer Assessment of Health Plans Survey Questions Pediatrician (N=126)................................................................................111 Table 16. A Comparison of the Provider Rating with a Transformed Provider Rating (N=126)........................................................................................112 Table 17. A Comparison of the Provide r Office Rating with a Transformed Provider Office Rating (N=126)..............................................................113 Table 18. Health Care U tilization Rates (N=126)...................................................114 Table 19. Factor Analysis of Bonding (N=126)......................................................115 Table 20. Bivariate Correlations with E xperiences of Care Ratings (N=126).........116 Table 21. Bivariate Correlations with In teraction Style Scale Scores (N=126)......117 Table 22. Bivariate Correlations with Transformed Health Care Utilization Data (N=126)...........................................................................................118 Table 23. Bivariate Correlations Am ong Potentially Moderating Variables (N=126)....................................................................................................119

PAGE 17

xiii Table 24. Transformed Provider Ra ting, Provider Office Rating and Interaction Style (N=126)........................................................................121 Table 25. Health Care Utilization by Transformed Provider and Provider Ratings (Unadjusted) (N=126).................................................................122 Table 26. Health Care Utilization, In teraction Styles, Provider/Provider Office Ratings (Adjusted)(N=126) .........................................................124 Table 27. Exploration of Potentially M oderating Variables on the Number of Well Child Care Visits Attended the First Year of Life, Controlling for Interaction Styles and Provi der Rating (Nine Separate Models: Black/Hispanic run together as dummy variables) (N=126) .................127 Table 28. Exploration of Potentially M oderating Variables on the Number of Well Child Care Visits Attended the First Year of Life (One Model) (N=126) ......................................................................................128 Table 29. Exploration of Potentially M oderating Variables on the Number of Well Child Care Visits Attended the First Year of Life, Controlling for Interaction Styles and Provi der Office Rating (Nine Separate Models: Black/Hispanic run t ogether as dummy variables) (N=126)....................................................................................................128 Table 30. Exploration of Potentially M oderating Variables on the Number of Well Child Care Visits Attended the First Year of Life (One Model) (N=126) ......................................................................................129 Table 31. Exploration of Potentially M oderating Variables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provi der Rating (Nine Separate Main Effects Models: Black/Hispanic run together as dummy variables) (N=126) ...................................................................................................130 Table 31-A Exploration of Potentially Moderating Variables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provide r Rating: Interaction of Birth Order (N=126).........................................................................................130 Table 31-B Exploration of Potentially Moderating Variables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of WIC/Healthy Start Participation (N=126)..............................................131 Table 31-C Exploration of Potentially Moderating Variables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of Mothers Feelings bout Doctors (N=126)..............................................................131

PAGE 18

xiv Table 31-D Exploration of Potentially Moderating Variables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of Maternal Bonding (N=126).....................................................................................133 Table 32. Exploration of Potentially M oderating Variables on the Number of Sick/Follow-up Visits Attended the Fi rst Year of Life (One Model) (N=126) ...................................................................................................133 Table 33. Exploration of Potentially M oderating Variables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Pr ovider Office Rating (Nine Separate Main Effects Models: Black/Hispanic combined dummy variables) (N=126) ..................................................................................138 Table 33-A Exploration of Potentially Moderating Variables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provide r Office Rating: Interaction of Birth Older (N=126)................................................................................139 Table 33-B Exploration of Potentially Moderating Variables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provide r Office Rating: Interaction of WIC/ Healthy Start Pa rticipation (N=126) .............................................139 Table 33-C Exploration of Potentially Moderating Variables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provide r Office Rating: Interaction of Mothers Feelings A bout Doctors (N=126).............................................140 Table 33-D Exploration of Potentially Moderating Variables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provide r Office Rating: Interaction of Maternal Bonding (N=126).....................................................................140 Table 34. Exploration of Potentially M oderating Variables on the Number of Sick/Follow-up Visits Attended the Fi rst Year of Life (One Model) (N=126) ...................................................................................................141 Table 35. Exploration of Potentially M oderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction St yles and Provider Rating (Nine Separate Models: Black/Hispanic as dummy variables) (N=126) ..........142 Table 35-A Exploration of Potentially Moderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of Childs Health Issues (N=126)...........................................................143

PAGE 19

xv Table 35-B Exploration of Potentially M oderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of Mothers Age (N=126)........................................................................143 Table 36. Exploration of Potentially M oderating Variables on the Number of Emergency Department Visits Attended the First Year of Life (One Model) (N=126)..............................................................................146 Table 37. Exploration of Potentially M oderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Office Rating (Nine Separate Models: Black/Hispanic as dummy variables) (N=126) ..........146 Table 37-A Exploration of Potentially Moderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of Childs Health Issues (N=126) ..........................................................147 Table 37-B Exploration of Potentially Moderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of Mothers Age (N=126)........................................................................147 Table 38. Exploration of Potentially M oderating Variables on the Number of Emergency Department Visits Attended the First Year of Life (One Model) (N=126)..............................................................................148 Table 39. Controlling for Interaction Styles and Provider, Exploration of Potentially Moderating Variables on the Number of Immunizations Attended the First Year of Life (N=126).................................................149 Table 40. Immunizations, Interaction St yles, and Provider Ratings (N=126) ........149 Table 41. Controlling for Interaction St yles and Provider Office, Exploration of Potentially Moderating Variables on the Number of Immunizations Visits Attended th e First Year of Life (N=126)..............150 Table 42. Immunizations, Interaction Styles, and Provider Office Ratings (N=126) ..................................................................................................150 Table 43. Recruitment of Study Participants (N=126)............................................216 Table 44. Pre-Pregnancy H ealth History (N=126)..................................................220 Table 45. Pregnancy Hea lth History (N=126).........................................................221 Table 46. Maternal Health Issu es During Delivery (N=126)..................................221 Table 47. Postnatal Health Care Visits (126)...........................................................222 Table 48. Prenatal Care Provider Helped Mother Choose a Pediatrician Prior to Delivery (N=126).................................................................................223

PAGE 20

xvi Table 49. Racial/Ethnic C oncordance Between Pediat ric Provider and Mother (N=126)....................................................................................................224 Table 50. Pediatric Health Care Transportation Issues (N=126).............................225 Table 51. Relationship Scal es Questionnaire (N=126)............................................227 Table 52. Consumer Assessment of Hea lth Plans Survey Questions-Pediatric Office (N=126)........................................................................................229 Table 53. Consumer Assessment of Health Plans Survey Questions Specialist Office (N=126)........................................................................235

PAGE 21

xvii List of Figures Figure 1. Maternal Interaction Style, Reported Experiences of Care and Pediatric Health Care Utilization Logic Model.........................................67 Figure 2. Maternal Interaction Style, Reported Experiences of Care and Pediatric Health Care Utilization Data Analysis Plan...............................84 Figure 3. Predicted Sick and Follow-Up Visits by Avoidant Interaction Style and Whether Target Child was Mo ther's First Controlling for Provider and Interaction Styles (Secure, Anxious, Avoidant).................133 Figure 4. Predicted Sick and Follow-Up Visits by Anxious Interaction Style and WIC/Healthy Start Participa tion Controlling for Provider and Interaction Styles (Secur e, Anxious, Avoidant)......................................134 Figure 5. Predicted Sick and Follow-Up Visits by Anxious Interaction Style and Feelings About Going to the Doctor Controlling for Provider and Interaction Styles (Secure, Anxious, Avoidant)................................135 Figure 6. Predicted Sick and Follow-Up Visits by Anxious Interaction Style and Bonding Issues Controlling for Provider and Interaction Styles (Secure, Anxious, Avoidant)...................................................................136 Figure 7. Predicted Emergency Depart ment Visits by Secure Interaction Style and Child Health Rating Controlling for Provider and Interaction Styles (Secur e, Anxious, Avoidant)......................................144 Figure 8. Predicted Emergency Depart ment Visits by Secure Interaction Style and Mother's Age Contro lling for Provider Office and Interaction Styles (Secur e, Anxious, Avoidant)......................................145 Figure 9. Fitted Sick/Follow-up Visits by Original Provider Ratings.....................239 Figure 10. Fitted Sick/Follow-up Visits by Transformed Provider Ratings ............239 Figure 11. Log of Visits by Original Provider Ratings.............................................240

PAGE 22

xviii Maternal Interaction Style, Re ported Experiences of Care And Pediatric Health Care Utilization Wendy Lauran Struchen Shellhorn ABSTRACT U.S. Immunization and well child-care rate s are below desired levels with lower income individuals being at higher risk for receiving inadequate care. To enhance the understanding of motivating factors to hea lth care utilization, this study explored relationships between a mothers interac tion style (secure, anxious, avoidant), her reported experiences with pediatric health car e and her childs utilization of pediatric health care. Participants included 126 US-born, Eng lish-speaking women with an infant 12 to18 months of age. Linear regression anal yses found no bivariate associations between maternal interaction style and reported expe riences of care. Poisson regression analyses measured associations of maternal interac tion style, reported experiences of care, and moderating variables with heal th care visits and immunizatio ns received. Main effect models found no associations between maternal intera ction style and reported experiences of care. Significan t associations were identified between provider ratings and sick visits. There were no associations be tween provider office ra tings and utilization rates. When interaction style and provider/p rovider office ratings were included in the model, high provider ratings (P<.05) and high anxious intera ction scores (p<.0001) were

PAGE 23

xix associated with more sick visits while highe r avoidant interaction style scores (p<.01) were associated with decreased use of sick visits. Multivariate modeling identified provider rating (p<.05) and anxious interaction score (p<.01) as main effects, childs health rating as a confounder, as well as target child being mothers first, WIC/Healthy Start participation, maternal bonding and feelings about going to the doctor acting as moderators to associatio ns between interaction style and sick/follow-up visits. Secure interaction style scores were associated with increased use of emergency department visits, contro lling for the confounding effects of maternal bonding and the moderating effects of childs health status and maternal age. Findings indicate that, in some cases, matern al interaction style is associated with how and when mothers access health care for their children. The confounders and moderators identified also highlight the need for more understanding regarding what motivates individuals. Finally, there were ra cial and ethnic differe nces including higher rates of avoidant interaction styles in Black, non-Hispanic mo thers. Predicting health care utilization patterns will help better target the specific needs of mothers and ultimately improve health outcomes.

PAGE 24

1 Chapter One Introduction The purpose of this study was to determin e the relationships between a mothers interaction style, her reported experiences w ith pediatric health care and utilization of pediatric health care services for her child. Health is a lifelong state of being with the early stages of life acting as the foundation fr om which most health potential and healthy habits begin. Because early recognition and intervention to address health issues can significantly impact an individua ls long-term health potentia l, regular well child health care visits beginning soon after birth are importa nt. Unfortunately, this preventive care is not always provided to a child, or at least not as often as is recommended. These lapses in health care utilization are due to a variety of issues in cluding maternal styles of interaction, previous maternal experiences w ith health care, availability of health insurance, individual differences among familie s and a host of other public health barriers (e.g. transportation, clinic hours, and wo rk schedules). Understanding the common factors that are associated w ith the expression of health car e behaviors, especially around the perinatal period, can promot e appropriate utilization of services by guiding system change to more adequately meet the needs of individuals.

PAGE 25

2 Importance of Preventive Health Care Early recognition of health issues is im portant because ther e are a variety of heritable and environmental factors that alte r health trajectories that can be prevented through their early identi fication and treatment. Preventive health care, also referred to as well child care, offers the opportunity to esta blish a historical r ecord of the childs development to either confirm normal de velopment or detect emerging problems (American Academy of Pediatrics, 2001). With all the benefits associated with the receipt of well ch ild care, it can be difficult to understand why the rates of h ealth screens, immu nizations and other preventive health behaviors are below reco mmended levels. For example, according to the Centers for Disease Control (2005), in 2003, 18% of children 19 to 35 months of age in the United States and 16% in Florida we re not up to date on their immunizations. Similarly, local data indicates that 15% of Hillsborough County and 16% of Pinellas County two year olds were not up to date on their immunizations (Flo rida Department of Health, 2005a). Racial and ethnic disparitie s exist in these immunization rates, with Asian and White children receiving the most immunizations, while Hispanic and Black children are receiving the fewest. There are al so socioeconomic status disparities with children living below poverty level having lo wer immunization rate s (Szilagyi et al., 2004). A key to addressing this issue of unde rutilization of we ll child care is understanding related issues. A variety of competing factors are associated with health care utilization behavior in cluding but not limited to education, economics, cultural differences, historical experiences and each in dividuals established pattern of behavior.

PAGE 26

3 Recently the Centers for Disease Control s ponsored a Community Guide to Preventive Services publication involving a systematic re view of the immunizati on literature (Guide to Community Preventive Services, 2005). The re sults of this review included evidencebased recommendations for which interv entions have been shown to improve immunization rates. These strategies include: client reminder/recall systems; multicomponent interventions and education; requir ements for child care or school attendance; reducing out-of-pocket expenses; expanding access; offering programs in WIC settings; home visits; and assessment a nd feedback for providers (Gui de to Community Preventive Services, 2005). Factors that Influence Utilization Behavior There has been a long standing debate over th e influences of nature versus nurture in regard to individual health and behavior Ultimately, literature supports the impact of both issues in the development of i ndividual traits including behavior. Genetics and Behavior Much of the research in the area of ge netic and behavior surrounds mental health issues, especially in relation to studying associated behaviors in mothers and their children as well as among siblings. One of th e most common findings is an association between genetics and anxiety disorders (Marks, 1986; Nationa l Institutes of Mental Health, 2000; Spence, Rapee, McDonald & Ingram, 2001). Genetic factors have also been found to be associated with depressi on, anxiety, and phobic disorders (Eley, Bolton, OConnor, Perrin, Smith & Plomin, 2003; Gill espie, Zhu, Heath, Hickie and Martin,

PAGE 27

4 2000). Finally, studies focusing specifically on ge netics and interacti on style suggest that anxious and avoidant st yles were associated with higher rates of alcohol use disorders as compared to individuals with secure intera ction styles (Vungkhanching, Sher, Jackson, & Parra, 2004). Attachment Theory, Interaction Styles and Behavior In addition to the genetic foundations of behavior, life experiences also play an important role in developing interaction styl es. One way to develop an understanding of motivating factors influencing interaction styl es is through the exploration of theoretical foundations of behavior. One f actor believed to be associat ed with the promotion of preventive health behaviors is the attach ment bond established between the mother and infant. Mother-child bonding is a process of developing attach ments that begins prior to birth. In addition to the stre ngth of the bonds, these att achments develop behavioral patterns that can express themselves in a variety of ways. John Bowlby (1969) theorized that a system of attachments between car egivers and infants was an evolutionary mechanism developed to promote the surv ival of the species. A strong positive relationship in attachment theory is referre d to as a secure attachment and requires a variety of features including predictability, responsiveness, intelligibility, supportiveness, and reciprocity of commitment (Brethert on, 1999). In a relationship where a strong positive attachment system has been established, the infant will seek close proximity to a caregiver. Conversely, if secure bonding has not taken place, infants may be ambivalent or avoidant in times of stress. The infant may also become anxious and cling to the mother (Klaus, Kennel, & Klaus, 1995).

PAGE 28

5 Although attachment theory began with explanations of mother-infant relationships and is generally limited to only a few close individuals, a life-course perspective has taken those attachment behaviors and extended them into adult relationships as well. More specifically, attachment th eorists have proposed four hypotheses about the development of attachme nt styles. The firs t hypothesis is that behavioral patterns start early in childhood in response to primary attachment figures. The second hypothesis is that these attachment styles remain relatively stable over time and can be applied across different settings. The third hypothesis is that these attachment styles influence adult relations hips. The final hypothesis states that as individuals have children of their own, these attachment styles will influence the behavioral patterns of the next generation (Simpson & Rholes, 1998). Adult Attachments Adult attachment theory is similar to mother-infant attachment bonds, where the motivation is that one indivi dual seeks out the proximity of another in order to feel comfort and security (Feeney & Noller, 1996) However, it differs from mother-infant bonds in regard to the issue of power in the relationship. A se cure mother-infant relationship has the mother in an authoritativ e role providing for the dependent infant. On the other hand, a secure adult-adult relationshi p generally does not involve one individual as having a permanent authoritative role over another but rather there are periods where either of the adults in the relationship can have greater authority. If one adult has persistent authority over anothe r, the relationship is consider ed to be abnormal (Feeney & Noller, 1996).

PAGE 29

6 Focusing more specifically on attachment styles within the health care arena, individuals with secure, avoida nt, or anxious attachment st yles may behave differently. Outside of the maternal and child health focu s, the attachment style between a patient and health care provider has been shown to have a significant association with patient behavior (Bultman & Svarstad, 2000; T hompson, Gee, Kotz, & Northrop, 2000). Because the interactions be tween individuals and health care providers often are relatively infrequent, there can be little opportunity for specific attachment bonds to develop. This lack of continuous care is fu rther exacerbated by the expansion of managed care systems where provider networks can ch ange every year and by practices that include greater numbers of physicians (S ultz & Young, 1999). This same attachment issue spans a variety of research arenas a nd prompted another way of conceptualizing relationship bonds. Although not assessing direct relationships, the measurement of a general attachment style that influences individual behavior when interacting with their environment is also useful. This general style of attachment behavior is referred to as an interaction style and is the focus of this study. Within the health care arena, theory i ndicates that those mothers with strong secure interaction styles tend to seek appropriate health car e when necessary and have the confidence to resolve minor issues indepe ndently. The theory also suggests that individuals with strong anxious styles of interaction tend to desire more support from their provider as their level of stress increases, such as a childs illness (Simpson & Rholes, 1998). Anxious mothers may over-utilize health care services in search of confidence that the health issues are being a ddressed adequately. Fina lly, individuals with strong avoidant styles of inte raction tend to distance themselves from others when they

PAGE 30

7 are distressed and may reject health care. This avoidant behavior co uld have an impact on the health of the child but it can also be costly in terms of lost productivity to the system of care if appointments are not kept (Simpson & Rholes, 1998). Reported Experiences of Care There are a number of issues that can dire ctly influence health utilization patterns. One such factor is an individua ls history of experiences with health care. If individuals experience poor customer service such as difficulties getting appointments, inconvenient office hours, office locations with limited accessibility, long wait times, and medical personnel who do not treat them with courtesy and respect, then the likelihood of a return visit decreased (Palfrey et al ., 2004). Making these experience s even more complicated is the fact that individual pe rceptions of the care provided can be altered based on past experiences and expectations placed upon the s ituation (Seid et al., 2001). It is due to these confounding issues that there has been a conceptual shift from researching satisfaction of care to that of anchoring questions on specific elements of care such as how often a patient was able to get care when they believed they needed it (Institute of Medicine of the National Academies, 2004a). In response to this shift towards experien ces of care, the Agency for Healthcare Research and Quality (AHRQ) established a set of measures and tools to help Medicaid, Medicare, public and private employers, as we ll as individual health plans collect and utilize information regarding health car e quality. The result was the Consumer Assessment of Health Plans (CAHPS) initia tive that includes not only standardized assessment instruments but also provides a Survey and Reporting Kit 2002 (Agency for

PAGE 31

8 Health Care Research and Quality, 2004b). This kit includes survey instruments, sample formats for developing reports and software to assist in data analysis, as well as guidance for implementing, reporting data and evaluating the results. This uniformity of measures provides the opportunity to esta blish a benchmarking database to facilitate sharing of results among the different users. In addition to smaller data collection initiatives, these same questions are being utilized in larger, nationally represen tative research efforts such as the AHRQ Medical Expenditure Panel Survey (Agency for Health Care Research and Quality, 2004a). In total, over 9 million people have access to the CAHPS resources (Agency for Health Care Re search and Quality, 2004b). Although these nationally-repres entative data sets can provide vast amounts of information regarding the gene ral population, they often lack the sample size to explore issues in specific sub-populati ons. This lack of sample size can be further limited if these larger organizations do not focus on specific target populations such as women in the perinatal period. As a result, there is a n eed for enhancing the experiences of care benchmarking database with smaller, targ eted samples of special populations. These enhancements allow researchers to recognize whether generalizability of the findings is strengthened or that additional care should be taken in generalizing from those larger data sets to special populations. Approaches to Enhancing Services. A literature review was conducted for the National Friendly Access program regarding friendly access to health care for low income mothers and babies (Albrecht, 2005) Studies indicated that access, providing non-medical support, understanding and limiti ng barriers, as well as having providers

PAGE 32

9 take womens experiences and attitudes into account were associated with utilization of well child care visits (Baldw in et. al., 1998; Buescher & Ward, 1992; Byrd et. al., 1998; Poland et. al., 1987). Furthermore, studies focusi ng on the use of prenatal pediatric visits indicated increased ra tes of breastfeeding, reduced emer gency departments visits, and improved providers perception of the physician-patient rela tionship (Serwint et. al., 1996). Associations between provider communi cation style, patient satisfaction and compliance were also identified. Other Factors Influencing Health Care Utilization Behaviors In addition to attachment, interaction styl es and previous experiences with health care, there are a variety of i ssues that can influence the receipt of health care. Some factors, such as income, education, hea lth literacy, language proficiency, race and ethnicity all play a role in health care utilization (Feen ey, 2000; Mickelson, Kessler, & Shaver, 1997). Even an individuals locus of control can influence utilization. Tinsley, Trupin, Owens, and Boyum (1993) found that wo men who perceived that they had some level of control over their heal th status were more likely to be compliant with medical recommendations. External resources also influence utilizat ion. For example, it has been known for decades that if an individual has no health insurance or has no usual source of care, it is less likely that adequate pr eventive health care services would be received (Novick, Mustalish, & Eidsvold, 1975). The same can be said for individuals with unsupportive families and social networks, those suffering from depression, and those having negative health experiences such as facing complicat ions during delivery. Fi nally, promotion of

PAGE 33

10 well childcare through programs such as H ealthy Start and the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) has been shown to positively influence utilization (Kenda l et al., 2002; Luman, McCa uley, Shefer & Chu, 2003). Statement of the Problem Although research has shown that ade quate preventive health care can significantly influence longterm health outcomes, more understanding about the underlying relationships is needed. National in itiatives and local studies are doing their part to answer some of these questions. Ho wever, the number of local initiatives that design their studies to parall el and build on the national e fforts are limited. Collecting information in a special population of fi rst time mothers will expand that knowledgebase. In addition, identifying the connection between interaction styles and reported experiences of care can provide a broader understanding of the previously unexplored issues. As noted earlier, the pa tient-provider attachment style has been shown to make an impact on patient behavior (Bultman & Svarstad, 2000; Thompson et al., 2000). Further exploration of this issue is n eeded to see if the associati on between attachment style and behavior exists among maternal interactions during the perinatal period. As a result, the purpose of this study was to determine the re lationships between a mothers interaction style, her reported experiences with pediatri c health care and utilization of pediatric health care services for her child.

PAGE 34

11 Research Questions 1. Is maternal interaction style (secure, avoidant, anxi ous) related to a mothers reported experiences with her childs pediatric health care provider since her childs birth? 2. Are reported maternal experiences with pe diatric health care related to pediatric health care utilization (number of health ca re visits and immunizations) during the first 12 months of a childs life? 3. Controlling for reported ma ternal experiences with pe diatric health care, is maternal interaction style (secure, avoidant, anxious) related to pediatric health care utilization (number of health care visits a nd immunizations) during th e first 12 months of a childs life? a) Controlling for reported maternal expe riences with pediatric health care, is an anxious maternal interaction style related to increased pediatric health care utilization during the first 12 months of a childs life as compared to a secure interaction style? b) Controlling for reported maternal expe riences with pediatric health care, is avoidant maternal interaction style re lated to decreased pediatric health care utilization during the first 12 months of a childs life as compared to a secure interaction style? 4. Are there variables (e.g., age, race, ethnicity, pre gnancy complications) that moderate the relationship of maternal interaction style (s ecure, avoidant, and anxious) and reported maternal experiences with pediatric health care on the utilization of pediatric

PAGE 35

12 health care (number of health care visits a nd immunizations) during th e first 12 months of her childs life? Research Hypotheses 1. The reported experiences of care with th e childs health care provider will be significantly different among wome n having differing interacti on styles (secure, avoidant, anxious). 2. Pediatric health care utilization will be significantly di fferent among women having different reported experiences with pedi atric health care providers as measured by the number of health care visits and immuni zations during the first 12 months of their childs life. 3. Controlling for reported maternal expe riences with pediatric health care, infant health care utilization (as measured by the number of h ealth care visits and immunizations during the first 12 months of a ch ilds life) will be si gnificantly different among women having different interaction styles (secure, avoidant, anxious). a) Controlling for reported maternal experiences with pediatric health care providers, an anxious maternal interaction style will be related to increased pediatric health care utiliz ation during the firs t 12 months of a childs life as compared to a secure interaction style. b) Controlling for reported maternal experiences with pediatric health care providers, avoidant maternal interacti on style will be rela ted to decreased pediatric health care utilization during th e first 12 months of a childs life as compared to a secure interaction style.

PAGE 36

13 4. There are variables (e.g., age, race, ethnicity, pregnancy complications) that moderate the relationship of maternal inte raction style (secure, avoidant, anxious) and reported maternal experiences with pediatric health care providers on the utilization of pediatric health care (number of health care visits and imm unizations) during the first 12 months of a childs life. Significance of the Study The interaction styles of mothers are believed to be associated with the subsequent utilization of hea lth care services. Affirming th e association and identifying factors that can influence that relationship, can guide staff training activities and suggest program modifications that could improve compliance with recommended health care services. Currently, research regarding inte raction styles, theore tical foundations of interaction styles, and influence of extraneous factors on health care utilization is limited. Definition of Terms Anxious Attachment: individuals with anxi ous attachment styles tend to become illat-ease or distressed during tim es of separation (Fraley, 2002). Attachment Theory: an approach to descri bing and explaining th e life-long evolution of bonding in close, personal rela tionships (Bartholomew, 1990). Avoidant Attachment: individuals with a voidant attachment styles do not become stressed during times of separation and will tend to avoid contact when in close proximity to others, especially in times of stress (Fraley, 2002).

PAGE 37

14 Construct Validity: The degree to which infe rences can legitimately be made from the operationalizations in your study to th e theoretical constr ucts on which those operationalizations were based (Trochim, 2004, p. 1). Content Validity: compares the operationalization [of the construct] against the relevant content domain for the construct (Trochim, 2004). Face Validity: refers to whether the in strument looks like it should measure the intended constructs (Trochim, 2004). Health: a state of physical, me ntal and social well-being and not merely the absence of disease or infirmity (World Health Organization, 2004, p. 1). Life-Course Perspective: behavioral pattern s developed over a life, spanning temporal and social contexts (Boss, Dohert y, LaRossa, Schumm, & Steinmetz, 1993). Health Care Provider: indivi dual providing primary source of care to an individual such as a midwife, physician, or physicians assistant. Infant: for the purpose of this study, the term infant focuses primarily on the first 12 months of life. However, in general the term refers to a childs first year of life. Interaction Style: a general pattern of interaction (secu re, anxious, avoidant) with other individuals based on atta chment theory but not refl ecting a relationship between specific individuals. Maternal Primary Health Care Provider: th e health care provider that the woman uses for non-pregnancy-related health care. Nurse Practitioner: a nurse with a graduate degree in advanced practice nursing. This allows him or her to provide a broad range of health care services (Medline Plus, 2006).

PAGE 38

15 Proximity: "the state or quality of being n ear; nearness in time, space, etc. (Gurlanik, 1986). Secure Attachment: the optimal way of dea ling with attachment, separation and loss in close personal relationship s. Secure adults find it re latively easy to get close to others and are comfortable depending on ot hers and having others depend on them (Hazan & Shaver, 1987). Self Efficacy: Individuals belief about thei r capabilities to produce effects which can then act as an activation to action. The level of motivation is reflected in the courses of action chosen, the intensity of those actions and the persistence of efforts (Bandura, 1994). States Child Health Insurance Program (SCH IP): a federally-funded low cost or free health insurance program for children (Cente rs for Medicare and Medicaid Services, 2004a). Well-Child Care: visits to a childs health care providers office at prescribed, regular intervals to receive health screenings, immunizations, and provide parents with anticipatory guidance and health educati on including nutritio n, risk avoidance, healthy lifestyles, and parenting skills development (Kanda, 2004).

PAGE 39

16 Chapter Two Literature Review The purpose of this study was to determin e the relationships between a mothers interaction style, her reported experiences w ith pediatric health care and utilization of pediatric health care services for her child. Understanding the cont ext from which this study arose can provide insight in to both the need for the resear ch as well as the driving forces behind its design. The literature revi ew begins by addressing the rationale for appropriate health care utilization and the barrie rs to receiving that care. The next section addresses the behavioral aspect s of attachment theory and ma ternal interaction styles as well as the relationships among competing behaviors. Measurem ent of health care quality is then explained through th e exploration of reported experiences of care. Also, the review addresses the reasons for selec ting a specific target population and the justification for using a faceto-face interview format for the study. Finally, there is a summary of the research to be conducted. Introduction The past few decades have been a time of great improvement in the health of Americans. One measure of this improvement is the infant mortality rate which declined from 26 per 1,000 births in 1960 to a low of 6.8 per 1,000 births in 2001 and then

PAGE 40

17 increased slightly to 7 per 1,000 births in 2002 (Centers for Disease Control, 2004; Kochanek & Martin, 2004; U.S. Departme nt of Health and Human Services, 2004). Similarly, Floridas infant mortality rate has stabilized at 7 per 1,000 (Florida Department of Health, 2005b). Looking specifically at the target counties, they have had similar decreases across history as well as increas es in recent years. Hillsborough County has increased from a low of 7.9 in 2000 to 8.9 in 2004 while Pinellas Countys rate has varied from 6.4 to 7.8 and back down to 6.0 during the same time span (Florida Department of Health, 2005b). Immunizations, on the other hand are at record high levels. Immunization rates from 2003 indicate that immunizations were up to date for 79% of children 19-35 months nationally and 81% in Florida defined by 4 DT P, 3 polio, 1 measles, 3 Hib, and 3 HepB doses (Centers for Disease Control, 2005) Immunization rates for Hillsborough and Pinellas Counties are simila r to the state (Florida Department of Health, 2005a). However, the trend in these advances has begun to plateau, indicating a need for new intervention strategies to be explored. In addition, these impr ovements in health care have not been uniform across racial, ethnic, a nd socioeconomic strata, leading to the development of specific objectives for reducing disparities within national initiatives such as Healthy People 2010 (U.S. Department of Health and Human Services, 2000; Institute of Medicine of the National Academies, 2003) To help push efforts toward achieving this national priority of decr easing disparities in health, new research frameworks are needed to more adequately capture the underlyi ng factors that influe nce these disparities (Carlson & Chamberlin, 2004). Understandi ng new approaches that also will be responsive to health disparities is necessary to continue to improve health outcomes.

PAGE 41

18 Health Care Utilization Patterns in the United States One of the best ways to improve health outcomes is to prevent problems from arising. Utilization of preventiv e health care services at any age can significantly impact an individuals life by detecti ng health issues early, when they are more likely to be successfully treated. For infants, early screening, dia gnosis and treatment of health issues is crucial. However, even though the need has been identified, many children go without receiving enough of this well childcare. Two of the major reasons for this lack of care is the absence of health insurance and the lack of a usual source of care. For example, according to the 2002 Medical Expenditure Pane l Survey data, there were approximately 5.3 million uninsured children in the United States representing 10.4% of the population under the age of 18 (Sim pson et al., 2004). More specifically, recent estimates of h ealth care utilization by infants in the United States indicate that 11.7% of infants had no source of health care and 22.3% had only publicly funded health insurance. These figures vary greatly by race with white, non-Hispanic infants having 11.7% with no insurance and 14.6% with only public insurance. African American infants have no insurance 11.5% of the time and have only public health insurance 43.3% of the time. Fina lly, Hispanic infants ha ve the lowest rates of coverage with 29.0% having no insuran ce and another 32.8% ut ilizing only public insurance (Simpson et al., 2004). The rates of uninsured children in Florida ar e slightly higher th an the rest of the nation but have been improving. According to the 2004 Florida Hea lth Insurance Study, 8.1% of children 0-4 years of age, 12.4% of children ages 5-9, and 14% of children ages 10-18 were uninsured in Florida (Agency for Health Care Administration, 2005).

PAGE 42

19 One approach to improving the rates of uninsured children whose families make too much money to qualify for Medicaid is through subsidized health insurance for children, such as Floridas KidCare Progr am. A study of KidCare found that when children were enrolled into the program, th e number of those children reporting a usual source of care increased from a pre-registra tion rate of 80% to 96% at 12 months postenrollment (Institute for Child Health Policy, 2004). Having health insurance is a key factor in having a usual source of care. When individuals receive care from th e same source, often referred to as a medical home, over time the continuity of care improves the abil ity of the provider to understand and better meet the individual needs of each patient. A medical home is a medical model from which health care service providers can partne r with families to help them achieve their maximum potential. It includes a seamless syst em of health care services that fosters collaboration and cooperation among all memb ers of the community in which the child and family live (Tonniges, Palfrey, & Mitchell, 2004, p.1472). Although definitions vary, there is some consensus that a medical home includes at least 5 major components: 1) a usual place for sick and well-child care; 2) a personal doctor or nurse; 3) in the event referrals to other health care providers are necessary, individuals should not experien ce difficulty receiving those re ferrals; 4) adequate care coordination among health care services; 5) and the care should be family-centered (American Academy of Pediatrics, Medical Home Initiatives for Children with Special Needs Project Advisory Committee, 2002). Ini tiatives such as the State Child Health Insurance Program and Medicaid are efforts to fund health care services and promote a

PAGE 43

20 medical home for families that would otherwis e not be able to afford health care (Office of the Law Revision Council, 2000). These advances have contributed to dec lines in the morbidity of individuals resulting from reductions in infectious diseas es and accidents, increased access to health care, and reductions in environmental agents (Institute of Medicine of the National Academies, 2004b). However, even though havi ng health insurance and a medical home is the foundation from which appropriate pr eventive health care ca n be accessed, access to those resources does not guarantee services will be sought or re ceived. According to Inkelas, Schuster, Olson, and Halfon (2004), only half of young children 4 to 35 months of age are reported to have a specific clinician for well-child care in the United States. Furthermore, the Medical Expenditure Pane l Survey data showed that only 82.4% of children under the age of 5 a nd only 70.1% of children under th e age of 18 received at least one health care office vis it in the past year, regardless of whether they had insurance (Simpson et al., 2004). The utilization of at l east one health care vi sit per year varied greatly by race and ethnicity (Simpson et al., 2004). Looking more specifically at infants r eceiving care through Medicaid, a study of Connecticuts Medicaid managed-care program f ound that, overall, babies did not receive their expected number of well-child care visi ts with African American (OR 0.49; 95% CI 0.37-.063) and Hispanic (OR 0.53; 95% CI 0.41-0.69) infants being less likely than Caucasian babies to receive ade quate care (Lee & Learned, 2002). The impact that this underu tilization of preventive h ealth care services has on overall health can be seen in the need to improve immunization rates in the United States. Nationally, 22% of children ages 19 to 35 m onths have not received all the recommended

PAGE 44

21 immunizations (Centers for Disease Control and Prevention, 2004). A variety of reasons have been given for why the children have not received all of their immunizations. One reason is a parents concern regarding the safe ty of the vaccines and the possible links to developing childhood illnesses such as the de velopment of autism (Abbotts & Osborn, 1993; United Press International, 2003). From a financing perspective, according to the Institute of Medicine, acti ons that have an impact on immunization rates include inadequate and unpredictable funding levels, lack of flexibility of national immunization policies, and limited involvement by the private sector (Chalk, 2004). The interest in enhancing immunization rates nationally prompted the Centers for Disease Control to sponsor a Community Gu ide to Preventive Services publication involving a systematic review of the immunization literature (Guide to Community Preventive Services, 2005). The results of this review included evidence-based recommendations for which interventions ha ve been shown to improve immunization rates. These strategies include: client reminder/recall systems; multi-component interventions and education; requirements fo r child care or school attendance; reducing out-of-pocket expenses; expanding access; offering programs in WIC settings; home visits; and assessment and feedback for pr oviders (Guide to Community Preventive Services, 2005). One example of the use of these strategies is Floridas institution of a state-wide Healthy Start initiative more th an a decade ago. This home visiting program works closely with the WIC program, both be ing administered through the county health departments. Floridas immunization rates ha ve increased from 66% of two year old being fully immunized in 1990 to 85% in 2003 with a high of 87% in 2000. Immunizations of two year ol ds in Hillsborough County increas ed from 75% in 1995 to a

PAGE 45

22 high of 85% in 2002 and then decreased to 78% in 2004 (Florida Depa rtment of Health, 2005a). Pinellas County rates increased from 79% to 81% over the same time period. In addition to these nationa lly-recognized strategies, local efforts to build upon these strategies are beginning to identify other potentially eff ective approaches. For example, El-Mohandes et al. (2003) incor porated a multi-factor approach into a community-based intervention program servi ng minority women and found that utilizing a self-efficacy model focusing on the knowle dge and beliefs of parents improved wellchild health care utilization. Self-efficacy refe rs to a persons belief regarding the ability to exert influence over events that affect their lives (B andura, 1994). El-Mohandes et al. (2003) found that by nine months of age, infant s in the intervention gr oup were more than twice as likely (OR 2.2; 95% CI 1.09-4.53) to have received their immunizations. Furthermore, those with 30 or more visits were more than three times more likely to have received their immunizations (OR 3.63; 95% CI 1.58-8.33). Other studies that have primarily concentrated on parent education without the focus on other competing factors, such as motivational issues, have not show n the same improvements in health care utilization (Oeffinger, Roaten, Hitchcock, & Oeffinger, 1992; Zuniga de Nincio et al., 2003). From a health promotion perspective, to facilitate the full and appropriate participation of individuals in the system of care compe ting factors associated with behavior must be better understood and addressed in the provision of health care services. Research studies in this area have also demonstrated that even without additional funding, minimal changes in operating procedur es can improve vaccination delivery. For example, one study of infants continually en rolled in managed care found that the proportion of children with up-to-date immuniza tions increased significantly to 87% as a

PAGE 46

23 result of their targeted in tervention (p<.05) (Minkovitz Belote, Higman & Weiner, 2001). However, there is still a great need for identifying, understanding and addressing the health disparities in imm unization rates among various underserved populations in order to achieve improved outcomes such as those outlined in Healthy People 2010 (Chu, Barker, & Smith, 2004). Another important preventive health service associated with well-child care visits is early and ongoing developmental screenings Screenings increas e the likelihood that health issues are iden tified soon after the point of mani festation, when interventions can prevent subsequent morbidity such as th e loss of developmental potential (American Academy of Pediatrics, Committee on Childre n with Disabilities, 2001). Unfortunately, these assessments are not consistently administ ered regardless of wh ether access to health care is an issue. Halfon et al. (2004) examin ed developmental assessment utilization of infants 10-35 months of age and found that approximately 57% of children received assessments. The impact of these assessments, or the lack of assessments, is multifaceted. In addition to those children receiving deve lopmental assessments being more likely to receive other developmental services, the parent s were more likely to report greater levels of satisfaction (6.9 vs 8.4, p<.0001) and have mo re favorable ratings of the interpersonal quality of well-child care (71.2 vs 59.1; p<.001). Well-Child Care and Immunization Guidelines As has been illustrated by the previous examples, the potential for preventing or at least minimizing health issues through early identification, assessment and treatment is great. However, with the rapid pace of advan cements it can be a challenge for parents,

PAGE 47

24 especially those with their fi rst child, just to keep up with the latest research and recommendations. It is for this reason that national organizations, such as the American Academy of Pediatric and the Centers fo r Disease Control, have established recommended guidelines for well child care. Preventive health care in children, often refe rred to as well-child care, addresses a variety of health issues including ongoing health screen ing, immunizations, and child safety in a systematic fashion. Ideally, pe diatric health care s hould begin with the selection of a pediatrician befo re the birth of the child. Pren atal care providers play can play an important role in this selection process. However, a survey of women who recently delivered a baby in the National Friendly Access Initiative found only limited prenatal involvement. When women were as ked whether their pren atal care provider helped them to find a health care provid er for their new baby, only 34% of women reported such assistance (Lawton and Rhea Chiles Center, 2005). Once born, the child should begin receiving regular health care screening, and immunizations. Simultaneously, caregivers should be receiving ongoing guidance regarding disease prevention and recognition. To support this ongoing effort of health monitoring, the American Academy of Pediatri cs has developed a set of guidelines and recommended intervals regard ing well child care visits during childhood [Appendix A]. Well child care offers the opportunity to esta blish a historical record of the childs development to either confirm normal deve lopment or detect emerging problems. For infants in their first year of life, these incl ude assessments at birth, two to four days, as well as in months one, two, four, six, nine a nd twelve (American Academy of Pediatrics, 2002).

PAGE 48

25 In addition to simply attending health care vi sits at these intervals, there are issues that are recommended to be addressed duri ng each visit. This periodic assessment includes a comprehensive interval history with anthropometric measurements, sensory screening, developmental apprai sal, physical examinations, la boratory tests, diagnostic procedures, immunizations, and a review of medications a nd drug reactions. In addition to the assessment of the child, communication with caregivers is also important. Providers discuss the findings from the physical examination and laboratory tests, provide anticipatory guidance regarding probl ems associated with normal development, assessment of interactions am ong family members and the ab ility to offer appropriate counseling regarding problems identified by the child, the parents, and/or the physician (Lee, Fitzgerald, & Ebel, 2003; Schor, 2004). Mon itoring a childs health status in such a thoughtful process helps maximize the benefits of the visits and reduces duplication through excessive visits for each health issue independently. Complimentary to the American Acad emy of Pediatrics Well-Child Care Guidelines, the immunization schedules for when various vaccinations are to be provided were established, in part, based on guidel ines set forth annually by the Centers for Disease Control and Prevention (2005) and leading to the Healthy People 2010 immunization goals (U.S. Department of H ealth and Human Services, 2000) [Appendix B]. Immunizations that are recommended to be provided by the first 12 months of life include: two Hepatitis B (HepB); three Dipthe ria, Tetanus, and Pertussis (DTaP); three Haemophilus Influenza b (Hib); two Inact ivated Polio Virus (IPV); and three Pneumococcal Virus (PCV). A third HepB a nd IPV can be administered between 6 and 18 months of age. In recognition of the untimely receipt of immunizations by many

PAGE 49

26 children, a Catch-Up schedule is also provided by the CDC for children whose immunizations have been delayed [Appendix C]. In addition to identifying the need for early identification, diagnosis, and treatment of health issues, research has also demonstrated that su ch early interventions significantly improve health outcomes. For exam ple, a study of Medicaid recipients from multiple states identified associations betw een preventive pediatric visits and adverse health outcomes (Hakim & Bye, 2001). The number of well-child visits had a positive association with a decrease of avoidable hospitalizations (HR), regardless of race, poverty level, or health st atus (California: HR 0.54, 95% CI 0.50-0.55; Georgia: HR 0.54, 95% CI 0.50-0.58; Michigan: HR 0.74, 95% CI 0.69-0.79). Based on the research, compliance with recommended preventive well -child visits and improved immunization rates should be reviewed to id entify key elements to enhanc ing utilization of preventive pediatric health care services. Health Care Utilization Barriers Despite the generally increasing rates of immunizations and the recognition of the importance of well child care, there are st ill disparities in utilization rates among individuals based on issues such as socio economic status, insurance status, and usual source of care, as well as race and ethnici ty (Moore & Hepworth, 1994; Nevin & Witt, 2002; Sambamoorthi & McAlpine, 2003). Accord ing to the National Health Disparities Report, 20% of children experience lapses in health coverage with Hispanic children being more likely (41%) than non-Hispanic White children (17%) to experience those lapses (Agency for Health Care Research a nd Quality, 2003). An anal ysis of data from

PAGE 50

27 the 1998 National Maternal and Child Health Survey found that 58% of White infants received the recommended number of well-chil d care visits in comp arison to only 35% of African American infants (African Ameri can versus White-OR 1.7; 95% CI 1.5-1.9) and 37% of Hispanic infants (Ronsaville & Hakim, 2000). In an other study, women of African American descent, those who had less than a high school education, were not married, had multiple children, were not par ticipating in WIC even though they were eligible and those below 50% of the federal poverty level we re less likely to take their children to receive adequate immunizations (Luman et al., 2003). Research has also found that even an individuals locus of cont rol can influence utilization. Tinsley, Trupin, Owens, and Boyum (1993) found th at women who perceived that they had some level of control over their health st atus were more likely to be compliant with medical recommendations. These findings indicate that ba rriers to seeking prev entive health care, interactions among factors, and cultural differences all need further exploration. Some barriers to access were identif ied in a study using the Nationallyrepresentative Medical Expe nditure Panel Survey for participants in managed care environments. Respondents reported going with out care, having no usual source of care and facing organizational barriers such as di fficulties in getting appointments and long waiting periods (Phillips, Mayer, & Aday, 2000) More specifically, Hispanic individuals reported more issues with obt aining care, going without care having a usual source of care, and being convinced that family memb ers could receive needed care. Hispanic respondents also reported more often that their health care provider did not provide them with needed information or listened to them carefully.

PAGE 51

28 As noted earlier, one of the st rongest factors in health ca re utilization is the ability to pay for the care through h ealth insurance including how that insurance is funded. However, the interplay between income and accessing adequate medical care is not linear. Families who can afford health insu rance are more likely to access care. Lowincome families, on the other hand, vary th eir participation ba sed on what supportive services are available to them. Families w hose incomes allow them to participate in subsidized programs such as Medicaid a nd the child health insurance program are provided more opportunities to participate in health care services. Some studies have illustrated utilization patterns similar to fa milies with private insurance while other studies still indicate lower rates. Those least likely to access well child care are individuals who make too much money to re ceive Medicaid and yet have no health insurance to help pay for the services (Gorman, Landale, & Oropesa, 2001; Lee & Learned, 2002; Slifkin, Freeman, & Silberman, 2002). The differences in utilization of health care span health care services and are influenced by a multitude of factors. For example, differences have been identified in the timeliness of the first newborn visit appoi ntment based on Medicaid and non-Medicaid appointments (p<.001), various health care prac tices (p<.001), and ma ternal age (p<.001) (Feinberg & Hicks, 2003). Another study found that in some populations children and adolescents enrolled in Medicaid, when co mpared with those served through private managed health care coverage, had significan tly lower immunization rates, lower wellchild visits, and fewer proce dures common for children of that age (p<.001) (Thompson, Ryan, Pinidiya, & Bost, 2003). In other populati ons, there were no significant differences between groups.

PAGE 52

29 Approaches to Enhancing Services. The National Friendly Access program c onducted a literature review regarding friendly access health care for low income mothers and babies (Albrecht, 2005). This review identified several dimensions of fr iendly access including those related to the health care system, the patient needing care, utilization and access i ssues, the nature of the patient-provider encounter patient satisfaction concer ns, and outcomes. Additional studies found issues that inhibited or prom oted utilization of preventive health care included that of access, providing non-me dical support, understanding and limiting barriers, and not taking women s experiences and attitudes into account (Baldwin et. al., 1998; Buescher & Ward, 1992; Byrd et. al., 1998; Poland et. al., 1987). Research regarding the use of prenatal pediatric visits indicated increased rates of breastfeeding among mothers, reduced attendance at emerge ncy departments, and improved physicians perception of the physician-patient relationshi p (Serwint et. al., 1996). Literature also indicates that there are associations betw een provider communication style, patient satisfaction and compliance. Rowland-Morin and Carroll (1990) found that physicians had higher levels of patient satis faction if they 1) were warm and courteous, 2) had active listening skills, 3) provided unsolicited in formation, 4) spent enough time explaining health issues including causes and treatme nts, 5) provided emotional support, and 6) trusted the patient. Health Care Disparities Although studies have repeated ly indicated that there ar e significant differences among racial and ethnic groups in their utiliza tion of health care se rvices, more research

PAGE 53

30 is still needed to understand the underlying ba rriers to care. For example, the proportion of minority women who receive adequate prenatal care is lower than for Caucasian women. Some of these differences can be attr ibuted to satisfaction with care (Saha, Arbelaez, & Cooper, 2003). However, in a prospective study of African-American women in a large managed care organizat ion, Handler, Rosenberg, Raube and Lyons (2003) found that women can be very satisfied with their prenatal care and still not receive an adequate number of prenatal visits. It has been well establishe d that many but not all of the racial and ethnic differences noted in the literat ure are related to socioeconom ic issues such as poverty, unemployment, education, a lack of consistent health insura nce and usual source of care (Aiken, Freed, & David, 2004; Diaz V.A. Jr., 2002; Fiscella, Franks, Doescher, & Saver, 2002; Jones, Cason, & Bond, 2002; Pasick, Stewart, Bird, & D'Onofrio, 2001; Sambamoorthi & McAlpine, 2003). However, understanding and isolating this socioeconomic impact from other factors can be difficult. For example, breast cancer screening research found that US-born wo men of Mexican descent had higher income levels, more education, and greater accu lturation than Mexican-born women. The USborn women were also more li kely to have health insura nce, receive breast health education, and were more motivated to pa rticipate in healthy behaviors (Borrayo & Guarnaccia, 2000). Some of these differences in the receipt of health care between racial and ethnic groups as well as among Hispanic groups are re lated to the ability to speak the primary language of the region (Derose & Baker, 2000; Fiscella et al., 2002). For example, one study found that non-English-speak ing Hispanics were less likely to have a physician

PAGE 54

31 than White non-Hispanic individuals (RR 0.77; CI 0.72-0.83) (Derose & Baker, 2000). If the caregiver does not speak the language of the provider, a se ries of new issues arise. There may be a lack of knowledge regardi ng the need to go, wher e to go, and how to communicate with providers. If th e caregiver brings an interpreter, that interpreter may be missing work or school. Furthermore, translat ion by untrained indivi duals is often less accurate than from a trained medical interprete r either through a lack of knowledge of the medical terms, hesitation in relaying accurate information, or through simple errors in translation (Laws, Heckscher, Mayo, Li, & Wilson, 2003). For example, instructions to take a medication every twenty-four hours could be misinterpreted as every 2 to 4 hours. Derose and Baker (2000) found that the impact of language barriers can be great with the association between the number of phys ician visits and having limited English proficiency being at a similar magnitude as individuals having no health insurance, no regular source of care or having poor health. Another key factor related to not seeking appropriate he alth care that is closely related to the language barrier is the lack of knowledge regard ing health care issues. For example, research has found that displayi ng a video-tape of Poison Control Center education material to low-income and Sp anish speaking populations in WIC clinics significantly improved knowledge, attitudes, be haviors and intentions regarding the poison control center (Kelly, Huffman, Mendo za, & Robinson, 2003). In another study, changes resulting from community outreach including education, pr ompting, and tracking parents in inner-city and s uburban neighborhoods in New York where racial and ethnic disparities in immunization rate s existed were made (Szilagyi et al., 2002). In three years, racial disparities in immuni zation rates were eliminated. In another study, researchers

PAGE 55

32 found that emergency department admi ssions involving Latino families were significantly associated (OR 3.3; 95% CI 1.4-7.76) with parental perceptions of acute need and that parent education could redu ce use (Lara et al., 2 003). Finally, a study of booster car seat use among Latino communitie s found that misinformation, the cost of booster seats, resistance by the child and unava ilability of shoulder belts were the main barriers to use (Lee et al., 2003). Beyond financial, language and health lit eracy barriers, there are a number of studies that identify culture, a lack of spir ituality, and a lack of holistic approaches to health care as barriers as well (American Academy of Pediatrics, Medical Home Initiatives for Children with Special Needs Project Advi sory Committee, 2004; Born, Greiner, Sylva, Butler, & Ahuluwalia, 2004). For example, researchers found that Mexican American women reported cultural preference for traditional ethnomedical alternative forms of health care, the level of employment and education, and dissatisfaction with primary car e as barriers to health care utilization in the US (Iniguez & Palinkas, 2003). Finally, another study expl oring genetic counse ling found that there was miscommunication resulti ng from too much medical ja rgon, the non-directive nature of counseling, misplaced cultural sensitivity inhibiting counselors, problems in translation and problems in trust (Browner Preloran, Casado, Bass, & Walker, 2003). Understanding the Culture According to the American Heritage Dicti onary (2005), culture is the totality of socially transmitted behavior patterns, arts, be liefs, institutions, a nd all other products of human work and thought. Culture exists in everyone, can differ significantly within and

PAGE 56

33 among various groups of individuals, and can cause miscommunications. Specific to this study, cultural barriers can arise based on the in teraction styles of health care providers, the location of services, and even the operati onal procedures of the provider sites. These differences come in a variety of forms a nd are often very subtle. For example, many religions have periods of fasting with time periods that may conflict with the recommended intake requirements and dosing sc hedules of prescription medications such as those that need to be taken on a fu ll stomach (Budda Sasana, 2004; Shaw, 1998; WPI Tech News, 2004). Issues arise either th rough untimely dosing, taking medication on an empty stomach resulting in gastrointestinal problems, or a complete lack of compliance regarding taking the medication. Another barr ier that may be seen from a cultural perspective is the level of trust with the medical profession. Experiences, such as the Tuskegee Study where cases of syphilis in Afri can American men were left untreated for decades to determine long terms effects of the disease, have made some individuals distrustful of the medical community (Jones, 1981). Even where the general assumption is that there are relatively few cultural differences, those differences still exist. For example, although White or Caucasian persons are often referred to as a relativel y homogeneous group, there are at least 53 different categories of European descent re presented by individuals considered to be White in the United States with more than half of those individuals from German, English or Irish descent (Giordano, & Mc Goldrick, 1996). In addition, because most White Americans have been in the United States for at least three generations, there is an assumption that their cultural differences ha ve faded into a shared cultural heritage. However, even the existence of a shar ed cultural heritages does not guarantee

PAGE 57

34 homogeneous emotions, thoughts, or loyalty to specific groups (Giordano & McGoldrick in McGoldrick, Giordino and Pierce, 1996) One issue that did take precedent among early settlers to the United States was the creation of a national identity that followed the Anglo-Protestant value system and English language. In addition, early settlers brought a belief that few external constraints exist to prevent individual success and failure is blamed on personal weakness. In contrast, immigration of pe rsons of African descent was not the result of choice but one of forced slavery where infant mortality was high and families were torn apart as family members were sold off. As part of th is practice of slavery, intentional efforts were made to eliminate any individual culture. In the end, as with ot her racial and ethnic groups, culture and spirituality have often rema ined part of the core of the individuals identity and survival skills (Black, 1996) Although persons of African descent came from a variety of regions with separate cultu res, they also have some similarities across cultures. They often place a great importance on the family and remain close with extended family members. Religion and spiritual ity is also highly valued whether it be Catholicism, Baptist, or Islam (Black, 1996). For the Hispanic or Latino cultures, a lthough there are differences in cultural heritage, there are also a number of co mmonalities across the populations including the Spanish language, high rates of Roman Catholic Church membership, spiritual values and a general willingness to sacrifice material possessions for those spiritual values. Personalism, or the valuing of inner qualities, differs from the more Americanized focus on achievement and is linked to the dignity of the individual and respect for authority. For example, Harwood (1992) found Latino mothers emphasized the childs ability to

PAGE 58

35 behave properly in public while Anglo mothers focused more on characteristics representing independence. The immigration of Caucasian and Black persons into the United States is relatively small in comparison to individuals with Hispanic ethnicity. The more recent immigration of so many Hispanic families brin gs along with it the cultural influence of their native country. These regi onal influences can often play a strong role in individual behaviors including those invol ving the use of medical care. For example, three major concepts in the Latino culture are simpatia, respeto, and fatalismo. Simpatia is a concept that values the expression of politeness or pleasantness in stressful situations and avoids hostile confrontations. As a result, when health care providers expre ss a relatively neutral attitude towards the family, it can be percei ved negatively resulting in distancing from the provider and non-compliance with care (Las siter & Baldwin, 2004). Actions that can promote simpatia include hand shaking, clos e distance during inte ractions, and taking actions that promote a warm personal rela tionship (Lassiter & Baldwin, 2004). Respeto refers to the respect given au thoritative roles such as that of the physician. Respeto may inhibit disclosure and drive a hesitancy on the pa rt of the family to ask questions, even if they do not understand the physician (La ssiter & Baldwin, 2004). In addition to demonstrating respect to the physician, Lati no families also expect to receive it, especially when the family member is older than the physician. If that reciprocal respect is not perceived, families may be less satisfied with their care. Another Latino cultural value is that of fatalismo, or the belief that th ere is little an individual can do to alter fate. Fatalismo can lead to decreased utilizati on of preventive health care and effective

PAGE 59

36 medical treatments. Incorporating the family s beliefs into the solutions can improve compliance (Lassiter & Baldwin, 2004). This study attempts to address the issue of culture in a number of ways. First, it uses instruments that have been validated across persons with a va riety of racial and ethnic origins (Agency for Health Care Re search and Quality, 2005). Furthermore, it restricts participants to t hose of Caucasian (White), Af rican American (Black), or Hispanic racial and ethnic groups who were bor n in the United Stat es and speak English. In addition, literature cited provi des some insight into how indi viduals of different racial and ethnic backgrounds may res pond. Finally, to help compen sate for cultural issues, questions were added to the survey instrume nt to help explain differences among groups. For example, consultation with an anthr opologist led to open-ended responses being added after the term close within the Relati onship Scales Questionn aire being used to measure interaction style. Effective Interventions In addition to identifying specific health disparities, research has also indicated that there are effective strategies for impr oving underutilization. The first step in allowing health care providers to better meet the indi vidual needs of the families they serve is through understanding and responding to the vari ous competing behavioral systems that are associated with access to care. Furthermor e, the act of listening to what the patient has to share during a visit not only provides useful health information but can also provide additional clues that would allow the he alth care provider to be more sensitive to the patients issues. Additionally, listening to the patient has been found to increase the

PAGE 60

37 level of satisfaction with care (Flocke, Mill er, & Crabtree, 2002; Hall, Horgan, Stein, & Roter, 2002). This increased satisfaction with care should, in turn, in crease the likelihood that the caregiver will seek further intera ctions with the provi der. Finally, cultural competence has been found to be a driving fo rce behind improvements in provider skills and patient satisfaction. For example, Beach et. al. (2005) conducted a systematic review of the literature and found th at cultural competence trai ning for providers improves professionals knowledge base, improves the attit udes and skills of pr ofessionals, and that cultural competence training im pacts patient satisfaction. In addition to behavioral change am ong providers, collaboration among programs can also improve appropriate health care u tilization. For example, studies have shown that promoting immunizations more inte nsively through the Special Supplemental Nutrition Program for Women, Infants, a nd Children (WIC) dramatically increased immunization coverage in Milwaukee (Shefe r et al., 2002). Along these same lines, the receipt of well-child care vi sits and other clinical prev entive services may also be improved (Shefer et al., 2002). Parental perceptio ns of care can also influence utilization. Busey, Schum and Meurer (2002) found that pare nts bringing their children to an innercity pediatric clinic reported that they be lieved well-child care was important, especially in reference to immunizations, growth and development issues, and ability to discuss behavioral issues. Parents also reported th at they preferred being provided written information for future reference. One significant change in the provisi on of health care for children was the introduction of subsidized health insurance fo r individuals not eligib le for Medicaid. The implementation of the Medicare, Medicaid, a nd State Child Health Insurance Program

PAGE 61

38 (SCHIP) Benefits Improvement and Pr otection Act of 2000 has increased the accessibility of health care for children in low-income families (United States Senate, 2004). For example, North Carolina reported an increase from 62% to 75% (P<.05) in well-child visits and an increas e from 68% to 78% (P<.05) for acute care in the private sector while the proportion of children with unmet medical ne eds declined from 20% to 2% (p<.05) (Slifkin et al., 2002). Factors that Influence Behavior There has been a persistent debate over th e influences of natu re versus nurture. Understanding the relationships between these tw o sets of factors is difficult. The use of twin studies, especially among those individu als separated early in their development, have provided a great deal of insight into the role of heredity and the role of environmental experiences. Advances in unde rstanding have also been made as a result of the human genome project research in recent years. Findings from this area of research supports the impact of both issues in the development of individual traits including behavior. Genetics and Behavior The advancements of genetic mapping re search efforts, such as the Human Genome Project, has increased interest by me ntal health researchers in more clearly identifying the genetic influences on mental health issues and behavior (National Institutes of Mental Health, 2000). Much of th is research involving behavior and genetics surrounds mental health issues, especially in relation to studying associated behaviors in

PAGE 62

39 mothers and their children as well as among siblings. One of the most common findings is an association between ge netics and anxiety disorders (Marks, 1986; National Institutes of Mental Health, 2000; Spence, Rapee, McDonald & Ingram, 2001). Additionally, Hudziak et. al., found a genetic influence in addition to environmental influences on Child Behavior Checklist Obsessive-Compul sive Scale Scores. Studies involving depression, anxiety, and phobic di sorders have also found an overlap between the two factors (Eley, Bolton, OConnor, Perrin, Smith & Plomin, 2003; Gillespie, Zhu, Heath, Hickie and Martin, 2000). A lthough some of the childhood depressive symptoms may be associated with maternal depression, not all of the variance is accounted for by the mother (Rice, Harold & Thapar, 2002). Ri ce, Harold, and Thap ar (2002) also found stronger genetic influences for boys than for girls. Finally, studies focusing specifically on genetics and interaction style suggest that anxious and avoidant st yles were associated with higher rates of alcohol use disorders as compared to individuals with secure interaction styles (Vungkhanchi ng, Sher, Jackson, & Parra, 2004). Attachment Theory and Health Care Utilization Although genetics may set the stage for futu re behaviors, environmental factors also play a strong role in guiding the development of thos e behaviors. Focusing more specifically on health care utili zation behaviors, it may not be clear why all individuals do not receive well child care given the expansi on of the extent of prevention efforts that have been implemented in recent years. Ther e are a variety of theories that attempt to explain this behavior. There is one theory that suggests attachment bonds influence helpseeking behaviors. The foundations of attachment theory can be seen in the work of John

PAGE 63

40 Bowlby who incorporated knowledge gained fro m the fields of evolutionary biology, ethology, developmental psychology, cognitive science and contro l systems theory (Bowlby, 1969). According to Bowlby, attachment refers to behaviors that result in a person desiring proximity to anot her individual, usually one that is con ceived as stronger or wiser. In addition to the strength of the mothe r-child bond, attachment theory focuses on the internalization and the temporal generaliz ation of infant-parent relationship styles while incorporating the concepts of cognitive science and control-systems theory (Boss et al., 1993). It is postulated that from birth, human s are social and have an inherent need to interact with a caregiver. Theory emphasi zes that a biologically-based desire for proximity evolved through the process of na tural selection whereby infants who stayed closer to their mothers were more likel y to survive to adul thood (Bowlby, 1969). Unlike various animal species, this moth er-child attachment bond does not occur immediately but develops slowly over the first six to nine months of life and occurs only between the infant and few individuals, usually caregivers. The result of these bonds is a synchronization of behavioral re sponses based on cues of the infant and of the caregiver (Klaus, Kennel &Klaus, 1995). Due to the ch anging threats and issues that arise as individuals become older and are exposed to more experiences, there is an increasing need for generalization to a broader range of behavioral cues and actions which Bowlby refers to as an attachment behavior syst em (Bowlby, 1969). This attachment behavior system allows the individual to be able to p redict the behavior of others and to plan ones own behavior to achieve relationa l goals (Feeney & Noller, 1996 p. 193).

PAGE 64

41 This variety of behaviors, compared to having just one behavioral association, allows for the flexibility to adapt to new situ ations later in life while still allowing for the progression toward the initial goal (Cassidy & Shaver, 1999). In addition, Stroufe and Waters (1977) noted that these behaviors do not have absolute and constant patterns but rather include an assortment of behaviors that have developed similar meanings and functions. There is also the belief that c ontinued adaptability in the development of attachment behavioral styles occurs regardless of whether th e relationship is positive and nurturing, as can be seen by the attachment dynamics that develop in families experiencing issues of child abuse a nd neglect (Ainswort h, 1967). Finally, Bowlby (1969) proposes a control system that prom otes homeostasis similar to that of a thermostat, the difference being that adjustme nts are continual rather than turning on and off to moderate the responses of wanting close proximity and wanting independence. When separation has become too great, the urge for proximity increases and when proximity has been achieved, the urge di ssipates, allowing for more desires of independence (Cassidy & Shaver, 1999). During the development of attachments, experiences can inhi bit the development of positive associations among certain behavior s and outcomes. For example, a child for whom requests for comfort and assistan ce are ignored could develop two opposing behavioral styles of acceptance and rejecti on. These early maladapt ive associations can reduce the flexibility of future behavioral as sociations incorporated into the attachment behavioral system (Bretherton, 1999). The development of attachment strategies results in a variety of working models for how an individual interacts with the envir onment. To help identify and describe these

PAGE 65

42 various working models, the Strange Situ ation Research approach was developed (Ainsworth, 1985). Strange Situ ation research refers to a series of studies where a mother and baby are joined in a room by an unfamiliar woman for short periods of time. At some time the mother leaves the room a nd then returns. The infants behavior during the mothers absence and reentry are then observed. Three primary reactions were identified through this research involving strange situations (Ainsworth, 1985). An infant whose mother was sensitive to cries and need for food and comfort during the first three months of life, generally welcomed its mothers return. In a playroom situation, these mothers we re more likely to allow the infant to play independently and intercede when the infant di splayed signs of distre ss. As a toddler, the child generally worked independently on probl em-solving tasks. Mothers intervened only when the child became stuck and asked for assi stance. This response is an indication of a secure attachment style. Contrary to these secure reactions, if the mother was more insensitive to the infants needs, the infant was more likely to reject the returning mother by snubbing her, looking, turning, walking away, or refusing interaction bids (Bretherton, 1999, p. 283). This reaction is referred to as insecure-avoida nt and is generally associated with mothers who provided less affectionate holding duri ng the first three months and frequently rejected bids for close bodily contact during th e last quarter of the fi rst year (Bretherton, 1999). In playroom situations, mothers play ed with their children when they were cheerful and withdrew if the infant displaye d negativity. These moth ers further reported to researchers that they dis liked bodily contact. As toddlers, insecure children tended to give up easily, whine, and their mothers tended not to provide assistance.

PAGE 66

43 Infants who responded ambivalently such as by allowing body contact while also displaying angry or resistant be havior is referred to as insecure-ambivalent and resulted from mothers being inconsistently sensitive at home, ignoring the infants signals but not rejecting close bodily contact (Ainsworth & Bell, 1969). From another perspective, avoidant infants were less likely to communicate with th eir parents nor seek bodily contact when stressed by separation while th e secure infants always remained close to their parents when unhappy (Klaus, Kennel & Klaus, 1995). Further development of the theory of a ttachment expanded these models to four by adding a fearful model that includes a negative sense of both self and others (Bartholomew, 1990). This fourth group was de scribed as insecure-d isorganized because the infant displays a combination of str ongly avoidant and resi stant behavior or disorganized behavior upon reunion (Main M. & Hesse E., 1990). Disorganized behavior may be a sudden cessation of behavior during a greeting and other be haviors that do not make sense. This fourth categ ory has not received the same validation as the first three, and parents of infants displaying these behavi ors differ in a variety of characteristics. Under the three category model, these individuals tend to be unclassifiable and are often excluded from research studies. Generally speaking, regional and cultural variability among individuals often can alter the expected dynamics of theoretical models. In the case of attachment theory, findings from Strange Situat ion research indicate th at although there are regional variations in the way individuals behave in these new situations, there is some level of cross-cultural validity in i ndustrialized nations (Boss et al., 1993). This consistency across groups acts to strengthen the structure of the model.

PAGE 67

44 Looking beyond just the mother-child bond, a ttachment research has provided one potential model for understanding how parental sensitivity to an infants needs evolves into the intergenerational patte rning of relationships. This expansion of attachment bonds to other individuals points towards the dynamic attributes of the human ecology theory where a variety of systems influence be havior (Simpson & Rholes, 1998). Additional postulations indicate that there is a hier archical and temporal development of the attachment behavior system in which each st age of development influences the next and that the working models of attachment rema in somewhat stable over time (Kerns, 1994; Low, 1991). It has been suggested that this attachme nt begins as early as the prenatal period and remains consistent after birt h (Levine, Tuber, Slade, & Ward, 1991). Furthermore, researchers found that self-report ed adult attachment st yles were strongly correlated with parent-ch ild bonding relationships (Ede lstein et al., 2004). This stability of attachment over time implie s that if an individual is able to have secure relationships with one car egiver, he or she is more lik ely to be able to develop a secure relationship with other caregivers. Fo r insecure persons, this development of behaviors also implies a type of self-fulfilli ng prophecy regarding the expectations of the individual toward others. As they place negative expectations on others, their own behaviors can then elicit nega tive behaviors from others resu lting in a validation of their initial expectations. This rec ognition of a life-course aspect of behavior led to further exploration and expansion of the model throughout life.

PAGE 68

45 Attachment Bond Now that the attachment behavioral syst em has been described, a closer look can be taken at the concept of an attachment bond. This bond is an internal ized characteristic where one individual has a bond to another individual, usually with someone who is perceived as stronger and wi ser (Ainsworth, 1989). This bond cannot be inferred by behaviors because most behaviors can be ut ilized by a number of systems for different reasons (Bretherton & Ainswort h, 1974). As a result, the st rength of the attachment behaviors should not be confus ed with the strength of an attachment bond (Cassidy & Shaver, 1999 p. 13). Furthermore, Bowlby propos es that the attachment bond is only one aspect of a mother-child relationship relating to security and protec tion during periods of stress (Cassidy & Shaver, 1999). The mother ma y also play the role of disciplinarian, playmate, or teacher without necessarily bei ng in conflict with the role in attachment. However, in situations of stress, the attach ment motivator is give n priority (Cassidy & Shaver, 1999). Conversely, an attachment bond cannot be assumed merely because an attachment component is present. For example, friendliness on the part of the infant to a stranger does not imply there is an attachme nt bond present. This example extends into relationships later in life such as through inte ractions with peers. Although most of the references in the literature discuss the mother-child relationship, it is no t the only attachment bond that can exist within the attachment behavioral system. The number of attachment relationships is relatively few in infancy and steadily grows over time. Fathers, siblings, other relatives, extended families and other non-related caregivers can also devel op such bonds although a hierarchy generally exists within these relations hips (Cassidy & Shaver, 1999). Th is hierarchy is associated

PAGE 69

46 with the time spent with the infant, the quality of care provided, emotional investment by the caregiver, and social cues (Colin, 1996). As the individual ages, additional attachments may be developed with mentors, sexual partners, and othe r central figures in a persons life (Cassidy & Shaver, 1999). Adult Attachments As noted earlier, although initially focu sed on the relationship between a child and caregiver, further development of attach ment theory expanded its scope into adult relationships. Shaver and Hazan (1993) iden tified similarities between the attachment relationships within children and those among adults For example, reciprocation, sensitivity and responsiveness are associ ated with the quality of both types of attachments, with secure individuals mo re likely to be happy and more adaptive. Furthermore, the attachment mechanism of proximity still applies. Additionally, separation between individuals increases stress and results in the in itiation of behaviors that lead towards improved proximity. Finally, there is increased sensitivity to approval from the attached individual regard ing the display of new discoveries. Patient-Provider Attachment and Health Care Utilization One type of adult-adult attachment is that bond between i ndividuals and their health care providers. The dynamics of this patient-health care provider attachment bond can vary greatly with the level of dominance by the health care provider. The health care provider generally has a higher level of knowledge regarding health issues and has the power in the relationship, such as to order tests or write prescriptions. Therefore, the

PAGE 70

47 provider can easily have a disp roportionately authoritativ e role in the relationship. However, according to the research, there is greater patient satisfaction with care when the relationship operates in a more collabo rative manner (Flocke, Miller & Crabtree, 2002). Attachment theory suggests that individuals with s ecure, anxious, or avoidant attachment styles will behave differently and that those behaviors extend into the health care arena. The theory implies that women with a secure attachment style tend to seek appropriate health care when necessary and ha ve the confidence to resolve minor issues independently. The theory also suggests that mothers with anxious styles of attachment will desire closer proximity to health care provi ders in times of stress, such as a childs illness. Anxious mothers may over-utilize health care services in search of confidence that the health issues are being adequately addressed. Finally, women with avoidant attachment styles would tend to be uncomfor table being close to others, including health care providers, in times of stress and may reject health care. This a voidant behavior can have an impact on the health of the child but it can also be costly to the system of care if appointments are not kept. Research of female patients using HMO funded health care services supports these hypotheses. This study found that anxiou s women tended to be overly dependent on providers, report more somatic symptoms (p<.03), and over-utili ze care (p<.003) in comparison to secure women (Ciechanowski, Walker, Katon & Russo, 2002). Interestingly, the number of symptoms and utilization patterns among avoidant individuals did not differ signi ficantly from secure individuals. Another study involving the self care and outcomes of diabetics found th at individuals with avoidant attachment

PAGE 71

48 styles were less likely to exer cise (p<.01), practice foot car e (p<.05), follow a proper diet (p=.001), and adhere to oral medications (p <.05) when compared to individual with secure attachment styles (C iechanowski et al., 2004). Res earchers also found that these differences were mediated through the patien t-provider relationship measured by using 3 questions from the Assessment of Chronic I llness Care. Outcome indicator found that individuals with anxiou s attachment styles had significan tly lower rates of glycosylated hemoglobin levels >8% compared to individuals with secure attachment styles (P<.05). To further explore attachment theory health behavior and health outcome, Thompson and Ciechanowski (2003) conducted a review of the lite rature. Researchers found that attachment theory can serve as a useful model for identifying important features of the patient-physician rela tionship and for providing an increased understanding of how to provide improved clinical care. The clinical relationships and health outcomes identified in these studies indicated the benefits to using a stepped approach to providing care to non-secure in dividuals requiring increased levels of communication such as telephone calls, remi nder postcards, and emails. Findings also indicate that services may need to be expanded to include not only the primary care provider but also nurse case manager, soci al workers or other supportive individuals working together to provide care (Ciechanowski et al., 2004). Understanding and implementing different management strategies, such as the ones noted above, can increase th e effectiveness of health ca re services for individuals with specific attachment styles. For example, individuals with avoida nt attachment styles may benefit from approaches that accommoda te the patients need for autonomy and interpersonal distance such as through increased flexibility regarding appointment

PAGE 72

49 duration and scheduling (Thompson & Ciech anowski, 2003). Patients with anxious attachment style may benefit from attention before the patient asks for it such as scheduled appointments and reminder cards. This care-eliciting approach allows the patient to believe they will receive support regardless of their symptoms. The provider should also be non-intrusive and consiste ntly responsive(Thom pson & Ciechanowski, 2003). Incorporating the various aspects of at tachment style into the patient-physician relationship can lead to impr oved patient care and enrich the clinical experience. Interaction Styles The measurement of attachment can take many forms from very detailed observations such as those involved with the Strange Situation approach to questionnaires regarding specific relations hips between two individuals. However, because research issues have arisen that cannot be addressed using these specific relationship measures, another approach has al so been developed and used over the years. This alternative approach resulted in the development of a more general assessment of the way in which individuals interact with others in their environment and is based on attachment theory. This general style of att achment behavior will be referred to as an interaction style and includes the use of the same secure, anxious and avoidant categorizations. This interaction style will be the focus of the current study since the interactions between individuals and health care providers often are relatively infrequent resulting in little oppor tunity for specific attachment bonds to develop. This lack of continuous care is further exacerbated by th e expansion of managed care systems where

PAGE 73

50 provider networks can change every year and by practices that include greater numbers of physicians (Sultz & Young, 1999). The measurement of interaction styles has been conducted for decades with moderate levels of reliability and validity of the measures across ra cially and ethnically diverse populations in the US as well as in other countries such as Italy (BuschRossnagel, Fracasso, & Vargas, 1994; Farma & Cortinovis, 2001). One of the largest studies, conducted in 1996, utili zed a interaction style measure, the Relationships Scales Questionnaire developed by Hazen and Shaver that will also be used in the current study (Mickelson, Kessler & Shaver, 1997). The Rela tionship Scales Ques tionnaire creates a continuous rating for each of the interact ion styles. Subsequently, many researchers identify the interaction style with the hi ghest score as the individuals dominant interaction style. Informati on regarding the inte raction style was collected from a nationally representative sample of 8,080 individuals. Findings indicated distributions of interaction styles that were different across racial and ethnic populations. However, the greatest differences were between Black and White respondents with Hispanic respondents falling in between the two. The study found that 61% of Whites, 58% of Hispanic and 51% of Black respondents were found to have dominant interaction styles defined as secure (p<.05; White different from Black different from Hispanic). Avoidant styles represented 25%, 21%, and 28% respectively (p< .05; White different from Hispanic). Anxious interaction styles were identified for 10% of Wh ite, 15% of Hispanic, and 16% of Black respondents (p<.05; White different than Black and Hispanic but no differences between Black and Hispanic). Fi nally, 4.1% of White, 5.2% of Hispanic and 5.9% of Black respondents could not be classifi ed with a dominant interaction style.

PAGE 74

51 These interaction styles can influenc e all aspects of life. For example, Ciechanowski et al. (2004), assessed the intera ction styles and choice of specialty studies by medical students, finding that the prevalence of interaction styles were similar to the general population and that secure student s were more likely to choose primary care compared to both anxious (5.9 OR) and avoida nt (2.4 OR) students. Studies have found a variety of differences in ba sed on other characteristics as well. Magai et al (2001) found that younger adults had more secure inte raction styles than elderly populations. Conversely, Broussard (1995) found that adoles cent mothers were much less likely to have secure interaction styl es (23.7%) than the more comm on rate of 55-65% found in other studies of the general population. The Role of Cognition and Context Although interaction styles can have a str ong relationship with behavior, they are not the only factors involved. There are ot her systems that can complicate behavioral responses, often due to conflicting goals. One such system is the need for exploration. The fear system is also strongly associated with behavior. Additionally, there is a need for humans to be sociable and to have a caregiver system. These other systems may provide competing drives or they may provi de a synergistic eff ect on the desire for proximity (Cassidy & Shaver, 1999). In relation to the utilization of health care, fear from a domestic violence situation may prevent the woman from leaving the house to take her child for care. Beyond competing systems, there are other i ssues within individuals that also are associated with attachment and interaction styl es. One internal characteristic is that of

PAGE 75

52 cognition. Bowlby proposed that cognitive compone nts such as mental representations of the object of attachment, the self and the environment are all heavily influenced by experiences and help to organize the a ttachment behavioral system (Bowlby, 1969). Bowlby believed that these working cognitive models offered individuals the ability to anticipate the future and make plans (C assidy & Shaver, 1999). As the attachment behavior system grows with new experiences the individual is better able to make decisions regarding which specific behavior s to use during different situations. The cognitive aspects of the attachment theory help the individual maintain and organize these responses. Emotion is another key factor affecting cognition and influencing attachment because many of the activitie s that promote the development of the attachment behavior system stem from emoti onal experiences such as love, pain, fear and anger (Cassidy & Shaver, 1999). Finally, context also plays an important role in both cognition and in the activation of attachment desires for proximity to the caregiver. This context can be in regard to the state of the i ndividual or to the environment (patient, caregiver, office) (Cassidy & Shaver, 1999). Of pa rticular importance are the lo cation and behavior of the caregiver. Furthermore, there is a great de al of complexity among the various factors associated with interactions ranging from mere proximity to mo re specific actions on the part of the caregiver. Variabili ty also exists in the outcome needed to terminate the urge for proximity. This too can range from mere proximity to a specific action.

PAGE 76

53 Other Behavioral Motivators Interaction styles do not operate in a v acuum but are in competition with other behavioral motivators. The literature has doc umented a number of other systems that are associated with behaviors. One of the key f actors is that of fina ncial constraints (Kim, Symons, & Popkin B. M., 2004; Wang, Gisondi, Golzari, van der Vlugt, & Tuuli, 2003). If the caregiver has limited financial resour ces, help-seeking actions can be limited to a certain extent. The child may not have health insurance to cover the costs or the caregiver may not be able to afford the co-payments required by the insurance plan. Transportation may be an issue, especially if other child ren in the family also require simultaneous supervision. Furthermore, many lower-paying jo bs do not offer health insurance, or may not offer sick leave from work to attend we ll-child visits. Additionally, some jobs, such as assembly line work, do not lend themselves to taking portions of a day off regardless of whether that time is paid leave. Alt hough a number of efforts have been made to alleviate some of these financial barriers, su ch as prenatal Medicaid, the States Child Health Insurance Program, and subsidized transportation, many barriers like those resulting from absence from work are not as readily addressed (Cen ter for Medicare and Medicaid Services, 2004a; Center for Me dicare and Medicaid Services, 2004b). In addition to the more structural barriers created by low socioeconomic situations, attitudinal barriers also are associ ated with behavior. For example, there are many instances where individuals receivi ng governmental subs idies are treated negatively by others due to the stigma of needing assistance. Finally, anecdotal reports from low-income prenatal patients in the H ealthy Start program in Florida have reported that they often felt so disrespected at some of their health care visits that they did not

PAGE 77

54 go back for care as often as was recommended by their health care provider (StruchenShellhorn, 2000). Efforts have been made in some discip lines to mask this issue of financial assistance to minimize the negative reacti on such as through the use of electronic benefits transfer cards (e lectronic food stamps) rather than paper coupons used previously (Food and Nutrition Service, 2004). School lunch programs offer similar electronic payment systems for all students so that those receiving free or reduced lunches are not as easily identified (Evolu tion ID Card Systems and Badge Supplies, 2004). Even with these efforts, individuals are still often treated diffe rently based on their ability to pay. Maternal Depression In addition to economic issues and th e resulting barriers that can inhibit participation in health care, there are individu al differences and other health issues that can also play a significant role. One issu e that has been found to have a strong relationship with the utilization of health care is maternal de pression. It is estimated that more than 20% of women experience some level of postpartum depression (McLennan & Kotelchuck, 2000). Furthermore, Zapata (2005 ) found that maternal depression scores declined progressively from one to 15 months after delivery and then increase at 24 months. Maternal depression impact s all aspects of the womans life including maternalchild bonding relationships, family functioning, and an impaired ability to adequately care for the child such as a reduction in the amount of hea lth care received (American

PAGE 78

55 College of Obstetrics and Gynecologist s, 2004; Nemours Foundation, 2004; Nicolson, 1998; Zimmer & Minkovitz, 2003). For exam ple, McLennan, Kotelchuck, and Cho (2001) used a nationally representative sa mple of women with newborn children and found that depressive symptoms were significantly associat ed with not breast-feeding (OR 1.25; 95% CI 1.06-1.47). Depression was also associated with the pregnancy being mistimed or unwanted (OR 1.40; 95% CI 1.19-1 .64), and the child having a poor health status versus excellent health status (O R 3.48; 95% CI 2.20-5.50). Additionally, maternal depression was associated with decreased util ization of child safety prevention practices such as the use of car seats (p<.0001), elec trical plug covers (p <.0001), and having syrup of ipecac (p<.0001) around the house (McLennan & Kotelchuck, 2000). The impact of maternal depression on children can extend even further as time goes on. For example, Zapata (2005) found an association between maternal depression and the childs level of social competence. However, the impact of maternal depression varied by the time of onset. The study also found that th is impact was only found for depressive symptoms severe enough to reac h the threshold of a depression diagnosis. As a result of the negative associations between maternal depression and healthy behaviors, efforts should be made to identify early signs of postpartum depression to help reduce long-lasting effects on the mother and child. Experiences of Care Reported by Patients As noted earlier, cognition a nd context can influence moth er-child and adult-adult relationships. The same can be said for thei r roles in the case of patient-health care provider relationships. For example, one mother may be satisfied with a phone call from

PAGE 79

56 the provider regarding a health related issue and in another case, a mother may want to not only see the doctor but also have a conf irmatory tests conducted and perhaps even receive a prescription before she feels comforta ble leaving the proximity of her caregiver. This patient satisfaction is also associated with a variety of other positive issues concerning health care. For example, patie nts who reported liking their physician had positive associations with better self-reported health (p<.01), more favorable ratings of providers (p<.01), and higher le vels of overall satisfaction (p <.01) (Hall et al., 2002). In addition to improving satisfaction, pa tient-health care provider interactions, location of the service facilit y, the type of funder and the ty pe of provider can also impact satisfaction with services and patient outcome s. For example, research indicates that variations in service delivery between publicly funded prenatal care locations and private office exist (Kotelchuck, Kogan, Alexande r, & Jack, 1997). Results found that the publicly funded sites provided more comprehens ive prenatal care serv ices, indicating that any generalized assumptions regarding equal access to care may not be accurate. In addition to site differences, researchers have al so identified differen ces in birth outcomes for women served in practices with single pr oviders versus groups of providers. Ickovics, et. al. (Ickovics et al., 2003) found that lowincome women served by public clinics in Atlanta, using groups of clinicians in one pr actice rather than singl e providers, had better birth outcomes. (p<.05) Given the variability of fact ors that influences satisfact ion with health care and ultimately with health outcomes, there is a need to explore the inter-relationships of these factors from a more standardized, criteri on-based approach (Mer rill & Allen, 2003; NgoMetzger, Legezda, & Phillips, 2004). There have been recent efforts to accomplish this

PAGE 80

57 task, starting with operationally defining th e issues in a more universally accepted manner. For example, although definitions vary, there is a general consensus that primary care should be accessible, continuous over time coordinated, communicated in a way the patient can understand, and based on the cumula tive knowledge of the patient and family (Seid et al., 2001). Measures of that care have shifted from ones based on satisfaction, representing expectations and preferences, to a measure of reported experiences of care based on a specifically prescribed criterion. This paradigm shift is important because satisfaction can vary greatly among individuals and does not provide specific strategies for improving the system (Sta rfield, Cassady, Nanda, Forrest, & Berk, 1998). As a result, deviations in the ratings of the criterion measures represent changes in quality and provide areas of improvement. Consumer Assessment of Health Plans Survey The push for standardized measures that can help compare findings across studies has been amplified by the need for measures of quality for evaluating the impact of increasing diversity and competitiveness of health care plans, such as managed care. In response, nearly a decade ago the Agency for Healthcare Research and Quality (AHRQ) began to establish a pool of questions and in struments to capture the reported experiences of consumers in health plans for both adults and children or parents of children. This Consumer Assessment of Health Plans (CAHPS) initiative developed an integrated set of standardized, valid and reliabl e questionnaires and other data collection instruments in both English and Spanish that have been in corporated into a number of data sets

PAGE 81

58 nationwide (Agency for Health Care Resear ch and Quality, 2004b; Weidmer, Brown, & Garcia, 1999). A subset of these questions was selected for use in this study. In addition to the instruments themselves, a CAHPS Survey and Reporting Kit 2002 was developed to aid in utilization a nd dissemination of the findings (Agency for Health Care Research and Quality, 2004b). This includes sample formats for reporting, software to assist in data analysis, as well as guidance for implementing, reporting and evaluating the results. Targeted users of this tool kit and database include Medicaid, and public and private employers as well as i ndividual health plans. Finally, the AHRQ has established a CAHPS benchmarking database to facilitate a shar ing of results among CAHPS users. At this point in time there ar e six years of data including adults and children receiving health care through comme rcial insurance, Medicaid, SCHIP, and Medicare (Centers for Medicare and Medi caid Services, 2004b). For the years 2003 and 2004 there were over 760,000 records in this database including 161,848 children. Most (117,240) of the child records were from child ren receiving health care services through Medicaid including 2,262 from the Florida Medicaid population (Agency for Health Care Research and Quality, 2004c). Initial CAHPS research of the field-test da ta used a factor analysis to delineate factor structures for all of the questions (Bender & Garfi nkle, 2001). Analyses identified a three-factor structur e including 1) quality of provider or staff communication, 2) timely access to quality health care, and 3) quality of plan administration. Slight differences in experiences among health care provider-types in dicated variability in the structure of their medical care delivery systems (Bender & Garfinkle, 2001). Differences focused on two questions. The loading of the question regarding ge tting a satisfactory doctor

PAGE 82

59 indicated that the provider factor was le ss focused for adult, privately insured respondents. The factor loading for getting th e care believed to be necessary may have resulted from some slight modifications of the questions during testing. For the purposes of the current study, the overa ll rating items was be used rather than developing composite scores. Finally, the focus on one h ealth care service delivery system attempts to minimize some of the variability identified in the field testing study. CAHPS research regarding i ndividual questions has found that shorter visits and missed or delayed care were associated with lower ratings (Halfon et al., 2004). Additional research indicated th at individuals of Asian descen t had lower ratings of their care across many questions but were also less li kely to change doctors because of their dissatisfaction. For example, Asian American individuals were less likely to receive counseling and less likely to report positive in teractions with doctors than Caucasians (Ngo-Metzger et al., 2004). Another study found that Hisp anics also reported lower ratings on the question regarding whether physicia ns listened to them carefully (Merrill & Allen, 2003). The cause of these differences is not known but a variety of racial disparities within health care services is suspec ted. In an attempt to control for this issue, the current study excludes persons of Asian de scent due to small their relatively small population. Furthermore, targeting clinics serv ing Medicaid participants increases the proportion of minorities being served at th e facility. This increased exposure may improve cultural sensitivity to patient needs. In addition to merely improving the quali ty of reported experiences of care, patient-health care provider interactions have also been found to impact outcomes.

PAGE 83

60 Flocke, Miller, and Crabtree (2002) found that physicians with a person-focused style of interacting had both higher pa tient satisfaction and patient outcomes, while physicians with high control styles of interacting had some of the lowest ratings (p<.001). Furthermore, the visit time for controlling physicians was, on average, two minutes shorter than for person-focused physicians. Some factors that are asso ciated with patient satisfaction include not provi ding enough information, not pr oviding explanations in a manner that is understandable to the patient and not providing enough time to answer questions (Keating et al., 2002). Another study addressing health care util ization is the Medical Expenditure Panel Survey (MEPS), an ongoing nationally repres entative survey of the US civilian noninstitutionalized population sampled from participants of the CAHPS study. The MEPS collects more detailed data regarding health care utilization, expend itures, source(s) of payment, quality, and insurance coverage (Age ncy for Health Care Research and Quality, 2004c). Data are collected through multiple contacts over a 2-year period from participating households. Efforts such as the CAHPS and MEPS ha ve found associations between reported experiences of care and known ri sk factors, such as race, ethnicity, and income levels (Simpson et al., 2004; Weinick, Jacobs, Stone Ortega, & Burstin, 2004). For example, one issue that may alter minority womens expe riences or perceptions is that of racial concordance between patient and providers which research has found to be positively associated with satisfaction (Laveist & Nuru-Jeter, 2002). Although the general findings illustrate disparities, most studies recomme nd further research regarding the underlying factors that influence those associations.

PAGE 84

61 Information gained from research effo rts such as the CAHPS and MEPS have been combined with the customer service e xpertise of Walt Disney World to develop a unique intervention strategy known as National Fr iendly Access. This maternal and child health initiative is a long term, research-based, community development and educational program to improve consumer access, use, sa tisfaction and outcomes (Lawton and Rhea Childs Center, 2004). Friendly Access is a c ooperative agreement with the Lawton and Rhea Chiles Center for Healthy Mothers and Ba bies and the Centers for Disease Control. Identifying A Target Population Exploring the underlying motiv ators to health utilizati on can be very difficult due to the multitude of factors involved as well as potential mediators and moderators that could influence those associations. For exam ple, age, socioeconomic status, insurance status and medical history can all impact health care utiliza tion (Moore & Hepworth, 1994;Szilagyi et al., 2004). This variability can threaten the internal validity of a study because if the dependent score is different among groups, the difference may be due to the independent variable or it may be due to the subject-related variable (Athabasca University, 2004). Appropriate se lection criteria can help minimize the variability among factors with the potential to th reaten internal validity. On the other hand, a sample that represents a sub-population that is too specific limits the ge neralizability of the findings beyond a small portion of the population, even though the identified associations may be strong. Balancing these two concerns is important for conducting a manageable and useful research study.

PAGE 85

62 In regard to this study, selection crite ria attempt to minimize differences in participant characteristics and individual histories. The study focused on mothers of infants who receive pediatric health care services through Medicaid funding. Selecting these women controlled for the influence of so cioeconomic status and yet still represent about half of all births in Florida (Age ncy for Health Care Administration, 2004) Furthermore, the lack of health insurance or breaks in health insu rance coverage impact utilization. Infants receiving Medicaid are no t re-assessed for eligibility until 12 months of age. By restricting participation to in fants who have been receiving Medicaid-funded health care services since birth removed the i ssue of interruptions in health care coverage. Another restrictive issue is whether to invol ve children with special health care needs who require more frequent health care vi sits due to their underlying health issues. Excluding infants with special health care ne eds can increase internal validity. Finally, selecting only US-born, English-speaking par ticipants helped to control for language issues and some of the cultural differences th at may limit the validity of the instruments. Advantages and Disadvantages of Face-to-Face Interviews Gathering data to conduct research can be conducted in a number of ways. Each of these approaches has its own strengths a nd weaknesses. For example, mail surveys are less expensive and allow for a larger number of respondents to be surveyed in a shorter period of time. Telephone surveys often have higher response rates than mail surveys. Finally, direct contact surveys, such as telephone surveys and f ace-to-face interviews, provide the ability to build more rapport and offer the flexibility to ask follow-up questions when needed (Dominowski & Bartholet, 2004).

PAGE 86

63 Beyond the administration issues for su rveying individuals, direct contact interviews also have advantages in maximizi ng the content of the surveys. For example, the response rate to each of the survey questi ons is higher via a direct contact format than self-administered due to researcher prompti ng (Dillman, 2000). This is increasingly true for open-ended questions due, in part, to the ability to probe furthe r. Additionally, selfadministered surveys, such as those conduc ted through the mail, allow the respondent to skip around the instrument when answering que stions. Direct contac t interviews provide increased control over the ordering of th e questions and allow the researcher to incorporate more complex skip patterns to the questions (Dillma n, 2000). Finally, direct contact surveys also can potentially reduce dist ractions, such as side conversations with others that can happen when comp leting mail surveys (Fowler, 1988). Using trained, calibrated interviewers in both face-to-face and telephone survey approaches enhances the reli ability and validity of survey responses as compared to having each respondent completing the survey in their own way, such as is the case for mail surveys (Salant & Dillman, 1994).These approaches allow flexibility to incorporate follow-up questions and allow interviewe rs to observe informative body language. There are also disadvantages to conduc ting face-to-face interviews. The greatest issue is usually the added cost. The need to use trained interviewers the time to address scheduling issues and travel as well as the added cost of transportation to the interview can all be avoided by using less intensive methods of data collection. Balancing the benefits with the costs is an importa nt step in the study design process. Once a data collection approach has been chosen, efforts need to be made to understand and maximize participant response rates. According to Dillman (2000), three

PAGE 87

64 issues influence response rates. First, resear chers must minimize the cost to participants for responding. Costs can include a number of issues including time, physical or mental effort, and risking embarrassment as well as any financial costs. Second, the researcher must maximize the rewards an individual re ceives for participation in the study. These rewards do not have to be fina ncial but may also come from satisfaction with helping to make positive changes or believing that indi vidual opinions matter. Finally, researchers must establish trust that those rewa rds will be delivered (Dillman, 2000). Summary of Research As noted earlier, the benefits of receiving preventive health care, especially in the early stages of life, can significantly impact the length and quality of life. Understanding the multitude of factors that infl uence the utilization of that care is paramount to quality improvement efforts. Although many of the factors that commonly influence health outcomes, such as race, ethnicity, and soci oeconomic status, have been recognized as influencing the receipt of health care, the underlying issues that dr ive these differences are not as well understood. This study attempte d to provide a more detailed exploration of these factors. First, to control for insurance status, cu ltural differences, and income level, the selection criteria focused on a hi gh-risk sub-population of all births in the counties. Understanding why some indivi duals in this more homogeneous population receive adequate health care while others do not can help enhance services to better meet the needs of all infants and th eir families. A multivariate modeling approach was used to help answer these questions and also allow e xploration of the inter actions among factors.

PAGE 88

65 Chapter Three Methods This chapter outlines the methods utilized in the study. It is organized in three sections: (a) the purpose of th e study; (b) the research ques tions and hypotheses; and (c) the methods. Purpose of the Study The purpose of this study was to determin e the relationships between a mothers interaction style, her reported experiences w ith pediatric health care and utilization of pediatric health care services for her child as measured by health care visits and immunizations. First, the association between mate rnal interaction style and reported experiences of care was assessed. Second, the association between reported experiences of care and health care utiliz ation was explored. Next, the study assessed the association of maternal interaction style, while controlling for experien ces of care, and health care utilization. Finally, the relationship of pot entially moderating factors was assessed. Funding and Other Resources Funding to conduct this study was obtaine d through the Pediatric Clinical Research Center of All Childrens Hospital and the University of South Florida (PCRC) (see Appendix D). Funding was available for one year. Resources paid for recruitment

PAGE 89

66 materials (postage, printing, laminating posters etc.), participant stipends, mileage and other costs. Remaining funds provided a small stipend for the principal investigator. The PCRC also made available biostatistics cons ultation upon request. Given the distance to the Tampa Campus of the University of S outh Florida, having a local resource for support was helpful. Design This study was based on a quantitative, cr oss-sectional design using face-to-face interviews (Figure 1.). A face-to-face intervie w format was chosen for the study for three primary reasons. First, reading the questions aloud controls for any literacy issues participants may experience. It has been well established that there are close relationships among an individuals income, other socio economic factors and level of education. Medicaid services are provide d to low-income individuals (Szilagyi et al., 2004). As a result, the assumption was made that many of these women had lower levels of education than the general population and that some would have literacy deficits. Face-to-face interviews ensured that all of the participants received the information consistently. Second, previous research targeting Medica id recipients indicat ed traditional mail surveys were less effective, especially in adolescent Medicaid populations, and that additional efforts are indicated in order to obtain a sufficient response rate (Brown, Nederend, Hays, Short, & Farley, 1999; Galla gher & Fowler, 2001). Third, the logistics of addressing HIPAA consent, completion of the survey and providing the incentive would be too great to administ er either within normal clinic operations or through a telephone format.

PAGE 90

67 Figure 1. Maternal Interaction Style, Re ported Experiences of Care and Pe diatric Health Care Utilization Purpose : identify relationships between a mothers interaction style, experiences of care and pediatric health care utilization. Figure 1. Maternal Attachment Style, Experiences of Care, and Pediatric Health Care Utilization Data Entry and Analysis :SPSS v12/SAS v9 Reliability: CronbachsAlpha Phase I: Data Reduction Descriptive, Bi-variateCorrelations Factor Analysis: Attachment BondDiagnostics Phase II: ANOVA Interaction Style & Experiences of CareDiagnostics Phase III: Poisson Regression Effects Interaction style, Experiences of Care and Health Care Utilization.Diagnostics Phase IV: Poisson Regression Assessing Moderating Factors 20-30 Minute Interview: Introduction, informed consent & medical release. Demographic/Health Questionnaire. Experiences of Care Questionnaire. Attachment Questionnaire. $15 gift certificate upon completion Health Care Data Abstraction: Clinic staff abstract data. Mail-Out Recruitment: ACHA/FMHI confidentiality protocols followed for mailing. Potential participants will call researchers. Study Preparation : Obtain USF IRB Consent Meet with a panel of experts Conduct pilot study Face-to-face Recruitment:Posters/fliers advertised study.Clinic Staff recruit in facility. Study Design :Quantitative Cross-Sectional Convenience Sample: 126 Face-to-face Interviews Medical record Abstraction 4 months data collectionSetting:Hillsborough and Pinellas CountiesPopulation:US-born English Lang. > 18 years old Black, White, Hispanic Child 12-18 months Child Receiving Medicaid Never in NICU/ No chronic health condition Outcome Variables# Well Child Care visits # Sick/Follow-up Visits # Emergency Depart. Visits # Immunizations Purpose : identify relationships between a mothers interaction style, experiences of care and pediatric health care utilization. Figure 1. Maternal Attachment Style, Experiences of Care, and Pediatric Health Care Utilization Data Entry and Analysis :SPSS v12/SAS v9 Reliability: CronbachsAlpha Phase I: Data Reduction Descriptive, Bi-variateCorrelations Factor Analysis: Attachment BondDiagnostics Phase II: ANOVA Interaction Style & Experiences of CareDiagnostics Phase III: Poisson Regression Effects Interaction style, Experiences of Care and Health Care Utilization.Diagnostics Phase IV: Poisson Regression Assessing Moderating Factors 20-30 Minute Interview: Introduction, informed consent & medical release. Demographic/Health Questionnaire. Experiences of Care Questionnaire. Attachment Questionnaire. $15 gift certificate upon completion Health Care Data Abstraction: Clinic staff abstract data. Mail-Out Recruitment: ACHA/FMHI confidentiality protocols followed for mailing. Potential participants will call researchers. Study Preparation : Obtain USF IRB Consent Meet with a panel of experts Conduct pilot study Face-to-face Recruitment:Posters/fliers advertised study.Clinic Staff recruit in facility. Study Design :Quantitative Cross-Sectional Convenience Sample: 126 Face-to-face Interviews Medical record Abstraction 4 months data collectionSetting:Hillsborough and Pinellas CountiesPopulation:US-born English Lang. > 18 years old Black, White, Hispanic Child 12-18 months Child Receiving Medicaid Never in NICU/ No chronic health condition Outcome Variables# Well Child Care visits # Sick/Follow-up Visits # Emergency Depart. Visits # Immunizations

PAGE 91

68 Study Setting This study included mothers with infants 12 to 18 months old who resided in Hillsborough and Pinellas Counties and whose pediatric health care was funded through Medicaid since birth. The number of infant s receiving Medicaid-funded health care services is relatively high in the target counties. As of July, 2004, Floridas Agency for Health Care Administration had more than 4,300 White non-Hispanic, Black nonHispanic, and Hispanic infants in Hillsbor ough and Pinellas Countie s 12 to 18 months of age that were in their database as having rece ived Medicaid services during the first 12 months of the childs life (Agency for Health Care Administration, 2004). In addition to the overall Medicaid popul ation, targeted recruitment efforts took place in both larger health care clinics as well as through the Pinellas Healthy Start Program. In Pinellas County, one of the prim ary pediatric health care providers included in this study is the Community Health Centers (CHC) of Pine llas County. The CHC health care facilities provide prenatal care, pediatric, family and internal medicine and family planning services throughout the c ounty. In Hillsborough County subjects were targeted through one of the c ountys largest Medicaid pediatri c clinics, the Health Park Clinic (formerly known as the Genesis Clinic), which is pr ovided through a collaborative effort between Tampa General and the Univ ersity of South Floridas College of Medicines Pediatric Department.

PAGE 92

69 HIPAA and the Protection of Human Subjects Prior to implementing this study, an applic ation was made to the University of South Floridas Institutional Review Board (IRB) (Appendix E. ) There was also the need to obtain permission from partnering agencies As part of this co llaborative process, Institutional Review Board a pproval also was obtained from the Florida Department of Health and Tampa General (Appendix F and G). These review processes both took just over two months to complete. All interviewi ng procedures were in accordance with HIPAA and USF IRB guidelines Protection of individuals rights and confidentiality is an important issue for any research using human s. In the case of this study, there were no physical and minimal psychological risks to participating in this research. Paper records have been stored in a locked file cabinet in a locked room at the USF College of Medicine, Department of Pediatrics. Furthermore, no identifying information was included in the electronic data used for analyses. Finally, electronic data were maintained on a password-protected com puter with identifiers separated from the study variables. Study Population US-born English-speaking women at least 18 years of age at the time of the interview who had a child 12 to 18 months of age and whose child began receiving pediatric health care services for that ch ild through Medicaid funding since birth were targeted for recruitment in the study. The i ssue of Medicaid eligibility since birth was chosen because once eligible, infants are not re-assessed for one year. Infants with gaps

PAGE 93

70 in health insurance coverage are likely to have lower utilizat ion during those breaks (Aiken et al., 2004). However, infants who have received Medicai d since birth should have no breaks in service, at least for insurance reasons. Furthermore, literature suggests that health disparities are influenced by la nguage and cultural issu es. Selecting only USborn English-speaking individuals reduced a ny cultural differences that may not be captured within the scope of this study. Furt hermore, the mothers included in the study needed to be the primary adu lt taking the child to health care visits. Exclusion criteria omit infants who have spent time in the ne onatal intensive care un it prior to going home from the hospital after birth. This criterion sc reened out children most likely to have chronic health conditions who t ypically attend more health vi sits than healthier children. In regard to the racial and ethnic dist ribution of study partic ipants, researchers have found that children of di fferent racial and ethnic back grounds are disproportionately represented in research. Black children are over represented in most research including clinical trials and potentia lly stigmatizing research while being under-represented in therapeutic research. Hispanic or Latino ch ildren are generally unde r-represented except for in stigmatizing research. Finally, White or Caucasian chil dren are often underrepresented in non-therapeutic research while over-represented in therapeutic research (Walsh & Ross, 2003). Attempts were made to ensure adequate racial and ethnic representation of infants in th is study, targeting one third of the study participants who were non-Hispanic White, one third who we re non-Hispanic Black, and one third who were Hispanic. Less than five (4.1%) percen t of all births in Hillsborough and Pinellas Counties (2001-2003) were to mothers of ra cial and ethnic backgrounds other than Hispanic, Caucasian and Black, making it diffi cult to recruit for the study as well as

PAGE 94

71 being able to make any meaningful race-spe cific conclusions (Flo rida Department of Health, 2004). As a result, women of other racial and ethnic bac kgrounds were excluded from the study. Sampling Framework A mixture of sampling frameworks was u tilized for this study. First, letters including a screening form were sent to 4,218 potentially eligible mo thers asking them to call researchers at the number provided(A ppendix H). Eighty-seven additional addresses were excluded because they were Department of Child and Families addresses indicating the children were not in the custody of the mother. The second approach was a convenience sample produced through particip ant contacts either face-to-face with researchers in the providers office, through dissemination of promoti onal materials in the community or referrals from health care professionals (Appendix I) Recruitment A multimodal recruitment process was c onducted through recruitment letters to potentially eligible mothers as well as advert isements (i.e., fliers and posters) and face-toface recruitment in the community. Recruitmen t lasted for a period of four months. For the purpose of this study, two cell phones were used, one with a Pinellas County and one with a Hillsborough County phone number so that calls were toll free for all mothers. A pilot study revealed that women who call a nd were sent to voicemail did not leave messages. As a result, phones were monito red during both day and evening hours seven days per week.

PAGE 95

72 Letter to Mothers. For the first approach, a list of potentially eligible infants was obtained from the Agency for Health Care Administration (AHCA), through the use of Medicaid enrollment files. This was accomp lished by including a description of this proposed study in the annual AHCA contract wi th the Louis de la Parte Florida Mental Health Institute (FMHI) at the University of South Florida to begin July 1, 2005. This contract allowed university researchers to use the Medicaid files to identify a list of names and addresses from potentia lly eligible infants for use in the recruitment process. The mailing labels followed the protocols for confidentiality set forth by the AHCA contract. For example, the mailing list file was made av ailable to researchers in a secure data room with no ability to make electr onic copies of the file. Labels were printed and placed on enveloped by study researchers. Eligible mothers who were interested in participating were asked in the letter to cont act researchers to lear n more about the study and schedule a time to be interviewed. See Appe ndix O for a more detailed description of the recruitment proce ss and lessons learned. Other Recruitment Approaches In addition to the letters, posters, fliers and other literature was posted in community facilities informing mothers of the opportunity to participate in the study. Faceto-face recruitment was conducted in the community as well. The dates and times of th ese recruitment periods varied to meet the needs of the different schedules of potential participants. Researchers wore USF identification, such as clothing with university logos, when r ecruiting participants face-to-face.

PAGE 96

73 Targeted Recruitment. Targeted recruitment attempted to increase Hispanic representation in the study and was conducte d in two ways. First, the second mailing, discussed earlier, targeted Pinellas County mothers focusing heavily on Hispanic mothers (See Appendix J). Additionally, individualized recruitment focused on facilities that served higher populations of Hispanic fa milies including Clearwaters Healthy Start Program and the Genesis/Health Park clinic. Staff within those agencies were asked to focus on referring Hispanic mothers to the study. Data Collection Instruments Data collection instruments included a sc reening form for inclusion in the study and three interview questionnaires administered in order of difficulty. Interviews lasted 20-40 minutes, depending on the mothers desi red level of communication. Demographic information was asked first because the que stions were minimally intrusive, were relatively easy to answer and acted as a re call mechanism regarding health-related issues. This recall helped participants answer the second set of questions that ask about their reported experiences of care. The format of these reported experien ces of care questions can be empowering to women since they ask their opinions regarding their interactions with the health care system. The interact ion style questions was asked last. These questions required the most t hought on the part of the partic ipant and may have appeared to be less relevant to the t opic at hand. By having these que stions last, it was anticipated that even individuals who experience difficulty would continue to answer because of the investment they had already placed in the survey with the questions.

PAGE 97

74 Many of the study questions were taken from the Consumer Assessment of Health Plans (CAHPS) survey protocols so that fi ndings from this study may be benchmarked against national data sets (Agency for Health Care Research and Quality, 2004b). Questions from the U.S. Census, the Medical Expenditure Panel Survey and the Friendly Access data collection protocols were also incorporated for the same reason (U. S. Bureau of the Census, 1996; Agency for Health Care Research and Quality, 2004c; Lawton and Rhea Chiles Center, 2004). Fina lly, maternal depression was measured through the use of a two question depre ssion screen (Whooley, Avins, Miranda & Browner, 1997). Screening Form Identification of potential pa rticipants from the desire d target population requires information that could only be obtained th rough self-report. Questi ons included maternal age, race, ethnicity, country of birth, langua ge proficiency, and age of child. In addition, the woman was asked whether she was the primar y caregiver that takes the child to health care visits, whether her child spent any time in the neonatal intensive care unit, whether Medicaid paid for the childs health care, and whether the child had any chronic health conditions. Only those women who answered al l of the questions in a manner consistent with the inclusion and exclus ion criteria were included. Demographic Questionnaire The demographic survey instrument in cluded a variety of background questions regarding the context from which each mother was reporting. It has standard

PAGE 98

75 demographic questions about the mother incl uding level of educa tion, race, ethnicity, marital status, and how many hours the mother worked each week (see Appendix K). Women were also asked about their health issues prior to their pregnancy, pregnancy conditions and complications, current health st atus, and her health care utilization since the birth of her child. Minkovitz, et al. ( 2002) found significant a ssociations between health service utiliza tion patterns for women and their ch ildren, suggesting that maternal utilization patterns needed to be considered when studying utilization patterns in child health care. Questions were also asked rega rding breastfeeding, mate rnal depression, fear of doctors, and the timing of the pregnancy. Information about the infant included overall health status, chronic health problems not pr eviously identified by the mother, and sleep disturbances. Health care pr ovider questions included tran sportation and access barriers, race concordance between the primary provider and the mother as well as whether the woman chose her health care provider before the baby was born. The instrument also asks questions regarding other services the fa mily may have been receiving such as WIC and Healthy Start. Finally, mo thers were asked about health care their child received outside the office of the childs primary heal th care provider (i.e., emergency department visits). In the event that th e health care visit was not a dded to the childs primary care record being abstracted, this self-report helped ensure that all utilization data was identified. Reported Experience of Care Questionnaire Over the years, the Agency for Healthcare Research and Quality has used a consistent set of reported experiences of car e questions that have been validated for

PAGE 99

76 individuals speaking both English and Spanis h and have been administered across a variety of national studies. A sample of these studies includes the Medical Expenditure Panel Survey (MEPS) and the Consumer Asse ssment of Health Pl ans Survey (CAHPS) (Agency for Health Care Research and Qualit y, 2004a; Agency for Health Care Research and Quality, 2004b). Questions focus on specific service issues such as When your child needed care right away for an illness or injury, how often did you get ca re as soon as you wanted? This questionnaire included 25 questions w ithin three sections (see Appendix L). The first section asked three questions re garding experiences involving the primary person who provided the care including the ab ility to choose the i ndividual providing the health care, whether the child had a personal provider, and a rating of that individual on a scale from zero to ten. There was also a que stion regarding whether the mothers learned from the pediatrician how to better managed their child(ren)s health care needs. The overall rating was the primary variable of interest for this section. The second section asked 17 questions re garding the operation of the entire provider office. This section addresses issues of regular and sick visits, wait times, whether staff treated the mothers with courtesy and respect, whether they listened to the mother carefully and showed respect for wh at the mother had to say. Three of these questions were in a yes or no format regard ing whether an action was initiated and were followed up by a rating of that services. The que stions were required for the format of the instrument but are collapsed into the rating for that action. Of the remaining 14 questions, ten use a format of never, sometimes, usua lly, always, and not applicable. One set of responses ranged from a big problem, a small problem, not a problem and not applicable.

PAGE 100

77 The last question was a rating from zero to ten. Understanding the scale for the various questions is needed to identify the most a ppropriate methods for reducing the data during analyses. The last section asks five questions regarding any specia lists the infant may have needed to see. Questions in this section addressed whether the child needed to see a specialist, the difficulty in seeing one, and a ra ting of that individual. Given the exclusion of children with chronic illnesses, analysis of these questions included an aggregated field indicating whether a specialist was seen. Relationship Scales Questionnaire The Relationship Scales Questionnaire (R SQ) instrument included 30 questions rated on a 7-point Likert-type scale ranging from Not at a ll like me [1] to Very much like me[7] (see Appendix M). Questions refe rence specific situations in which the woman was asked to indicate how she w ould respond. This Relationship Scales Questionnaire (RSQ) and was developed usi ng Hazen & Shavers (1993) attachment measure as well as the Relationship Questionna ire and the Adult Attachment Scale used by Collins and Read (Collins & Read, 1990; Griffin & Bartholomew, 1994). This instrument offered the ability to modify the terminology depending on the relationship type of interest. The measure was designed to be a set of continuous rating scores for the different interaction styles but can be categorized, if necessary.

PAGE 101

78 Health Care Utilization Data Abstraction Form In addition to the interview, the num ber of sick, follow-up, well child, and emergency department visits was obtained fr om the medical record. Abstraction from the record is important because studies have found that recall bias for certain information is high and that collecting information from c onsistent collection sources, such as that contained in medical records, is more reli able (Bolton, Holt, Ross, Hughart, & Guyer, 1998). Abstraction occurred in a number of wa ys. Some clinics subcontract out their abstraction services to a third party, some us ed medical records staff, some used nurses and several clinics allowed researcher s to abstract the data directly. In addition to health care utilization, in formation was also obtained regarding any accidents the child experienced and whether the child had ever been identified as experiencing failure-to-thrive. Both accidents and a medical diagnosis of failure-to-thrive have been associated with mother-child bonding and was used as part of a surrogate mother-child bonding measure (to be discussed later) (see Appendix N). The infants immunization records were also extracted from medical records. According to the Centers for Disease Cont rol, the immunizations that should be administered by the first 12 months of life incl ude: two or three Hepa titis B (HepB), three Diptheria, Tetanus, and Pertussis (DtaP); th ree Haemophilus Influenza b (Hib); two or three Inactivated Polio Virus (IPV); and th ree Pneumococcal Virus (PCV). The optional HepB and IPV vaccinations were excluded from the analyses.

PAGE 102

79 Instrument Validity and Reliability Procedures Assessment of Validity Internal validity was addressed primarily through thoughtful sele ction criteria that attempted to control for a numbe r of factors such as socioe conomic status and funding for health insurance. Efforts were also made to strengthen the extern al validity of the study such as by using standardized measures. Alt hough some of the sta ndardized instruments were used in their entirety, others were subsets of questions such as the CAHPS instruments items used for the experiences of care questionnaire. A few of the questions included in this study were developed specifica lly for this study and were not included in a standardized instrument. Once these study-sp ecific questions were developed and the instrument was drafted, a se ries of efforts were made to determine whether the new questions as well as the combin ation of questions was valid. Panel of Experts To ensure the content and wording of the questions were consistent with the needs and issues of participants, the study design and survey instruments were reviewed for face, content and construct validity. This review was conducted with committee members as well as other experts including university faculty from within the college, within the university and nationally. An anthropologist from the University of South Florida reviewed the research design and instrumenta tion to ensure that the cultural differences among participants would not lead to invalid me asures. As a result of this conversation, it was determined that restri cting the study to US-born, English-speaking women would

PAGE 103

80 lessen the impact of cultural differences. This e xpert review also led to the inclusion of some follow-up questions to the RSQ items related to closeness to help interpret the data. Furthermore, six personnel working directly with clients, including three nurses, one physician, and 2 care coordinators were consulted about the study design and the instrument items. These experts were helped re fine terminology used in the instrument so that it reflects the words they use during thei r interactions with the mothers. The method was also modified to incorporate the recru itment process into the standard operation of participating facilities. Finally, a leading national maternal and child health expert familiar with attachment theory in his research, Milton Kotelchuck, was consulted and recommended some modifications in the survey questions and study variables. Once the questions were de veloped, two previous Medi caid mothers now working in the public health field were asked to pr ovide feedback regarding the study items. One of the mothers was Hispanic and spoke E nglish as her second language. A qualitative process known as cognitive interviewing was to used during the administration of the interview. More specifically, a two-stage conc urrent process of admi nistering the survey by reading all of the questions, receiving a response, and probing for additional information was used. The respondents were th en asked to identify the thought processes involved in understanding and answeri ng each question (Willis, 1999). Additional changes in wording were made as a result of these conversations such as referring to well child care visits as check-up visits. No ch anges were made to the standardized CAHPS and RSQ instruments.

PAGE 104

81 Pilot Study Determining the feasibility of the stu dy design and establishment of the face validity of the survey instruments was comple ted during a pilot-test phase. First, a 5 day pilot study using a face-to-face recruitmen t methodology was conducted in the Pinellas Park and Johnnie Ruth Clark Community Heal th Center clinics. One individual was recruited through this process. The cognitiv e testing process was repeated for this respondent and identified problems with a fi ve question relationshi p measure previously included in this study design. This finding led to the elimination of those five questions from the study and thus using only the Rela tionship Scales Questionnaire to measure interaction styles. Reasons for the lack of recruitment duri ng this pilot study are varied. Delays in the IRB process placed the timing of the st udy during the slower winter holiday season (December, 2004). Furthermore, the inclusion criteria was too narro w (12-15 months of age). There were a number of women with infants coming in for their 18 month well child care visits but very few were scheduled for their 15 month well child care visit. The women that were scheduled were often no s hows for their appointme nts, an issue that was reported by the clinics as representing 15%30% of their appointments on any one day. It is for these reasons that the inclusi on criteria was expanded to include infants from 12-18 months of age for the final study. Due to the low recruitment in the fi rst study, a second pilot study was conducted in February, 2005. This second study expanded the recruitment methodology to include a variety of approaches targeted at women ta king infants to the Pinellas Park and Johnnie Ruth Clarke Community Health Centers. R ecruitment included 100 letters sent from the

PAGE 105

82 Community Health Centers to mothers of in fants age 12-18 months who were receiving pediatric health care services through Medicaid. Mothers we re given 2 weeks in which to respond. In addition, posters were placed in each of the exam rooms at both clinics. Fliers were also placed throughout the clinics. The recruitment results were mixed, i ndicating the need for more than one recruitment approach in the data collection process. Thirteen of the 100 letters were returned due to incorrect addr esses. Of the remaining 87 letters, 5 (6%) women contacted researchers and were interviewed. Two wo men were recruited through the fliers. The seven women who participated in the second pilot study were dispersed throughout the study area and were served by the two target clinics as well as two that had transferred to St. Petersburg Pediatrics a provider with eight clinics located throughout south county. Two of the women we re currently pregnant, one had recently delivered her second child, and one had an older child in addition to the target infant. The racial distribution of respondents was four White, two Black, and one Hispanic. Most (6) of the women were single, had no serious hea lth problems (6), had not planned their first pregnancy (6) and participated in WIC and/or Healthy Start (6) postnatally. Advertising materials were also piloted at the Gr eenwood Community Health Resource Center Free Clinic in Clearwater, Florida. A total of ten participants were provided with three different versions of the letters inviting mothers to participate in the study. Respondents were asked to rank all of the invitations in order of preference as well as provide verbal feedback regarding how the different layouts made them feel. The most preferred design was used in the study.

PAGE 106

83 Assessment of Reliability The internal consistency of responses was measured to assess the reliability of the data using the split-half reliability measure for the Relationship Scal es Questionnaire, and its subcomponents. The split-ha lf reliability measure, commonly used for cross-sectional data, divides the sample of instrument items in half and correlates the responses between the two halves. Cronbachs Alpha conceptually represents the aver age of all possible split-half reliability estimates with an alpha of one being a perfect association and a zero representing no association (Trochim, 2004). It em total correlations were also reviewed. Data Analysis As noted in the literature review, there are a number of factors that are associated with the interaction styles, re ported experiences of care and th e utilization of health care services. It has also been reported that th ere is a need for more research into the underlying factors that influence these asso ciations. To accomplish this, a variety of statistical procedures including multivariat e models were used. These statistical approaches accomplished three primary task s. The first was to describe the study population. The second task was that of da ta reduction. The th ird was to identify associations among variables, using multip le factors simultaneously. A variety of approaches have been developed to aid in the exploration of relationships among variables. Each approach has specific variable formats as well as situational or contextual realms for which they can be appropriate ly used. Three primary methodologies were used, including correlation coefficients, a factor analysis and Poisson regression analyses.

PAGE 107

84 Data were entered into a Microsoft Excel database maintained by the principal investigator. Data analysis was cond ucted using SPSSv12 and SAS v9.1 statistical software packages. There were four phases to this analysis (Figure 2). Phase one was a data reduction effort. Phase two explored the associations between maternal interaction style and reported experiences of care. Ph ase three assessed th e associations among maternal interaction style and reported experi ences of care with heal th care utilization. Phase four explored moderating factors that are associated with maternal interaction style, reported experiences of care and health care utilization.

PAGE 108

85 Figure 2. Maternal Attachment, Reported Experiences of Care and Pediatric Health Care Utilization 2) Factor AnalysisBonding Composite : (1=Yes, 0=No) Breast Feeding (1/0) Intendedness(1/0) Depressed (1/0) Failure to Thrive (1/0) Accidents (1/0) Sleep Problems (1/0) 1) UnivariateStatisticsFrequencies Means Standard deviations Variance Kurtosis Skewness. 1) Immunizations=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Well Child Care Visits=intercept+AX(x+Av(x)+PR(x)+PO(x)+error 2) Sick/Follow-up Visits=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Emergency Department Visits=intercept+AX(x)+AV(x)+PO(x)+Po(x)+error(Ax=anxious attachment, AV=avoidant attachment, PR=provider rating, PO=Provider office rating) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI)Phase 3: The influence of Attachment Style & Reported Experiences of Care on Health Care Utilization -Poisson Regression 1) Pediatric Provider Score= intercept + attachment (x) + error 2) Pediatric Provider Office Score=intercept + attachment (x) + error.Phase 2: Attachment Style & Reported Experiences of Care-ANOVA Phase 4: Factors that Moderate the Influence of Attachment Style& Reported Experiences of Care on Health Care Utilization-Poisson Regression Phase I: Data Reduction I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error WC=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+ BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error SF=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error ED=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error Example of Interaction Model:I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA (x)+OS(x)+FT(x)+FP(x)+AV*WC(x)+error(I=Immunizations, WC=well child care visits, SF=sick/Follow-up Visits, ED=Emergency Department Visits) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI) Pre-Analysis Calculations:Attachment categorization: Aggregate 30 questions using standardized scoring method Diagnostics: Outliers Diagnostics: Outliers Influential Observations Multicolinearity Study Variables : (MA) Mom age (18-50) (AA) Af. Am. (white referent)(1/0) (HE) Hispanic (white referent) (1/0) (FT) Full time Employment (1/0) (MH) Mothers overall health (1-10) (BO) Birth Order (Continuous) (FP) Feelings about physicians (1-10) (CH) Childs overall health (1-10) (PP) Rating of Physician (0-10) (OS) Other services (WIC, HS)(1/0) (PO) Rating of doctor office (0-10) (AV) Avoidant (Secure referent) (1/0) (BC) Bonding Composite (continuous) (AX) Anxious (Secure referent) (1/0)Figure 2. Maternal Attachment Style, Reported Experiences of Care, and Pediatric Health Care Utilization Data Analysis 3) Bi-VariateStatisticsSpearman Correlation Coef. (Ordinal) Pearson Correlation Coef. (Continuous) Cramers V (Nominal) (Output: Source, Degrees of Freedom, Sums of Squares, Mean Square, F stat., p-value, OR & 95% CI) 2) Factor AnalysisBonding Composite : (1=Yes, 0=No) Breast Feeding (1/0) Intendedness(1/0) Depressed (1/0) Failure to Thrive (1/0) Accidents (1/0) Sleep Problems (1/0) 1) UnivariateStatisticsFrequencies Means Standard deviations Variance Kurtosis Skewness. 1) Immunizations=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Well Child Care Visits=intercept+AX(x+Av(x)+PR(x)+PO(x)+error 2) Sick/Follow-up Visits=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Emergency Department Visits=intercept+AX(x)+AV(x)+PO(x)+Po(x)+error(Ax=anxious attachment, AV=avoidant attachment, PR=provider rating, PO=Provider office rating) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI)Phase 3: The influence of Attachment Style & Reported Experiences of Care on Health Care Utilization -Poisson Regression 1) Pediatric Provider Score= intercept + attachment (x) + error 2) Pediatric Provider Office Score=intercept + attachment (x) + error.Phase 2: Attachment Style & Reported Experiences of Care-ANOVA Phase 4: Factors that Moderate the Influence of Attachment Style& Reported Experiences of Care on Health Care Utilization-Poisson Regression Phase I: Data Reduction I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error WC=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+ BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error SF=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error ED=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error Example of Interaction Model:I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA (x)+OS(x)+FT(x)+FP(x)+AV*WC(x)+error(I=Immunizations, WC=well child care visits, SF=sick/Follow-up Visits, ED=Emergency Department Visits) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI) Pre-Analysis Calculations:Attachment categorization: Aggregate 30 questions using standardized scoring method Diagnostics: Outliers Diagnostics: Outliers Influential Observations Multicolinearity Study Variables : (MA) Mom age (18-50) (AA) Af. Am. (white referent)(1/0) (HE) Hispanic (white referent) (1/0) (FT) Full time Employment (1/0) (MH) Mothers overall health (1-10) (BO) Birth Order (Continuous) (FP) Feelings about physicians (1-10) (CH) Childs overall health (1-10) (PP) Rating of Physician (0-10) (OS) Other services (WIC, HS)(1/0) (PO) Rating of doctor office (0-10) (AV) Avoidant (Secure referent) (1/0) (BC) Bonding Composite (continuous) (AX) Anxious (Secure referent) (1/0)Figure 2. Maternal Attachment Style, Reported Experiences of Care, and Pediatric Health Care Utilization Data Analysis 3) Bi-VariateStatisticsSpearman Correlation Coef. (Ordinal) Pearson Correlation Coef. (Continuous) Cramers V (Nominal) (Output: Source, Degrees of Freedom, Sums of Squares, Mean Square, F stat., p-value, OR & 95% CI) 2) Factor AnalysisBonding Composite : (1=Yes, 0=No) Breast Feeding (1/0) Intendedness(1/0) Depressed (1/0) Failure to Thrive (1/0) Accidents (1/0) Sleep Problems (1/0) 1) UnivariateStatisticsFrequencies Means Standard deviations Variance Kurtosis Skewness. 1) Immunizations=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Well Child Care Visits=intercept+AX(x+Av(x)+PR(x)+PO(x)+error 2) Sick/Follow-up Visits=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Emergency Department Visits=intercept+AX(x)+AV(x)+PO(x)+Po(x)+error(Ax=anxious attachment, AV=avoidant attachment, PR=provider rating, PO=Provider office rating) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI)Phase 3: The influence of Attachment Style & Reported Experiences of Care on Health Care Utilization -Poisson Regression 1) Pediatric Provider Score= intercept + attachment (x) + error 2) Pediatric Provider Office Score=intercept + attachment (x) + error.Phase 2: Attachment Style & Reported Experiences of Care-ANOVA Phase 4: Factors that Moderate the Influence of Attachment Style& Reported Experiences of Care on Health Care Utilization-Poisson Regression Phase I: Data Reduction I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error WC=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+ BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error SF=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error ED=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error Example of Interaction Model:I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA (x)+OS(x)+FT(x)+FP(x)+AV*WC(x)+error(I=Immunizations, WC=well child care visits, SF=sick/Follow-up Visits, ED=Emergency Department Visits) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI) Pre-Analysis Calculations:Attachment categorization: Aggregate 30 questions using standardized scoring method Diagnostics: Outliers Diagnostics: Outliers Influential Observations Multicolinearity Study Variables : (MA) Mom age (18-50) (AA) Af. Am. (white referent)(1/0) (HE) Hispanic (white referent) (1/0) (FT) Full time Employment (1/0) (MH) Mothers overall health (1-10) (BO) Birth Order (Continuous) (FP) Feelings about physicians (1-10) (CH) Childs overall health (1-10) (PP) Rating of Physician (0-10) (OS) Other services (WIC, HS)(1/0) (PO) Rating of doctor office (0-10) (AV) Avoidant (Secure referent) (1/0) (BC) Bonding Composite (continuous) (AX) Anxious (Secure referent) (1/0)Figure 2. Maternal Attachment Style, Reported Experiences of Care, and Pediatric Health Care Utilization Data Analysis 3) Bi-VariateStatisticsSpearman Correlation Coef. (Ordinal) Pearson Correlation Coef. (Continuous) Cramers V (Nominal) (Output: Source, Degrees of Freedom, Sums of Squares, Mean Square, F stat., p-value, OR & 95% CI) 2) Factor AnalysisBonding Composite : (1=Yes, 0=No) Breast Feeding (1/0) Intendedness(1/0) Depressed (1/0) Failure to Thrive (1/0) Accidents (1/0) Sleep Problems (1/0) 1) UnivariateStatisticsFrequencies Means Standard deviations Variance Kurtosis Skewness. 1) Immunizations=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Well Child Care Visits=intercept+AX(x+Av(x)+PR(x)+PO(x)+error 2) Sick/Follow-up Visits=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Emergency Department Visits=intercept+AX(x)+AV(x)+PO(x)+Po(x)+error(Ax=anxious attachment, AV=avoidant attachment, PR=provider rating, PO=Provider office rating) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI)Phase 3: The influence of Attachment Style & Reported Experiences of Care on Health Care Utilization -Poisson Regression 1) Pediatric Provider Score= intercept + attachment (x) + error 2) Pediatric Provider Office Score=intercept + attachment (x) + error.Phase 2: Attachment Style & Reported Experiences of Care-ANOVA Phase 4: Factors that Moderate the Influence of Attachment Style& Reported Experiences of Care on Health Care Utilization-Poisson Regression Phase I: Data Reduction I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error WC=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+ BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error SF=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error ED=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error Example of Interaction Model:I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA (x)+OS(x)+FT(x)+FP(x)+AV*WC(x)+error(I=Immunizations, WC=well child care visits, SF=sick/Follow-up Visits, ED=Emergency Department Visits) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI) Pre-Analysis Calculations:Attachment categorization: Aggregate 30 questions using standardized scoring method Diagnostics: Outliers Diagnostics: Outliers Influential Observations Multicolinearity Study Variables : (MA) Mom age (18-50) (AA) Af. Am. (white referent)(1/0) (HE) Hispanic (white referent) (1/0) (FT) Full time Employment (1/0) (MH) Mothers overall health (1-10) (BO) Birth Order (Continuous) (FP) Feelings about physicians (1-10) (CH) Childs overall health (1-10) (PP) Rating of Physician (0-10) (OS) Other services (WIC, HS)(1/0) (PO) Rating of doctor office (0-10) (AV) Avoidant (Secure referent) (1/0) (BC) Bonding Composite (continuous) (AX) Anxious (Secure referent) (1/0)Figure 2. Maternal Attachment Style, Reported Experiences of Care, and Pediatric Health Care Utilization Data Analysis 3) Bi-VariateStatisticsSpearman Correlation Coef. (Ordinal) Pearson Correlation Coef. (Continuous) Cramers V (Nominal) (Output: Source, Degrees of Freedom, Sums of Squares, Mean Square, F stat., p-value, OR & 95% CI) 2) Factor AnalysisBonding Composite : (1=Yes, 0=No) Breast Feeding (1/0) Intendedness(1/0) Depressed (1/0) Failure to Thrive (1/0) Accidents (1/0) Sleep Problems (1/0) 1) UnivariateStatisticsFrequencies Means Standard deviations Variance Kurtosis Skewness. 1) Immunizations=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Well Child Care Visits=intercept+AX(x+Av(x)+PR(x)+PO(x)+error 2) Sick/Follow-up Visits=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Emergency Department Visits=intercept+AX(x)+AV(x)+PO(x)+Po(x)+error(Ax=anxious attachment, AV=avoidant attachment, PR=provider rating, PO=Provider office rating) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI)Phase 3: The influence of Attachment Style & Reported Experiences of Care on Health Care Utilization -Poisson Regression 1) Pediatric Provider Score= intercept + attachment (x) + error 2) Pediatric Provider Office Score=intercept + attachment (x) + error.Phase 2: Attachment Style & Reported Experiences of Care-ANOVA Phase 4: Factors that Moderate the Influence of Attachment Style& Reported Experiences of Care on Health Care Utilization-Poisson Regression Phase I: Data Reduction I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error WC=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+ BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error SF=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error ED=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error Example of Interaction Model:I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA (x)+OS(x)+FT(x)+FP(x)+AV*WC(x)+error(I=Immunizations, WC=well child care visits, SF=sick/Follow-up Visits, ED=Emergency Department Visits) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI) Pre-Analysis Calculations:Attachment categorization: Aggregate 30 questions using standardized scoring method Diagnostics: Outliers Diagnostics: Outliers Influential Observations Multicolinearity Study Variables : (MA) Mom age (18-50) (AA) Af. Am. (white referent)(1/0) (HE) Hispanic (white referent) (1/0) (FT) Full time Employment (1/0) (MH) Mothers overall health (1-10) (BO) Birth Order (Continuous) (FP) Feelings about physicians (1-10) (CH) Childs overall health (1-10) (PP) Rating of Physician (0-10) (OS) Other services (WIC, HS)(1/0) (PO) Rating of doctor office (0-10) (AV) Avoidant (Secure referent) (1/0) (BC) Bonding Composite (continuous) (AX) Anxious (Secure referent) (1/0)Figure 2. Maternal Attachment Style, Reported Experiences of Care, and Pediatric Health Care Utilization Data Analysis 3) Bi-VariateStatisticsSpearman Correlation Coef. (Ordinal) Pearson Correlation Coef. (Continuous) Cramers V (Nominal) (Output: Source, Degrees of Freedom, Sums of Squares, Mean Square, F stat., p-value, OR & 95% CI) 2) Factor AnalysisBonding Composite : (1=Yes, 0=No) Breast Feeding (1/0) Intendedness(1/0) Depressed (1/0) Failure to Thrive (1/0) Accidents (1/0) Sleep Problems (1/0) 1) UnivariateStatisticsFrequencies Means Standard deviations Variance Kurtosis Skewness. 1) Immunizations=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Well Child Care Visits=intercept+AX(x+Av(x)+PR(x)+PO(x)+error 2) Sick/Follow-up Visits=intercept+AX(x)+AV(x)+PR(x)+PO(x)+error 2) Emergency Department Visits=intercept+AX(x)+AV(x)+PO(x)+Po(x)+error(Ax=anxious attachment, AV=avoidant attachment, PR=provider rating, PO=Provider office rating) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI)Phase 3: The influence of Attachment Style & Reported Experiences of Care on Health Care Utilization -Poisson Regression 1) Pediatric Provider Score= intercept + attachment (x) + error 2) Pediatric Provider Office Score=intercept + attachment (x) + error.Phase 2: Attachment Style & Reported Experiences of Care-ANOVA Phase 4: Factors that Moderate the Influence of Attachment Style& Reported Experiences of Care on Health Care Utilization-Poisson Regression Phase I: Data Reduction I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error WC=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+ BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error SF=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error ED=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA(x)+OS(x)+FT(x)+FP(x)+error Example of Interaction Model:I=a+AX(x)+AV(x)+PP(x)+PO(x)+AA(x)+BO(x)+HE(x)+CH(x)+BC(x)+MH(x)+MA (x)+OS(x)+FT(x)+FP(x)+AV*WC(x)+error(I=Immunizations, WC=well child care visits, SF=sick/Follow-up Visits, ED=Emergency Department Visits) (Output: Degrees of Freedom, Waldparameter estimate, standard error, WaldCI) Pre-Analysis Calculations:Attachment categorization: Aggregate 30 questions using standardized scoring method Diagnostics: Outliers Diagnostics: Outliers Influential Observations Multicolinearity Study Variables : (MA) Mom age (18-50) (AA) Af. Am. (white referent)(1/0) (HE) Hispanic (white referent) (1/0) (FT) Full time Employment (1/0) (MH) Mothers overall health (1-10) (BO) Birth Order (Continuous) (FP) Feelings about physicians (1-10) (CH) Childs overall health (1-10) (PP) Rating of Physician (0-10) (OS) Other services (WIC, HS)(1/0) (PO) Rating of doctor office (0-10) (AV) Avoidant (Secure referent) (1/0) (BC) Bonding Composite (continuous) (AX) Anxious (Secure referent) (1/0)Figure 2. Maternal Attachment Style, Reported Experiences of Care, and Pediatric Health Care Utilization Data Analysis 3) Bi-VariateStatisticsSpearman Correlation Coef. (Ordinal) Pearson Correlation Coef. (Continuous) Cramers V (Nominal) (Output: Source, Degrees of Freedom, Sums of Squares, Mean Square, F stat., p-value, OR & 95% CI)

PAGE 109

86 Phase One: Data Reduction Phase One had three components. First, de terminations were made regarding the distribution of the data elements for future statistics as well as wh ether the responses to multiple questions could be collapsed. The second component involved using a factor analysis to reduce the data where appropriate to increase the power of future analyses. Finally, bivariate associations were assessed between the in dependent variables and the dependent variables. Study Variables. The primary dependent variable s included a continuous number of immunizations as well as a continuous num ber of pediatric health care visits attended in the first 12 months of life falling within three types of visi ts (well child care, sick/follow-up visits, and emer gency department visits). The primary independent variable included interaction style. The authors of the RSQ recommend that continuous variables for each interaction style be used. Interaction style can also be categorized one variable (s ecure, anxious, avoidant). Previous studies of adult interaction styles indicat e that approximately 55-59% of adults are classified as secure, 25% as avoidant, and 11-20% as anxious. The remaining 5-10% were unclassifiable (Mickelson et al., 1997; Hazan & Shaver, 1987). There are a series of decision-making criteria for cate gorizing interaction style (M ickelson et al., 1997). First, the category with the highest rating becomes the interaction style. If Secure is tied for having the highest score with another category, the other category will be identified as the interaction style. If three categories are equally rated high then the interaction style is

PAGE 110

87 unclassifiable and will be excluded from the analyses. For the purposes of this study, the distribution of the data led to the use of thr ee continuous variables to measure each of the three interaction styles (secu re, anxious, avoidant). Two other independent variables include th e ratings of experiences of care with pediatric health care provider and a separate rating for the entire provider office. In addition to the independent variables, severa l potential moderators were included in the study: race/ethnicity (White non-Hispanic, Black non-Hisp anic, Hispanic), maternal rating of her overall health status (1-10) and of her childs health status (1-10), participation in other health promotion programs (yes/no: participation in WIC or Healthy Start), a bonding composite score, birt h order, maternal f eelings about doctors, maternal age, employme nt status (<30 hours, > 30 hours per week), immunizations, and the number of health care visits noted in the clinic medical record. If the mother reported emergency department visits not noted in the medical record, those visits were included in the count. Component One: Descriptive Statistics. Phase one data reduction efforts began with descriptive statistics necessary for determining the normality of the data including frequencies, means, variances, standard de viations, kurtosis and skewness measures. Variables with insufficient data were removed from the study or coll apsed into variables that could be included. For example, the fo llowing two questions would be merged into one: Since your child was born, did you call a doctors office or clinic during regular office hours to get help or advice for your self (Yes/No)? and Since your child was

PAGE 111

88 born, when you called during regular office hour s, how often did you get the help or advice you needed (Never, Sometimes, Us ually, Always, I didnt call for help)? The skewed distribution of so me continuous variables, su ch as the ratings of the provider and the providers office, warrant ed transformation. As a result, variable responses were reversed (i.e., 1=10, 2=9, 3=8, etc) and the square root was taken. These transformed variables were then used in further analyses. Component Two: Factor Analysis. A factor analysis wa s used to develop a composite score for the maternal-infant bond. Ther e were six variables that literature has shown to be associated with maternal in fant bonding including br eastfeeding, maternal depression, a diagnosis of failure-to-thri ve, a high propensity for accidents, the intendedness of the pregnancy, and infant sleep problems (Feldman, Weller, Leckman, Kuint, & Eidelman, 1999). Dichotomous variab les were created for each factor. The first factor, maternal depression, was identified by a respondent saying yes to both maternal depression questions. Breastfeeding was coded as a yes if the woman breastfed at least one month. Failure-to-thrive is a medical diagnosis that is recorded in the medical charts if an infant is not gaining we ight at the expected rate, base d on growth charts. Accidental injuries reported in the medical charts were used by the clinic staff to provide their professional opinion regarding whether the childs propensity for having accidents. Infants identified as medium or high propens ity for accidents were coded as a yes. Infant sleep issues were identified if the mo ther reports frequent problems with sleeping. Finally, a timing of the pregnancy that was either later in life or never was considered an

PAGE 112

89 unintended pregnancy (Goldberg, 1991; Mertin, 1986; O'Callaghan & Hull, 1978; Rapley, 2002; Ricci, Giantr is, Merriam, Hodge, & Doyl e, 2003; Scher, 2001). A factor analysis is a procedure specifi cally suited for using measures obtained on a number of variables and reduci ng them into a smaller number of created variables. This procedure is useful when ther e is the belief that redundancy in the variables exists. More specifically, it is believed th at there are variables highly correlated with one another, possibly because the variables are measuring the same construct. In addition, it is assumed that a group of latent factors that have not been observed account for the correlations among the observed va riables and if the latent va riables were held constant, the partial correlations of the observed vari ables would all become zero (ACITS, 1995). Ultimately, reduction into the principle component s of the construct w ould retain most of the variance in the observed variables while re ducing the number of variables required in the model (Hatcher & Stepanski, 2001). The factor analysis differs from the similar principle components analysis by assumi ng an underlying causal structure among the variables exists. In addition to simply reducing the number of variables, the use of a composite score can provide more reliable estimates by pooling the information that the items have in common (Tricare, 2004). Another advant age this scaled score has over a single response score is that it better represents th e concept, providing more information and greater statistical power for the purposes of hypothesis testing. There are four general criteria for using a multi-trait scaling score. First, there mu st be convergent validity as illustrated by the internal consistency of each item being linearly related to the total score for other items in that group. Second, discrimi nant validity should be demonstrated by

PAGE 113

90 items correlating higher with the construct than it correlates with other constructs. Third, the items in the same scale should contain the same proportion of information about the construct as demonstrated by equal item-tota l correlates (Tricare 2004). If this does not occur, items should be weighted. Finally, items measuring the same construct should have approximately equal variance so that st andardization of scores prior to combining them is unnecessary. As noted earlier, Cronbach s alpha and item total correlations were used to estimate the reliability of the responses. The factor analysis model included failure to thrive (FT), breastfeeding (BF), accidents (AC), sleep disturbances (SD), maternal depression (MD), and the timing of the pregnancy (TP). Assuming all si x variables cluster into th e same factor loading, the following model would result: C1=b1(FT) + b2(BF) + b3(AC) + b4 (SD) + b5(MD) + b6(TP) Component Three: Bivariate Correlations. The next step was to compute bivariate associations among vari ables to identif y relationships of inte rest. Variables that significantly correlated with the dependent variables were included in the multivariate model. Variables that were significantly co rrelated with both the dependent variables and the primary independent variables (interaction style, provider rating and provider office rating) were assessed to determine the exte nt to which the unique variance of each variable contributes to th e model during the Phase four analyses. This assessment included adding an interaction variable to the multivariate model.

PAGE 114

91 Phase Two: Interaction Styl e and Experiences of Care Once the data were reduced, hypothesis te sting began. This next phase of the analysis involved the determination of any a ssociations between the independent variable of maternal interaction style and the dependent variables of reported experiences of care with the childs health care provider and the provider office. As noted earlier, the reported experiences of care measures incl uded ratings for the provider, the provider office and involvement with specialists. Since this study filtered out infants with chronic illness and those who spent time in the ne onatal intensive care unit, the need for specialists was minimized. As a result, th e information regarding specialists was dichotomized to control for their involvement in other analyses but we re not used in this phase. The remaining scores (0-10) for the health care provider and the provider office score for the pediatric health care provider were transformed and became the primary continuous dependent variables for the first hypothesis. Since both the dependent and independent variables were continuous, linear regression analyses were used to test relationships between the variables. The assumptions of a linear regression include th e existence of data elements that are normally distributed and independent, that associ ations are linear and there is a constant variance of the error terms (homoscedasticity) (Kleinbaum Kupper & Muller, 1988). A linear regression estimates the expected value of a dependent variable based on the value(s) of independent variable(s). The assumptions for a linear regression include the existence of data elements that are normally distributed and independent, that associations are linear and there is a constant variance of the error terms

PAGE 115

92 (homoscedasticity) (Kleinba um, Kupper & Muller, 1988) The following linear regression models were calculated: 1) Provider Score= intercep t + secure (x) + avoidant (x) + anxious (x) + error. 2) Provider Office Score=inter cept + secure (x) + avoidant (x) + anxious (x) + error. It should be noted that diagnostics iden tified a moderately high (r=.68) correlation between provider and provider office. Th is level of correlation would introduce multivariate co-linearity issues in to the mode l resulting in spurious associations or lack of associations. This issue often leads to one of the variables being excluded from the analyses. However, the association of st udy variables with the ratings of both the provider and the provider office is of part icular interest. As a result, provider and provider office ratings were modeled sepa rately for all of the study hypotheses. Phase Three: Interaction Style, Ex periences of Care, and Utilization For this study, health care utilization wa s represented by four different variables including well-child, sick/fo llow-up, emergency department visits and immunizations. Because the data represented a count of events, the following SAS v9.1 Poisson regression models were used (I=Imm unizations, W=Well Child Care Visits, S=Sick/Follow-up Visits, Pp=provider, Po=Provider Office): 1) Well Child Care=Intercept+Provider+S ecure(x)+Avoidant(x)+Anxious(x)+error 2) Sick/Follow-up Intercept+Provider+S ecure(x)+Avoidant(x)+Anxious(x)+error 3) Emergency Dept=Intercept+Provider+S ecure(x)+Avoidant(x)+ Anxious(x) +error

PAGE 116

93 4) Immunization=Intercept+ Provider+Secure(x)+Avoida nt(x)+Anxious(x)+error 5) Well Child Care=Intercept+Office+Secu re(x)+Avoidant(x)+Anxious(x)+error 6) Sick/Follow-up=Intercept+Office+S ecure(x)+Avoidant(x)+Anxious(x)+error 7) Emergency Dept=Intercept+Office+Secu re(x)+Avoidant(x)+Anxious(x) +error 8) Immunization=Intercept+Office+Secu re(x)+Avoidant(x)+Anxious(x)+error The PRM is a Generalized Linear Mo del (GLM) with an assumed Poisson distribution for a variable Y using a log li nk. Although analyses can be modeled using a GLM procedure using an identity link, it is more common to model the log of the mean that is always positive and can take any real value. The following is the form of the PRM: log x where exp x)=ea(eb)x. As the model illustrates, an increase of one unit in X impacts the influence of e on on a multiplicative scale. Furthermore, if b>0 then the mean increases as X increases. Conversely, if b<0, the mean decreases as X increases (Agresti, 1996). One issue that can arise from response counts being used rather than a true Poisson distribution is that the variability can be greater. A common cause for increased variability is heterogeneity among study subjec ts. This variability leads to a variance being greater than the mean rather than th e two being equal as is the case in a true Poisson distribution (Agresti, 1996, p. 80). This effect is refe rred to as over-dispersion and is common in binomial and Poisson mode ling. Unlike ordinary regression models that have a separate variance parameter for the mean, the binomial and Poisson distributions base their varian ce as a function of the mean.

PAGE 117

94 If the distribution of responses were mo re variable than a standard Poisson distribution, the response vari ance would be proportional to the mean rather than equal and adjustments would be made to the es timates (Agresti, 1996, p. 80). A number of approaches to control for over-dispersion ex ist. However, the most straightforward approach is to adjust the standard error. The adjusted standard error (ASE) can be computed through a modification of the expected value of X2. Dividing X2 by the degrees of freedom results in a estimated proportiona lity constant. Using GLM software, the ASE can be computed by multiplying the GLM values by the scaling factor X2/df. It should be noted that standard maximum likelihood pa rameter estimates are still applicable and inferences are made in a traditional fashion using the ASE (Agresti, 1996). The appropriate method for performing si gnificance tests of null hypotheses about parameters produced using count data is the Wald statistic. It is one of the more simplified estimate procedures and uses the large-sample normality of maximum likelihood estimates. The Wald statistic divide s the parameter estimate by the ASE and is represented as: z= ^/ASE. Once parameters have been estimated, the model must be checked using goodness of fit stat istics to determine the adeq uacy of the model (Agresti, 1996). The final estimated model parameters include the degrees of freedom, parameter estimate, standard error, Wa ld 95% confidence limits, chi-square test and p-value. Phase Four: Assessing Moderating Factors Once the initial model including maternal interaction style, reported experiences of care and health care utiliz ation were conducted, testing fo r potentially and moderating variables was conducted. To accomplish this a two-step process using all of the

PAGE 118

95 independent variables was used. First each of the nine potentially moderating variables was added to the full main effects model one at a time using a SAS v9.1 Poisson regression analysis. Next a saturated model including an inte raction term for each variable by the three interaction styles was added to determ ine whether the inclusion of the interaction terms significantly added to the model. This comparison of the main effects model and the saturated model was conducted by using the difference of the log likelihoods. This log likelihood difference is th e likelihood ratio test value wi th the degrees of freedom being the difference in the number of variab les included in the two models. The Wald statistic follows the chi square distribu tion (Stokes, Davis & Koch, 2000). For these analyses the three interaction terms per mode rator leads to three degrees of freedom. In the case of race/ethnicity, two dummy variab les were used (Black, non-Hispanic, and Hispanic) resulting in the addi tion of six interaction terms and six degrees of freedom. The corresponding critical values for the chi square distri bution are 12.59 for 6 degrees of freedom and 7.82 for 3 degrees of freedom. The modeling of testing this hypothesis resu lted in a total of 144 separate models (9 potential moderators *main effect or saturated model* four health ca re utilization measures provider or office) wh ich can be found in Appendix O. The following is an example of the model structure for this hypothesis: Main Effects Model: Health Care Utilization variable= Intercept+Provider+Secure(x)+Avoidant (x)+Anxious(x)+Moderator (x)+error

PAGE 119

96 Saturated Model: Health Care Utilization variable= Intercept+Provider+Secure(x)+Avoidant(x)+Anxious(x)+Moderator (x)+Moderator*Secure + Moderator* Avoidant+Moderator*Anxious +error Sample Size Calculation Now that the various statistical methods to be utilized in the study have been described, the calculation of sample size and power can be discussed. First, a factor analysis is a complicated process requiring at least 100 cases and having a general rule of thumb requiring 10 cases per variable (ACITS, 1995). There were up to 6 variables that were included in the f actor analysis, resulting in a minimum sample size of 100. Calculating the sample size for the multiple regression analyses included an alpha of 0.05, beta of 0.20, and R2 of 0.10, and 13 independent variables to be included in the model resulted in a minimum sample si ze of 120 cases (Cohen, 1988). Since the regression sample size was the analytic techni que requiring the largest sample size, it was the overall minimal sample size required for the study.

PAGE 120

97 Chapter Four Results This study examined the relationships among interaction style, reported experiences of care and utiliza tion of health care services. The results of this study are provided in the order of the research ques tions and phases of analyses described in Chapter Three. Recruitment of Study Sample Recruitment resulted in 139 women being interviewed for the study. However, 13 women were ultimately excluded from the study due to ineligibility or incomplete records (Table 1). In the case of seven, discussions during the interview revealed that the child was not eligible for the study. Included in th is figure was one woman who had adopted her grandchild, and another child who entered in to foster care before data was abstracted. Of the remaining 132 eligible women, attempts to obtain complete medical records for children of six women were unsuc cessful. The primary reasons for this loss of data were twofold. First, women changed from one pediat rician to another w ithout transferring the medical record. Second, subsequent efforts to contact mothers to obtain new releases of medical information from the original pediat rician were unsuccessful. For example, one mother provided the names of two pediat ricians she claims treated her child.

PAGE 121

98 Table 1 Final Study Population (N=139) Hillsborough Pinellas Total N % N % N % Completed Interviews 6949.67050.4 139 100.0 White 2942.03651.4 65 46.8 Black 3043.52738.6 57 41.0 Hispanic 1014.5710.0 17 12.2 Excluded: Ineligibility 45.834.3 7 5.0 Excluded: Incomplete Records 57.211.4 6 4.3 Final Sample Size 6047.66652.4 126 100.0 In the case of one mother, the pediatrician provided services to the womans other children but not the target child. A second provider acknowledged seeing the child just after birth but had not seen the child again. Finally, one mo thers provider refused to release medical information even though the mother had given appropriate authorization for the information. It should be noted that when compared to the retained cases, dropped cases were not statisti cally different in regard to maternal race, education, marital status or age (Table 2). A full descri ption of the recruitment process can be found in Appendix O.

PAGE 122

99 Table 2 Comparison of Study Participants with Comp lete Data Versus Those Excluded (N=139) Number Excluded % Excluded ChiSquare(df) Pattern of Finding Maternal Race White Non-Hispanic 4 6.2 .286(2) Not Significant Black Non-Hispanic 8 14.0 Hispanic 1 5.9 Maternal Education <12th Grade 6 17.1 .136(3) Not Significant 12th Grade/GED 4 8.0 Some College 2 5.4 > Bachelors Degree 1 5.9 Marital Status Married/Live-in 8 1.3 .159(1) Not Significant Single 5 6.5 One Way ANOVAs Mean SD F(df) Pattern of Finding Maternal Age (Years) Dropped 26.1 6.2 .663 (1) Not Significant Included 26.9 6.5 Note Includes Widowed, Divorced and Separated Data Analysis Phase One: Data Reduction and Transformation The intent of the first phase in the anal ysis was to reduce th e data into its key components. This reduction process began w ith descriptive stat istics of the study variables. Exploring the characteristics of these variables allows for aggregation and elimination of some data as well as identif ying any variables that require transformation prior to use in further analyses.

PAGE 123

100 Component One: Descriptive Statistics Maternal Demographics. Respondents were recruited from Hillsborough (66, 52.4%) and Pinellas (60, 47.6%) C ounties (Table 3). The racial and ethnic distribution of study participants included 61 (46.2%) White (non-Hispan ic), 49 (38.9%) Black (nonHispanic), and 16 (12.7%) Hispanic. Table 3 County of Residence by Race/Ethnicity (N=126) Race/Ethnicity Hillsborough Pinellas Total N % N % N % White NH 26 43.3 35 53.061 48.4 Black NH 24 40.0 25 37.949 38.9 Hispanic 10 16.7 6 9.116 12.7 Total 66 52.4 60 47.6126 100.0 Note NH: Non-Hispanic In regard to marital status, more than half of the women reported either being married (45, 35.7%) or living with a partner (27, 21.4%) at the time of the interview, while ten (7.9%) were divorced, widowed, or separated and 44 (34.9%) were never married (Table 4). There were statistically significant differences (p<.0001) among racial and ethnic groups regarding whether or not the women were in more permanent relationships (married or live-in partner). White non-Hispanic (45, 73.8%) and Hispanic (12, 75.0%) women were most likely to be in a relationship while Black non-Hispanic women were least likely to be married or have a live-in partner (15, 30.6%).

PAGE 124

101 Table 4 Marital Status (N=126) N % Single 44 34.9 Married 45 35.7 Live-in Partner 27 21.4 Widowed/Divorced/Separated 10 7.9 N % Chi-Square (df) Pattern of Finding Married/Live-in by Race/Ethnicity White Non-Hispanic 45 73.8 .000(2) Black Non-Hispanic 15 30.6 Hispanic 12 75.0 Black women were less likely to be married or have a live-in partner. Levels of education of st udy participants varied greatly with 29 (23.0%) mothers having less than a high school education, 46 (36.5%) having a diploma or GED, 35 (27.8%) having earned some colle ge credits but had no degree and 16 (12.7%) earned a bachelors degree or higher (Table 5). The leve ls of education were significantly different among women with varying ra cial and ethnic backgrounds (p<.05). Black non-Hispanic (13, 26.5%) women were less likely to have ach ieved more than a high school diploma or GED than White non-Hispanic (30, 49.2%) and Hispanic (8, 50.0%) women. There were also differences by marital status (p<.05) w ith women who did not have a live-in partner or husband (16, 29.6%) being less likely to have obtained at least a high school diploma or GED than those with a partner (35, 48.6%).

PAGE 125

102 Table 5 Maternal Education (N=126) N % Highest Grade Completed < High School 29 23.0 High School Diploma/GED 46 36.5 Some College 35 27.8 > Bachelors Degree 16 12.7 N% ChiSquare(df) Pattern of Finding Categorized: >12th Grade/GED 5140.5 By Race/Ethnicity White Non-Hispanic 3049.2 .039(2) Black Non-Hispanic 1326.5 Hispanic 850.0 Black women received less college education. By Marital Status Married/Live-in Partner 3548.6 .043(1) Single* 1629.6 Women without a live-in partner received less college education. Note Includes Widowed, Divorced, and Separated Mothers ranged in age from 18 to 46 y ears with an average age of 26.9 years (Table 6). Seven (5.6%) of women had children in their 40s, while also having children currently in their teens or twenties. None of these olde r mothers had planned on having another baby. Maternal age is associated with a number of other f actors included in this study. For example, women with at least some college were significantly (p<.05) older (28.5 years) than those who completed 12 or fewer years of education (25.8 years). Mothers with live-in partners (28.0 years) we re significantly older (p<.05) than single mothers (25.4 years). Similarly, mothers with multiple children were significantly (p<.001) older (28.5 years) than mothers w ith only one child (24.5 years). Finally,

PAGE 126

103 women with pre-pregnancy health problems (30.4 years) were also significantly (p<.01) older than those with no h ealth problems (26.3 years). Table 6 Maternal Age (N=126) Mean SDMin.Max.Kurtosis Skew Maternal Age 26.9 6.51846.040 .216 One Way ANOVAs Mean Years SD Significance Pattern of Finding By Education No College 25.86.5.021 Some College 28.56.3 Women with some college were older. By Marital Status Married/Live-in Partner 28.06.4.034 Single/Widow/Divorced/Separated25.46.5 Mothers with live-in partners were older. By Having Multiple Children One Child 24.56.1.001 Multiple Children 28.56.4 Mothers with only one child were significantly younger. By Pre-pregnancy Health Problems No Problems 26.36.2.010 Pre-pregnancy health problems 30.47.5 Older mothers had more prepregnancy health problems. Maternal health issues can influence health care utiliz ation. Self-reported overall ratings of health indicated that most wome n in the study believed they were relatively healthy with an average rating of 7.9 where one equates to poor health and ten equates to excellent health (Table 7). However, health ratings ranged from two to ten with some mothers having serious health problems such as ovarian cancer and severe epilepsy. Mothers who experienced pre-pregnancy h ealth conditions reported lower (p<.001)

PAGE 127

104 overall health ratings (6.7) than mothers with no pre-pregnancy health conditions (8.1). In addition to physical problems, 21 (16.7%) of mothers indicated experiencing postpartum depressive symptoms recently as indicated by mothers answering yes two both depression screen questions. Table 7 Current Maternal Health Status (N=126) MeanSDMin.Max.KurtosisSkew Overall Health Rating 7.91.8210.056-.769 One Way ANOVAs MeanSigni ficancePattern of Finding By Pre-pregnancy Issues No Health Problems 8.1.001 Pre-Pregnancy Problems 6.7 Mothers with prepregnancy health issues report lower current health ratings. N % Depression Screening: Depressed 21 16.7 An individuals feelings about docto rs can also influence health care utilization. The women in th is study used a rating rangi ng from one Enjoy going to the doctor to ten Dislike going to the doctor. The average rating was 5.4 (SD 2.6, Kurtosis -7.0, Skew .17). One mother reported she puts off her own care but will take her children regardless of her feelings. However, the same mother did not show up for eight scheduled pediatric visits during the ch ilds first year of life. A nother woman liked going to the doctor for herself because it was a brief resp ite from her children. The mother reported enjoying being able to sit quietly in the waiting room to read a magazine. Another issue that can influence a mother s attitude during the childs first few years of life is whether the child was planned to be part of the fam ily. In the case of study

PAGE 128

105 respondents, more than half (76, 60.3%) of the infants born we re either wanted later in the mothers life (48, 38.1%) or not intended at all (28, 22.2% ) (Table 8). Twelve (9.5%) mothers had wanted their children earlier a nd 39 (30.2%) of the infants were wanted at the time they were born. Table 8 Planning of Pregnancy for Target Child (N=126) Yes % Intendedness of Pregnancy Wanted Earlier 12 9.5 Wanted at that time 38 30.2 Wanted later in life 48 38.1 Not Planned 28 22.2 Having other children also can make it mo re difficult to manage preventive health care activities such as well child care visits and immunizations. This is especially true for subsequent pregnancies because of the need to attend prenatal care visits in addition to the standard well child visits and any sick visits. For this st udy, the majority of participants (76, 60.3%) had more than one child (Table 9). More specifically, 50 (39.7%) mothers had one child, 67 (53.2%) had at least one child olde r than the target infant, nine (7.1%) gave birth to a subsequent child in addition to the target infant, one (0.8%) had both older and younge r children, and 11 (8.7%) we re pregnant with an additional child at the time of the intervie w. One mother had an 18 month old target child, a nine month old and three week ol d twins. Mothers averaged 2.3 children.

PAGE 129

106 Table 9 Pregnancy History (N=126) N % Number of Children Target Child Only 50 39.7 Older child(ren) and Target Child Only 67 53.2 Target Child and Younger child(ren) only 8 6.3 Older and Younger Children 1 0.8 Number who were pregnant (with or without other children) 11 8.7 MeanSDMin.Max.Kurtosis Skew Number of Children 2.31.31108.4 2.1 Target Child Health Issues. Data regarding birth wei ght was not provided by all pediatric offices with 55 (43.7%) cases with missing data (Table 10). However, children with birth weight data weighed, on average, 3,281 grams at birth. Birth weights ranged from 1,792 to 4,345 grams. Gestational ag e ranged from 34 to 42 weeks. Table 10 Child Health Issues (N=126) NMeanSDMin.Max. Kurtosis Skew Birthweight 713281.657417924345 -.244-.239 Gestational Age (wks) 4638.91.93442 1.00-1.12 Months Breastfed (> 1mo) 677.05.1118 -.411.852 Yes%No % Breastfed at Least One Month 6753.259 46.8 Breastfed at Least Three Months 5542.971 56.3 Frequent Sleeping Problems 86.3118 93.7 Failure to Thrive 32.4123* 97.6 Propensity for Accidents 64.8120* 95.2 Child Health Issues 129.5116 90.6 N% (df) Pattern of Finding Married/Live-in Partner 34.1.029 Single/Widow/Divorced/Separated 916.7 Single mothers reported more child health issues. MeanSD Min. Max. Kurtosis Skew Childs Health Rating 9.1.126110 9.8-2.67 Note: No or Missing

PAGE 130

107 For infants, breastfeeding can have a vari ety of benefits. Sixty-seven (53.2%) in the study breastfed at least one month, 55 ( 42.9%) of women in the study breastfed at least three months and the average length of time breastfeeding wa s seven months. An additional nine women tried br eastfeeding but were unsuccessful (Table 10). Statewide, 68% of mothers receiving Medica id reported breastfeeding with 9% breastfeeding at least three months (Florida Department of Health, 2004). Some signs that there are issues with the child include sleep problems, failure to thrive and a propensity for accidents. Eight (6.3%) mothers in this study reported the infant had frequent sleeping problems (Table 10). Additionally, thre e infants (2.4%) were identified as weighing below what is reco mmended on a standard growth chart and six (4.8%) infants were noted by providers as having a higher than normal propensity for accidents. Looking at the childs overall hea lth status, few (12, 9.5%) children had noteworthy health problems (Table 10). The ove rall health rating for children averaged 9.1 on a continuum from one to ten where one re presents poor health and ten represents excellent health. Twelve (9.4%) children had health ratings be low eight and were considered to have some health issues. One potential barrier to r eceiving pediatric healthcare is working a fulltime job. The majority of mothers were not employe d (81, 64.3%) (Table 11). Of the 45 (35.7%) women who were employed, they averaged 31.7 hours per week with 16 (35.6%) of the mothers working less than 30 hours per week. No mothers indicated that work issues prevented or inhibited them from taking their children to health care visits.

PAGE 131

108 Table 11 Employment and Daycare (N=126) N % Unemployed 81 64.3 Had a Job 45 35.7 Worked <30 hours 16 35.6 Worked >30 hours 29 64.4 N Mean SD Min. Max. Kurtosis Skew Weekly Work Hours 45 31.7 10.17 50 -.03 -.8 Weekly Daycare Hours 31 35.4 9.65 45 -.02 -.8 Yes % Daycare Requires Proof of Immunizations 28 90.3 When working, and in the case of this study where many of the mothers were in school, daycare centers were often used. Th irty-one (24.6%) women put their child in daycare for an average of 35.4 hours per week (Table 11). Of the 31 women who used daycare services, 28 (90.3%) women indicated that their daycare facility required proof of immunizations. Intervention programs can also influen ce attendance at hea lth care visits. For example, infants receiving WIC are required to keep up to date with their immunizations. Statewide, 74% of women receiving prenat al care services th rough Medicaid reported using WIC (Florida Department of Hea lth, 2004). In the cu rrent study, 111 (88.1%) women reported receiving WIC (Table 12). Wo men having no college education were more likely (p<.01) to be involved with WIC (94.7%) than women with at least one year of college (78.4%). Another progr am that encourages utilizati on of preventive health care services is Healthy Start. Thirty-seven (29.4%) women re ported being involved with Healthy Start although not all of the women reported particip ating in the Healthy Start program for the full 12 months of the childs life.

PAGE 132

109 Table 12 Other Child-Serving Programs (N=126) N % Received WIC 111 88.1 Received Healthy Start 37 29.4 N % ChiSquare(df) Pattern of Finding WIC by Education No College 7194.7 .010 Some College 4078.4 Fewer mothers with some college education used WIC services. Additional survey questions addressing ma ternal health history and pediatric health care issues are presented in more detail in Appendix P and Q. Interaction Style. The original intent of the st udy was to explore differences among women with anxious, avoidant and secure attachment styles. However, only three (2.4%) of the women recruited and interviewe d were categorized with a predominantly anxious rating (Table 13). The majority of women (72, 57.1%) were categorized as avoidant with the remaining 51 (40.5%) of women being categorized as secure. This differs from the national literature which in dicates that approximately 55% is secure, 25% is avoidant and 15% is anxious. To enha nce the power of statis tical analyses, it was decided to include the conti nuous rating scales rather th an using the dichotomized secure/avoidant variable. Mothers scores av eraged 22.3 for the secure attachment scale, 24.0 for the avoidant scale and 15.0 for the anxi ous scale. For a more detailed description of specific Relationship Scales Questionnaire items see Appendix R.

PAGE 133

110 Table 13 Maternal Interaction Style (N=126) Primary Interaction Style Secure Avoidant Anxious N%N% N % 5140.57257.1 3 2.4 MeanSDMinMax Kurtosis Skew Secure Attachment Score 22.34.21233 -.345 .072 Avoidant Attachment Score 24.04.31135 .007 -.072 Anxious Attachment Score 15.04.1425 -.231 -.122 The Cronbach Alpha measure of internal consistency for the Relationship Scales Questionnaire was .78. The protocol for calcu lating the three sub-scales of secure, anxious, and avoidant interac tion styles does not use all 30 questions. Looking only at those questions used in the subscales, the corr elation coefficient drops to .36 (Table 14). Specific sub-scale reliability alphas were .35 (secure), .48 (avoidant) and .26 (anxious). Table 14 Item Total Correlations (N=126) Item Cronbachs Alpha All questions in subscales: 2, 3, 6, 8, 9, 10, 15, 16, 19, 22, 25, 26, 28 .36 Secure Items: 3, 9, 10,15,28 .35 Avoidant Items: 2, 6, 19, 22, 26 .48 Anxious Items: 6, 8, 16, 25 .26 Note : Ratings reversed These coefficients are slightly lower than was reported in the psychometric testing of the RSQ subscales (Griffin and Bartholomew, 1994). Additionally, Griffin and Bartholomew (1994) acknowledged that Cronbach alpha coefficients we re slightly lower than ideal (i.e., secure alpha=.41 and avoida nt alpha=.70). To try to explain these low alphas, the researchers noted that the scales still demonstrated high convergent validity

PAGE 134

111 (.57 or higher) and hypothesized that corre lations were low because two different orthogonal dimensions were being combin ed within one subscale. Griffin and Bartholomew (1994) also suggested that secu re interaction style, having the lowest convergent validity, was most sus ceptible to self-report bias. Reported Experiences of Care Questions. Table 15 includes the responses of the reported experience of care quest ions pertaining to the child s health care provider. Most (103, 82.7%) mothers indicated no problem findi ng a suitable doctor or nurse for their child (Table 15). An additiona l 16 (12.7%) mothers indicated having a small problem and seven (5.6%) reported having a big problem. Six (4.8%) mothers reported interviewing or taking their child to appointments involving thre e or four providers before settling on one they liked. Table 15 Consumer Assessment of Health Plans Survey Questions Pediatrician (N=126) Problem Big Small No N % N % N % Problems Getting a Personal Provider for Child: 7 5.6 16 12.7 103 82.7 N % Child has no Personal Provider 17 13.5 Child has a Personal Provider: 109 86.5 Mother educated by Provider 112 88.9 Mean SD Min Max Kurtosis Skew Childs Personal Provider Rating: 8.7 2.0 1 10 6.13 -2.27 National Data Study Data National CAHPS Categories % N % Low (0-6) 10 14 11.1 Medium (7-8) 25 28 22.2 High (9-10) 65 84 66.7

PAGE 135

112 Seventeen (13.5%) mothers indicated that they did not believe they had one provider that usually saw their child, seeing wh atever doctor was on call at the time of the visit (Table 15). These women were asked to reflect upon the doctor they saw most often when answering the provider rating questions The overall rating of the provider averaged 8.7 on a scale where zero represen ts poor and ten represents excellent. The distribution of these ratings are similar to those found in the National CAHPS Benchmarking Database. Finally, most (112, 88.9%) mothers indicated th at educational conversations with the provider made mothers feel like they were bett er able to manage their childs care. For a more detailed description of specific expe riences of care questi ons, see Appendix S. The overall provider ratings were con tinuous but negatively skewed and had a kurtosis that exceeded 1.0. As a result, pr ovider ratings were inverted (i.e., 0=11, 1=10=1) to reverse the skew. Data were th en transformed by taking the square root of the rating. Once transformed, data were mo re similar to a normal distribution with the kurtosis reduced from 6.13 to 1.72 and reduci ng the skew from -2.27 to 1.36 (Table 16). Although the transformation of the data di d not reduce the kurtosis below 1.0, the regression analysis is somewh at robust to this issue. Table 16 A Comparison of the Provider Rating with a Transformed Provider Rating (N=126) Mean SD Min Max Kurtosis Skew Provider Rating: 8.7 2.0 1 10 6.13 -2.27 Transformed Provider Rating: 1.4 .54 1 3.3 1.72 1.36

PAGE 136

113 When asked to provide an overall rating of the providers office, mothers average rating was 8.5 on a continuum ranging from zer o to ten where zero representing the worst possible care and ten representing the best Because of the non-normal distribution (excess skew and kurtosis) the same process of reversing the ratings and transforming the data by taking the square root was perfor med for the provider office rating. Table 17 illustrates the changes in skewness of the provider rating from -1.78 to .90 after the transformation. Additionally, the kurtosis was re duced from 4.5 to .53. It should be noted that although odds ratios ar e typically presented for re gression analyses, this transformation process makes interpretation of the odds ratio diffi cult and will not be presented in the analysis of the data (see Appendix T). Table 17 A Comparison of the Provider Office Rating with a Transformed Provider Office Rating (N=126) Mean SD Min Max Kurtosis Skew Provider Rating: 8.5 1.80 10 4.50 -1.78 Transformed Provider Rating: 1.5 .521 3.3.53 .90 Health Care Utilization. The average number of va ccinations provided to each child was 11.7 of the 13 recommended immuni zations with only 73 (57.9%) receiving all 13 of the vaccinations (Centers for Dis ease Control, 2005) (Table 18). The most commonly missed vaccinations include the pne umococcal vaccine (missed shots: 10[1st shot], 18 [2nd shot], 40 [3rd shot]) and the HepB vaccine (missed shots: 7, 14, 30). In regard to well child care visi ts, an average of 5.8 visits wa s achieved with 51 (40.4%) of the infants receiving all of the recommended visits (at least seven of eight). Infants

PAGE 137

114 averaged 5.0 sick and follow-up visits with the primary reasons for those visits being: otitis media, upper respirator y infections, and diarrhea. Fi nally, emergency department (ED) visits were made by 43 (34.1%) of the infants included in the study with 20 (15.9%) infants having more than one ED visit. Table 18 Health Care Utilization Rates (N=126) MeanSDMinMax Kurtosis Skew Well Child Care Visits (7-8 Ideal) 5.81.518 .66 -.86 Sick Visits 5.03.9017 -.05 .76 Emergency Department Visits 0.50.804 1.90 1.52 Immunizations(Up to date:13 shots)** 11.72.6013 10.38 -3.09 Missing Immunizations 1st 2nd 3rd Diptheria, Tetanus, Pertusis 3 8 16 Haemophilus influenza type b 4 7 26 Pneumococcal 10 18 40 Inactivated Polio Virus 3 9 43* Hepatitis B 7 14 30* Note 3rd Shot given between 6-18 months not included in shot count. Note ** Three mothers refused immunizations a nd were excluded from further analyses. Component Two: Factor Analysis The development of a composite mothe r-child bonding score involved a factor analysis of the dichotomized factors of infa nt sleep disturbance, failure-to-thrive, timing of the pregnancy, breastfeeding longer th an one month, an infa nts propensity for accidents, and maternal depres sion. A factor analysis of th ese six variables resulted in three factors with Eigen values greater th en one (Table 19). The following model was originally tested: C1=b1(AC) + b2(UP) + b3(SD) + b4 (MD) + b5(BF) + b6(FT).

PAGE 138

115 Table 19 Factor Analysis of Bonding (N=126) Six Variable Component Matrix* Four Variable Component Matrix Factor Factor Cronbachs Alpha 1 2 3 1 Eigen Value 1.291.111.041.26 .30 Propensity for Accidents (AC) -.18 .68 -.12 Unintended Pregnancy (UP) .70 -.24 .10.48 Infant Sleep Disturbance (SD) .53 .30 -.17.54 Maternal Depression (MD) .11 .50 .27.46 Breastfeed <1 Month (BF) .13 -.01 .51.27 Failure to Thrive (FT) -.20 .02 .73 MeanSD Min. Max. Kurtosis Skew Bonding Score 0 1 -1.18 3.73 2.6 1.3 Four-Variable Bonding One Way ANOVAs Mean SD F(df) Pattern of Finding White Non-Hispanic -.26 0.814.44 (2) Black Non-Hispanic .30 1.12 Hispanic .07 1.06 White mothers had the strongest bonds and Black mothers had the weakest bonding scores. Note Varimax Rotation The data regarding failure to thrive and propensity for accidents was obtained from pediatric provider offices. Individuals who abstracted the da ta had a variety of backgrounds including some having no medical tr aining. As a result, the reliability of the data for these two questions was questionable. Th is data quality issue can also be seen in the identification of three separate factors resulting from six vari ables in the factor analysis. Therefore, these two variables we re dropped from the analyses. Conducting a factor analysis on the remaining four va riables (unintended pregnancy, maternal depression, lack of breastfeeding, and infant sl eep problems) resulted in one factor that represented the four variable s having a Eigen value of 1.26 (Table 19). The model tested was: C1= b1(UP) + b2(SD) + b3 (MD) + b4( BF). Using the four-factor weighted

PAGE 139

116 composite score, white (non-Hispanic) mothers had the most positive bonding relationship while black (non-Hi spanic) had the least positive bonds with their children. It should be noted that the Cronbachs Alpha measure of internal consistency was .30, below what is accepted as adequate (.80) for psychosocial research. Component Three: Bivariate Associations Bivariate Pearson correlation coeffici ents were calculated between each independent variable and the two measures of experiences of care fo r the provider and the office (Table 20). The provider office rating (.679, p<.01) was highly correlated with the provider rating. Maternal age wa s correlated with the provid er rating (r=-.184, p<.05) and the provider office rating (r=-.180, p<.05). The difference between provider and provider office rating correlations was the strong associ ation between whether the mother liked (1) or disliked (10) going to the doctor (r=.235, p<.01) and the provider office rating. Table 20 Bivariate Correlations with Experi ences of Care Ratings (N=126) Provider Rating Provider Office Rating Provider Rating NA .679** Bonding Factor 1 -.061 -.128 First Child @ -.028 -.010 Black non-Hispanic @ .018 -.089 Hispanic@ .082 .105 Working at least 30 hours per week@ -.128 -.010 Overall Maternal Health Rating .008 -.032 Feelings About Doctors .159 .235** Overall Child Health Rating .042 .054 WIC or Healthy Start Involvement -.048 -.069 Maternal Age -.184* -.180* Note @: Yes=1, 0=No; P<.05, ** P<.01

PAGE 140

117 Bivariate Pearson correlation coefficients also were calculated between each independent variable and the three measures of interacti on style (secure, anxious, and avoidant) (Table 21). The level of secure in teraction style was signi ficantly correlated with the mothers overall health rating (r=.202, p<.05). The avoidant interaction style score was correlated with a mothers race/ ethnicity being Black (r=.298, p<.001). Finally, the level of anxious interaction style was corre lated with both overall maternal health (r=.216, p<.05) and child health ratings (r=-.216, p<.05). Table 21 Bivariate Correlations with Intera ction Style Scale Scores (N=126) Interaction Style Secure Avoidant Anxious Bonding Factor 1 -.048 .099 .102 First Child* .030 -.047 .151 Black non-Hispanic* -.140 .298** .081 Hispanic* .005 .041 -.092 Working at least 30 hours per week* .004 .109 -.155 Overall Maternal Health Rating .202* -.076 -.216* Feelings About Doctors -.024 .042 .085 Overall Child Health Rating .133 -.119 -.216* WIC or Healthy Start In volvement -.078 -.065 -.044 Maternal Age .007 -.026 -.158 Note a: Yes=1, 0=No; P<.05 Bivariate Pearson correlation coefficien ts calculated between each independent variable and the four measures of health care utilization (we ll child care visits, sick/follow-up visits, emergency department visits, and immunizations) were notably different (Table 22). For example, as mate rnal bonding issues decreased (lower bonding score) the number of well child care visits increased (r=-.315, p<.01). Additionally, Black non-Hispanic mothers were less likely to take their child to as many well child care visits

PAGE 141

118 when compared to White and Hispanic mo thers (r=-.238, p<.01). Mothers whose first child was the studys target child were more likely to take their children to recommended well child care visits compared to moth ers with older children (r=.178, p<.05). When looking at the number of sick or follow-up visits, lower ratings of the childs overall health (r=-.197, p< .05) were associated with more visits (Table 22). The number of times a mother took her child to the emergency room wa s correlated with the maternal bonding measure (r=.180, p<.05). Finall y, the number of immunizations a child received was associated with whether the mother worked more than 30 hours per week (r=-.253, p<.01). Table 22 Bivariate Correlations with Transformed Health Care Utilization Data (N=126) Well Child Visits Sick/ Follow-up Visits Emergency Department Visits Immunizations Bonding Factor 1 -.315**.020 .180* -.060 First Child* .178* .120 .072 .154 Black non-Hispanic -.238**-.108 .046 -.016 Hispanic .062 .016 -.011 .012 Working at least 30 hours/week .041 .019 .019 -.253** Overall Maternal Health Rating -.128 .043 .007 -.058 Feelings About Doctors -.148 -.028 .106 -.019 Overall Child Health Rating -.014 -.197* -.170 -.017 WIC or Healthy Start Involvement -.126 -.092 .055 .048 Maternal Age -.003 .019 -.169 -.154 Note P<.05, ** P<.01 The last set of correlatio ns explored the relations hips among variables that potentially moderated the associations betw een the primary independent and dependent variables (Table 23). Black non-Hispanic race and Hispanic Ethnicity are highly

PAGE 142

119 correlated (r=-.304, p<.01) because they are dummy variables for race/ethnicity with White non-Hispanic as the referent group. Bl ack non-Hispanic race/ethnicity was also negatively correlated with child s overall health rating (r=-.197, p<.05), and maternal age (r=-.201, p<.05). Black race was also positively correlated with having maternal bonding issues (r=.244, p<.01), and working at leas t 30 hours per week (r=.183, p<.05). Hispanic ethnicity was not correlated with any other variables. However, maternal age also was correlated with whether the child was the mo thers first (r=-.362, p<.01) and the childs overall health status (r=.182, p<.05). Additi onally, maternal bonding was associated with WIC or Healthy Start participation (r=.249, p<.01) childs overall hea lth status (r=-.180, p<.05) and the mother working at leas t 30 hours per week (r=.272, p<.01). Finally, childs overall health rating also was correlate d with whether the moth er worked at least 30 hours per week (r=-.241, p<.01). Table 23 Bivariate Correlations Among Potentia lly Moderating Variables (N=126) Hisp. First CH MH Bond WIC Work Feel Age Black -.304** -.099 -.197* -.016 .244** .042 .183* -.004 -.201* Hispanic .055 .040 -.052 .027 .147 -.039 -.044 .119 First .028 .008 -.008 .021 -.054 -.111 -.362** CH .052 -.180* -.078 -.241** .003 .182* MH -.103 -.162 .007 -.062 -.067 Bond .249** .272** .068 -.065 WIC -.032 .155 -.134 Work -.045 .060 Feel .036 Note P<.05, ** P<.01 Legend (Hisp.: Hispanic Ethnicity; First: Targ et Child First Child; CH: Childs health Rating, MH: Mothers Health Rating; Bond: Maternal Bonding Work: work >30 hours per week, Feel: Feelings about Doctors; WIC: WIC and/or Healthy Start; Age: Mothers Age at Interview)

PAGE 143

120 Data Analysis Phase Two: Interact ion Style and Experiences of Care Research Question 1: Is maternal interac tion style related to a mothers reported experiences with her child s pediatric health care To explore the relationships between ma ternal interaction style and reported experiences of care, the following two linear re gression models were used to test this hypothesis: 1) Provider Score=intercept + secure (x) + avoidant (x) + anxious (x) + error. 2) Provider Office Score=inter cept + secure (x) + avoidant (x) + anxious (x) + error. Including only the three in teraction style scores in the model, there were no associations between the transformed provide r ratings and the inte raction style scores with significance levels ranging from .350 for secure scores to .822 for avoidant scores (Table 24). Similarly, no associations we re found between the provider office and the different interaction style scores with signi ficance levels ranging from .404 for avoidant scores and .877 for anxious scores. Potential reasons for this lack of association are discussed in Chapter five.

PAGE 144

121 Table 24 Transformed Provider Rating, Provider Of fice Rating and Interaction Style (N=126) Interaction Style B SE df t-valueSig. R2 Provider Rating Intercept 1.537.452 13.40.001 .01 Secure -.011.012 1-.94.350 Avoidant .003.012 1.23.822 Anxious .005.013 1.41.686 Provider Office Rating Intercept 1.841.439 14.19<.0001 .01 Secure -.003.012 1-.26.796 Avoidant -.010.012 1-.84.404 Anxious -.002.013 1-.16.877 Research Question 2: Are reported maternal experiences with pediatric health care related to pediatric health care utilization dur ing the first 12 months of a childs life? To test this hypothesis, ei ght separate Poisson regre ssion models were run to included the provider or provide r office rating and one health care utilization measure: 1) Well Child Care=Intercept + Provider + error 2) Sick/Follow-up= Inter cept + Provider + error 3) Emergency Dept=Inter cept + Provider + error 4) Immunization=Intercep t + Provider + error 5) Well Child Care=Intercept + Office + error 6) Sick/Follow-up= Intercept + Office + error 7) Emergency Dept=Int ercept + Office + error 8) Immunization=Intercept + Office + error

PAGE 145

122 As noted in Chapter three, the distri bution of the provider and office ratings indicated the need to transform the data. This transformation process reversed the scale so that a lower rating was better. Therefore, the significant correlation of the provider rating (B=-.189, p<.05) with the number of sick or fo llow-up visits a child received indicates that better provider ratings were associated w ith more sick visits. Neither the number of well child care visits, emergency department visits, nor immunizations were significantly associated with provider ratings. Furthe rmore, no significant associations were found between provider office ratings and any hea lth care utilization measures (Table 25). Table 25 Health Care Utilization by Transformed Provider and Provider Ratings (Unadjusted) (N=126) Dependent Variable IV B SE dfWald CI Limits Wald Sig. Goodness of fit @ Well Child Care Provider -.047.0711-.185.091.44 .506 .40 Sick/Follow-up Provider -.189.0801-.346-.0315.52 .019* 3.00 Emergency Dept Provider .004.2311-.448.456.00 .987 1.32 Immunizations Provider .008.0491-.087.104.03 .865 .22 Well Child Care Office .011.0711-.013.151.02 .876 .40 Sick Follow-up Office -.023.0781-.176.129.09 .765 3.02 Emergency Dept Office .191.2281-.025.638.70 .402 1.32 Immunizations Office .015.0511-.084.114.08 .773 .22 Note Pearson Chi-Square (value/df) Phase Three: Interaction Style, Experien ces of Care, and Health Care Utilization Research Question 3. Controlling for reporte d maternal experiences with pediatric health care, is maternal interaction style related to pediatric health care utilization during the first 12 months of a childs life?

PAGE 146

123 This hypothesis has two separate su bcomponents, both having a predicted direction for the association. First, control ling for reported maternal experiences with pediatric health care, is an anxious matern al interaction style related to increased pediatric health care utilization during the first 12 months of a childs life? Second, controlling for reported maternal experiences with pediatric health care, is avoidant maternal interaction style related to decrease d pediatric health care utilization during the first 12 months of a childs life? The models used to test this hypothe sis and its subcomponents include: 1) Well Child Care=Intercept+Provider+S ecure(x)+Avoidant(x)+Anxious(x)+error 2) Sick/Follow-up Intercept+Provider+S ecure(x)+Avoidant(x)+Anxious(x)+error 3) Emergency Dept=Intercept+Provider+S ecure(x)+Avoidant(x)+Anxious(x)+error 4) Immunization=Intercept+ Provider+Secure(x)+Avoida nt(x)+Anxious(x)+error 5) Well Child Care=Intercept+Office+Secu re(x)+Avoidant(x)+Anxious(x)+error 6) Sick/Follow-up=Intercept+Office+S ecure(x)+Avoidant(x)+Anxious(x)+error 7) Emergency Dept=Intercept+Office+Secu re(x)+Avoidant(x)+Anxious(x) +error 8) Immunization=Intercept+Office+Secu re(x)+Avoidant(x)+Anxious(x)+error Controlling for provider and provider o ffice ratings, there we re no significant associations identified between well child ca re visits, emergency department visits or immunizations and interaction style scores (Table 26). However, in addition to the provider rating being associated with the number of sick and follow-up visits (p<.01; OR=.82), avoidant interaction st yle was negatively correlated with the number of visits

PAGE 147

124 (p<.01; OR=.97). This finding supports the al ternative hypothesis (3 a) that women with higher avoidant interaction style scores would te nd to take their child to fewer health care visits. In addition, anxious inte raction ratings were positivel y correlated with the number of sick and follow-up visits (p<.001; OR =1.05). This, too, supports the alternative hypothesis (3b) that anxious inte raction styles would tend to ta ke their child to more sick and follow-up visits. Models containing provider office ratings found no significant associations between office ra ting and utilization of sick/follow-up. However, avoidant interaction style scores were negatively associated (p<.01 ; OR=.97) with sick/follow-up visits. Conversely, anxious inte raction style scores were pos itively associated (p<.0001; OR=1.67) with the number of sick/follow-up visits. Table 26 Health Care Utilization, Interaction Styles, Provider/Provider Offic e Ratings (Adjusted) (N=126) Poisson Regression B SE df Wald CI Wald Sig. GF/OR Provider Ratings Well Child Care GF=.40 Intercept 1.859.36611.1322.56625.55 <.0001 Doctor -.047.0711-.186.092.44 .508 Secure -.002.0091-.020.017.03 .865 Avoidant .004.0091-.014.023.20 .654 Anxious -.006.0101-.025.013.37 .545 Sick/Follow-up GF=2.87 Intercept 1.854.39311.0852.62322.3 <.0001 Doctor -.201.0821-.362-.0416.07 .014 Secure -.002.0101-.022.017.05 .816 Avoidant -.029.0101-.049-.0098.33 .004 ** Anxious .051.0111.030.07222.13 <.0001 ***

PAGE 148

125 Table 26 Continued. Emergency Dept. GF=1.33 Intercept -.7131.2001-3.0661.639.35 .552 Doctor -.022.2311-.474.430.01 .923 Secure -.019.0311-.080.041.39 .535 Avoidant .010.0331-.051.072.11 .744 Anxious .018.0001-.046.082.31 .579 Immunizations GF=.22 Intercept 2.412.25811.9062.91787.35 <.0001 Doctor .009.0491-.105.105.04 .850 Secure .002.0071-.015.015.11 .736 Avoidant -.000.0071-.013.013.00 .948 Anxious .002.0071-.012.015.07 .793 Office Ratings Well Child Care GF=.40 Intercept 1.754.37311.2042.48522.17 <.0001 Office .012.0721-.128.152.03 .864 Secure -.001.0091-.019.017.01 .912 Avoidant .004.0091-.014.023.20 .653 Anxious -.006.0101-.025.013.40 .529 Sick/Follow-up GF=2.92 Intercept 1.583.3981.8022.36315.79 <.0001 Office -.020.0781-.173.133.07 .798 Secure -.000.0101-.020.019.00 .972 Avoidant -.030.0101-.050-.0108.76 .003 ** Anxious .051.0111.029.07221.44 <.0001 *** Emergency Dept. GF=1.33 Intercept -1.1341.2291-3.5431.275.85 .356 Office .205.2291-.244.653.80 .372 Secure -.018.0311-.079.042.35 .555 Avoidant .012.0311-.049.074.15 .701 Anxious .019.0331-.046.084.32 .571 Immunizations GF=.22 Intercept 2.398.26511.8792.91782.05 <.0001 Office .015.0511-.085.115.09 .769 Secure .002.0071-.011.015.11 .741 Avoidant -.000.0071-.013.013.00 .973 Anxious .002.0071-.012.015.07 .785 Note GF: Goodness of Fit Pearson Chi-Squa re (value/df) p<.05, ** p<.01, *** p<.001

PAGE 149

126 Phase Four: Moderating Factors Research Question 4. Are there variables that moderate the relationship of maternal interaction style and reported maternal exper iences with pediatric health care on the utilization of pediatric health care duri ng the first 12 months of her childs life? To examine this fourth hypothesis more closely, a variety of analyses were conducted. For each potentially moderating variable, a main effects model including the moderator, the transformed provider or provi der officer rating, a nd interaction style (secure, avoidant, and anxious) scores was r un. A subsequent interaction model was then performed adding interaction terms between one moderator at a time and the three interaction style scores to the main effect s model. Finally, the difference in the log likelihood estimates for both the main effect s and the saturated m odels was doubled and compared to critical values ( =.05) for a 2 distribution to test for significance. Potentially moderating variables include : Black non-Hispanic(B)/Hispanic (H) [dummy coding], first child (FC), childs ove rall health status (CH), mothers overall health status (MH), WIC/Healthy St art Participation (WC), Working > 30 hours per week (WK), feelings about going to the doctor (F D), maternal age (MA) and maternal bonding (MB) were added to the models. The following is an example of the model structure for this hypothesis (see Appendix U): Main Effects Model : Health Care Utilization variable= Intercept+Provider+Secure(x)+Avoidant (x)+Anxious(x)+Moderator (x)+error Saturated Model : Health Care Utilization variable= Intercept+Provider+Secure(x)+Avoidant(x)+Anxious(x)+Moderator (x)+Moderator*Secure + Moderator* Avoidant+Moderator*Anxious +error

PAGE 150

127 Examining the main effect models related to well child care visits, controlling for provider rating and interacti on style scores, Black non-Hisp anic race/ethnicity (p<.01) and maternal bonding scores (p<.05) were identified as the on ly two potentially moderating variables that were si gnificantly correlated with visits (Table 27). There were no significant interactions among any of th e potentially moderating variables and the number of well child care visits (Appendix U.1 to Appendix U.9). Table 27 Exploration of Potentially Moderating Va riables on the Number of Well Child Care Visits Attended the First Year of Life, Cont rolling for Interaction Styles and Provider Rating (Nine Separate Models: Black/Hispan ic run together as dummy variables) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Models (Model #) Black (1)-.157.0851-.324.011 3.36 .007 ** Hispanic (1)-.027.1171-.256.202 .05 .819 Target Child Mothers First (2).105.0761-.043.253 1.94 .164 Childs Health Rating (3)-.004.0271-.056.049 .02 .896 Mothers Health Rating (4)-.021.0211-.062.020 1.00 .317 WIC/Healthy Start (5)-.104.1111-.321.114 .87 .350 Mother Worked > 30 Hours (6).004.0911-.173.182 .00 .961 Mothers Feelings (7)-.013.0151-.042.015 .83 .361 Mothers Age (8)-.001.0061-.013.010 .06 .813 Mothers Bonding Score (9)-.090.0401-.167-.012 5.15 .023 Note p<.05, ** p<.01 When Black non-Hispanic race, Hispanic ethnicity, and maternal bonding were added to the main effects model including provi der and the three interaction style scores, no variables were associated w ith the number of well child care visits received (Table 28). Similar results were found when inco rporating the relationship of the entire providers office rather than the individual provider (Table 29-30; Appendix U.10-U.18).

PAGE 151

128 Table 28 Exploration of Potentially Moderating Va riables on the Number of Well Child Care Visits Attended the First Year of Life (One Model) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Intercept1.800.36611.0822.517 24.14 <.0001 Provider-.058.0721-.198.083 .65 .421 Secure-.003.0091-.021.015 .10 .755 Avoidant.009.0101-.010.029 .88 .349 Anxious-.005.0101-.024.014 .28 .600 Black-.114.0881-.287.059 1.68 .165 Hispanic.003.1181-.228.233 .00 .982 Maternal Bonding Score-.078.0411-.158.002 3.62 .057 Table 29 Exploration of Potentially Moderating Va riables on the Number of Well Child Care Visits Attended the First Year of Life, Cont rolling for Interaction Styles and Provider Office Rating (Nine Separate Models: Black/Hi spanic run together as dummy variables) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Models (Model #) Black (1)-.035.1171-.264.194 .09 .064 Hispanic (1)-.158.0851-.326.009 3.42 .767 Target Child Mothers First (2) .107.0761-.041.255 1.99 .158 Childs Health Rating (3)-.005.0271-.058.048 .03 .855 Mothers Health Rating (4)-.021.0211-.063.020 1.04 .308 WIC/Healthy Start (5) -.099.1111-.317.119 .79 .375 Mother Worked > 30 Hours (6).012.0911-.164.188 .02 .892 Mothers Feelings (7)-.016.0151-.045.013 1.17 .280 Mothers Age (8)-.001.0061-.012.011 .01 .935 Mothers Bonding Score (9)-.088.0401-.166-.010 4.91 .027 Note p<.05

PAGE 152

129 Table 30 Exploration of Potentially Moderating Va riables on the Number of Well Child Care Visits Attended the First Year of Life (One Model) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Intercept1.731.37211.0022.460 21.67 <.0001 Provider Office-.010.0721-.152.133 .02 .893 Secure-.002.0091-.021.016 .06 .799 Avoidant.009.0101-.010.029 .86 .355 Anxious-.006.0101-.024.014 .32 .570 Black-.117.0881-.290.055 1.78 .183 Hispanic.006.1181-.224.225 .00 .962 Maternal Bonding Score-.076.0411-.156.005 3.41 .065 Examining the models related to sick and follow-up visits, controlling for provider rating and inter action style scores, foun d childs overall hea lth rating, (p<.01), and WIC/Healthy Start particip ation (P<.05) to be two poten tially moderating variables that were significantly correlated with visits (Table 31). When test ing interaction terms, the childs overall health rating did not have a significant interaction term with any of the interaction style scores. However, the main va riable of childs overall health status was included in the full model to control for confounding. When the remaining potential interactions were tested, f our factors acted as modera tors (Tables 31-A to 31-D; Appendix U.19 to Appendix U.27).

PAGE 153

130 Table 31 Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating (Nine Separate Main Effects Models: Black/ Hispanic run together as dummy variables) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Black-.136.0931-.318.047 2.12 .146 Hispanic.086.1271-.163.334 .46 .500 Target Child Mothers First.110.0821-.051.271 1.81 .179 Childs Health Rating-.070.0251-.118-.021 7.86 .005 ** Mothers Health Rating.042.0231-.004.008 3.19 .074 WIC/Healthy Start-.230.1151-.456-.004 3.97 .046 Mother Worked > 30 Hours.190.0971-.080.298 1.28 .257 Mothers Feelings-.007.0161-.038.024 .18 .667 Mothers Age.005.0061-.008.017 .50 .481 Mothers Bonding Score.001.0391-.076.077 .00 .989 Note p<.05, ** p<.01 Table 31-A Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of Birth Order (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept1.324.5541.2382.411 5.71 .017 Provider-.192.0821-.352-.032 5.50 .019 Secure-.101.0131-.036.016 .60 .439 Avoidant.007.0151-.023.036 .20 .657 Anxious.036.0161.005.067 5.04 .025 Target Child Mothers First.829.7711-.6832.34 1.15 .283 First Child*Secure.016.0211-.025.056 .58 .446 First Child*Anxious.028.0221-.015.072 1.61 .204 First Child*Avoidant-.064.0211-.104-.024 9.64 .002 ** Log Likelihood(Main Effect Model) 393.9019 Log Likelihood (Saturated Model) 399.7203 Difference in Log Likelihoods 5.8184 *2= 11.6@ Note @: Critical value for of .05=7.82; p<.05, ** p<.01

PAGE 154

131 Table 31-B Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of WIC/Healthy Start Participation (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept 5.551.26713.0718.032 19.24 <.0001 Provider -.187.0841-.351-.022 4.96 .026 Secure -.030.0251-.080.019 1.44 .230 Avoidant -.079.0311-.140-.017 6.30 .012 Anxious -.059.0391-.136.018 2.24 .135 WIC/Healthy Start -3.954.3291-6.560-1.349 8.85 .003 ** WIC/HS*Secure .0330.0281-.024.084 1.16 .281 WIC/HS*Anxious .116.0411.036.197 8.08 .005 ** WIC/HS*Avoidant .052.0331-.013.118 2.49 .115 Log Likelihood (Main Effect Model) 394.8837 Log Likelihood (Saturated Model) 400.4060 Difference in Log Likelihoods 5.5223 *2= 11.04@ Note @: Critical value for of .05=7.82; p<.05, ** p<.01 Table 31-C Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of Mothers Fee lings About Doctors (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Intercept1.398.9631-.4903.286 2.11 .147 Provider-.209.0841-.374-.043 6.12 .013 Secure.037.0251-.013.087 2.15 .142 Avoidant-.011.0231-.057.034 .24 .626 Anxious-.003.0231-.047.041 .02 .886 Mothers Feelings .091.1691-.239.422 .29 .588 Feeling*Secure-.071.0051-.016.002 2.50 .114 Feelings*Anxious.011.0041.003.019 6.83 .009 ** Feelings*Avoidant.004.0041-.012.003 1.28 .258 Log Likelihood(Main Effect Model) 393.0933 Log Likelihood (Saturated Model) 399.1433 Difference in Log Likelihoods 6.0500 *2= 12.10@ Note @: Critical value for of .05=7.82; p<.05, ** p<.01

PAGE 155

132 Table 31-D Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of Mate rnal Bonding (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept 1.852.39711.0742.63021.75 <.0001 Provider -.201.0831-.363-.0385.84 .016 Secure -.003.0101-.023.016.10 .752 Avoidant -.029.0101-.049-.0098.00 .005 ** Anxious .052.0111.030.07322.14 <.0001 *** Mothers Bonding Score .577.3741-.1571.3112.38 .123 Bonding*Secure -.016.0111-.038.0071.88 .170 Bonding*Anxious -.031.0111-.053-.0108.19 .004 ** Bonding*Avoidant .012.0101-.007.0311.43 .231 Log Likelihood (Main Effect Model) 393.0009 Log Likelihood (Saturated Model) 398.5198 Difference in Log Likelihoods 5.5189*2= 11.04@ Note @: Critical value for of .05=7.82; p<.05, ** p<.01, *** p<.001 Looking more closely at th e interaction, birth order moderated the association between avoidant interaction style scores and utilization of sick and follow-up visits. Mothers with other children t ook their target child for sick and follow up visits at relatively the same rate, regard less of their anxious interaction style score. In comparison, mothers where the target child was her first took the child to fewer sick-follow-up visits as their avoidant score increased (Figure 3).

PAGE 156

133 Predicted Sick and Follow-Up Visits by Avoidant Interaction Style and Whether Target Child was Mother's First Controlling for Provider Office and Interaction Styles (Secure, Anxious, Avoidant)0 2 4 6 85 7 9 11 13 15 1 7 19 21 23 2 5 27 29 31 33 35Avoidant Interaction ScorePredicted Visits First Child Not First Child Figure 3. Predicted sick and follo w-up visits by avoidant interaction style and whether target child was mothers first controlling for provider and interact ion styles (secure, anxious, avoidant) A moderator to the association betwee n anxious interaction style scores and utilization of sick and follow-up visits was pa rticipation in WIC or Healthy Start. Women who participated in WIC/Healthy Start are predicted to use fewer sick and follow-up visits with those having higher anxious intera ction scores being sli ghtly more likely to attend more visits (Figure 4). Women who di d not participate in WIC/Healthy Start are predicted to attend more sick and followup visits with those having higher anxious interaction scores being less lik ely to attend as many visits as those with lower anxious interaction scores.

PAGE 157

134 Predicted Sick and Follow-Up Visits by Anxious Interaction Style and WIC/Healthy Start Participation Controlling for Provider and Interaction Styles (Secure, Anxious, Avoidant)0 3 6 9 12 15 18 12345678910111213141516171819202122232425 Anxious Interaction ScorePredicted Visits No WIC WIC Figure 4. Predicted sick and follow-up visi ts by anxious interaction style and WIC/Healthy Start participation c ontrolling for provide r and interaction styles (secure, anxious, avoidant) A mothers general feelings about going to doctors also acted as a moderator to the association between anxious interactions style scores and utilization of sick and follow-up visits. For women with increasingly anxious interac tions style scores, the more a mother expressed negative feelings about going to the doctor (i.e., higher feeling scores) the more sick and follow-up visits th e child was predicted to attend (Figure 5).

PAGE 158

135 Predicted Sick and Follow-Up Visits b y Anxious Interaction St y le and Feelings About Going to the Doctor Controlling for Provider and Interaction Styles (Secure, Anxious, Avoidant)0 20 40 60 80 100 120 140 12345678910111213141516171819202122232425 Anxious Interaction ScorePredicted Visits 2 4 6 8 10 Figure 5 Predicted sick and follow-up visi ts by anxious interaction style and feelings about going to the doctor controlling for provider and interaction styles (secu re, anxious, avoidant) The last moderator to the association of anxious interaction styl e scores with sick and follow-up visit utilization was the level of maternal bonding. It is predicted that mothers with lower anxious interaction sc ores and stronger bonding scores (lower bonding score) take their childre n to fewer sick and follow-up visits (Figure 6). However, mothers with high anxious interaction scores and strong bonds are more likely to take their children to greater number s of sick and follow-up visits.

PAGE 159

136 Predicted Sick and Follow-Up Visits by Anxious Interaction Style and Bonding Issues Controlling for Provider and Interaction Styles (Secure, Anxious, Avoidant)0 1 2 3 4 5 6 7 8 9 10 11 1234567891011121314151617181920212223 Anxious Interaction ScorePredicted Visits -1 0 1 2 3 Figure 6 Predicted sick and follow-up visi ts by anxious interaction style and bonding issues controlling for provi der and interaction styles (secure, anxious, avoidant) Adding the confounding and moderating variab les as well as the interaction terms resulted in a model containing 21 variables. Provider rating (p<.05), anxious interaction style score (p<.01), feelings about doctors (p<.05), and WI C/Healthy Start participation (p<.0001) were main effects (Table 32). In a ddition, the moderating effects of birth order (p<.05) on avoidant interacti on style scores with sick and follow-up visits were maintained. The moderating effects of WIC or Healthy Start participation (p<.01) on associations of anxious intera ction style scores with sick and follow-up visit utilization

PAGE 160

137 were maintained. Additional moderation of avoida nt interaction style sc ores with sick and follow-up visits by WIC/Healthy Start partic ipation (p<.01) also were identified. Maternal feelings about going to the doctor no longer modera ted the effect of anxious interaction styles scores on sick and follow-up visits but moderation of secure (p<.01) and avoidant (p<.05) did become significant. Finally, the moderating effects of maternal bonding (p<.01) on associations of anxious interaction style were maintained. Table 32 Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life (One Model) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept 4.3361.61111.1787.493 7.24 .007 Provider -.197.0901-.377-.017 4.59 .032 Secure .034.0461-.060.121 .43 .512 Avoidant -.033.0411-.113.048 .63 .427 Anxious -.106.0391-.182-.030 7.51 .006 ** Childs Health Rating -.062.0271-.116-.009 5.20 .023 First Child 1.052.8791-.6712.774 1.43 .232 WIC/Healthy Start -5.2381.4601-8.100-2.377 12.88 .000 *** Feelings About Doctors .461.2011.066.856 5.24 .022 Bonding .672.4201-.1511.494 2.56 .110 First Child Secure -.103.0241-.057.036 .19 .665 First Child *Anxious .042.0251.007.090 2.85 .091 First Child *Avoidant -.057.0231-.102-.012 6.08 .014 WIC/HS Secure .050.0321-.013.113 2.45 .118 WIC/HS Anxious .093.0411.013.173 5.25 .022 WIC/HS Avoidant .098.0381.024.172 6.70 .010 Bonding Secure -.017.0131-.043.008 1.83 .177 Bonding Anxious -.031.0111-.052-.010 8.54 .004 ** Bonding Avoidant .010.0111-.012.032 .77 .379 Feelings Secure -.014.0051-.024-.003 6.34 .012 Feelings Anxious .009.0041-.000.017 3.66 .056 Feelings Avoidant -.011.0051-.021-.002 5.46 .020 Note Goodness of Fit: Pearson Chi-Square (value/df) 2.71; p<.05, ** p<.01, *** p<.001

PAGE 161

138 The original sample size was calculated for a maximum of thirteen variables in the full model. Therefore, although the result s of the analyses indicate significant associations, there is not enough statistical power to say that these associations occurred for reasons other than rando m chance (Table 32). Similar findings resulted from the modeli ng of the provider office with interaction style ratings and potential moderators (Tables 33 to 34; Appendix U.28 to Appendix U.40). Again, there were too many variable s for the model given the sample size. Table 33 Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Office Rating (Nine Separate Main Effects Models: Black/Hispanic combined dummy variables) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Black-.142.0931-.325.041 2.30 .129 Hispanic.057.1261-.191.304 .20 .653 Target Child Mothers 1st.114.0821-.047.274 1.93 .164 Childs Overall Health Rating-.075.0251-.124-.026 8.99 .003 ** Mothers Overall Health Rating.040.0241-.006.086 2.91 .088 WIC/Healthy Start Participation-.216.1161-.443.011 3.49 .062 Mother Worked > 30 Hrs/Week.138.0961-.050.325 2.06 .151 Mothers Feelings of Doctors-.012.0161-.044.020 .51 .474 Mothers Age.007.0061-.006.019 1.19 .275 Mothers Bonding Score.005.0401-.072.083 .02 .891 Note ** p<.01

PAGE 162

139 Table 33-A Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Office Rating: Interaction of Birth Order (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept 1.637.5151.6272.647 10.09 .002 Provider Office .001.0791-.153.155 .00 .989 Secure -.012.0131-.037.014 .80 .372 Avoidant .004.0151-.025.034 .09 .766 Anxious .033.0161.002.064 4.27 .039 Target Child Mothers 1st .510.7551-.9691.989 .46 .499 First Child*Secure .025.0201-.015.065 1.48 .224 First Child*Anxious .032.0221-.021.076 2.02 .156 First Child*Avoidant -.061.0211-.101-.021 8.79 .003 ** Log Likelihood(Main Effect Model) 390.8215 Log Likelihood (Saturated Model) 396.8498 Difference in Log Likelihoods 6.0283 *2= 12.05@ Note @: Critical value for of .05=7.82; p<.05, ** p<.01 Table 33-B Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Office Rating: Interaction of WIC/Hea lthy Start Participation (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept 2.3471.1301.1334.561 4.32 .038 Provider Office -.041.0831-.203.122 .24 .624 Secure -.032.0261-.082.018 1.55 .214 Avoidant -.080.0321-.142-.018 6.46 .011 Anxious -.068.0391-.144.008 3.06 .080 WIC/Healthy Start -4.2081.321-6.760-1.657 10.45 .001 ** WIC/HS*Secure .034.0281-.021.089 1.48 .224 WIC/HS*Anxious .126.0411.046.205 9.57 .002 ** WIC/HS*Avoidant .053.0331-.012.119 2.54 .111 Log Likelihood (Main Effect Model) 391.5158 Log Likelihood (Saturated Model) 397.9420 Difference in Log Likelihoods 6.4262 *2= 12.85@ Note @: Critical value for of .05=7.82; p<.05, ** p<.01

PAGE 163

140 Table 33-C Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Office Rating: Interaction of Mothers Feelings About Doctors (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept 2.327.8691.6254.030 7.18 .007 Provider Office -.019.0841-.183.146 .05 .825 Secure .028.0261-.022.079 1.23 .268 Avoidant -.011.0231-.057.035 .23 .631 Anxious -.001.0221-.054.034 .19 .662 Mothers Feelings .025.1681-.303.354 .02 .880 Feeling*Secure -.005.0051-.014.004 1.25 .264 Feelings*Anxious .012.0041.004.020 8.70 .003 ** Feelings*Avoidant -.005.0041-.012.003 1.42 .233 Log Likelihood (Main Effect Model) 390.1139 Log Likelihood (Saturated Model) 395.9973 Difference in Log Likelihoods 5.8783 *2= 11.76@ Note @: Critical value for of .05=7.82; ** p<.01 Table 33-D Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Office Rating: Interaction of Ma ternal Bonding (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept 1.808.37911.0652.55222.72 <.0001 Provider Office -.011.0801-.169.146.02 .889 Secure -.001.0101-.020.019.01 .926 Avoidant -.029.0101-.049-.0098.20 .004 ** Anxious .051.0101.029.07221.13 <.0001 *** Mothers Bonding Score .524.3741-.2081.261.97 .1604 Bonding*Secure -.012.0111-.034.0101.19 .275 Bonding*Anxious -.034.0111-.055-.0129.19 .002 ** Bonding*Avoidant .013.0101-.007.0321.65 .199 Log Likelihood(Main Effect Model) 389.8659 Log Likelihood (Saturated Model)395.4847 Difference in Log Likelihoods 5.6188*2= 11.24@ Note @: Critical value for of .05=7.82; p<.05, ** p<.01, *** p<.001

PAGE 164

141 Table 34 Exploration of Potentially Moderating Vari ables on the Number of Sick/Follow-up Visits Attended the First Year of Life (One Model) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Intercept 4.8921.56611.8227.9619.75 .002 Provider Office -.038.0931-.219.144.17 .684 Secure .007.0461-.083.098.02 .875 Avoidant -.032.0411-.113.049.59 .444 Anxious -.119.0381-.194-0449.73 .002 ** Childs Health Rating -.071.0271-.124-.0186.86 .009 ** First Child .681.8581-1.0012.363.63 .427 WIC/Healthy Start -5.4901.4411-8.315-2.66614.52 .000 *** Feelings About Doctors .394.2011.000.7883.84 .050 Bonding .649.4161-.1661.4642.44 .189 First Child*Secure .003.0231-.042.048.01 .906 First Child*Anxious .040.0251.008.0882.69 .101 First Child*Avoidant -.052.0231-.097-.0085.22 .022 WIC/HS Secure .059.0321-.004.1213.39 .066 WIC/HS Anxious .104.0401.025.1836.63 .010 WIC/HS Avoidant .097.0381.019.1626.00 .014 Bonding Secure -.016.0131-.041.0091.59 .207 Bonding Anxious -.033.0111-.054-.1179.36 .002 ** Bonding Avoidant .011.0111-.012.033.87 .350 Feelings Secure -.011.0051-.022-0004.27 .039 Feelings Anxious .009.0041.000.0174.07 .044 Feelings Avoidant -.011.0051-.021-.0025.30 .021 Note Goodness of Fit: Pearson Chi-Square (value/df) 2.73; p<.05, ** p<.01, *** p<.001 Examining the models related to emergency department visits, controlling for provider rating and interaction style scores, mothers age (p<.05), and maternal bonding score (p<.05) were identifi ed as two potentially mode rating variables that were significantly correlated with visits (Tab le 35). Maternal bo nding did not have a significant interaction with any of the in teraction style scores. However, the bonding variable was included in the full model to control for confounding.

PAGE 165

142 Table 35 Exploration of Potentially Moderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating (Nine Separate Models: Bl ack/Hispanic as dummy variables) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Black.096.2801-.453.645 .12 .732 Hispanic.015.4071-.783.813 .00 .971 Target Child Mothers First.222.2511-.271.714 .78 .377 Childs Overall Health Rating-.130.0701-.269.008 3.41 .065 Mothers Overall Health Rating.029.0721-.112.170 .17 .683 WIC/Healthy Start Participation.299.4311-.5451.143 .48 .487 Mother Worked > 30 Hrs/Week.088.2971-.495.670 .09 .768 Mothers Feelings of Doctors.063.0481-.032.157 1.68 .195 Mothers Age-.046.0221-.089-.004 4.52 .034 Mothers Bonding Score.227.1071-.018.436 4.52 .034 Note p<.05 When the remaining potential interactions were tested, the childs overall health status and maternal age acte d as potential moderators to the associations between provider ratings and interacti on style scores with the numbe r of emergency department visits attended (Tables 35-A and 35B; Appendix U.41 to Appendix U.49).

PAGE 166

143 Table 35-A Exploration of Potentially Moderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of Childs Health Issues (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept -15.3417.7811-30.592-.090 3.89 .049 Provider -.109.2411-.582.363 .21 .650 Secure .692.2451.2111.173 7.94 .005 ** Avoidant -.069.2281-.515.378 .09 .764 Anxious .164.1581-.146.474 1.08 .299 Childs Health Rating 1.642.8571-.0363.321 3.68 .055 Childs Health*Secure -.078.0271-.130-.025 8.34 .004 ** Childs Health*Anxious -.017.0181-.052.018 .94 .333 Childs Health*Avoidant .008.0251-.041.056 .10 .753 Log Likelihood (Main Effect Model) -106.5343 Log Likelihood (Saturated Model) -101.0785 Difference in Log Likelihoods 5.4558 *2= 10.91@ Note @: Critical value for of .05=7.82; p<.05, ** p<.01 Table 35-B Exploration of Potentially Moderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interacti on of Mothers Age (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept -9.8104.9541-19.520-.100 3.92 .048 Provider -.097.2341-.556.362 .17 .679 Secure .347.1381.077.617 6.36 .012 Avoidant .038.1531-.262.337 .06 .806 Anxious .157.1421-.122.436 1.22 .269 Mothers Age .372.1961-.013.756 3.59 .058 Age*Secure -.014.0051-.024-.004 7.91 .005 ** Age*Anxious -.006.0061-.017.005 1.30 .254 Age*Avoidant -.001.0061-.013.011 .03 .857 Log Likelihood(Main Effect Model) -105.5969 Log Likelihood (Saturated Model) -101.0117 Difference in Log Likelihoods 4.5852 *2= 9.17@ Note @: Critical value for of .05=7.82; p<.05, ** p<.01

PAGE 167

144 Interaction terms indicated that, alth ough women with higher secure interaction scores were more likely to take their childre n to the emergency room, this was less true for children with better (higher) overall health ratings (Figure 7). Si milarly, women with higher secure interaction scores who were olde r were less likely to take their children to the emergency department than comparab le women who were younger (Figure 8). Predicted Emer g enc y Department Visits b y Secure Interaction Style and Child Health Rating Controlling for Provider and Interaction Styles (Secure, Anxious, Avoidant)0 5 10 15 20 121314151617181920212223242526 Secure Interaction ScorePredicted Visits 2 4 6 8 10 Figure 7. Predicted emergency department vi sits by secure interaction style and child health rating controlling for provider and in teraction styles (secure, anxious, avoidant)

PAGE 168

145 0 100 200 300 400 500 15182124273033 Secure Interaction ScorePredicted Visits 18 24 30 36 45Predicted Emergency Department Visits by Secure Interaction Style and Mother's Age Controlling for Provider Office and Interaction Styles (Secure,Anxious,Avoidant) Figure 8. Predicted emergency department visits by secure interaction style and mothers age controlling for provider and interaction styles (secu re, anxious, avoidant) Adding confounders and moderators with their interaction terms resulted in a model with thirteen variables (Table 36). Us ing this model, women with higher secure interaction style scores were more likely to take their child to the emergency department (p<.01; OR=2.16). However, this association is moderated by age with younger mothers being more likely to take their child to the emergency department (p<.05; OR=.99). Similar results were found when incorporat ing the relationship of the entire providers office rather than the individual provid er (Tables 37-38; Appendix U.46 to U.56).

PAGE 169

146 Table 36 Exploration of Potentially Moderating Variables on the Number of Emergency Department Visits Attended the First Year of Life (One Model) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept -18.6019.1511-36.536-.665 4.13 .042 Provider -.110.2431-.587.367 .20 .651 Secure .771.2631.2561.28 7 8.59 .003 ** Avoidant .026.2881-.538.590 .01 .929 Anxious .165.1841-.195.525 .81 .369 Maternal Age .304.1981-.084.693 2.36 .124 Maternal Bonding .169.1151-.056.395 2.17 .141 Childs Health Status 1.181.9241-.6292.99 2 1.64 .201 Childs Health*Secure -.054.0291-.112.003 3.40 .065 Childs Health*Anxious -.013.0191-.050.024 .47 .494 Childs Health*Avoidant .005.0261-.045.056 .05 .832 Maternal Age*Secure -.012.0061-.023-.001 4.39 .036 Maternal Age*Anxious -.002.0061-.015.010 .14 .710 Maternal Age*Avoidant -.003.0061-.015.009 .21 .647 Note p<.05, ** p<.01 Table 37 Exploration of Potentially Moderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Office Rating (Nine Separate Mode ls: Black/Hispanic as dummy variables) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Black.104.2811-.448.656 .14 .712 Hispanic-.013.4061-.810.783 .00 .974 Target Child Mothers 1st.226.2511-.266.718 .81 .368 Childs Overall Health Rating-.139.0711-.278.000 3.84 .050 Mothers Overall Health Rating.033.0731-.110.175 .20 .654 WIC/Healthy Start Participation.139.4301-.5261.164 .55 .459 Mother Worked > 30 Hrs/Week.096.2941.-.480.673 .11 .744 Mothers Feelings of Doctors.053.0501-.045.150 1.13 .288 Mothers Age-.044.0221-.087-.001 3.96 .047 Mothers Bonding Score.247.1091.033.461 5.13 .024 Note p<.05

PAGE 170

147 Table 37-A Exploration of Potentially Moderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interaction of Childs Health Issues (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept -14.7707.6351-29.735.195 3.74 .053 Provider Office .153.2431-.0324.630 .39 .530 Secure .636.2361.1741.10 7.28 .007 ** Avoidant -.077.2291-.525.371 .11 .738 Anxious .193.1661-.132.518 1.36 .244 Childs Health Rating 1.535.8411-.1133.184 3.33 .068 Childs Health*Secure -.071.0261-.122-.021 7.64 .006 ** Childs Health*Anxious -.021.0191-.057.016 1.24 .266 Childs Health*Avoidant .009.0251-.040.058 .13 .722 Log Likelihood(Main Effect Model) -105.9743 Log Likelihood (Saturated Model) -100.9903 Difference in Log Likelihoods 4.9840 *2= 9.97@ Note @: Critical value for of .05=7.82. Table 37-B Exploration of Potentially Moderating Variables on the Number of Emergency Department Visits Attended the First Year of Life, Controlling for Interaction Styles and Provider Rating: Interacti on of Mothers Age (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Intercept -11.0625.0571-20.974-1.151 4.79 .029 Provider Office .216.2351-.245.676 .84 .359 Secure .359.1391.086.631 6.66 .001 ** Avoidant .044.1541-.259.346 .08 .778 Anxious .179.1451-.105.464 1.53 .216 Mothers Age .401.1991-.011.791 4.07 .044 Age*Secure -.015.0051-.025-.005 8.19 .004 ** Age*Anxious -.001.0061-.018.004 1.63 .202 Age*Avoidant -.001.0061-.014.011 .05 .830 Log Likelihood(Main Effect Model) -105.5343 Log Likelihood (Saturated Model) -100.6893 Difference in Log Likelihoods4.8450 *2= 9.69@ Note @: Critical value for of .05=7.82; p<.05, ** p<.01

PAGE 171

148 Table 38 Exploration of Potentially Moderating Variables on the Number of Emergency Department Visits Attended the First Year of Life (One Model) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Intercept -18.5978.9891-36.215-.978 4.28 .039 Provider Office .257.2551-.242.756 1.02 .313 Secure .709.2531.2141.204 7.88 .005 ** Avoidant .026.2871-.537.590 .01 .928 Anxious .214.1911-.160.587 1.26 .262 Maternal Age .354.2041-.046.754 3.02 .083 Maternal Bonding .200.1171-.028.429 2.94 .086 Childs Health Status .78.9061-.7802.755 1.17 .280 Childs Health*Secure -.044.0291-.100.012 2.37 .124 Childs Health*Anxious -.017.0201-.055.022 .71 .398 Childs Health*Avoidant .007.0251-.043.057 .07 .787 Maternal Age*Secure -.013.0061-.024-.002 4.99 .025 Maternal Age*Anxious -.003.0061-.016.009 .25 .614 Maternal Age*Avoidant -.003.0061-.015.009 .27 .601 Note p<.05, ** p<.01 Examining the models related to immuni zations, controlling for provider rating and interaction style scores, no other variables were significantly correlated with visits (Table 39). There were also no significant interactions among a ny of the potentially moderating variables and the number of well immunizations (Appendix U.55 to Appendix U.63). Similarly, the saturated mode l had no significant correlations (Table 40). Results for the provider office main eff ects and saturated mode ls were also not significant (Table 41 and 42; Appendix U.64 to U.72).

PAGE 172

149 Table 39 Controlling for Interaction Styles and Provid er, Exploration of Po tentially Moderating Variables on the Number of Immunizations Atte nded the First Year of Life (nine separate models) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Black-.001.0601-.117.116 .00 .991 Hispanic.006.0831-.157.168 .00 .945 Target Child Mothers 1st.040.0531-.064.145 .57 .452 Childs Overall Health Rating-002.0191-.039.035 .01 .926 Mothers Overall Health Rating-.005.0151-.034.024 .11 .745 WIC/Healthy Start Participation.024.0831-.139.187 .08 .772 Mother Worked > 30 Hrs/Week-.082.0651-.209.045 1.60 .207 Mothers Feelings of Doctors-.001.0101-.022.019 .02 .891 Mothers Age-.003.0041-.011.005 .53 .465 Mothers Bonding Score-.001.0271-060.044 .09 .764 Table 40 Immunizations, Interaction Styles and Provider Ratings (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Intercept2.412.25811.9062.918 87.35 <.0001 Provider .009.0491-.087.105 .04 .850 Secure.002.0071-.011.015 .11 .736 Avoidant-.000.0071-.013.013 .00 .948 Anxious.002.0071-.012.015 .07 .793

PAGE 173

150 Table 41 Controlling for Interaction Styles and Prov ider Office, Exploration of Potentially Moderating Variables on the Number of Immuni zations Visits Attende d the First Year of Life (Nine separate models) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Black.001.0601-.116.118 .00 .992 Hispanic.005.0831-.158.168 .00 .951 Target Child Mothers 1st-.040.0531-.064.144 .56 .453 Childs Overall Health Rating-.002.0191-.039.035 .01 .924 Mothers Overall Health Rating-.005.0151-.034.025 .10 .758 WIC/Healthy Start Participation.026.0831-.138.189 .09 .759 Mother Worked > 30 Hrs/Week-.082.0641-.208.044 1.63 .202 Mothers Feelings of Doctors-.082.0111-.023.019 .03 .857 Mothers Age-.003.0041-.011.005 .51 .477 Mothers Bonding Score-.001.0271-.060.045 .08 .782 Table 42 Immunizations, Interaction Styles, an d Provider Office Ratings (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Intercept2.3998.26511.8972.917 82.05 <.0001 Provider Office.015.0511-.085.115 .09 .769 Secure.002.0071-.011.015 .11 .741 Avoidant-.000.0071-.013.013 .00 .973 Anxious.002.0071-.012.015 .07 .785

PAGE 174

151 Chapter Five Synthesis of Research Findings The results of this study illustrate the f act that there is still much that we do not know about what motivates individuals to utiliz e health care. What is known is that the relationships among maternal interaction style, reported experiences of care and utilization of health care services are multif aceted. Furthermore, significant factors, such as maternal bonding and feelings about going to the doctor, indicate that the way in which we seek to develop that additional knowledge needs to be expanded. This study has provided insight into these relations hips while also leaving many questions unanswered. The chapter begins by looking at the findings and imp lications regarding utilization of health care serv ices, experience of care ratings, maternal interaction style, additional study findings and limitations of th e study. Finally, implications for action and recommendations for future research are outlined. Utilization of Health Care Services Findings regarding the utilization of pediat ric health care services were mixed. As a result, discussions regarding services will be separated by service type.

PAGE 175

152 Utilization of Well Child Care Visits Finding/Implication Few differences existed among infants rega rding well child care visits. There are a number of potential reasons for this lack of diversity in utilization rates. There are a variety of social pressures, such as from doctors, family members, service providers (WIC/Healthy Start), and governme ntal regulations regarding receipt of assistance that encourages utilization. For example, WIC counselors ask mothers to provide immunization records during renewal or receipt of services. There is also a ceiling effect for the eight recommended well child care visits that limits variability in the data. Data recording issues may have also been i nvolved (see Appendix V). Finally, the limited distribution of the data due to the ceiling eff ect may require the use of a truncated Poisson regression analysis, the calcula tion of which is not availabl e within existing statistical software. For a more detailed discussion of the issue, see Appendix W. Utilization of Sick and Follow-up Visits Finding/Implication The utilization of pediatric sick and fo llow-up visits had a va riety of significant findings. Ratings of providers increased and higher child health ratings decreased the predicted number of visits. Furthermore, in addition to WIC/Healthy Start participation being a significant factor, there was a signifi cant interaction between WIC/Healthy Start participation and women with higher anxious interaction. These findings indicate that, based on the interaction style of the part icipating women, support programs working with women can have a different impact in helping to reduce the number of sick and follow-up visits the child utilizes. It may be that the WIC appointments and Healthy Start care coordination provides the anticipatory gu idance and education needed without going

PAGE 176

153 to a health care facility. It could also be that the nutriti onal supplements provided through WIC are helping the child to maintain a better health status, requiri ng fewer sick visits. This finding is supported by previous studies that have demonstrated WIC participation decreased the likelihood that th e infant would experience iron deficiencies and increased the likelihood that the infant experienced adequate weight gain (Altucher, Rasmussen, Barden & Habicht, 2005; Black, et al., 2004; Owen & Owen, 1997). To further understand the impact of WIC, a study conducted by Black et al., 2004, divided the non-WIC users into two groups, those who reportedly did not need WIC services and those who reported barriers to accessing services. The finding of improved health outcomes were betw een the non-WIC participants facing barriers and WIC participants. Exploring these differences in the current study was beyond the scope of the data collected. However, there were moth ers who reported not needing WIC, mothers who used WIC but only for a fe w months, and mothers who us ed WIC services for a full year. Therefore, it is possibl e that separating out mothers w ho did not believe they needed WIC from those who faced barriers, the influence of WIC and/or Healthy Start participation would be even greater than iden tified. Understanding th is differential effect can help guide enhancements to these progr ams to address the unique needs of mothers with varying interaction st yles, thereby expanding the impact of such programs. The moderating effects of birth order on si ck and follow-up visits indicates that utilization patterns of first-time mothers with high avoidant intera ction style scores is lower than for first-time mothers with lo wer avoidant interacti on scores and mothers with older children. This may result from mothers having multiple children developing a better understanding regarding when to take their children to a health care provider.

PAGE 177

154 Anticipatory guidance for first-time mothers, especially those having high avoidant interaction scores, can help encourage mother s to bring their children in to health care providers when they are sick. The interaction between a mo thers feelings about doctors and interaction style is unclear. This factor was not si gnificant in main effects models, while interaction models indicated a moderating effect of feelings about doctors on anxious interaction style scores. However, the full model including all interaction terms identified significant negative associations of feelings about docto rs with avoidant and secure interaction scores but not with anxious scores. Although th e sample size for the full model is less than adequate, the presence of these associations indicates a need for further exploration of the relationships. The last moderating factor associated w ith sick and follow-up visits is maternal bonding. Women with more positive bonding and lo wer anxious interaction scores take their children to less visits. Women with more bonding issues and women with higher anxious interaction scores take their children to more visits. Th is variability in utilization patterns may indicate different levels of attent iveness to the childs needs. For example, Baydar (1995) found that children under the age of two who were mistimed and unwanted children, a issue includ ed in the bonding factor used in this study, received fewer resources and learning opportunities than children who were planned. This deprivation of developmental opportunities limited skills de velopment. Being aware of potential bonding problems can help promote early intervention. For more discussions regarding maternal bonding a nd feelings about going to the doctor, see Appendix X.

PAGE 178

155 Utilization of Emergency Department Visits Finding/Implication Mothers with higher secure interaction st yle scores took their children to more emergency department visits (B=.771, p= .003) (Table 36). A lthough it was hypothesized that anxious women would take their child to more emer gency department visits, a hindsight perspective of the at tachment theory constructs su ggests that anxious mothers would take their child to th e provider office more quickl y, thereby averting delay in health care until the point where emergency care is needed. There was a significant interaction between maternal age and secure interaction style score. The higher utilization rate s for younger mothers makes intuitive sense because those mothers are also less likely to have other children from which they would have learned how to provide health care th emselves. Another significant interaction term was between the childs overall health rating an d secure interaction style. However, when placed in the full model including maternal age, this interaction term was no longer statistically significant. Add itionally, bivariate associations between maternal age and whether the target child was the mothers fi rst child were significant (Table 23). Both factors have a certain level of inexperience wi th children associated with them. It is not difficult to understand that those with less experience would use more care and be more likely to take their child to the emergenc y department. Study findings supporting this association have been found in the literatur e previously. Lee, Friedman, Ross-Degnan, Hibberd & Goldman, (2003) found that younger pa rents (<30) were more than ten times as likely (OR:10.0 [1.6-64.3]) to use additi onal ambulatory care than older women. Data quality issues regarding the use and reporting of emergency department visits are similar to those discussed for the well child care visits. This is especially true

PAGE 179

156 when it comes to having emergency department records being forwarded to pediatricians. If the mother is not clear about whom th e information should be forwarded, which is sometimes intentional, the primary care pr ovider may never know of the visit. For example, some parents take their children to different emerge ncy departments and different pediatricians to avoi d detection of maltreatment issues that may be occurring (Friedlaender, Rubin, Alpern, Mande ll, Christian & Alessandrini, 2005). Utilization of Immunizations Finding/Implication Immunizations are important because when enough individuals within a community have been immunized against a part icular illness, the likelihood that an epidemic of that disease will break out in the community is small (Lilienfeld & Stolley, 1994). The proportion of a population needing this herd immunity to protect the community varies for each di sease. For example, it has b een suggested that 70% of a population were immunized against measles or 80-85% of the population for rubella (Lilienfeld & Stolley, 1994). The impact of immunization programs can be seen in the relatively low number of individuals reported to have acquired the disease. Duri ng the 2004 calendar year, the Florida Department of Health (2004) reported seven cases of the mumps, one case of measles, no cases of rubella and 90 cases of pe rtussis. Hepatitis B is a disease for which immunization strategies were developed more recently, resulting in a lower level of herd immunity. During the same 2004 calendar year, Florida reported 501 cases of acute and 824 cases of chronic Hepatitis B.

PAGE 180

157 Emphasizing the use of vaccinations to ma intain the health of individuals and communities is an ongoing challenge. Currently, th e system is addressing this issue in a variety of ways from outreach programs to combining multiple vaccines into one shot. The results of previous and ex isting immunization effo rts can be seen in the proportion of children in Florida who are up to date on their immunizations According to the Centers for Disease Control (2005), 81% of Florida s children and 79% of the nations children 19-35 months of age were up to da te on their immunizations in 2003. This study found no significant differen ces regarding the specific research hypotheses measuring associations with immu nizations due, in part to most infants receiving the recommended vaccinations (p<.01 ). One potential reason for the lack of findings is the use of Catch-Up strategies by providers for mothers who have delayed immunizing their children. Many children recei ved their 13 vaccinations within four to six different visits. However, if the mother had not initiated immunizations within the first few months of life, data indicated that the provider gave the child all the recommended vaccinations (DTaP, Polio, Pneumococcal, HepB, and HIB) at each visit until they were up to date. This Catch-Up process indicates both the attentiveness of the providers to the n eeds of the child as well as a need to better understand the different vaccination patterns of infants. This lack of differences may also have resulted from a selection bias where mothers who did not ge t their child immunized did not volunteer for the study. Further exploration of these patterns with in this study is limited for two primary reasons. First, although the actual dates of i mmunizations were coll ected, it was not the

PAGE 181

158 primary focus of this study. Second, not all of the records contained exact dates for each event. For a more detailed discussi on of the issue, see Appendix X. Experience of Care Ratings The original intent of the reported expe riences of care questio ns was to respond to the conceptual shift from researching satisfac tion of care to that of anchoring questions on specific elements of care (Institute of Me dicine of the National Academies, 2004a). The failure to reject the nu ll hypothesis for Research Question 1 may be a testament to the success of the anchori ng process in removing subjectivity from the provider and provider office ratings. Other patterns of fi ndings regarding the experiences of care ratings for providers and their offices were different and are discussed separately. Provider Ratings of Experience s of Care Finding/Implication The study found that more positive ratings of providers incr eased the likelihood that mothers brought their children in for si ck and follow-up visits. Additionally, mothers told interviewers that they wanted to be involved in the decision-making process for treatment and appreciated providers w ho took the time to listen, provided thorough explanations and provided anticipatory guida nce. Mothers reportedly did not respond well to providers who told them what to do or spoke to them in a scolding manner. Interestingly, the studys selection biases re sulted in this populati on having more positive ratings of providers and yet statistical differe nces in the use of care were still identified. This may be an indication that the differen ce in health care utilization patterns may be even greater in more generalizable populations.

PAGE 182

159 These findings indicate that improving mothers experiences of care has the potential to change the way in which they ut ilize pediatric health care for their infants. The challenge is to identify approaches and techniques that improve maternal experiences that do not also place an exceptional bur den on providers. For additional discussion regarding this issue and comparison with national CAHPS data, see Appendix Y. Provider Office Ratings of Experi ences of Care Finding/Implication As with the provider rati ngs, responses to the indi vidual questions regarding reported experiences of care with the provide r office indicated that mothers included in this study generally had more positive rati ngs than those found in the national CAHPS data set. Findings of this study did not id entify significant associations between provider office ratings and health care utilization ra tes but did find a high correlation between provider office rating and provider rating, a factor that was associated with health care utilization. This strong correla tion indicates that improvements in the provider office ratings has the potential to influence a mothe rs overall experiences of care. For a more detailed comparison of Study and na tional CAHPS data, see Appendix Y. Maternal Interaction Style Finding/Implication The proportion of individuals having specifi c dominant interacti on styles within a population is relevant when providing s upports that meet individual needs.

PAGE 183

160 High Anxious Interaction Scores The proportion of respondents identified with a dominant anxious interaction style was lower than that found in population-ba sed studies (Mickelson, Kessler, & Shaver, 1997). One possible reason for the lower anxiou s ratings is the restriction of women by spouses, boyfriends, and family members not to participate in the study. According to attachment theory, women with high anxious interaction scores would seek closer proximity to persons of authority. This can lead to becoming involved with partners who are more controlling. In the case of this study, several women who expressed interest in the study and scheduled appointments canceled their appointments later citing familial restrictions. One woman even sounded somewh at fearful about a researcher coming by her home. High Avoidant Interaction Scores The proportion of women with high avoidant scores in th is study is larger (57.1%) than the national population-based estimat es (25%) (Mickelson, Kessler, & Shaver, 1997). There was also a significant positive association between Black non-Hispanic race/ethnicity and avoidant intera ction scores (p<.01) (Table 21). Even if one takes into consideration the potential self-selection bias of participants in this study, a dom inant avoidant interaction styl e rate more than twice that of other populations is important to consider because the context from which individuals operate. Furthermore, the patterns of behavior they exhibit can greatly differ among those with various dominant intera ction styles. For example, in a different study of 193 unmarried undergraduate college students, av oidant respondents re ported fewer and less

PAGE 184

161 intense love relationships while anxious re spondents reported more frequent but less enduring love relationships. In comparison, indi viduals with dominant secure interaction styles reported more loving a nd satisfying relationships. A dditionally, participants with dominant avoidant interaction styles were mo re likely to engage in casual (uncommitted) sex than individuals with other dominant interaction styles (F eeney, Noller & Patty, 1993). It should be noted that a subset of this study (N=85) particip ating in a subsequent diary-writing exercise also found that avoi dant females and anxious males were least likely to report having sex during the study period. Although the sample size was small, it is an indication that more research needs to be done to explore differences in the influence of interaction style on behavior. Impact of Differing Interaction Scores The impact of interaction style on individual behaviors extends beyond the respondent to individuals around them. More sp ecifically, these behaviors can have an influence on the childs life as well. For example, an internet survey of 5,000 predominantly White (77.7% [Black 6%, Hi spanic 4%]) respondents found that high avoidant interaction styles we re associated with high leve ls of avoidance of former partners (Davis, Shaver & Vernon, 2003). Re searchers found some evidence that those with avoidant interaction st yles could successfully suppres s the stress related to the breakup if they could avoid direct reminde rs of the relationship (Davis, Shaver & Vernon, 2003; Fraley et al., 1998; Fraley & Shaver, 1997). This avoidant behavior is also signifi cant when children are involved because of the benefits of having the fath er being involved in the child s life. Mothers control much

PAGE 185

162 of the access to children when the father is not in the home. It is common for mothers, especially when the father was never married to the mother, to restrict access to the child with the rationale being that the father has not paid enough ch ild support or that he would have a negative effect on the child. By in creasing the level of understanding of where these negative feelings towards the father of the child are coming from may allow service providers (pediatricians, WIC/H ealthy Start staff, etc) to help mothers resolve some of these barriers so that father-ch ild relationships can be improved. Maternal interaction style can also have an influence on her child(ren)s health outcomes and feelings about health care providers. Edelstein, et al. (2004), studied childrens reactions to inoculat ions based on parental intera ction style. Researchers found that children having parents with high avoidant interaction scores we re more distressed during inoculations than parents with low avoidant interaction scores. Additionally, parents with high avoidant in teraction scores were less re sponsive to the high distress expressed by the children. Conversely, parents with low avoidant interaction scores were more responsive to the level of distress in their children. Th ese differences in responses to distress were independent of the childs temperament and parental personality. Looking beyond the current relationships, inte raction styles have intergenerational patterns and consequences. Developing a be tter understanding of the issues and responding to them from an attachment theory perspective may help target the underlying issues rather than just the immediate behavi ors being expressed. This in turn, could have a lasting impact on the mother, the child a nd the perpetuation of appropriate health behaviors.

PAGE 186

163 One aspect of the associations between in teraction style and health care utilization that was not captured in this study was the number of times women did not show up for their scheduled appointments. Although women with higher avoidant interaction style ratings did not take their children to as many sick or follow-up visits, this may not mean they did not make appointments to see the pr ovider. There were several instances where researchers received information (anecdotal ly and through researcher medical record abstraction) that the num ber of times mothers did not show up for scheduled appointments was high for some women. This pattern of not showing up for scheduled appointments negatively impacts the child by no t receiving the care. It also impacts the provider practice by limiting the number of clie nts they can see in a day and prevents other mothers from getting appointments as soon as they wanted for their children. Interaction Style Reliability Issues There may be issues with the validity of the interaction style measure in this population indicated by low levels of intern al consistency. The instruments authors suggest that self-report bias may play a role. This variability may also be the result of the difficulties faced by individuals lacking the fi nancial resources to be self-sufficient. Many of these women reported having to rely on others for things that many people take for granted such as transportation, housing, f ood and other resources. Frequent requests for assistance from others can lead mothers to have feelings of helplessness and a desire to regain control over their lives. Constant requests for assistance also increases the chances that others will make mistakes, forg et to follow-through, or become frustrated with the requests. Regardless of the mothers ba sic interaction style, this lack of control

PAGE 187

164 can lead to stronger maternal responses to the RSQ items regarding dependency and selfsufficiency. This also could e xplain the higher levels of avoi dant interaction style scores. Additional Study Findings In addition to the findings of the specif ic research questions and hypotheses, a variety of other findings were identified during the process of this study. For example, a number of significant bivariate correlations between race/ethni city and other factors were identified. However, when race was included in a more complex model containing other independent variables, confounders, and/or moderators, these differences became nonsignificant. For a better unders tanding of the issues, statis tics beyond bivariate analyses should be used. Another finding was that the health care syst em and providers appeared to be responsive to the needs of infants dem onstrated by: 1) mothers being able to change providers easily, 2) providers accelerating immunization schedules for infants who did not begin receiving immunization soon after birth, and 3) moth ers reporting being able to get appointments as soon as they needed. A dditionally, the study experienced difficulties in recruiting women of Hispanic ethnicity. The benefit of enhanced continuity of care resulting from having an electronic medical record was also identified. Issues were found regarding the accuracy of the marital status measure and baby spacing. Finally, the variability of clinic environments and the potential impact they have on patient satisfaction is discussed. For a more detailed description of these i ssues, see Appendix Z.

PAGE 188

165 Limitations of the Study There were a number of limitations to th is study. First, internal validity is threatened due to the study not being an experimental design. Second, there were a number of generalizability issues related to narrow inclusion/exclusion criteria, low response rates, especially in the Hispanic population, and the voluntary nature of the recruitment. Third, the study design was lim ited by the specific items on the survey instruments and a ceiling effect regarding th e number of well child care visits and immunizations had a narrow range (8 and 13) of possible responses. Fourth, self-reported data can suffer from recall bias. Fifth, the sample size limited the number of variables to be analyzed to 13. For more details re garding study limitations see Appendix Z. Implications for Action Receiving adequate preventive health care in the first years of life, such as immunizations and developmental screenings, can set the stage for a childs level of health and functioning into adulthood. Moth ers representing minority groups, whether they are based on race, ethnicity or socio economic status, are more likely to delay preventive health care. These delays are also often associated with increased sick visits and emergency department usage. Besides bein g costly to families and society, reactive health care does not support the development of good health for the rest of the childs life. Results from this study help provide a deeper level of understanding of the issues that promote or inhibit mothers traditionally served by Medicaid from getting the most appropriate care for their child(ren).

PAGE 189

166 While understanding the need to provide mo re individualized care, there is also the reality that health care a nd other social service resources are often scarce and need to be allocated in the most efficient manne r possible. Suggesting that new resources, whether they are in the form of staff time or additional costs, be added to the system without taking something else away are not well-received by those providing or paying for services. The findings from this study can help promote such a balance. The following recommended actions address changes that can be made at all levels and across systems. Medicaid Changes This study found that only a small portion of the infants medical record, usually just the immunization schedule, transferred from one provider to anothe r. This continuity of care is an integral part of providing the most appropriate health care services and reducing duplication of services This continuity of information can also help detect issues such as maltreatment that may be concealed by changes in providers. Earlier identification and intervention of all types of health issues can prevent more serious and long-term issues from arising. To help ensure the transfer of medical record information, the Medicaid program could require the that a copy of the whole medical record be transferred from one clinic to another when provider changes are made Because of the added cost of copying, a small billable charge such as $15-$20 could be added to the billing system. The costs of such a requirement should be offset by the savings resulting from not duplicating tests and other health care services across different providers.

PAGE 190

167 An even more efficient approach to providing Medicaid se rvices would be through a web-based electronic medical record similar to that used for immunizations developed by the Florida Department of H ealth (2005). This would provide easy access to the individuals full medical record to be used in clinical deci sion-making as well as provide a repository of data from which Medicaid could conduc t program evaluation activities more easily. In regard to any additional information that should be collected, there are some data elements that are currently being collected by providers that would be beneficial for Medicaid to have access. For example, the issues included in the bonding measure developed for this study have been associ ated with a number of health outcomes. Monitoring these indicators can help the sy stem respond more quickly to issues and provides information for better decision-ma king. For example, whether the pregnancy was intended (i.e., Did you intend to have this infant earlier, at the time you did, later in life, or other?) could be reported at the fi rst pediatric visit. Additionally, whether the mother was breastfeeding and administration of a 2-question depression screen at well child care visits could be recorded. Fina lly, a diagnosis of failure-to-thrive, high propensity for accidents, and sleeping problems all are indicators of medical issues but can also alert providers to potential maltreatment. Additionally, although not re lated to the research que stions, there were vast differences in the environments of the clin ics visited. The most extreme being a barren room with grey walls containing only black plas tic chairs. Medicaid should consider a set of minimum standards for clinic waiting room s. For example, at least one toy for which children could entertain themselves. The most popular toy in the clin ics visited for this

PAGE 191

168 study was a table emitting colored beads st rung through stiff colored wires. The paint/wallpaper on the walls also should be in viting to patients and families. The clinic waiting area is the patients first exposure and experience with the health care system. Making that experience more positive can im prove the overall experi ence with care, promote the relationship between provider a nd patient, and ultimately improve health outcomes. Health Care Providers Educating providers and their office staff about what was learned in this study can help to develop an understanding of the mothe rs preferred mode of interaction. In doing so, anticipatory guidance and other parent educa tion activities can be provided in the most efficient manner. For example, mothers with high avoidant interaction styles may reject parent education classes but may pr efer a video/DVD of parenting tips such as those that are often played in clinic waiting rooms. Conversely, a mother with anxious interaction styles may prefer the classe s where she can ask questions and receive additional validation from instructors about the best action to ta ke with her child. In addition to education, knowledge of a mothers interaction style can help providers to predict servi ce utilization behavior. For example, women with higher anxious scores generally attend more sick/fol low-up visits. As a result, they may benefit from involvement in programs such as Hea lthy Start which are specifically designed to address the individual needs of mothers and infants. That one-on-one attention could reduce the number of visits to the health care system for minor heal th issues, freeing up that time such as for other patients or professional development activities.

PAGE 192

169 Conversely, women with higher avoidant in teraction scores are more likely to attend fewer health care visits Staff may want to take proactive actions to improve compliance with recommend follow-up visits such as by making sure the mother receives a reminder call about the appointment. These moth ers also may benefit from involvement of programs such as Healthy Start who provide additional supports and encouragement for the women to take their chil dren to health care visits. In some cases, Healthy Start has even sent a nurse to the hom e to provide immunizati ons in order to help meet their own immunization objec tives (Struchen-Shellhorn, 2000). In recent years, some of the Healthy St art programs have co-located Healthy Start care coordinators within health care clinic s. There are a number of efficiencies and benefits for families, health care providers, and Healthy Start care coordinators. Mothers can work with their care coordinators either before or after their health care visits, allowing for more discussions regarding health issues and reducing the need to schedule separate appointments. Clinic staff can benefit by guiding the Healthy Start Care coordinators to provide any additional parent education resources for which the mother may benefit. Finally, Healthy Start Care coor dinators can benefit by seeing more mothers in shorter periods of time by reducing drive tim e to home visits. This time-savings can be applied to families needing more intensive serv ices, to serving additional families, or to professional skills developm ent of care coordinators. Finally, as noted earlier, the clinic waiting room is the first exposure a family has to the health care provider and can set the st age for the remainder of the visit. Making the waiting rooms more inviting, offering sick and well child areas where possible, and

PAGE 193

170 providing activities to occupy time (books, simple toys, etc) can all begin to establish a positive rapport between the family and the clinic. Other Programs There is a large body of lit erature regarding individua l learning styles and the benefits of teaching using those styles (W inn & Vesper, 2005). Similarly, individuals have differing preferences for how they like to interact with others, assimilate new information and develop new skills, regardless of whether or not the individual is at high risk for specific health outcomes and beha viors. Programs that pr ovide support services to women and children such as parent educa tion programs, WIC and Healthy Start can all benefit from increased understanding of what motivates or inhibits mothers from utilizing preventive health care services. The use of an abbreviated interaction style measure using 5 questions can be incorporated into the set of paperwork completed for a new patient. The information gained from this screen could then be used to help guide program activ ities that best meet the needs and prefer ences of mothers. Finally, this study provides another source of support fo r the benefits of breast feeding, a component of the studys maternal bonding score. There is a need to continue to promote the use of breast feeding as well as a need for more accurately measuring the extent to which mothers are breast feeding. That information can then be used to assist in further program evaluation and decision-making.

PAGE 194

171 Recommendations for Future Research This exploratory study expanded the know ledge base regarding significant relationships that can help guide actions. It also raised more questions about what motivates mothers to take their children to pe diatric health care visi ts that need to be explored. Replication of Study Using Other Populations First, the narrowly defined population provi ded the power to detect differences in factors beyond what was already known in th e literature (i.e., controlled for health insurance and socioeconomic status). However, the same selection processes also limited the generalizability of the st udy to a small group of women. As a result, this study should be replicated using other populations to de termine whether the same findings apply. Replication should include the number of no s hows on the part of the mother. It also should gather more detailed health care utiliz ation data for the moth er both prenatally and postnatally. Maternal Bonding Maternal bonding has been cited thro ughout the maternal and child health literature as being a associated to health behaviors and outcomes. Unfortunately, the existing measures of maternal bonding ar e resource-intensive and are not easily incorporated into larger studies. The mate rnal bonding measure found to be a significant factor in this study uses data that is normally collected on patients through the provision of standard health care. The bonding measur e was developed specifi cally for this study

PAGE 195

172 and was not put through rigorous testing of its psychometric properties. The development of a brief maternal bonding screen such as either the four or six variable measures in this study whose sound psychometric properties ha ve been demonstrated would be very useful in expanding the knowledge base be yond its current measurement limitations. Measurement of Feelings, Attitudes, and Perceptions The statistical significance of the measure of a mothers feelings about doctors while controlling for provider ratings indicate s more needs to be understood regarding the underlying issues that lead individuals to have specific feelings, att itudes, and perceptions of health care providers and systems. Devel oping a better understanding of these issues can help systems provide more individualized care to patients and their families. Incorporation of Provide rs in Research Activities A hurdle that greatly limited not only th e generalizability of the study through recruitment issues was the lack of participa tion by providers in the process. The limited opportunities for recruitment such as thos e resulting from compliance with HIPAA guidelines has led to the extensive use of the Medicaid mailing list as a way to recruit study participants. Researchers may ultimatel y be targeting the small percentage of Medicaid participants who choos e to be involved. As a result, too much research may be relying on the same respondents making the generalizability of the findings much narrower than was intended. In the long run, not expanding to other sources of recruitment could lead to bias ed results in the literature.

PAGE 196

173 The selection bias of the low response ra tes from this approach need to be balanced by studies that can have greater randomization. Future research should include the active involvement of providers in the re search activities such as by having a staff member within the agency being part of the research team. This would allow for recruitment opportunities that are less likely to interfere with the daily operation of the clinics. It would also allow for easier access to the individuals entire medical record, an issue that also caused barriers for this study. Family Planning and Baby Spacing Family planning and baby spacing is another area in need of additional emphasis, if not in the area of researc h, at least in the area of pr ogram evaluation. A number of women had subsequent pregnanc ies too early. More than one indicated to the researcher that the infant was not wanted and that requests for assist ance in family planning fell short. Some mothers were late in their ch ild bearing years and were surprised by the pregnancy. Finally, women like the one who ha d five infants in 26 months expressed no issues with having so many children. Fr om another perspective, 60% of study participants and 64% of Medica id participants who gave birth in the state reported the pregnancy was unintended (Florida Depart ment of Health, 2004). Additionally, the Florida Vital Statistics Annual Report (2004) noted 91,710 abortions, with 90,315 of those abortions being for personal choice reas ons rather than for a specific medical reason. Women who do not want childr en at a particular point in time are often not in a position to adequately care for th e child. Given the enormity of th is issue, it is in the best

PAGE 197

174 interest of children, families, and society to help women who do not want to become pregnant to have access to services to prev ent pregnancy and education about the health benefits of spacing their children further apart. Electronic Medical Record Finally, there were a number of issues regarding the collec tion of health care utilization data from pediatricians that hi ghlighted the need fo r electronic medical records. However, the technology and the security safeguards fo r such systems is still in its infancy. As more locations develop electr onic medical records, researchers should use that information to identify and promote their sp ecific benefits to individuals such as the continuity of care that was often lost wh en mothers changed from one provider to another. In summary, this study was intended to explore a new way of thinking about old problems. It sought to identify the relati onships among interac tion style, reported experiences of care and pediatric health ca re utilization. However, this study has also highlighted the need for more understanding re garding what motivates individuals when just knowing what should be done is not enough. Finally, true to its purpose, this study found that interaction style and reported expe riences of care were associated with how mothers utilized pediatric health care.

PAGE 198

175 References Abbotts, B., & Osborn, L. M. (1993). Immuni zation status and reasons for immunization delay among children using public health immunization clinics. American Journal of the Disabled Child, 147(9), 956-958. ACITS. (1995). Usage note: stat-53: factor analysis using SAS proc factor. Retrieved August 18, 2004, from: http://www.utexas.edu/cc/docs/stat53.html. Agency for Health Care Administration. (2004) Age band market pene tration analysis as of July, 2004. Retrieved July 21, 2004, from www.fdhc.state.fl.us/mchq/ managed_health_care/mhmo/doc s/mc_enroll/enrjul2004.xls. Agency for Health Care Administration. (2005). Characteristics of the non-elderly uninsured. Retrieved December 2, 2005, from www.ahca.myflorida.com/ medicaid/research/projects/fhis2004 /pdf/ characteri stics_of_the_nonelderly_uninsured_030305.pdf. Agency for Health Care Research and Qu ality. (2003). National Health Disparities Report. Washington, D. C.: U. S. Depart ment of Health and Human Services. Retrieved August 1, 2004, from: http://www.ahcpr.gov/qual/nhdr03/ nhdrsum03.htm. Agency for Health Care Research and Qu ality. (2004a). Overview of the medical expenditure panel survey. Re trieved March 4, 2004, from http://meps.ahrq.gov/WhatIsMEPS/Overview.htm. Agency for Health Care Research and Qu ality. (2004b). CAHPS and the national CAHPS benchmarking database. Retrieved July 10, 2004, from http://www.ahcpr.gov/qual/cahpfact.htm. Agency for Health Care Research and Qu ality (2004c). (n.d.). Annual benchmarking of the national CAHPS benchmarking database : 2000 annual report. Retrieved August 1, 2004, from www.ahcpr.gov/qual/ncbd2000/NCBDrepa.htm. Agency for Health Care Research and Qua lity (2005). Overview of CAHPS Instruments. Retrieved December 2, 2005 from www.ahrq.gov/qual/cahps. Agresti, A. ( 1996). An introduction to categorical data analysis New York: Wiley.

PAGE 199

176 Aiken, K. D. Freed, G. L., & David, M. M. (2004). When insu rance is not static: Insurance transitions of low income children and implications for health and health care. Ambulatory Pediatrics, 4(3), 234-243. Ainsworth, M. D. S. (1967). Infancy in Uganda: Infant care and the growth of attachment. In J. Cassidy, & P. Shaver (eds), Handbook of attachment: theory, research and clin ical applications (p. 4). New York: Gilford Press. Ainsworth, M. D. S. (1985). Patterns of attachment. Clinical Psychologist, 38(2), 27-29. Ainsworth, M. D. S. (1989). Attachments beyond infancy. American Psychologist 44, 709-716. Ainsworth, M. D. S., & Bell, S. M. (1969). Some contemporary patterns in the feeding situation. In P. Boss, W. Doherty, R. R. LaRossa, W. Schumm, & S. Steinmetz (eds), Sourcebook of family theories and methods: A contextual approach New York, NY: Plenum Press. Albrecht, T. (2005). Portal to portal. Na tional Friendly Access Retrieved 3/20/05 from www.etfap.com/member/. Altucher K., Rasmussen K.M., Barden E.M., & Habicht J.P. (2005). Predictors of improvement in hemoglobin concentration among toddlers enrolled in the Massachusetts WIC program. Journal of the American Dietetic Association, 105(5):709-15. American Academy of Pediatrics. (2002). Guidelines for Health Supervision III (3rd ed.). Elk Grove Village, IL: American Academy of Pediatrics. American Academy of Pediatrics, Committ ee on Children with Disabilities. (2001). Developmental surveillance and scre ening of infants and young children. Pediatrics 108, 192-196. American Academy of Pediatrics, Medical Ho me Initiatives for Children with Special Needs Project Advisory Committee. (2002) Policy statement: the medical home. Pediatrics 110(1), 184-186. American College of Obstetrics and Gyn ecologists. (n.d.). Post Partum Depression. Retrieved March 4, 2004, from http://www.medem.com/ medlb/article_detaillb.cfm?artic le_ID=ZZZSLRXO97C&sub_cat=4. American Heritage College Dictionary 4th Edition (2002). Boston. Houghton Mifflin. Athabasca University. (2004). Psychology 404: Experimental ps ychology. Retrieved November 20, 2004 from http://psych.ath abascau.ca/html/Validity/index.shtml.

PAGE 200

177 Baldwin, L.M., Larson E.H., Connell, F.A., Nordlund, D., Cain, K.C., Cawthon M.L., Byrns P., & Rosenblatt, R.A. (1998). The effect of expanding Medicaid prenatal services on birth outcomes. American Journal of Public Health 88, 1623-1629. Bandura, A. (1994). Self-efficacy. In V. S. Ramachaudran (ed), Encyclopedia of human behavior (Reprinted in H. Friedman [Ed. ], Encyclopedia of mental health. San Diego: Academic Press, 1998) (Vol. 4, pp. 71-81). New York: Academic Press. Bartholomew, K. (1990). Avoidance in in timacy: an attachment perspective. Journal of Social and Personal Relationships 7, 147-178. Baydar, N. (1995). Consequences for childre n of their birth planning status. Family Planning perspectives, Nov-Dec; 27(6):228-34, p. 245. Beach M.C., Price E.G., Gary T.L., Robinson K.A., Gozu A., Palacio A., et al.. (2005). Cultural competence: a systematic review of health care pr ovider educational interventions. Medical Care, 43(4), 356-373. Bender, R.H., & Garfinkle, S.A. (2001). Differe nces in the structure of CAHPS measures among the medicare fee-for-service, a nd privately insured populations. Health Services Research 36(3), 489-508. Black, L., McGoldrick M., & Pearce, J.K. (1996). Ethnicity and family therapy (2nd ed.). New York: Guilford Press. Black, M.M., Cutts, D.B., Frank, D.A., Geppe rt, J., Skalicky, A., Levenson S., et al. (2004). Special supplemental nutrition program for women, infants and children participation and infants growth and health: a multi-site surveillance study. Pediatrics, 114(1):169-76. Bolton, P., Holt, E., Ross, A., Hughart, N., & Guyer, B. (1998). Estimating vaccination coverage using parental recall, vacc ination cards, and medical records. Public Health Report 113(6), 521-526. Born, W., Greiner, K. A., Sylva, E., Butler, J., & Ahuluwalia, J. S. (2004). Knowledge, attitudes, and beliefs about end-of-life care among innercity African Americans and Latinos. Journal of Palliative Medicine 7(2), 247-256. Borrayo, E.A., & Guarnaccia, C.A. (2000). Di fferences in Mexican-Born and US-born women of Mexican descent regarding factors related to breast cancer screening behaviors. Health Care Women International 21(7), 599-613. Boss, P., Doherty, W., LaRossa, R., Sc humm, W., & Steinmetz, S. (1993). Sourcebook of family theories and methods: a contextual approach New York, NY: Plenum Press.

PAGE 201

178 Bowlby, J. (1969). Attachment and loss: Vol.1. Attachments New York: Basic Books. Bretherton, I. (1999). Theoretical contributi ons from developmental psychology. In J. Cassidy, & P. Shaver (eds), Handbook of attachment: Theory, research and clinical applications (p. 283). New York: Gilford Press. Bretherton, I., & Ainsworth, M. D. S. (1974). Re sponses of one-year-old s to a stranger in a strange situation. In J. Cassidy, & P. Shaver (eds), Handbook of attachment: Theory, research and clinical applications (p. 283). New York: Gilford Press. Broussard, E. R. (1995). Infa nt attachment in a sample of adolescent mothers. Child Psychiatry and Human Development 25(4), 211-219. Brown, J.A., Nederend, S.E., Hays, R.D., Shor t, P.F., & Farley, D.O. (1999). Special issue in assessing care of Medicaid recipients. Medical Care 3(3 suppl), MS79-88. Browner, C.H., Preloran, H.M., Casado, M. C., Bass, H.N., & Walker, A.P. (2003). Genetic counseling gown awry: Miscommunica tion between prenatal genetic service providers and Mexican -origin clients. Social Science Medicine 56(9), 1933-1946. Budda Sasana. (2004). Religious significance of fasting. Retrieved March 15, 2004, from http://www.saigon.com/~anson/ebud/whatbudbeliev/214.htm. Buescher, P.A., & Ward, N.I. (1992). A comparison of low birth weight among Medicaid patients of public health department s and other providers of prenatal care in North Carolina and Kentucky Public Health Reports 107, 54-59. Bultman, D.C., & Svarstad, B.L. (2000). E ffects of physician communication style on medication beliefs and adherence with antidepressant treatment. Patient Education and Counseling 40(2), 173-185. Busch-Rossnagel, N.A., Fracasso, M.P., & Varg as, M. (1994). Reliabili ty and validity of Q-sort measure of attachment security in Hispanic infants. Hispanic Journal of Behavioral Sciences 16(2), 240. Busey, S., Schum, T.R.& Meurer, J.R. (2002).Parental perceptions of well-child care visits in an inner-city clinic. Archives of Pediatric Adolescent Medicine 156(1), 62-66. Bush, G.W. (2004). Promoting innovation and competitiveness: president bushs technology agenda. Retrieved 12/12/05 from:www.whitehouse.gov/infocus/ technology. Byrd, T.L., Mullen, P.D., Selwyn, B.J., & Lorimor, R. (1996). Initiati on of prenatal care by low-income Hispanic women in Houston. Public Health Reports 111, 536-540.

PAGE 202

179 Capterra. (2005). Electronic medical record software directory. Re trieved 12/1/05 from: www.capterra.com/electronic-medical-record-software. Carlson, E.D., & Chamberlin, R.M. (2004). Th e black-white percep tion gap and health disparities research. Public Health Nursing 21(4), 372-379. Cassidy, J., & Shaver, P. (1999). Handbook of attachment: theo ry, research and clinical applications New York: Guildford Press. Centers for Disease Control. ( 2004). National Centers for Health Statistics Fastats: Infant Health. Retrieved 12/2/05 from www.cd c.gov/nchs/fastats/infant_health.htm. Centers for Disease Control and Prevention. (2005). Recommended child and adolescent immunization schedule 2004. Retrieved December 2, 2005, from http://www.cdc.gov/nip/recs/child-schedule.htm. Centers for Disease Control and Preven tion. (2005a). Combination vaccines for childhood immunizations. Morbidity and Morta lity Weekly Report May 14, 1999; 48(RR05):1-15. Centers for Medicare and Me dicaid Services. (2004a). Me dicaid Services. Retrieved March 15, 2004, from www.cms.hhs.gov/medicaid/mservice.asp. Centers for Medicare and Me dicaid Services. (2004b). Welc ome to the state childrens health insurance program. Retrieved March 15, 2004, from http://www.cms.hhs.gov/s chip/about-schip.asp? Chalk, R. (2004). Calling the s hots: Institute of Medicine report on immunization finance policies and practices. Retr ieved October 10, 2004, from www.ahcpr.gov/qual/ ncbd2000/NCBDrepa.htm. Chu, S.Y., Barker, L.E., & Smith, P.J. (2004) Racial/ethnic disparities in preschool immunizations: united states 1996-2001. American Journal of Public Health, 94(6), 973-977. Ciechanowski P., Russo J., Katon W., Von Ko rff M., Ludman E., Lin E., Simon G., & Bush T. (2004). Influence of patient attach ment style on self-care and outcomes in diabetes. Psychosomatic Medicine Sep-Oct; 66(5):720-8. Ciechanowski PS, Walker EA, Katon WJ, Russo JE. (2002). Attachment theory: a model for health care utilization and somatization. Psychosomatic Medicine Jul-Aug; 64(4): 660-7. Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers.

PAGE 203

180 Colin, V.L.(1996). Human attachment J. Cassidy & P. Shaver. New York: McGraw-Hill. Collins, N.L., & Read, S.J. (1990). Adult a ttachment, working models and relationship quality in dating couples. Journal of Personality and Social Psychology 58, 644-663. Davis, D., Shaver, P.R. & Vernon, M.L. ( 2003). Physical, emotiona l, and behavioral reactions to breaking up: th e roles of gender, age, emotional involvement, and attachment style. Personality and Social Psychology Bulletin, 29(7): 871-884. Derose, K. P., & Baker, D. W. (2000). Lim ited English proficiency and Latinos' use of physician services. Medical Care Research Review 57(1), 76-91. Diaz, V.A. Jr. (2002). Cultural factors in preventive care: Latinos. Primary Care 29(3), 503-517. Dillman, D. (2000). Mail and internet surveys: the tailored design method New York. John Wiley & Sons. Dominowski, P., & Bartholet, A. (2004). The listener survey toolkit: Advantages and disadvantages of mail and telephone surveys. Retrieved October 2, 2004, from http://www.wksu.org/toolki t/chapter4/section1.html. Edelstein, R.S., Alexander, K.W ., Shaver, P.R., Schaff, J.M., Quas, J.A., Lovas, G.S. et al. (2004). Adult attachment style and pa rental responsiveness during a stressful event. Attachment and Human Development 6(1), 31-52. Eley TC, Bolton D, O'Connor TG, Perrin S, Sm ith P, Plomin R. (2003). A twin study of anxiety-related behaviors in pre-school children. Journal of Child Psychology and Psychiatry, Oct; 44(7): 945-60. El-Mohandes, A.A., Katz, K.S., El-Khorazata y, M.N., McNeely-Johnson, D., Sharps, P. W., Jarrett, M. H. et al. ( 2003). The effect of a parent ing education program on the use of preventive pediatric health care se rvices among low income, minority mothers: A randomized controlled study. Pediatrics 111(6 pt 1), 1324-1332. Evolution ID Card Systems and Badge Supplie s. (2004). ID Card Systems provide more than Student ID's, they help schools create a secure learning e nvironment. Retrieved March 15, 2004, from http://www.evolution-1.com/student.htm. Farma, T., & Cortinovis, I. (2001). Three north-American attachment questionnaires: reliability of the measures and their use in the Italian context. Ricerche Di Psicologia 24(1), 75-97. Feeney, J., & Noller, P. (1996). Adult Attachment Thousand Oaks, CA: Sage.

PAGE 204

181 Feeney, J. A. (2000). Implications of attach ment style patterns of health and illness. Child: Care, Health and Development 26(4), 277-288. Feinberg, A.N., & Hicks, W.B. (2003). Pati ent compliance with the first newborn visit appointment. Journal of Perinatology 23(1), 37-40. Feldman, R., Weller, A., Leckman, J. F., Kuint, J. & Eidelman, A.I. (1999). The nature of the mothers tie to her infant: Matern al bonding under conditions of proximity, separation, and potential loss. Journal of Child Psychology and Psychiatry, 40(6), 929-939. Fiscella, K. Franks, P., Doescher, M.P., & Sa ver, B.G. (2002). Disparities in health care by race, ethnicity, and language among insured: Findings from a national sample. Medical Care 40(1), 52-59. Flocke, S.A., Miller, W.L., & Crabtree, B.F. (2002). Relationships between physician practice style, patient satisfaction, and attributes of primary care. Journal of Family Practice 51(10), 835-840. Florida Department of Health. (2004). Me dicaid participation among women who are pregnant in Florida, 20002001. Retrieved 11/23/05 from: www.doh.state.fl.us/disease_ctrl/epi/prams.htm. Florida Department of Healt h. (2004a). Resident live birt hs. Retrieved October 10, 2004, from www.floridacharts.com/charts/re port.aspx?domain=03&indNumber=0025. Florida Department of Health (2005). Percent of two year old children fully immunized. Retrieved 3/3/05 from: http://www.floridacharts.com/charts/ report.aspx?domain=03&IndNumber=0075. Florida Department of Health (2005a). Week ly Disease Table. Retrieved 1/19/06 from: www.doh.state.fl.us/disease_ctrl/ epi/disease _table/2005_weeks/ dt_week482005frame2.html. Florida Department of Health (2005b). Infant Mortality Rate (0-364 days from birth). Retrieved 3/3/05 from: http://www.floridacharts.com/charts/ Domain2.aspx?Domain='03'. Florida Department of Health (2005c). Fl orida SHOTS: state online tracking system. Retrieved 12/1/05 from: www.flshots.org. Food and Nutrition Service. (2005). Food Stamp Program. Un ited States Department of Agriculture. Retrieved March 4, 2005 from http://www.fns.usda.gov/fsp/. Fowler, F. (1988). Survey research methods Newbury Park, CA: Sage Publications.

PAGE 205

182 Fraley R.C. & Shaver P.R.(1997). Adult a ttachment and the suppression of unwanted thoughts. Journal of Personality and Social Psychology 73, 1080-1091. Fraley R.C. ( 2002). Attachment stability fr om infancy to adulthood: Meta-analysis and dynamic modeling of developmental mechanisms. Personality and Social Psychology Review 6(2), 123-151. Friedlaender E.Y., Rubin D.M., Alpern E.R., Mandell D.S., Christian C.W., & Alessandrini E.A. (2005). Patterns of health care use that may identify young children who are at risk for maltreatment. Pediatrics Dec 116, 1303-1308. Gallagher P. M. & Fowler F. J.Jr. (2001). Seventh Conference on Health Survey Research Methods. DHHS Publication No. (PHS) 01-1013. Hyattsville, MD: National Center for Health Statistics. Gaudino J. (2005). Suggestions regarding meas uring health care ut ilizations. (Personal Communication). Gavin N.I., Adams E.K., Hartmann K.E., Bene dict M.B., & Chireau M. (2004). Racial and ethnic disparities in the use of pre gnancy-related health care among Medicaid pregnant women. Maternal and Child Health Journal, Sep 8(3), 113-26. Gillespie N.A., Zhu G., Heath A.C., Hickie I.B., & Martin N.G. (2000). The genetic etiology of somatic distress. Psychological Medicine Sep; 30(5):1051-61. Gilmore, G.J. (2005). DoD demonstrates global electronic medi cal records system. American Forces Information Servi ce News Articles. Nov. 11, 2005. Retrieved 12/2/05 from: www.dod. mil/news/nov2005/20051121_3403.html. Giordano J. & McGoldrick M. African Americ an Families. In M. McGoldrick, J. K. Pearce, & J. Giordano Ethnicity and Family Therapy New York: Guildford. Glauber J.H. (2003). The immunization deliver y effectiveness assessment score: a better immunization measure? Pediatrics ; 112(1):e39-e45. Goldberg, S. (1991). Recent developments in attachment theory and research. Canadian Journal of Psychiatry 36(6), 393-400. Gorman, B. K., Landale, N. S., & Oropesa, R. S. (2001). Poverty, insurance, and wellbaby care among mainland Puerto Rican children. Social Biology 48(1-2), 67-85. Griffin, D. W., & Bartholom ew, K. (1994). The metaphysic s of measurement: The case of adult attachment. In K. Bartholomew, & D. Perlman (eds), Advances in Personal Relationships (Vol. 5pp. 17-52). Lon don: Jessica Kingsley.

PAGE 206

183 Guide to Community Preventive Services. (2005).Vaccine preventable Disease. Retrieved 4/1/05 from: http:// www.thecommunityguide.org/vaccine/. Gurlanik, D. (1986). Websters new world dictionary: second college edition New York, NY: Simon & Schuster. Hakim, R.B., & Bye, B.V. (2001). Effectivene ss of compliance with pediatric preventive care guidelines among Medi caid beneficiaries. Pediatrics 108(1), 90-97. Halfon, N., Regalado, M., Sareen, H., Inkelas, M., Reuland, C ., Glascoe, F. et al. (2004). Pediatrics 113(6): 1926-1933. Hall, J.A., Horgan, T.G., Stein, T.S., & Ro ter, D.L. ( 2002). Li king in the physicianpatient relationship. Patient Education and Counseling 48(1), 69-77. Handler, A., Rosenberg, D., Raube, K., & L yons, S. (2003). Satisfaction and use of prenatal care: their relationship among Af rican-American women in a large managed care organization. Birth 30(1), 23-30. Harwood, R.L.(1992). The influence of cultura lly derived values on Anglo and Puerto Rican mothers perceptions of attachment behavior. Child Development 63(4),822-39. Hatcher, L., & Stepanski, E. (2001). A step-by-step approach to using the sas system for univariate and multivariate statistics Cary, NC: SAS Institute. Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology 52, 511-524. Hudziak J.J., Van Beijsterveldt C.E., Altho ff R.R., Stanger C., Rettew D.C., Nelson E.C., Todd R.D., Bartels M., Boomsma D.I. (2004). Genetic and environmental contributions to the Child Behavior Check list Obsessive-Compulsi ve Scale: a crosscultural twin study. Archives of General Psychiatry Jun; 61(6): 608-16. Ickovics, J.R., Kershaw, T.S., Westdahl, C., Rising, S.S., Klima, C., Reynolds, H. et al. (2003). Group prenatal care a nd preterm birth weight: results from a matched cohort study at public clinics. Obstetrics and Gynecology 105(5 pt1), 1051-1057. Iniguez, E., & Palinkas, L.A. (2003). Vari eties of health se rvices utilization by underserved Mexican American women. Journal of Health Care for the Poor and Underserved 14(1), 52-69. Inkelas, M., Schuster, M., Olson, L., Par k, C., & Halfon, N. (2004). Continuity of primary care clinician in early childhood. Pediatrics 113(6 suppl), 1985-1987.

PAGE 207

184 Institute for Child Health Policy. (2004). The Florida KidCare Evaluation Series: Florida KidCare Evaluation Report Retrieved 3/01/04 from: http://www.ichp.ufl.edu/ documents/KidCareReport2004_Final.pdf. Institute of Medicine of the National Academies. (2003). Unequal treatment Washington D.C.: National Academy Press. Institute of Medicine of the National Academies. (2004a). Crossing the quality chasm: A new health system for the 21st century, calling for fundam ental change in the health care system Washington DC: National Academy Press. Institute of Medicine of the National Acad emies. (2004b). Childrens health, the nations wealth: Assessing and improving child health. Retrieved July 6, 2004b, from www.iom.edu/report.asp?id=21082. Jones, J.(1981). Bad blood: The Tuskegee syphilis experiment New York, NY:Free Press. Jones, M. E. Cason, C. L., & Bond, M. L. (2002). Access to preventi ve health care: is method of payment a barrier to immigrant Hispanic women? Womens Health Issues 12(3), 129-137. Kanda, M. (2004). Well-Child Care: what is it and why is it importa nt? Retrieved March 16, 2004, from http://hsnrc.org/He althInst/042203_Kanda-bw.pdf. Keating, N.L., Green, D.C., Kao, A.C., Ga zmarian, J.A., Wu, V.Y., & Cleary, P.D. (2002). How are patients specific ambulator y care experiences related to trust, satisfaction, and considering changing physicians? Journal of General Internal Medicine 17(1), 29-39. Kelly, N.R. Huffman, L.C., Mendoza, F.S., & Robinson, T.N. (2003). Effects of a videotape to increase use of poison c ontrol centers by low-income and Spanishspeaking families: A randomized control trial. Pediatrics 111(1), 21-26. Kendal, A.P., Peterson, A., Manning, C., Xu, F., Neville, L.J., & Hogue, C. (2002). Improving the health of infants on Medi caid by collocating special supplemental nutrition clinics with managed care provider sites. American Journal of Public Health 92(3), 399-403. Kerns, K.(1994). A developmental model of the relations between mother-child attachment and friendship. In R. Erber & R. Gilmour (ed), Theoretical frameworks for personal relationships (pp.129-156). Hillsdale, NJ:Lawren ce Erlbaum Associates, Inc. Kim, S., Symons, M., & Popkin B.M. (2004). Contrasting socioecono mic profiles related to healthier lifestyles in China and the United States. Journal of Epidemiology, 159(2), 184-191.

PAGE 208

185 Klaus, M. H. Kennel, J. H., & Klaus, P. H. (1995). Bonding: Building the foundations of secure attachment and independence New York, NY: Addison-Wesley Publishing Company. Kleinbaum D.G., Kupper L.L. & Muller, K.E. Applied regression analysis and other multivariate methods, second edition. Belmont California: Duxbury Press. Kochanek, K. D. and Martin, J. A. (n.d.). Supplem ental analyses of recent trends in infant mortality. Retrieved July 4, 2005, from www.cdc.gov/nchs/products/ pubs/hestats/infantmort/infantmort.htm. Kohane I.S., Greenspun P., Fackler J ., Cimino C. & Szolovits P. (1996). Journal of the American Medical Information Association ; 3(3):191-207. Kotelchuck, M., Kogan, M.D., Alexander, G.R ., & Jack, B.W. (1997). The influence of site of care on the content of pr enatal care for low income women. Maternal and Child Health Journal 1(1), 25-34. Lara, M., Duan, N., Sherbourne, C., Half on, N., Leibowitz, A., & Brook, R.H. (2003). Children's use of emergency departments for asthma: Persistent barriers or a cure need? Journal of Asthma 40(3), 289-299. Lassiter, J.H., & Baldwin, J.H. (2004). Hea lth care barriers for latino children and provision of culturall y competent care. Journal of Pediatric Nursing 19(3), 184-192. Laveist, T.A., & Nuru-Jeter, A. (2002). Is doctor-patient race concordance associated with greater satisfaction with care? Journal of Health and Social Behavior, 43(3), 296-306. Laws, M.B. Heckscher, R., Mayo, S.J., Li, W., & Wilson, I.B. (2003). A new method for evaluating the quality of medical interpretation. Medical Care 42(1), 71-80. Lawton and Rhea Chiles Center. (2005). Did your prenatal care provider help you find a health care provider for your new baby? Personal communication. Retrieved 4/8/05. Lawton and Rhea Chiles Center. (2004). Na tional Friendly Access Program publication. Retrieved August 4, 2004, from http://www.chilescenter.org/ newsletters/FAccess.pdf. Lee G.M., Friedman J.F., Ross-Degnan D., Hibberd P.L. & Goldman D.A. (2003). Misconceptions about colds and predicto rs of health se rvice utilization. Pediatrics Feb;111(2):231-36. Lee, J.W., Fitzgerald, K., & Ebel, B.E. (2003). lessons for increasing awareness and use of booster seats in a Latino community. Injury Prevention 9(3), 268-269.

PAGE 209

186 Lee, M. A., & Learned, A. (2002). Well-baby care in Connecticuts Medicaid managedcare program. Connecticut Medicine 66(9), 515-521. Levine, L.V., Tuber, S.B., Slade, A., & Ward, M.J. (1991). Mothers mental representations and their relations hip to mother-infant attachment. Bull Menninger Clinic 55(4), 454-469. Lilienfeld D. E. & Stolley P.D. (1994). Foundations of epidemiology: third edition. New York, NY. Oxford University Press, Inc. Low, N. (1991). Infant attachment: what we know Washington, D. C.: US Department of Health and Human Services. Luman, E.T., McCauley, M.M., Shefer, A., & Chu, S.Y. (2003). Maternal characteristics associated with vaccination of young children. Pediatrics 111(5 pt 2), 1215-1218. Magai, C., Cohen, C., Milburn, N., Thorpe B., McPherson, R., & Peralta, D. (2001). Attachment styles in older european american and african american adults. The Journals of Gerontology Series B: Psyc hological Sciences and Social Sciences, 56(s28-s35). Main M., & Hesse E. (1990). Parents unres olved traumatic experi ences are related to infant disorganized attachment status: Is frightened or fr ightening parental behavior the linking mechanism? In P. Boss, W. Doherty, R. R. LaRossa, W. Schumm, & S. Steinmetz (eds), Sourcebook of family theories and methods: A contextual approach New York, NY: Plenum Press. Marks, I.M. (1986). Genetics of fear and anxiety disorders. British Journal of Psychiatry Oct; 149: 406-18. Review. Matsuoka, N., Uji, M., Hiramura, H., Chen, Z ., Shikai, N., Kishida, Y., et al. (2005). Adolescents attachment style and early expe riences: a gender difference. Archives of Womens Mental Health. Oct 12, 2005 Online Publication. McLennan, J.D., & Kotelchuck, M. (2000). Pa rental prevention practices for young children in the context of maternal depression. Pediatrics 105(5), 1090-1095. McLennan, J.D., Kotelchuck, M., & Cho, H. (2001). Prevalence, persistence, and correlates of depressive symptoms in a national sample of mother and toddlers. Journal of the American Academy of Child and Adolescent Psychiatry 40(11), 13161323. Medline Plus. (2006). Medical encyclopedi a: nurse practitioner (NP) profession. Retrieved 1/25/06 from http://www.nlm.nih.gov/medlineplus/ency/article/001934.htm.

PAGE 210

187 Merrill, R., & Allen, E. (2003). Racial and et hnic disparities in satis faction with doctors and health providers in the united states. Ethnic Disparities 13(4), 492-498. Mertin, P.G. (1986). Maternal-infant atta chment: A developmental perspective. Australian New Zealand Journal of Obstetrics and Gynaecology 26 (4), 280-283. Mickelson, K.D., Kessler, R.C., & Shaver, P.R. (1997). Adult attachment in a nationally representative sample. Journal of Personality and Social Psychology ,73(5),1092-106. Minkovitz, C.S., Belote, A.D., Higman, S.M., S.J.R., & Weiner, J.P. (2001). Effectiveness of a pediatric-based interv ention to increase vaccination rates and reduce missed opportunities. Archives of Pediatrics and Adolescent Medicine, 155(3), 382-386. Minkovitz, C.S., O'Campo, P.J., Chen, Y.H ., & Grason, H.A. (2002). Associations between maternal and child health stat us and patterns of medical care use. Ambulatory Pediatrics 2(2), 85-92 Moore, P. & Hepworth, J.T. (1994). Use of perinatal and infant health services by Mexican-American Medicaid enrollees. Journal of the American Medical Association 272(4), 297-304. National Institutes of Mental Health. (2000). Neurobiology of anxiety and fear: challenges for genomic science of the new millennium. Biologic Psychiatry ; 48: 1144-1146. Nemours Foundation. (2005). Post Partum Depression and Caring for your baby. Retrieved March 4, 2005, from http://kidshealth.org/par ent/ pregnancy_newborn/ home/ppd_baby.html. Nevin, J. E., & Witt, D.K. (2002). Well child and preventive care. Primary Care, 29(3), 543-555. Ngo-Metzger, Q., Legezda, A., & Phillips, R. (2004). Asian Americans reports of their health care experiences. Resu lts of a national survey. Journal of General Internal Medicine 19(2), 111-119. Nicolson, P. (1998). Post-natal depression: Psychology science, and th e transition to motherhood New York: Routledge. Novick, L.F., Mustalish, A., & Eidsvold, G. ( 1975). Converting child health stations to pediatric treatment centers. Medical Care 13(9), 744-752. O'Callaghan, M.J., & Hull, D. (1978). Failure to thrive or failure to rear? Archives of the Disabled Child 53(10), 788-793.

PAGE 211

188 Oeffinger, K.C., Roaten, S.P., Hitchcock, M. A., & Oeffinger, P.K. (1992). The effect of patient education on pediat ric immunization rates. Journal of Family Practice, 35(3), 228-293. Office of the Law Revision Council. Medi care, Medicaid, a nd SCHIP Benefits Improvement Protection Act of 2000. Vol. PL 106-554, Sec. 1(a)(6). 114 Stat. 2763, 2763A-463. Owen, A.L. & Owen, G.M.(1997). Twenty years of wic: a review of some effects of the program. Journal of the American Diet etic Association; 97(7): 777-82. Palfrey, J.S., Solfis, L.A., Davidson, E.J., Liu, J. Freeman, L., & Ganz, M.L. (2004). The pediatric alliance for coordinated care: Evaluation of a medical home. Pediatrics 113(5 suppl), 1507-1516. Pasick, R.J., Stewart, S.L., Bird, J.A., & D'On ofrio, C.N. (2001). Quality of date in multiethnic health surveys. Public Health Reports 116(1 suppl), 223-243. Phillips, K.A., Mayer, M.L., & Aday, L.A. (2000). Barriers to care among racial/ethnic groups under managed care. Health Affairs 19(4), 65-75. Poland, M.L., Ager, J.W., & Olson, J.M. (1987). Barriers to receiving adequate prenatal care. American Journal of Ob stetrics and Gynecology 157, 297-302. Rapley, G. (2002). Keeping mothers and babies together-breastfeeding and bonding. Midwives 5(10), 332-334. Ricci, L., Giantris, A., Merriam, P., Hodge, S ., & Doyle, T. (2003). Abusive head trauma in Maine infants: medical, child pr otective and law enforcement analysis. Child Abuse and Neglect 27(3), 271-283. Rice F, Harold GT, Thapar A. (2002). Asse ssing the effects of age, sex and shared environment on the genetic aetiology of depression in childhood and adolescence Journal of Child Psychology and Psychiatry Nov; 43(8): 1039-51 Ronsaville, D.S., & Hakim, R.B. (2000). Well child care in the United States: racial differences in compliance with guidelines. American Journal of Public Health, 90(9), 1436-1443. Rosenberg T.J., Raggio T.P., & Chiasson M. A. (2005). A further examination of the epidemiologic paradox: birth outcomes among Latinas. Journal of the National Medical Association Apr;97(4):550-6.

PAGE 212

189 Rowland-Morin, P.A. & Carroll, J.G. (1990). Verbal communication skills and patient satisfaction: A study of doctor-patient interviews Evaluation and Health Professions 13, 168-185. Saha, S., Arbelaez, J.J., & Cooper, L.A. (2003) Patient-physician relationships and racial disparities in the qual ity of health care. American Journal of Public Health, 93(10), 1713-1719. Salant, P., & Dillman, D. (1994). How to conduct your own survey John Wiley & Sons. Sambamoorthi, U., & McAlpine, D.D. (2003). Racial, ethnic, socioeconomic, and access disparities in the use of preventive services among women. Preventive Medicine 37(5), 475-484. Scher, A. ( 2001). Attachment and sleep: a st udy of night waking in 12 month old infants. Developmental Psychobiology 38(4), 274-285. Schor, E. (2005). The future of well child care visits, as if it really mattered. Retrieved March 4, 2005, from http://www.aap.org/Peds-21/schor %20well-child%20care.ppt. Seid, M., Varni, J.W., Bermudez, L.O., Zi vkovis, M., Nelson, M., & Kurtin, P.S. (2001). Parents perceptions of primary care: meas uring parents experiences of pediatric primary care quality. Pediatrics 108(2), 264-270. Senanayake P. (1982). Child spacing and child health. Draper Fund Report Dec; 11:5-8. Serwint, J., Wilson, M.E.H., Vogelhut, J. W., Repke, J.T., & Seidel, H.M. (1996). A randomized controlled trial of prenatal pediat ric visits for urban low-income families. Pediatrics 98, 1069-1075. Shaver, P.R., & Hazan, C. (1993). Adult romant ic attachment: theory and evidence. In: D. Pearlman, & W. Jones (eds), Advances in Personal Relationships (pp. 29-70). London: Jessica Kingsley. Shaw, T. (1998). The burden of the flesh: Fasting and sexuality in early Christianity Minneapolis: Fortress Press. Shefer, A.M., Fritchley, J., Stevenson, J., L yons B., Friedman, R., Hopfensperger, D. et al. (2002). Linking WIC and immunization serv ices to improve preventive health care among low-income children in WIC. Journal of Public Health Management and Practice 8(2), 56-65. Shen J.J., Tymokow C., & MacMullen N. (2005). Disparities in maternal outcomes among four ethnic populations. Ethnic Di sparities. Summ er;15(3):492-7.

PAGE 213

190 Simpson, J. A., & Rholes, W.S. (1998). Attachment theory and close relationship New York: Guilford Press. Simpson, L. Zodet, M.W., Chevarle y, F.M., Owens, P.L. Dougherty, D., & McCormick, M. (2004). Health care for ch ildren and youth in the united states: 2002 report on trends in acces s, utilization, quality, and expenditures. Ambulatory Pediatrics 4(2). Slifkin, R.T., Freeman, V.A., & Silberman, P. (2002). Effect of the north carolina state childrens health insurance program on beneficiary access to care. Archives of Pediatrics and Adolescent Medicine 156(12), 1223-1229. Spence S.H., Rapee R., McDonald C., Ingr am M. (2001). The structure of anxiety symptoms among preschoolers. Behav Res Ther Nov; 39(11):1293-316. Starfield, B., Cassady, C., Nanda, J., Forre st, J., & Berk, R. (1998). Consumer experiences and provider per ceptions for managed care. Journal of Family Practice 46, 216-226. Stokes, M.E., Davis, C.S. & Koch, G.G. (2000) Categorical Data An alysis using the SAS system 2nd Edition: Chapter 12 Poisson Regression. SAS Institute Inc. Cary, NC. Stroufe, L.A., & Waters, E. (1977). Attach ment as an organizational construct. Child Development 48, 1184-1199. Struchen-Shellhorn, W. (2000). Pinellas county health department: In terviews with care coordinators working with healthy start program prenatal participants regarding prenatal care satisfaction. (Personal communication). Sultz, H.A., & Young, K.M. (1999). Health care USA: Unde rstanding its organization and delivery (2nd) New York, NY: Aspen Publications. Szilagyi, P.G., Dick, A.W., Klein, J D., S hone, L.P. Zwanziger, J., & McInerny, T. (2004). Improved access and quality of care af ter enrollment in the New York State Childrens Health Insurance Program(SCHIP). Pediatrics 113(5), 395-404. Szilagyi, P.G., Schaffer, S., Shone, L., Ba rth, R., Humiston, S.G., Sandler, M. et al. (2002). Reducing geographic disparities in childhood immunizat ion rates by using reminder/recall interventions in urban primary care practices. Pediatrics, 110(5), 58. Thompson, J.W., Ryan, K.W., Pi nidiya, S.D., & Bost, J.E. (2003). Quality of care for children in commercial a nd Medicaid managed care. JAMA 11, 1486-1493.

PAGE 214

191 Thompson, D. & Ciechanowski, P.S. (2003) Attaching a new understanding to the patient-physician relationship in family practice. Journal of the American Board of Family Practice May-Jun; 16(3): 219-26. Thompson, M., Gee, S.L.P.A., Kotz, K., & Northrop, L. (2000). Healthy loyalty promotion visits for new enrollees: resu lts of a randomized controlled trial. Patient Education and Counseling 42(1), 53-65. Tinsley, B.J., Trupin, S.R., Owens, L., & B oyum, L.A. (1993). The significance of womens pregnancy-related locus of cont rol beliefs for adherence to recommended prenatal health regimens and pregnancy outcomes. Journal of Reproductive and Infant Psychology 11, 97-102. Tonniges, T., Palfrey, J., & Mitchell, M. (2004). History of the medical home concept. Pediatrics 113(5). Tricare. (2004). Department of defense reli ability and validity an alysis. Retrieved 2004, from www.tricare.osd.mil/tricaresurveys/rel_val.html. Trochim, W.M. (2004). Measuremen t. Retrieved July 24, 2004, from www.socialresearchmethods.net/kb/constval.htm. U. S. Bureau of the Census. (1996). Population projections of th e United States by age, sex, race, and Hispanic origin: 1995 to 2050 (Current Population Reports, P251130). Washington, DC: U.S. G overnment Printing Office. U. S. Department of Health and Human Serv ices. (2004). Health c onditions health care. Retrieved July 26, 2004, from aspe.os.dhhs.gov/hsp/97trends/hc1-1a.htm. U.S. Department of Health and Human Services. (2000). Healthy People 2010: Understanding and improving health (2nd ed.). Washington, DC: U.S. Government Printing Office. United Press International. (2003). Vaccine sa fety parental concern. Retrieved January 13, 2004 from www.nlm.nih.gov/medlin eplus/news/fullstory_15145.html. United States Senate. (2004). 106th Congress 2D Session Amendments to the Medicare, Medicaid, and SCHIP Balanced Budget Refinement Act. Retrieved July 13, 2004, from www.senate.gov/~finance/bipa2000.pdf. Vungkhanching M., Sher K.J., Jackson K.M., Pa rra G.R. (2004). Relation of attachment style to family history of alcoholism a nd alcohol use disorder s in early adulthood. Drug Alcohol Dependency Jul 15; 75(1): 47-53.

PAGE 215

192 Walsh, C.W., & Ross, L.F. (2003). Are minority children underor overrepresented in pediatric research? Pediatrics 112(4), 890-895. Wang, N. E. Gisondi, M. A., Golzari, M., van der Vlugt, T. M., & Tuuli, M. (2003). Socioeconomic disparities are negatively associated with pediatric emergency department aftercare compliance. Academy of Emergency Medicine ,10(11), 1278-84. Weidmer, B. Brown, J., & Garcia, L. (1999). Translating the CAHPS 1.0 survey instruments into Spanish. Medical Care 37(3 suppl), MS89-MS96. Weinick, R. Jacobs, E., Stone, L., Ortega, A ., & Burstin, H. (2004). Hispanic healthcare disparities: challenging the myth of a monolithic Hispanic population. Medical Care 42(4), 313-320. Whooley MA, Avins AL, Miranda J, Browner WS. (1997). Case-finding instruments for depression: Two questions are as good as many. Journal of General Internal Medicine ; 12: 439-445 Willis, G.B. Cognitive interviewing: a how to guide. Short course presented at the 1999 meeting of the American Statistical Association Research Triangle Institute. Winn J.M. & Vesper V.G. (2005). Journal of Nuclear Medicine Technology Vol. 33(4):210-213. World Health Organization. (2004). Internationa l Statistical Classification of Diseases and Related Health Problems. Retrieved March 15, 2004, from www.who.int/whoint/whosis /ICD10/descript.htm. WPI Tech News. (2005). The wisdom of Ra madan: reasons behind religious fasting. Retrieved March 4, 2005, from http://www.wpi.edu/News/TechNews/ article.php?id=177. Zapata, L.B. (2005). Associations between maternal depression and child social competence and display of problem behaviors : a longitudinal inves tigation of direct, indirect and moderating effects Dissertation, University of South Florida,Tampa, FL. Zimmer, K.P., & Minkovitz, C.S. (2003).Maternal Depression:a n old problem that merits increased recognition by child healthcare practitioners. Pediatrics 15(6), 636-40. Zuniga de Nincio, M.L., Nader, P.R., Sawy er, M.H., DeGuire M., Prislin, R., & Elder, J.P. (2003). A prenatal prevention study to improve timelines of immunization initiation in Latino Infants. Journal of Community Health 28 (2), 151-165.

PAGE 216

193 Appendices

PAGE 217

194 Appendix A. Well-Child Visits American Academy of Pediatrics R ecommended Well-Child Care Schedule S S DENTAL REFERRAL ANTICIPATORY GUIDANCE Injury Prevention Violence Prevention Sleep Positioning Counseling Nutrition Counseling * * * * * PROCEDURES-PATIENTS AT RISK Lead Screening Tuberculin Tests Cholesterol Screening STD Screening Pelvic Exam PROCEDURES-GENERAL Hereditary/Metabolic Screening Immunization Hematocrit or Hemoglobin Urinalysis PHYSICAL EXAMINATION DEVELOPMENTAL/ BEHAVIORAL ASSESSMENT O O O S S S S S S S S S S S S S S S S S S S S O SENSORY SCREENING Vision Hearing MEASUREMENTS Height and Weight Had Circumference Blood Pressure HISTORY Initial/Interval 4y 3y 24mo 18mo 15mo 12mo 9mo 6mo 4mo 2mo By 1mo 2-4D NEWBORN PRENATAL AGE EARLY CHILDHOOD INFANCY KEY: = to be performed; = to be performed for patients at risk; S = subjective, by history; O = objective, by a standard testing method; = the range during which a service may be provided, with the dot indicating the preferred age. S S DENTAL REFERRAL ANTICIPATORY GUIDANCE Injury Prevention Violence Prevention Sleep Positioning Counseling Nutrition Counseling * * * * * PROCEDURES-PATIENTS AT RISK Lead Screening Tuberculin Tests Cholesterol Screening STD Screening Pelvic Exam PROCEDURES-GENERAL Hereditary/Metabolic Screening Immunization Hematocrit or Hemoglobin Urinalysis PHYSICAL EXAMINATION DEVELOPMENTAL/ BEHAVIORAL ASSESSMENT O O O S S S S S S S S S S S S S S S S S S S S O SENSORY SCREENING Vision Hearing MEASUREMENTS Height and Weight Had Circumference Blood Pressure HISTORY Initial/Interval 4y 3y 24mo 18mo 15mo 12mo 9mo 6mo 4mo 2mo By 1mo 2-4D NEWBORN PRENATAL AGE EARLY CHILDHOOD INFANCY S S DENTAL REFERRAL ANTICIPATORY GUIDANCE Injury Prevention Violence Prevention Sleep Positioning Counseling Nutrition Counseling * * * * * PROCEDURES-PATIENTS AT RISK Lead Screening Tuberculin Tests Cholesterol Screening STD Screening Pelvic Exam PROCEDURES-GENERAL Hereditary/Metabolic Screening Immunization Hematocrit or Hemoglobin Urinalysis PHYSICAL EXAMINATION DEVELOPMENTAL/ BEHAVIORAL ASSESSMENT O O O S S S S S S S S S S S S S S S S S S S S O SENSORY SCREENING Vision Hearing MEASUREMENTS Height and Weight Had Circumference Blood Pressure HISTORY Initial/Interval 4y 3y 24mo 18mo 15mo 12mo 9mo 6mo 4mo 2mo By 1mo 2-4D NEWBORN PRENATAL AGE EARLY CHILDHOOD INFANCY KEY: = to be performed; = to be performed for patients at risk; S = subjective, by history; O = objective, by a standard testing method; = the range during which a service may be provided, with the dot indicating the preferred age.

PAGE 218

195 Appendix B. Immunization Schedule

PAGE 219

196 Appendix C. Catch-up Immunization Schedule Catch-up schedule for children age 4 months through 6 years Minimum Interval Between Doses Dose 1 to Dose 2Dose 2 to Dose 3Dose 3 to Dose 4 4 wk 4 wk 4 wk 4 wk44 wk 4 wk 8 wk (and 16 wk after first dose)6 mo 4 wk26 mo1 Dose 1 (Minimum Age) DTaP (6 wk) IPV (6 wk) HepB3(birth) MMR (12 mo) Varicella (12 mo) Hib5(6 wk) 4 wk:if first dose given at age <12 mo 8 wk (as final dose):if first dose given at age 12-14 moNo further doses needed:if first dose given at age> 15 mo4 wk6: if current age <12 mo8 wk (as final dose)6:if current age> 12 mo and second dose given at age <15 moNo further doses needed:if previous dose given at age> 15 moPCV7: (6 wk) 8 wk (as final dose):this dose only necessary for children age 12 mo y who received 3 doses before age 12 mo 4 wk:if first dose given at age <12 mo and current age <24 mo8 wk (as final dose):if first dose given at age> 12 mo or current age 24-59 moNo further doses needed:for healthy children if first dose given at age> 24 mo4 wk:if current age <12 mo8 wk (as final dose):if current age> 12 moNo further doses needed:for healthy children if previous dose given at age> 24 mo8 wk (as final dose):this dose only necessary for children age 12 mo y who received 3 doses before age 12 moCatch-up schedule for children age 4 months through 6 years Minimum Interval Between Doses Dose 1 to Dose 2Dose 2 to Dose 3Dose 3 to Dose 4 4 wk 4 wk 4 wk 4 wk44 wk 4 wk 8 wk (and 16 wk after first dose)6 mo 4 wk26 mo1 Dose 1 (Minimum Age) DTaP (6 wk) IPV (6 wk) HepB3(birth) MMR (12 mo) Varicella (12 mo) Hib5(6 wk) 4 wk:if first dose given at age <12 mo 8 wk (as final dose):if first dose given at age 12-14 moNo further doses needed:if first dose given at age> 15 mo4 wk6: if current age <12 mo8 wk (as final dose)6:if current age> 12 mo and second dose given at age <15 moNo further doses needed:if previous dose given at age> 15 moPCV7: (6 wk) 8 wk (as final dose):this dose only necessary for children age 12 mo y who received 3 doses before age 12 mo 4 wk:if first dose given at age <12 mo and current age <24 mo8 wk (as final dose):if first dose given at age> 12 mo or current age 24-59 moNo further doses needed:for healthy children if first dose given at age> 24 mo4 wk:if current age <12 mo8 wk (as final dose):if current age> 12 moNo further doses needed:for healthy children if previous dose given at age> 24 mo8 wk (as final dose):this dose only necessary for children age 12 mo y who received 3 doses before age 12 mo Centers for Disease Control and Prevention ( 24H htt p ://www.cdc. g ov/ni p /recs/child-catchu p ppt )

PAGE 220

197 Appendix D.

PAGE 221

198 Appendix E.

PAGE 222

199

PAGE 223

200 Appendix F.

PAGE 224

201

PAGE 225

202 Appendix G.

PAGE 226

203 Appendix H. We are trying to learn how your childs clinic services can be made better. This includes learning more about the parents and what they want from the clinics. The clinics serve families with children of all ages who use all kinds of services. To start learning about the issues we want to look first at mothers of infants 1218 months old. Please answer the questions below. If your answers were all in the shaded area (an answer of No to the first question and Yes to all of the rest) you are someone we would like to talk to for about 30 minutes. $15.00 gift certificates are being given for your time. Screening Questions Does your child have a health problem, like asthma or a heart problem, that needs extra clinic visits? Yes No Changed order of Yes/No answers Are you the mother of the child? No Yes Do you usually take your child to clinic visits? No Yes Are you at least 18 years of age? No Yes Are you African American, White, or Hispanic/Latino? No Yes Were you born in the United States? No Yes Is your child between 12 and 18 months of age? No Yes Did your child get health care through Medicaid during the first year of life? No Yes Did your child go home with you when you left the hospital after your delivery? No Yes (For example: child spent no time in the neonatal intensive care unit [NICU]) Health Care Improvement Survey

PAGE 227

204 Appendix I.

PAGE 228

205 Appendix J.

PAGE 229

206

PAGE 230

207 Appendix K. Demographic Survey Instrument The survey begins by asking some questions about you, your general health and your childs health. The next questions ask you to tell me about your feelings and opinions about the health care services you and your child got from the pediatric clinic. The last set of questions asks about how you like people to have relationships with you. Do you have any questions before we start? [If yes] answer [If no] Okay, lets start. How did you hear about the study: _______________ What is your zipcode? _______________ How many children do you have? _____ What are their ages? _____________________ Are you currently pregnant? No Yes How long have you been pregnant? __________ What clinic does your 12-18 month old child use? ___________________________ What is the highest grade of school that you have finished? ___________________ [prompts] HS Diploma, GED, some college/no degree, AA/AS, BA/BS, BA/BS, MA/MS Technical Degree/Certificate: ___________________________________ How would you describe your self? White African American Hispanic (Country family came from) _________ What is your age? _______________ How would you describe your living or dating relationship? ______________________ For example, are you: Married Have a Live-in partner Divorced Widowed Single/Never Married Separated Other: ________ About how many hours per week do you work? _________ [I dont work] How many hours a week does your child spend in daycare? _____ [Does not attend daycare] Are you required to show proof of immunizations to your daycare provider? Yes No Before getting pregnant, did you have any health problems that made you go to the doctor/physicians assistant/nurse practitioner more often (ie, asthma, diabetes)? __________________________________________________________________________ Did you have any health problems when you were pregnant? For example (circle all yes) [Prompts] Anemia Cardiac Disease Lung Disease Diabetes High Blood Pressure Eclampsia Renal Disease RH Sensitization Uterine Bleeding Incompetent Cervix Other Problem: _____________________________________________________________________ During the birth of your baby did you have any health problems? (for example): C-section Excessive bleeding Seizure Labor>20 hours Breech Fetal Distress Other Complications: _____________________________________________________________________________ ID Number: Comments

PAGE 231

208 In general, how would you rate your overall health now? Poor 1 2 3 4 5 6 7 8 9 10 Excellent Did you choose to: breastfeed, bottle-feed or did you do both? If breastfed, how long ? ______________ During the past month, have you often been bothered by feeling down, depressed or hopeless? Yes No During the past month, have you often been bothered by little interest or pleasure in doing things? Yes No About how many times have you been to your doctor/nurse practitioner/physicians assistant since you had your baby? _____ Why :______________________________________________________________ How would you rate your feelings about doctors and other health care providers? Enjoy going to the doctor Neutral about doctors Dislike going to the doctor 1 2 3 4 5 6 7 8 9 10 When did you plan to have this baby? Earlier At the time you did Later Other________________ In general, how would you rate your childs overall health now? Poor 1 2 3 4 5 6 7 8 9 10 Excellent Does your child have any problems sleeping? No Occasionally Frequently Explain : _______________________________________________________________________ Have you received any agency services since having your baby such as WIC or Healthy Start? No Yes : List :________________________________________________________________ Is your childs doctor/nurse practitioner/physicians assistant the same race/ethnicity as your race/ethnicity? Yes No Did your prenatal care provider help you pick a pediatric provider before your babys birth? Yes No Did you get your prenatal care at the same clinic office that your child gets care? Yes No How do you get to your childs health care appointments (drive, bus, ride from friend, taxi)? ______________ Is it difficult for you to get to your childs health care visits? Yes No Why? ___________________________ Besides clinic staff scheduling visits, does anyone else decide when you take your child to health care visits? Yes No Who/How? ____________________________________________________________________ Does anyone usually go with you to your childs health care visits? Yes No Who? _______________________ Does anyone else take your child to health care visits when you cant ? Yes No Who? ___________________ Has your child received care from any other health care providers? Yes No For What reason? _______________________________________________________________________ How many times did you take your child to the other health care providers? _________ That is the end of this group of questions. Comments

PAGE 232

209 Appendix L. CAHPS 2.0 Child Core Questions The next group of questions refers to your experiences with your childs health care such as your childs doctor or nurse practitioner, the clinic office, and any specialists you have tried to take your child to see. Do not include care your child got when you stayed overnight in a hospital. Pediatric Office With the choices your childs health plan gave you, how much of a problem, if any, was it to get a personal doctor or nurse for your child you are happy with? A big problem A small problem Not a problem My child didnt get a choice Do you have one person you think of as your childs personal doctor or nurse? If your child has more than one personal doctor or nurse, choose the person your child sees most often. Yes No We want to know your rating of your childs personal doctor or nurse If your child has more than one personal doctor or nurse, choose the person your child sees most often. Use any number from 0 to 10 where 0 is the worst personal doctor or nurse possible, and 10 is the best personal doctor or nurse possible. How would you rate your childs personal doctor or nurse now? 0 1 2 3 4 5 6 7 8 9 10 Did conversations/education from your childs health care provider make you feel like you were able to manage your childs care? Yes No Experience with Health Care Provider Office Did you call a doctors office or clinic during regular office hours to get help or advice for your child? Yes No When you called during regular office hours, how often did you get the help or advice you needed for your child? Never Sometimes Usually Always I didnt call for help or advice for my child during regular office hours in the last 12 months A health provider could be a general doctor, a specialist doctor, a nurse practitioner, a physician assistant, a nurse, or anyone else you would see for health care. Did you make any appointments for your child with a doctor or other health provider for regular or routine health care? Yes No How often did your child get an appointment for regular or routine health care as soon as you wanted? Never Sometimes Usually Always My child didnt need an appointment for regular or routine care in the last 12 months. Did your child have an illness or injury that needed care right away from a doctors office, clinic or emergency room? Yes No When your child needed care right away for an illness or injury, how often did your child get care as soon as you wanted? Never Sometimes Usually Always My child didnt need care right away for an illness or injury in the last 12 months. Not counting times you went to an emergency room, how many times did your child go to a doctors office or clinic? None 1 2 3 4 5 to 9 10 or more Worst personal doctor or nurse Best personal doctor or nurse My child didnt have a personal doctor or nurse.

PAGE 233

210 How much of a problem, if any, was it to get the care for your child that you or your doctor believed necessary? A big problem A small problem Not a problem My child had no visits in the last 12 months. How much of a problem, if any, were delays in your childs health care while you waited for approval from your childs health plan? A big problem A small problem Not a problem My child had no visits in the last 12 months. In the last 12 months, how often did your child wait in the doctors office or clinic more than 15 minutes past the appointment time to see the person your child went to see? Never Sometimes Usually Always My child had no visits in the last 12 months. How often did office staff at your childs doctors office or clinic treat you with courtesy and respect ? Never Sometimes Usually Always My child had no visits in the last 12 months. How often were office staff at your childs doctors office as helpful as you thought they should be? Never Sometimes Usually Always My child had no visits in the last 12 months. How often did your childs doctors or other health care providers listen carefully to you ? Never Sometimes Usually Always My child had no visits in the last 12 months. How often did your childs doctors or other health providers explain things in a way you could understand? Never Sometimes Usually Always My child had no visits in the last 12 months. How often did your childs doctors or other health providers show respect for what you had to say ? Never Sometimes Usually Always My child had no visits in the last 12 months. How often did your childs doctors or other health care providers spend enough time with your child? Never Sometimes Usually Always My child had no visits in the last 12 months. We want to know your rating of all your childs health care from all doctors and other providers Use any number from 0 to 10 where 0 is the worst health care possible, and 10 is the best health care possible. How would you rate all your childs health care? 0 1 2 3 4 5 6 7 8 9 10 Specialists Office When you answer the next questions, do not include dental visits. Specialists are doctors like surgeons, heart doctors, allergy doctors, skin doctors, and others who specialize in one area of health care. Did you or a doctor think your child needed to see a specialist? Yes No How much of a problem, if any, was it to get a referral to a specialist that your child needed to see? A big problem A small problem Not a problem My child did not see a specialist. Did your child see a specialist? Yes No We want to know your rating of the specialist your child saw most often including a personal doctor if he or she was a specialist. Use any number from 0 to 10 where 0 is the worst specialist possible, and 10 is the best specialist possible. How would you rate your childs specialist now? 0 1 2 3 4 5 6 7 8 9 10 Was the specialist your child saw most often the same doctor as your childs personal doctor? Yes No My child did not have a personal doctor or my child didnt see a specialist in the last 12 months. Worst health care possible Best health care possible My child hadnovisits My child did not see a specialist in the last 12 months. Comments

PAGE 234

211 Appendix M. Relationship Scales QuestionnaireUsing the same one to seven scale of one for not at all like me to a seven for very much like me, please rate how much you believe each sentence best describes your feelings about close relationships. Not at all like me Somewhat like me Very much like me 1. I find it difficult to depend on other people. 1 2 3 4 5 6 7 2. It is very important to me to feel independent. 1 2 3 4 5 6 7 3. I find it easy to get emotionally close to others. 1 2 3 4 5 6 7 4. I want to merge completely with another person. 1 2 3 4 5 6 7 5. I worry that I will be hurt if I allow myself to become too close to others. 1 2 3 4 5 6 7 Describe what you mean by close: 6. I am comfortable without close emotional relationships. 1 2 3 4 5 6 7 7. I am not sure that I can always depend on others to be there when I need them. 1 2 3 4 5 6 7 8. I want to be completely emotionally intimate with others. 1 2 3 4 5 6 7 9. I worry about being alone. 1 2 3 4 5 6 7 10. I am comfortable depending on other people. 1 2 3 4 5 6 7 11. I often worry that romantic partners don't really love me. 1 2 3 4 5 6 7 12. I find it difficult to trust others completely. 1 2 3 4 5 6 7 13. I worry about others getting too close to me. 1 2 3 4 5 6 7 14. I want emotionally close relationships. 1 2 3 4 5 6 7 Describe what close means to you: 15. I am comfortable having other people depend on me. 1 2 3 4 5 6 7 16. I worry that others don't value me as much as I value them. 1 2 3 4 5 6 7 17. People are never there when you need them. 1 2 3 4 5 6 7 18. My desire to merge completely sometimes scares people away. 1 2 3 4 5 6 7 19. It is very important to me to feel self-sufficient. 1 2 3 4 5 6 7 20. I am nervous when anyone gets too close to me. 1 2 3 4 5 6 7 21. I often worry romantic partners won't want to stay with me. 1 2 3 4 5 6 7 22. I prefer not to have other people depend on me. 1 2 3 4 5 6 7 23. I worry about being abandoned. 1 2 3 4 5 6 7 24. I am somewhat uncomfortable being close to others. 1 2 3 4 5 6 7 25. I find that others are reluctant to get as close as I would like. 1 2 3 4 5 6 7 26. I prefer not to depend on others. 1 2 3 4 5 6 7 27. I know that others will be there when I need them. 1 2 3 4 5 6 7 28. I worry about having others not accept me. 1 2 3 4 5 6 7 29. Romantic partners often want me to be closer than I feel comfortable being. 1 2 3 4 5 6 7 30. I find it relatively easy to get close to others. 1 2 3 4 5 6 7

PAGE 235

212 Before we end, I want to ask if you have any questions for me about the study? [add comments] Finally, I would like to thank you for helping us with by answering this survey. We are hopeful that what we learn in the next few months will help make services more friendly for the people who use them. Have a nice day.

PAGE 236

213 Child Medical Record Abstraction Questions Date of Birth: Bi rth weight (grams): Gender: Gestational age at birth (weeks): Well Child Care Visits (infants age at visit)/date Birth Visit: 2-4 Day Visit: 1 Month Visit: 2 Month Visit: 4 Month Visit: 6 Month Visit: 9 Month Visit: 12 Month Visit: Other Medical Office Visits: Date Reason Provider Dates of Immunizations or age of child when immunized: Diptheria, Tetanus, Pertussis (DTP): (3) Haemophilus influenza type b (Hib): (3) Pneumococcal: (3) Inactivated Poliovirus (IPV): (2 or 3) Hepatitis B: (2 or 3) Hospitalizations/Emergency Room Visits: Visit 1: Date: Medical provider referred to hospital: Y N Reason: Visit 2: Date: Medical provider referred to hospital: Y N Reason: Has this child ever been identified as Failure to thrive? Y N Would you rate this childs propensity for accidents/injuries as being: High Medium Low/None noted Does this child have a chronic illness that requires more frequent medical visits (congenital anomalies, asthma, sickle cell disease, or a heart condition)? Y N Appendix N.

PAGE 237

214 Maternal Medical Record Abstraction Questions These questions target the prenatal care se rvices received during the gestational period for the target child. Gestational age (in weeks) at first prenatal care visit : ___________ Number of prenatal care visits : __________ ID Number:

PAGE 238

215 Appendix O. Recruitment of Study Sample A modification of the Dillman approach was used for this study (Dillman, 2000). A recruitment letter and a scre ening form for the study was sent to mothers of infants 1218 months of age who received pediatric heal th care through Medicaid. Attempts were made to personalize these letters, such as by signing them individually and providing postage stamps rather than metered mail mark s on the letters. Letters were sorted by zip code and mailed in bundles of approximately 500 per week so that interviews could be scheduled without delay. To balance recr uitment across regional and racial/ethnic populations a second letter was mailed to all (233) Hispanic women and 717 White nonHispanic or Black non-Hispanic women in Pinellas County. Approximately one month after the first letter was sent, a second letter was mailed to non-respondents in an attempt to increase the response rate. Women who contacted resear chers were provided with an overview of the study. If the woman was interested in participati ng further, an appointment was scheduled to interview her at a convenient location in the community. Th e interview began with the researcher reading the informed consent to the mother, obtaining consent, and having the mother sign release of medical information forms. At the completion of the interview, participants received a $15.00 gift certificate. It was believed that the gift certificate would act as an incentive for pa rticipation without being so gr eat that participants would feel economically coerced into participating.

PAGE 239

216 Study recruitment and data collectio n occurred from August 1, 2005 through November 30, 2005, and included mailing invitations to participate, disseminating fliers, receiving referrals from local health care providers and from word-of-mouth among friends. Although a variety of recruitment me thods were attempted, the most successful was the mailing of 4,218 letters to potential participants inviting them to become involved in the study (Table 43). Due to the mobility of this population and the age of the Medicaid mailing list being nearly one y ear old, one thousand a nd sixty (25.1%) letters were returned to researchers as non-deliverable. These rates were higher for Pinellas County (409, 30.5%) than for Hillsborough County (651, 22.6%). Of the remaining 3,158 letters, 100 (3.2%) ultimately le d to interviews. There were an additional 18 women who called the researcher, scheduled appointmen ts, were not home fo r the interview and would not return follow-up phone calls. An additional four women scheduled appointments only to cancel them soon afte r, indicating that th eir husband or boyfriend would not allow them to be interviewed. Table 43 Recruitment of Study Participants (N=126) Hillsborough Pinellas Total N %N % N Letters First mailing 2,87668.21,34231.8 4,218 Return to Sender (1st Mail) 65122.640930.5 1,060 Repeat Mailing 0950 950 Fliers/Posters Number of Locations 035 35 Recruited Letters 57574343 100 Friends 1062.5637.5 16 Professional Referrals 212.51487.5 16 Fliers and Posters 007100 7 Total Number Recruited 6949.67050.4 139

PAGE 240

217 The majority of participants recru ited through the first mailing were from Hillsborough County. To increase the number of respondents, an additional 950 second mailings were sent to resident s of Pinellas County. It is no t possible to separate out the exact number of individuals recruited thr ough the first or the second mailing due to lengthy lag time between receip t of letters and calls to researchers. However, it is believed that approximately al l but ten women were recrui ted from the first mailing. Additionally, sixteen mothers were r ecruited through refe rrals by health professionals such as clinic staff, Healthy Start caseworkers and h ealth education staff (Table 43). Sixteen other part icipants were recruited thr ough friends who had learned of the study and forwarded the information. Th e least successful recruitment method was the placement of posters and fliers throughout the communit y. Thirty-five posters with tear-tab phone numbers were placed in la undromats as well as being posted in six pediatric clinics, four social service provider locations a nd two community centers. Seven women were recruited through these posters, th ree of whom were dropped from the study because it was later determined that their children were ineligible. The original intent of the recruitment was to obtain a stratified sample from an equal number of respondents with in the three racial and ethn ic groups targeted. However, initial recruitment efforts elic ited only 14 mothers of Hispan ic or Latino ethnicity. To compensate for this low response, specific recruitment efforts were targeted at the Hispanic population. These additional efforts included mailing 233 of the 950 letters included in the second mailing to all Hi spanic mothers in Pinellas County on the Medicaid mailing list. Professional staff working in areas with high populations of Hispanic clients, such as the Clearwater location of the Pine llas County Health

PAGE 241

218 Department and the Genesis Health Park C linic in Hillsborough County, were also asked to help recruit more Hispanic mothers to the study. These efforts resulted in an additional three Hispanic women being referred by hea lth professionals. Prof essional staff noted that nearly all of their Hispanic Medicaid clients did not speak English, excluding them from the study. This study also highlighted recruitment and generalizability issues for researchers. Recruitment approaches have been limited by HIPAA regulations, ma king it difficult to recruit for small studies such as this one. Th ere is an increasing te ndency to use internetbased surveys because they can provide anonymity and data entry is done by the respondent, decreasing the workload for res earchers. However, underserved populations often do not have computers an d/or internet access prompti ng researchers to concentrate on traditional approaches such as the Medica id mailing list used for this study. During the interviews a number of women indicated that they were either tired of receiving solicitations for surveys especially when th ey were not offered any compensation for their time. Other women indicated that they enjoy completing surveys and do so as often as possible. Future studies should make attemp ts to deviate from these types of traditional recruitment paths so that the base of literatu re does not become restricted to only a small, unique portion of the population of interest. Another recruitment issue is the involvement of mi nority populations. The is especially important for the Hispanic populat ion that represents and even-increasing portion of the population. With this growth comes the need to understand the unique context from which Hispanics operate. In rega rd to health issues, Hispanic individuals often have better outcomes, such as in infant birth weight (Rosenberg, Raggio &

PAGE 242

219 Chiasson, 2005). While in other cases, they are experiencing increased he alth risk such as in diabetes (Shen, Tymkow & MacMullen, 2005) It takes years, or even decades, to build a body on knowledge on a topic. Focu sing more efforts now may provide the understanding needed to prevent health di sparities from devel oping in the future.

PAGE 243

220 Appendix P. Maternal Health History Most (106, 84.1%) mothers did not report experiencing health problems before, during, or after their pregnancy (Table 44). For women who experienced health issues prior to getting pregnant, the most common i ssues were asthma (8) and hypertension (3). Epilepsy, allergies, diabetes, ova rian cysts were also noted. The number of children in the home was positively associated with having a pre-pregnancy health issues (p<.05). Table 44 Pre-Pregnancy Health History (N=126) N% Mother Experienced No Pre-Pregnancy Health Problems 10684.1 Mother Experienced Any Pre-Pr egnancy Health Problem(s) 2015.9 Type of Pre-Pregnancy H ealth Problems (duplicated count): Asthma 86.3 Hypertension 32.4 Epilepsy 21.6 Other (allergy, diabetes, ovarian cysts, etc.) 43.2 Mean SD Significance Pattern of Finding By log (Number of Children) .013 No Health Problems .54.55 Health Problems .88.48 Mothers with multiple children had more pre-pregnancy health issues. Once pregnant, 47 (37.3%) were diagnosed with health issues that could impact the mother or child (Table 45). The most comm on issues were hyperten sion (15), being at high risk for poor pregnancy outcomes (8) such as previous poor perinatal outcomes,

PAGE 244

221 gestational diabetes (6), prem ature labor (3), anemia (3), and gall bladder problems (3). Placenta previa (2), sickle cell trait (2) a nd other conditions (14) were also reported. Table 45 Pregnancy Health History (N=126) N % No Pregnancy-related Health Problems 79 62.7 Experienced pregnancy-related health problems 47 37.3 Health Problems Iden tified (duplicated count): Hypertension 15 11.9 Identified high ri sk (previous miscarry, etc.) 8 7.1 Gestational Diabetes 6 4.8 Premature labor 3 2.4 Anemia 3 2.4 Gall bladder problems 3 2.4 Placenta Previa 2 1.6 Sickle Cell Trait 2 1.6 Other 14 11.1 There were 44 (34.9%) women who reported so me type of added risk during their target childs delivery (Table 46). Three ch ildren (2.4%) were breec h, three (2.4%) were reported to have cord problems, three (2.4 %) had infections, and three (2.4%) were in fetal distress. Additionally, thirty-four (27.0%) women utili zed a cesarean section for a variety of reasons.

PAGE 245

222 Table 46 Maternal Health Issues During Delivery (N=126) N % Mother Experienced No Health Problems During Labor 82 65.1 Mother Experienced health problem(s) during labor 44 34.9 Health Problems Iden tified (duplicated count): C-Section 34 27.0 Breech 3 2.4 Cord Problems 3 2.4 Infection/Meconium Aspiration 3 2.4 Fetal Distress 3 2.4 Induced-Hypertension 3 2.4 Infant Large Size 2 1.6 Since the birth of their target child, th e number of health care visits attended by mothers ranged from zero to more than 30 (Table 47). Separating mothers depending on their subsequent pregnancies, the average number of visits ranged from 4.2 for mothers with no subsequent pregnancies, 12.6 visits for mothers currently pregnant and 17.7 visits for mother giving birth to at least one more child. Table 47 Postnatal Health Care Visits (N=126) MeanSDMin.Max.KurtosisSkew Maternal Postnatal Health Visits5.96.50301.311.50 No Subsequent Pregnancies 4.25.10308.02.7 Pregnant at Interview 12.66.2420-1.5-0.2 Subsequent Birth 17.73.61220-0.9-1.1

PAGE 246

223 Appendix Q. Pediatric Health Care Issues Ideally, the relationship between a mother her child and their pediatrician should start prenatally. One way to become connect ed with a pediatrici an is through guidance from the mothers prenatal care provider. One in five (26, 20.6%) of the women interviewed indicated that th eir prenatal provider helped them choose a pediatrician (Table 48). Additionally, 14 (11.1%) women r eceived prenatal care at the same health care facility that thei r child currently attends. Mothers who did not have a live-in partner or husband (9/72, 31.5%) were recommended a pedi atrician more often than those with a partner (17/54, 12.5%) (p<.05). Table 48 Prenatal Care Provider Helped Mother Choose a Pediatrician Prior to Delivery (N=126) N % Prenatal Provider Helped Pick Pediatrician 26 20.6 Prenatal and Pediatric Care at Same Clinic 14 11.1 N % ChiSquare(df) Pattern of Finding By Marital Status Married or Live-in Partner 9 12.5 .013(1) Single/Widowed/Divorced/ Separated 17 31.5 Mothers without livein partners received more help picking a pediatrician. Racial and ethnic concordance between families and providers occurred in more than one third (50, 39.7%) of women intervie wed (Table 49). There were significant

PAGE 247

224 differences in racial concordance among mothers of different racial and ethnic backgrounds. Hispanic mothers (62.5%) were mo re likely (p<.01) to see a Hispanic provider than White non-Hispanic (45.9%) or Black non-Hispanic (24.5%) mothers. Table 49 Racial/Ethnic Concordance Between Pedi atric Provider and Mother (N=126) N % ChiSquare(df)Pattern of Finding Pediatrician Race Concordance 50 39.7 By Single/Widow/Divorced/Separated 17 31.5 By Maternal Race/Ethnicity White Non-Hispanic (N=65) 28 45.9 .01(2) Black Non-Hispanic (N=56) 12 24.5 Hispanic (N=17) 10 62.5 Hispanic mothers most likely to see Hispanic provider. Black mothers least likely to see Black provider. In addition to selecting a pediatric hea lth care provider, there is a need for reliable, convenient transpor tation to the visits. Most (81, 64.3%) of the mothers interviewed drove to visits while an add itional 24 (19.1%) were given rides by family members or friends (Table 50). Few mothers (11, 8.7%) reported taking the bus, walking to visits (9, 7.1%) or taking cabs (2, 1.6%). Women who experienced health problems during the target pregnancy (68.4%) were more lik ely (p<.01) to report that it was hard to get to the infants pediatric health care visi ts than mothers who di d not experience health problems during the pr egnancy (31.6%).

PAGE 248

225 Table 50 Pediatric Health Care Transportation Issues (N=126) N % No Transportation Difficulties 107 84.9 Transportation is difficult 19 15.1 N%ChiSquare(df) Pattern of Finding By Health Issues During Pregnancy No Problems 67.6.01(1) Health Problems 1327.7 Mothers with health issues during pregnancy were more likely to report transportation problems. N % Mode of Transportation Drive 81 64.3 Ride 24 19.1 Bus 11 8.7 Walk 9 7.1 Medicaid Cab 2 1.6 Others control when visits occur 6 4.8 Others go to visit with mother and child 65 51.6 Others take child to doctor 49 38.9 Note (Duplicated Counts) One in seven of the women interviewed reported difficulties getting to health care visits due to reliance on othe rs while six women (4.8%) repor ted feeling like the person giving the ride has control over when visits occur (Table 50). More than half (51.6%) of mothers reported being accompanied by another person to health care visits regardless of whether that additional person was providing the transportation to the visit. Furthermore, although 38.9% of the mothers reported having so meone else to take their child at least once, mothers reported that they preferred not to have others take their child to health care visits in their place. When it was nece ssary to send the child with someone else, it was usually the father of the baby (27) or one of the grandmothers (26).

PAGE 249

226 Appendix R. Maternal Interaction Style: Re lationship Scales Questionnaire Table 51 includes the responses to sp ecific Relationship Scales Questions used to calculate the overall interaction rating. Looking at specific items regardless of their attachment rating, the i ssue with the strongest averag e rating was the need for the mother to feel independent (6.2) followed by th e need to be self-sufficient (5.8). The least characteristic statements included a desire to merge completely that scared others away (2.5), a concern that romantic partners won t want to stay with me (2.5), and worrying about being abandoned (2.5).

PAGE 250

227 Table 51 Relationship Scales Questionnaire (N=126) Mean SD 1. I find it difficult to depend on other people. 4.3 2.0 2. It is very important to me to feel independent. 6.2 1.2 3. I find it easy to get emotionally close to others. 4.6 1.8 4. I want to merge completely with another person. 4.1 1.9 5. I worry I will be hurt if I allow myse lf to become too close to others. 3.9 2.0 6. I am comfortable without clos e emotional relationships. 3.5 2.0 7. I am not sure that I can always de pend on others to be there when I need them. 4.4 1.9 8. I want to be completely emotionally intimate with others. 3.6 1.6 9. I worry about being alone. 3.3 2.1 10. I am comfortable depend ing on other people. 2.9 1.7 11. I often worry that romantic pa rtners don't really love me. 2.7 2.0 12. I find it difficult to trust others completely. 4.2 2.0 13. I worry about others getting too close to me. 2.9 1.8 14. I want emotionally close relationships. 4.4 1.8 15. I am comfortable having ot her people depend on me. 5.0 1.8 16. I worry that others don't value me as much as I value them. 3.9 2.0 17. People are never there when you need them. 3.4 1.9 18. My desire to merge completely sometimes scares people away. 2.5 1.7 19. It is very important to me to feel self-sufficient. 5.8 1.5 20. I am nervous when anyone gets too close to me. 2.7 1.7 21. I often worry romantic partners won't want to stay with me. 2.5 1.9 22. I prefer not to have other people depend on me. 2.9 1.8 23. I worry about being abandoned. 2.5 2.0 24. I am somewhat uncomfortable being close to others. 2.8 1.7 25. I find that others are reluctant to get as clos e as I would like. 3.0 1.7 26. I prefer not to depend on others. 4.7 2.1 27. I know that others will be there when I need them. 4.5 1.8 28. I worry about having others not accept me. 2.9 1.9 29. Romantic partners often want me to be closer than I feel comfortable being. 2.7 1.8 30. I find it relatively easy to get close to others. 4.5 1.9

PAGE 251

228 Appendix S. Reported Experiences of Care In addition to the standard CAHPS ques tions, conversations with the mothers identified specific examples of issues behind their ratings of their experiences of care. Some mothers indicated that demographic ch aracteristics were an issue for them. For example, two mothers reported that their initial pediatrician spoke limited English prompting them to change. Another mother not ed that she took her child to a specialist and felt uncomfortable that the doctor was older. She had a number of age-related concerns such as the quality of his eye si ght and how current he was with the best treatments available. Conversely, another moth er indicated that she liked having an older doctor in a small practice rather than a less-experi enced provider. Beyond physical characteristics, mothers id entified a number of behaviors that influenced their experiences with pediatrician s. Mothers appreciated accessibility to their pediatricians with most (92, 73.0%) reporting being able to get regular or routine appointments as soon as they wanted a nd (92 of 113, 81.4%) receiving immediate care for an illness or injury when needed (Table 52). One woman noted that her pediatrician is always there when you need her. Another noted that he always sees my child no matter how many times I call.

PAGE 252

229 Table 52 Consumer Assessment of Health Plans Survey Questions Pediatric Office (N=126) Never Sometimes Usually Always NA N%N% N %N%N% Received the needed advice when calling doctors office 44.01616.0 21 21.05959.02620.6 Received well child appointments as soon as she wanted: 10.81411.1 19 15.19273.000.0 Received care when wanted for ill child 32.743.5 14 12.49281.41310.3 Waited >15 minutes past the appointment time 2822.24233.3 23 18.33326.2 Office staff treated mother with courtesy and respect 10.832.4 24 19.09877.8 Office staff as helpful as mother thought they should be 21.62217.5 26 20.67660.3 Providers listen carefully to mother 10.8129.5 37 29.47660.3 Providers explain things so mother could understand 00.0118.7 14 11.110180.2 Providers showed respect for what mother had to say 10.886.3 15 12.710180.2 Health care providers spent enough time with child 21.61814.3 24 19.08265.1 Problem Big Small No NA N % N % N % N % Problem getting care believed necessary 1 0.8 1 0.8 124 98.4 0 0.0 Problem waiting for plan approval 1 0.8 2 1.6 123 97.6 0 0.0 Mean SD Min Max Kurtosis Skew Rating of Childs Provider Office 8.5 1.8 0 10 4.5 -1.78 N % Rating <7 13 10.3 Rating 7,8 36 28.6 Rating 9,10 77 61.1

PAGE 253

230 Thoroughness of the visit was noted as an important feature. Mothers reported that they liked it when providers took the ti me, going step by step to problem solve and letting mothers explain issues before offeri ng a solution. Mothers also appreciated when the doctors took the time to explain existing or potential health problems, what to expect and how to respond to symptoms as they arise. Some of the positive descriptions mothers provided of their pediatricians included thos e that were knowledgeable, friendly, non-judgmental, and laid back. The open-mindedness demonstrated by many providers helped mothers to be more involved in decisions regarding their ch ilds care, allowing them to work more like a team. Mothers also liked feeling they could ask any question without the provider making her feel stupid, even when aski ng what now seems like common sense questions. By listening carefully to the con cerns and needs of the mother, the provider doesn't suggest things to mom that she wouldn't like. One mother appreciated the pediatricians willingness to try new things if the first course of treatment didn't work. In contrast to these positive attributes, some mothers reported feeling that their pediatricians didn't try to explain things, they told you [mothers] what to do and wouldn't answer questions. Moth ers also reported receiving comments in a negative tone such as "you should have called me first before you went to the ER." Two mothers indicated that some providers were rude, one mother said there was no personal one-onone care, and one mother re ported that the doct or looked down on you. In one case a doctor repeatedly pushed the mother to change HMOs but di d not explain why. An issue that had praises and critics on either side was in regard to how aggressive to be when diagnos ing and treating the ch ilds health issues. For example, one

PAGE 254

231 mother indicated she liked that her pediatrician was not an al armist in regard to running tests, another believed the pe diatrician was too quick to se nd her child to a specialist and a third mother liked the providers precauti onary approach (getting labs, x-rays) to helping diagnose and treat her child. Reported Experiences of Care Provider Office Questions. Being able to get health care information quickly is important to mothers. Most mothers (98.4%) reported no problems getting care they believed was necessa ry for their child or with having to wait for their insurance plan approval (97.6%) (T able 52). These rates are much higher than data in the National CAHPS Benchmarking Da tabase (79% and 59% respectively). The selection criteria that exclude chronically ill children and the younger age of the children most likely account for these difference. Once access to care has been established, when asked how often mothers called to get help or advice for their child, 59 (59.0 %) reported always receiving the desired advice, 21 (21.0%) reported usua lly receiving it, and 20 (20.0 %) reported they never or only sometimes received the advice (Table 52 ). These rates are somewhat lower than those reported from the National CAHPS Benchmarking Database (62% always, 23% Usually, and 15% Never or Sometimes). The ability to make well-c hild care appointments in a timely manner is important to mothers. Most (92, 73.0%) women reported always getting them as soon as they wanted, 19 (15.1%) said they usually get them, and 15 (11.9%) reported never or only sometimes getting appointments when they wanted them (Table 52). Similarly, of the 113 women who reported needing to schedule a sick visit, 92 (81.4%) repo rted getting one as

PAGE 255

232 soon as they wanted while 14 (12.4%) reported they usually get them, and seven (6.2%) reporting that they never or only sometimes get appointments as soon as they wanted. Once an appointment is has been made, th e next step is to actually receive the medical care. The majority (70, 55.5%) of mother s reported rarely, if ever, having to wait more than 15 minutes past their scheduled appointment time as comp ared to the national average of 49% (Table 52). Even so, long waiting periods are not uncommon. One quarter (26.2%) of all mothers in the study repo rted that they always waited more than 15 minutes past their scheduled appointment a nd 23 (18.3%) reported they usually waited. Interactions with Staff. The rapport that develops be tween a mother and health care staff is important. In the case of mo thers in the study, most (98, 77.8%) reported staff always treated them with courtesy and respect while 24 (19.0%) were usually respected and only four (3.2%) mo thers reported being treated with little or no courtesy and respect (Table 52). Even though mothers re ported providers showed respect for what mother had to say, they were less likely to repor t that providers listen ed to them carefully. In all, 76 (60.3%) of mothers re ported that providers always li stened to them carefully, 37 (29.4%) usually listened, and 13 (10.3%) some times or never respected their words. Nationally, 70% reported provide rs showed respect for what mothers said, 22% usually showed respect, and 8% sometime s or never showed respect. Health Care providers took care in e xplaining things in a way mothers could understand (101, 80.2% always, 14, 11.1% usua lly, and 11, 8.7% sometimes/never) (Table 52). Two-thirds ( 82, 65.1%) of mothers reported th at providers always spent enough time with their child while 24 (19.0%) usually did, and 20 (15.9%) only

PAGE 256

233 sometimes or never spent the time. Finall y, 76 (60.3%) of mothers thought staff were always as helpful as they should be wh ile 26 (20.6%) were usually thoughtful and 24 (19.1%) were either sometimes or never as thoughtful as they thought staff should be. In addition to the standard questions, ope n-ended responses identified other issues that were important to mothers. Several mo thers (7) reported that when they called the providers office, they were put on hold too l ong, that it took to long to return phone calls and that they did not always like the answer the clinics provided. One mother reported becoming frustrated when the clinic woul d not answer questions over the phone and would tell her to come to the clinic or go to the emergency department. Another respondent indicated that her pediatricians o ffice doesnt get her in immediately so she goes to the emergency department first, then her pediatrician has to see her child the next day. Provider Behavior and Attitudes. There were a number of provider behaviors and attitudes that appealed to mothers in the study. For example, five mothers reported liking it when providers were attentive and listen ed to them. One liked that her provider proactively provided explanations for issues and what to expect. Additionally, three indicated they liked it when the office cal led them to follow-up from an emergency department visit or an illness. Mothers also indicated that they liked it when their pediatricians were responsive to needs, talked to them like they were friends, made the mothers feel comfortable, and asked how thei r child was doing. One mo ther reported that her pediatrician was obviously very attached to her children and tr eated them like they were the pediatricians own children.

PAGE 257

234 In addition to the positive feedback, moth ers also freely shared issues that bothered them. Three mothers reported that they did not like it when they felt staff were not listening to them carefully and five wo men reported feeling rushed. Mothers also did not like it when they believed providers were not compassionate or not interested in their child. One mother noted her pediatrician didn t seem like he wanted to be there. To further illustrate this perceived lack of in terest, one mother indicated that after seven months of seeing her child, th e pediatrician still did not know the childs name. Another mother noted that nurses lacked concern re garding sick children and they "take their sweet time" doing their job. Finally, an eight m onth pregnant mother who just moved to a new city, went to the health department to receive prenatal care and was told it would take five weeks to get an appointment. The sa me woman reported that staff were rude to her with their arms crossed in front of them as they talked. She went to another provider that used midwives and was very pleased with her care. Diagnostic problems that left a strong and lasting impression also were experienced by some mothers. Two women reported having th eir children diagnosed with serious medical concerns (spina bifida, tw isted testicle) and were referred to the emergency departments only to learn that neith er diagnosis was correct. Furthermore, in the case of the twisted testicle, the mother indicated that the doctors office never contacted the mother regarding the test results. Reported Experiences of Care Specialist Questions. Due, in part, to the restricted inclusion and exclusion crit eria, few infants (23, 18.3%) s ought care from a specialist (Table 53). Most (22, 95.7%) mothers reported no problem getting a referral to see a

PAGE 258

235 specialist. Of the reasons for seeing a specialist, the majority were for minor issues such as hearing screens and dermatology issues. Of those who did see a spec ialist, their ratings averaged 8.7 based on a continuum wher e zero is poor and ten is excellent. Table 53 Consumer Assessment of Health Plans Sur vey Questions Specialist Office (N=126) Problem Big Small No Not Needed N % N% N % N % Problem getting a referral to see a specialist for child 1 4.3% 0 0.0% 2295.7 % 103 81.7 % Yes No N % N % Child saw a specialist 23 18.3% 103 81.7% Specialist was also child s doctor. 1 4.3% 22 95.7% Mean SD Min Max Kurtosis Skew Rating of specialist seen most 8.7 1.8 4 10 1.67 -1.52 Office Environment. In addition to the question re garding staff behaviors and attitudes, mothers provided open-ended info rmation regarding the environment of the offices that impacted their feelings about h ealth care offices. Several mothers comments highlighted the dramatic difference in the qua lity and character of clinic facilities. Visits made by researchers to 53 particip ating pediatric clinic s in Pinellas and Hillsborough counties allowed researchers to get a better understanding of these differences. Some offices were barren, having solid color walls, such as a light grey-blue, plastic chairs and nothing else in the room. On the other e nd of the spectrum, one office had separate sick rooms, well-child rooms, infant rooms and even a small room with a rocker for a mother to breastf eed her child. The more appealin g offices also had colorful murals on the walls, or at least some wallpaper borders, as well as a few toys for children

PAGE 259

236 that could be easily disinfected. Three mother s indicated that their kids would get bored waiting in an often crowded, barren office. The one feature mothers liked most was a separate sick and well children waiting area. In many of these clinics there were plexiglass walls that, even in the case of smaller waiting rooms, allowed the office to not feel too enclosed but that protected children attending well child visits from being exposed to sick children. Le ss than half of the clinics had separate waiting areas. Additiona lly, the issue most noted by mothers (n=10) was that the office was too busy, and that they hating the long waiting times due to overbooking. Some environmental issues cannot be changed. For example, two mothers reported liking smaller practices because the larger ones were too impersonal. Similarly, mothers did not like rigid clin ic procedures often found in the larger clinics such as clinics requiring mothers to call ahead to make an appointment to see the doctor prior to coming into the office. One mother reported that she brought her child to the office and wanted an appointment. The clinic told her sh e needed to call first and was not given an appointment at that time. When she went home and called, she was able to make an appointment for later that day. Another issue is the speed at which medical records can be accessed. Some offices store medical records at other locations requiring time to obtain the records. One mother reported not liki ng the lag time between requesting medical records and receiving them. Finally, one woma n reported feeling uncom fortable with the other families attending the clinic. Although the various Medicaid-funded insura nce plans have diffe rent procedures, most plans allow women to change providers immediately with the longest wait for a

PAGE 260

237 change to occur taking one to two months. It appeared that women are aware of this flexibility. In nearly every instance when a mother indicated dissatisfaction with her pediatrician, she simply changed to another one. More than one mother reported changing doctors repeatedly with some changing up to four times with their current child or a previous child. There were three instances where insurance changes required the mothers to change doctors because of which doctors were included on each insurance plan. One mother indicated problems finding a provide r who took new patients. It was reported by at least one provider that the Medicaid re imbursement did not cover existing expenses and so they stopped accepting new patients until they could restructure their office in a way that their expenses could be covered.

PAGE 261

238 Appendix T. Calculation of and Odds Ratio The Odds Ratio for these regression model coefficients can be calculated. However, the interpretation of these ratios is complicated due to the transformation process of reversing the scores and taking th e square root of the original rating. For example, the odds ratio for the transformed provider rating in rela tion to the number of sick/follow-up visit is (OR=.83). This odds ratio would need to be interpreted based on the square root of a provider rating were a lower score was better. Although the amount of change is not easily conceptualized, the association between more positive ratings and increased visits can be seen in graphical re presentations. For example, Figure 9 plots the predicted number of visits based on the or iginal provider rating where higher scores represent more positive ratings of providers. Figure 10 represents the transformed data where ratings were reversed and Figure 11 repr esents of the number of predicted visits by the original ratings. The similarity in regres sion lines demonstrates a more linear pattern indicating that the model was strengthened by the transformation process but that the directionality of the asso ciations have not changed.

PAGE 262

239 Fitted Sick/Follow-up Visits by Original Provider Ratings 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 024681012 Original Provider RatingsFitted Visits Figure 9. Fitted Sick/Follow-up Visits by Original Provider Ratings Fitted Sick/Follow-up Visits by Transformed Provider Ratings 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 01234 Log RatingsFitted Visits Figure 10. Fitted Sick/Follow-up Visits by Transformed Ratings

PAGE 263

240 Log of Visits by Original Provider Ratings -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 024681012 Original RatingsLog of Visits Figure 11. Log of Visits by Or iginal Provider Ratings

PAGE 264

241 Appendix U. Hypothesis 4: Potentially Moderating Factors Model Structure and Results Well Child Care Visits/Provider Black non-Hispanic(B)/Hispanic(H) W=Intercept+Pro+Sec(x)+Avo( x)+Anx(x)+B(x)+H(x)+error W= Intercept+Pro+Sec(x)+Avo(x)+Anx( x)+B(x)+H(x)+B*Sec+B*Avo+B*Anx+ B*Sec+B*Avo+B*Anx +error Target Child was Mothers First Child (FC) W=Intercept+Pro+Sec(x)+A vo(x)+Anx(x)+FC(x)+error W= Intercept+Pro+Sec(x)+Avo(x)+Anx(x) +FC(x)+FC*Sec+FC*Avo+FC*Anx+error Childs Overall Health Status (CH) W=Intercept+Pro+Sec(x)+A vo(x)+Anx(x)+CH(x)+error W= Intercept+Pro+Sec(x)+Avo(x)+Anx(x) +CH(x)+CH*Sec+CH*Avo+CH*Anx+error Mothers Overall Health Status (MH) W=Intercept+Pro+Sec(x)+A vo(x)+Anx(x)+MH(x)+error W=Intercept+Pro+Sec(x)+Avo(x)+Anx(x) +MH(x)+MH*Sec+MH*Avo+MH*Anx+error WIC/Healthy Start Participation (WC) W=Intercept+Pro+Sec(x)+A vo(x)+Anx(x)+WC(x)+error W=Intercept+Pro+Sec(x)+Avo(x)+Anx(x) +WC(x)+WC*Sec+WC*Avo+WC*Anx+error Work 30+ Hours Per Week (WK) W=Intercept+Pro+Sec(x)+A vo(x)+Anx(x)+WK(x)+error W=Intercept+Pro+Sec(x)+Avo(x)+Anx(x)+WK(x)+WK*Sec+WK*Avo+WK*Anx+error Feelings About Going to the Doctor (FD) W=Intercept+Pro+Sec(x)+A vo(x)+Anx(x)+FD(x)+error W= Intercept+Pro+Sec(x)+Avo(x)+Anx( x)+FD(x)+FD*Sec+FD*Avo+FD*Anx+error Maternal Age (MA) W=Intercept+Pro+Sec(x)+A vo(x)+Anx(x)+MA(x)+error W=Intercept+Pro+Sec(x)+Avo(x)+Anx(x) +MA(x)+MA*Sec+MA*Avo+MA*Anx+error Maternal Bonding (MB) W=Intercept+Pro+Sec(x)+A vo(x)+Anx(x)+MB(x)+error W=Intercept+Pro+Sec(x)+Avo(x)+Anx(x) +MB(x)+MB*Sec+MB*Avo+MB*Anx+error (The Models above were repeated for Pr ovider Office as well as for Sick/Follow-up Visits, Emergency Department Visits, and Immunizations.)

PAGE 265

242 Appendix U-1. Well Child Care Visits, Interaction Style, and Provider Ratings by Race/Ethnicity (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.046.0711-.186.094 .41.520 Secure-.003.0091-.021.015 .10.752 Avoidant.010.0101-.001.029 .92.338 Anxious-.007.0101-.026.013 .46.497 Black-.157.0851-.324.011 3.36.007 Hispanic-.027.1171-.256.202 .05.819 Saturated Model Provider-.039.0721-.180.103 .29.592 Secure-.007.0131-.033.019 .28.594 Avoidant.024.0141-.004.052 2.80.094 Anxious-.002.0141-.030.026 .02.884 Black.155.8121-1.4371.747 .04.849 Hispanic.580.9731-1.3282.488 .36.551 Black*Secure.022.0211-.020.064 1.07.301 Black*Anxious-.010.0271-.051.031 .23.629 Black*Avoidant-.027.0231-.072.018 1.38.240 Hispanic*Secure-.013.0261-.063.038 .24.624 Hispanic*Anxious.024.0381-.050.098 .39.531 Hispanic*Avoidant-.028.0261-.078.022 1.19.275 Log Likelihood (Main Effect Model) 559.9512 Log Likelihood (Saturated Model) 563.1867 Difference in Log Likelihoods 3.2355 *2=6.47

PAGE 266

243 Appendix U-2. Well Child Care Visits, Interaction Style, a nd Provider Ratings by Birth Order (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.044.0711-.184.095 .39.533 Secure-.002.0091-.021.016 .05.818 Avoidant.005.0091-.013.024 .33.565 Anxious-.008.0101-.028.011 .72.396 Target Child Mothers First Child.105.0761-.043.253 1.94.164 Saturated Model Provider-.050.0721-.192.092 .48.488 Secure.001.0121-.022.025 .01.911 Avoidant.007.0141-.020.033 .24.624 Anxious-.007.0141-.034.020 .26.608 Target Child Mothers First Child.396.7221-1.021.812 .30.583 First Child*Secure-.009.0201-.048.029 .22.639 First Child*Anxious-.002.0201-.040.037 .01.938 First Child*Avoidant-.003.0191-.040.035 .02.889 Log Likelihood (Main Effect Model) 559.1385 Log Likelihood (Saturated Model) 559.2503 Difference in Log Likelihoods .1118 *2=.22

PAGE 267

244 Appendix U-3. Well Child Care Visits, Interaction Style, and Provider Ratings by Childs Health Rating (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider-.047.0711-.185.093 .43.514 Secure-.002.0091-.020.017 .02.875 Avoidant.004.0091-.014.023 .20.658 Anxious-.006.0101-.025.013 .38.536 Childs Overall Health Rating-.004.0271-.056.049 .02.896 Saturated Model Provider-.054.072 1-.196.087 .57.452 Secure.023.0881-.148.195 .07.792 Avoidant.032.0781-.121.184 .17.682 Anxious-.067.0671-.199.065 1.00.317 Childs Overall Health Rating.011.3221-.620.641 .00.973 Childs Health Rating*Secure-.003.0091-.021.056 .09.769 Childs Health Rating*Anxious.007.0071-.007.021 .88.349 Childs Health Rating*Avoidant-.003.0081-.019.013 .13.717 Log Likelihood (Main Effect Model) 558.1816 Log Likelihood (Saturated Model) 558.7752 Difference in Log Likelihoods .5936 *2=1.19

PAGE 268

245 Appendix U-4. Well Child Care Visits, Interaction Style, and Provider Ratings by Mothers Health Rating (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider-.044.0711-.183.095 .39.533 Secure-.000.0091-.019.019 .00.999 Avoidant.005.0091-.014.029 .23.632 Anxious-.008.0101-.027.012 .60.437 Mothers Overall Health Rating-.021.0211-.062.020 1.00.317 Saturated Model Provider-.038.0721-.018.103 .27.601 Secure.010.0461-.080.101 .05.822 Avoidant-.016.0451-.105.073 .12.724 Anxious-.012.0411-.092.069 .08.779 Mothers Overall Health Rating-.065.2061-.469.339 .10.753 Mothers Health Rating*Secure-.001.0061-.013.010 .05.818 Mothers Health Rating*Anxious.001.0051-.009.011 .02.892 Mothers Health Rating*Avoidant.003.0061-.008.014 .22.639 Log Likelihood (Main Effect Model) 558.6687 Log Likelihood (Saturated Model) 558.9515 Difference in Log Likelihoods .2828 *2=.57

PAGE 269

246 Appendix U-5. Well Child Care Visits, Interaction Style, and Provider Ratings by WIC/Healthy Start Participation (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.050.0711-.189.089 .50.480 Secure-.003.0091-.021.016 .07.792 Avoidant.004.0091-.015.022 .15.703 Anxious-.006.0101-.025.013 .41.524 WIC/Healthy Start-.104.1111-.321.114 .87.350 Saturated Model Provider-.056.0721-.198.085 .61.435 Secure.016.0241-.031.062 .42.516 Avoidant-.032.0281-.087.022 1.35.245 Anxious.005.0381-.070.080 .02.893 WIC/Healthy Start-.4121.1981-2.7591.936 .12.731 WIC/HS*Secure-.021.0261-.072.031 .62.431 WIC/HS*Anxious-.013.0401-.091.064 .11.739 WIC/HS*Avoidant.040.0301-.018.098 1.85.174 Log Likelihood (Main Effect Model) 558.5986 Log Likelihood (Saturated Model) 559.9683 Difference in Log Likelihoods 1.3700 *2=2.74

PAGE 270

247 Appendix U-6. Well Child Care Visits, Interaction Style, and Provider Ratings by Mothers Employment Status (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider-.046.0711-.186.093 .42.516 Secure-.002.0091-.020.017 .03.866 Avoidant.004.0101-.015.023 .19.666 Anxious-.006.0101-.025.014 .34.561 Mother Worked > 30 Hrs/Week.004.0911-.173.182 .00.961 Saturated Model Provider-.046.0721-.187.095 .41.523 Secure.001.0101-.020.021 .00.962 Avoidant.003.0111-.019.025 .07.785 Anxious-.006.0121-.029.018 .23.631 Mother Worked > 30 Hours/Week.142.9131-1.6481.93 .02.877 Work*Secure-.010.0241-.057.037 .19.665 Work *Anxious.001.0221-.042.043 .00.974 Work *Avoidant.003.0221-.040.047 .02.877 Log Likelihood (Main Effect Model) 558.1742 Log Likelihood (Saturated Model) 558.3193 Difference in Log Likelihoods .1451 *2=.29

PAGE 271

248 Appendix U-7. Well Child Care Visits, Interaction Style, and Provider Ratings by Mothers Feelings About Providers (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider-.037.0721-.178.104 .26.608 Secure-.002.0091-.020.017 .02.876 Avoidant.004.0091-.014.023 .21.649 Anxious-.005.0101-.024.014 .28.594 Mothers Feelings About Providers-.013.0151-.042.015 .83.361 Saturated Model Provider-.040.0741.185.106 .28.595 Secure-.004.0231.050.042 .03.869 Avoidant-.015.0221.057.027 .49.485 Anxious.005.0201.035.045 .06.805 Mothers Feelings About Providers-.077.1551.381.227 .25.519 Feeling*Secure.000.0041.008.009 .01.931 Feelings*Anxious-.002.0041.010.005 .38.538 Feelings*Avoidant.004.0041.003.011 1.05.305 Log Likelihood (Main Effect Model) 558.5907 Log Likelihood (Saturated Model) 559.1807 Difference in Log Likelihoods .5900 *2=1.18

PAGE 272

249 Appendix U-8. Well Child Care Visits, Interaction Styl e, and Provider Ratings by Mothers Age (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider-.050.0721-.191.091 .48.488 Secure-.002.0091-.020.017 .03.859 Avoidant.004.0091-.014.023 .21.650 Anxious-.006.0101-.025.013 .40.526 Mothers Age-.001.0061-.013.010 .06.813 Saturated Model Provider-.054.0721-.196.087 .57.452 Secure.021.0421-.062.103 .24.624 Avoidant.013.0431-.072.097 .09.768 Anxious-.034.0421-.117.048 .66.418 Mothers Age.008.0541-.098.115 .02.879 Age*Secure-.001.0021-.004.002 .27.601 Age*Anxious.001.0021-.002.004 .44.504 Age*Avoidant-.000.0021-.004.003 .03.858 Log Likelihood (Main Effect Model) 558.2008 Log Likelihood (Saturated Model) 558.6474 Difference in Log Likelihoods .4466 *2=.89

PAGE 273

250 Appendix U-9. Well Child Care Visits, Interaction Style, and Provider Ratings by Mothers Bonding Score (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.058.0711-.197.081 .67.413 Secure-.002.0091-.020.061 .04.843 Avoidant.006.0101-.013.024 .37.544 Anxious-.005.0101-.024.015 .21.645 Mothers Bonding Score-.090.0401-.167-.012 5.15.023 Saturated Model Provider-.062.0711-.202.008 .75.385 Secure-.004.0091-.022.015 .16.687 Avoidant.005.0101-.014.024 .27.600 Anxious-.006.0101-.025.014 .32.571 Mothers Bonding Score-.002.3811-.750.745 .00.995 Bonding*Secure-.010.0011-.032.012 .77.380 Bonding*Anxious-.005.0111-.026.016 .19.667 Bonding*Avoidant.008.0011-.011.028 .69.407 Log Likelihood (Main Effect Model) 560.8403 Log Likelihood (Saturated Model) 561.7527 Difference in Log Likelihoods .9124 *2=1.82

PAGE 274

251 Appendix U-10. Well Child Care Visits, Interaction Style, a nd Provider Office Ratings by Race/Ethnicity (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.006.0721-.136.147 .01.938 Secure-.002.0091-.021.015 .07.792 Avoidant.010.0101-.010.029 .93.336 Anxious-.007.0101-.026.012 .51.475 Black-.158.0851-.326.009 3.42.064 Hispanic-.035.1171-.264.194 .09.767 Saturated Model Provider Office.022.0731-.121.165 .09.763 Secure-.006.0131-.032.019 .23.629 Avoidant.024.0151-.004.053 2.93.087 Anxious-.002.0141-.030.025 .03.866 Black.1725.80911.413 1.758 .05.831 Hispanic.610.97311.298 2.52 .39.531 Black*Secure.225.0211-.020.064 1.10.294 Black*Anxious-.010.0211-.050.031 .21.646 Black*Avoidant-.028.0231-.073.017 1.51.219 Hispanic*Secure-.014.0261-.065.036 .30.583 Hispanic*Anxious.023.0381-.051.097 .37.541 Hispanic*Avoidant-.028.0261-.078.022 1.18.277 Log Likelihood (Main Effect Model) 559.7447 Log Likelihood (Saturated Model) 563.0870 Difference in Log Likelihoods 3.3423 *2=6.68

PAGE 275

252 Appendix U-11. Well Child Care Visits, Interaction Style, and Provider Office Ratings by Birth Order (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider Office.013.0721-.127.154 .04.851 Secure-.002.0091-.020.017 .03.866 Avoidant.006.0091-.013.024 .34.560 Anxious-.009.0101-.028.011 .76.382 Target Child Mothers First Child.107.0761-.041.255 1.99.158 Saturated Model Provider Office.013.0721-.129.155 .03.854 Secure.001.0121-.022.024 .01.930 Avoidant.006.0141-.020.033 .22.638 Anxious-.008.0141-.035.019 .33.563 Target Child Mothers First Child.316.7111-.0771.71 .20.656 First Child*Secure-.007.0191-.045.031 .13.722 First Child*Anxious-.001.0201-.039.038 .00.978 First Child*Avoidant-.002.0191-.039.035 .01.915 Log Likelihood (Main Effect Model) 558.9599 Log Likelihood (Saturated Model) 559.0241 Difference in Log Likelihoods .0811 *2=.16

PAGE 276

253 Appendix U-12. Well Child Care Visits, Interaction Style, a nd Provider Office Ratings by Childs Health Rating (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.013.0721-.128.153 .03.857 Secure-.001.0091-.019.017 .01.925 Avoidant.004.0091-.014.023 .20.657 Anxious-.006.0101-.026.013 .43.514 Childs Overall Health Rating-.005.0271-.058.048 .03.855 Saturated Model Provider Office.002.0741-.143.146 .00.984 Secure.013.0881-.160.185 .02.887 Avoidant.032.0781-.121.185 .17.683 Anxious-.068.0681-.201.066 .99.320 Childs Overall Health Rating-.015.32271-.648.618 .00.963 Childs Health Rating*Secure-.002.0091-.020.017 .03.870 Childs Health Rating*Anxious.007.0071-.008.021 .85.357 Childs Health Rating*Avoidant-.003.0081-.019.013 .13.719 Log Likelihood (Main Effect Model) 557.9830 Log Likelihood (Saturated Model) 558.4888 Difference in Log Likelihoods .5058 *2=1.02

PAGE 277

254 Appendix U-13. Well Child Care Visits, Interaction Style, and Provider Office Ratings by Mothers Health Rating (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.010.0721-.131.150 .02.891 Secure.001.0091-.018.019 .00.954 Avoidant.005.0091-.014.023 .23.634 Anxious-.008.0101-.027.011 .65.420 Mothers Overall Health Rating-.021.0211-.063.020 1.04.308 Saturated Model Provider Office.014.0721-.127.155 .04.850 Secure.010.0461-.081.100 .04.832 Avoidant-.020.0051-.108.068 .20.658 Anxious-.012.0411-.093.070 .08.780 Mothers Overall Health Rating-.080.2051-.482.322 .15.696 Mothers Health Rating*Secure-.001.0061-.012.010 .04.836 Mothers Health Rating*Anxious.001.0051-.009.011 .02.898 Mothers Health Rating*Avoidant.003.0061-.008.014 .32.573 Log Likelihood (Main Effect Model) 558.4813 Log Likelihood (Saturated Model) 558.8310 Difference in Log Likelihoods .3497 *2=.70

PAGE 278

255 Appendix U-14. Well Child Care Visits, Interaction Style, and Provider Office Ratings by WIC/Healthy Start Participation (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.008.0721-.133 .148 .01.916 Secure-.002.0091-.020 .016 .04.840 Avoidant.004.0091-.015 .022 .14.704 Anxious-.007.0101-.026 .013 .44.506 WIC/Healthy Start Participation-.099.1111-.317 .119 .79.375 Saturated Model Provider Office-.023.0751-.169 .124 .09.762 Secure.015.0241-.032 .064 .38.535 Avoidant-.035.0281-.090 .021 1.50.221 Anxious.002.0381-.072 .077 .00.954 WIC/Healthy Start Participation-.5381.1821-2.854 1.779 .21.649 WIC/HS*Secure-.019.0261-.071 .032 .53.466 WIC/HS*Anxious-.011.0391-.088 .067 .07.789 WIC/HS*Avoidant.042.0301-.016 .101 2.01.156 Log Likelihood (Main Effect Model) 558.3515 Log Likelihood (Saturated Model) 559.7068 Difference in Log Likelihoods 1.3553 *2=2.71

PAGE 279

256 Appendix U-15. Well Child Care Visits, Interaction Style, and Provider Office Ratings by Mothers Employment Status (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.012.7151-.128.153 .03.863 Secure-.001.0091-.019.017 .01.912 Avoidant.004.0101-.015.023 .18.975 Anxious-.006.0101-.025.014 .35.556 Mother Worked > 30 Hours/Week.012.0901-.164.188 .02.892 Saturated Model Provider Office.019.0731-.124.162 .07.795 Secure.001.0101-.019.022 .02.895 Avoidant.003.0111-.019.025 .06.802 Anxious-.005.0121-.029.018 .19.666 Mother Worked > 30 Hours/Week.213.9101-.5712.00 .05.815 Work*Secure-.012.0241-.059.035 .25.617 Work *Anxious-.001.0221-.044.042 .00.950 Work *Avoidant.004.0221-.040.047 .03.872 Log Likelihood (Main Effect Model) 557.9755 Log Likelihood (Saturated Model) 558.1465 Difference in Log Likelihoods .1710 *2=.342

PAGE 280

257 Appendix U-16. Well Child Care Visits, Interaction Style, and Provider Office Ratings by Mothers Feelings About Provider Of fices (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.032.0741-.113.179 .19.665 Secure-.001.0091-.019.017 .01.911 Avoidant.005.0091-.014.023 .23.629 Anxious-.005.0101-.024.014 .29.592 Mothers Feelings About Providers-.016.0151-.045.013 1.17.280 Saturated Model Provider Office.037.0771-.113.188 .24.627 Secure-.009 .0241-.055.037 .14.704 Avoidant-.016.0221-.058.027 .53.465 Anxious.004.0241-.036.044 .04.848 Mothers Feelings About Providers-.113.1551-.417.191 .53.465 Feeling*Secure.002.0041-.007.010 .12.730 Feelings*Anxious-.002.0041-.009.005 .20.587 Feelings*Avoidant.004.0041-.003.011 1.15.283 Log Likelihood (Main Effect Model) 558.5519 Log Likelihood (Saturated Model) 559.1561 Difference in Log Likelihoods .6042 *2=1.21

PAGE 281

258 Appendix U-17. Well Child Care Visits, Interaction Style, and Provider Office Ratings by Mothers Age (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.011.0731-.132.154 .02.878 Secure-.001.0091-.019.017 .01.911 Avoidant.004.0091-.014.023 .20.652 Anxious-.006.0101-.026.013 .40.526 Mothers Age-.001.0061-.012.011 .01.935 Saturated Model Provider Office.011.0731.133.154 .02.886 Secure.021.0421.061.103 .26.610 Avoidant.014.0431.071.098 .10.754 Anxious-.032.0421.115.051 .58.446 Mothers Age.011.0551.100.118 .04.834 Age*Secure-.001.0021.004.002 .28.597 Age*Anxious.001.0021.002.004 .38.537 Age*Avoidant-.000.0021.004.003 .04.841 Log Likelihood (Main Effect Model) 557.9697 Log Likelihood (Saturated Model) 558.3717 Difference in Log Likelihoods .4020 *2=.80

PAGE 282

259 Appendix U-18. Well Child Care Visits, Interaction Style, and Provider Office Ratings by Mothers Bonding Score (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office-.007.0721-.148.135 .01.926 Secure-.001.0091-.020.017 .02.883 Avoidant.006.0101-.013.024 .34.560 Anxious-.005.0101-.024.014 .24.622 Mothers Bonding Score-.088.0401-.166-.010 4.91.027 Saturated Model Provider Office-.010.0731-.153.134 .02.896 Secure-.003.0091-.022.015 .11.744 Avoidant.005.0101-.014.024 .25.616 Anxious-.006.0101-.025.013 .37.542 Mothers Bonding Score-.010.3821-.758.738 .00.979 Bonding*Secure-.009.0111-.031.013 .66.418 Bonding*Anxious-.005.0111-.026.056 .25.615 Bonding*Avoidant.009.0101-.011.028 .72.400 Log Likelihood (Main Effect Model) 560.5045 Log Likelihood (Saturated Model) 561.3787 Difference in Log Likelihoods .8725 *2=1.75

PAGE 283

260 Appendix U-19. Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Race/Ethnicity (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.211.0831-.373-.048 6.47.011 Secure-.004.0101-.023.016 .13.718 Avoidant-.025.0111-.046-.005 5.69.017 Anxious.052.0111.034.074 22.26<.0001Black-.136.0931-.318.047 2.12.146 Hispanic.086.1271-.163.334 .46.500 Saturated Model Provider-.207.0831-.370-.045 6.25.012 Secure.010.0141-.016.037 .57.451 Avoidant-.014.0151-.043.016 .83.361 Anxious.050.0161.019.081 10.23.001 Black.994.8661-.7032.69 1.32.251 Hispanic.0171.1141-2.1652.20 .00.988 Black*Secure-.028.0231-.073.016 1.59.207 Black*Anxious-.017.0231-.063.029 .54.464 Black*Avoidant-.010.0251-.060.039 .16.688 Hispanic*Secure-.025.0341-.085.034 .69.406 Hispanic*Anxious.085.0441-.001.172 3.73.053 Hispanic*Avoidant-.027.0261-.077.024 1.07.302 Log Likelihood (Main Effect Model) 394.7650 Log Likelihood (Saturated Model) 398.9104 Difference in Log Likelihoods 4.1454 *2=8.29

PAGE 284

261 Appendix U-20. Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Birth Order (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.200.0821-.361-.039 5.95 .015 Secure-.003.0101-.023.016 .11 .735 Avoidant-.028.0101-.048-.008 7.64 .006 Anxious.048.0111.026.070 18.97 <.0001Target Child Mothers First Child.110.0821-.051.271 1.81 .179 Saturated Model Provider-.192.0821-.352-.032 5.50 .019 Secure-.101.0131-.036.016 .60 .439 Avoidant.007.0151-.023.036 .20 .657 Anxious.036.0161.005.067 5.04 .025 Target Child Mothers First Child.829.7711-.6832.34 1.15 .283 First Child*Secure.016.0211-.025.056 .58 .446 First Child*Anxious.028.0221-.015.072 1.61 .204 First Child*Avoidant-.064.0211-.104-.024 9.64 .002 Log Likelihood (Main Effect Model) 393.9019 Log Likelihood (Saturated Model) 399.7203 Difference in Log Likelihoods 5.8184 *2= 11.6

PAGE 285

262 Appendix U-21. Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Childs Health Rating (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider-.185.0821-.346-.024 5.06 .025 Secure.001.0101-.019.021 .01 .925 Avoidant-.031.0101-.051-.011 9.08 .003 Anxious.045.0111.023.066 16.43 <.0001Childs Overall Health Rating-.070.0251-.118-.021 7.86 .005 Saturated Model Provider-.182.0841-.347-.018 4.74 .030 Secure-.063.0831-.227.010 .58 .448 Avoidant-.001.0771-.151.149 .00 .987 Anxious-.045.0671-.176.087 .44 .506 Childs Overall Health Rating-.315.3161-.933.304 1.00 .319 Childs Health Rating*Secure.007.0091-.011.024 .59 .442 Childs Health Rating*Anxious.010.0071-.005.024 1.80 .180 Childs Health Rating*Avoidant-.003.0081-.019.013 .14 .706 Log Likelihood (Main Effect Model) 396.6592 Log Likelihood (Saturated Model) 397.7515 Difference in Log Likelihoods 1.0923 *2= 2.18

PAGE 286

263 Appendix U-22. Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Mothers Health Rating (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider-.205.0821-.364-.045 6.30 .012 Secure-.006.0101-.026.014 .33 .563 Avoidant-.030.0101-.049-.010 8.55 .004 Anxious.055.0111.033.076 24.38 <.0001 Mothers Overall Health Rating.042.0231-.004.088 3.19 .074 Saturated Model Provider-.205.0821-.367-.044 6.21 .013 Secure-.048.0511-.147.051 .90 .342 Avoidant-.033.0511-.132.067 .41 .523 Anxious.075.0491-.022.171 2.29 .130 Mothers Overall Health Rating-.039.2241-.477.400 .03 .863 Mothers Health Rating*Secure.005.0061-.007.017 .73 .392 Mothers Health Rating*Anxious-.003.0061-.014.009 .21 .648 Mothers Health Rating*Avoidant.000.0061-.012.013 .00 .949 Log Likelihood (Main Effect Model) 394.6240 Log Likelihood (Saturated Model) 395.2164 Difference in Log Likelihoods .9984 *2= 2.0

PAGE 287

264 Appendix U-23. Sick/Follow-up Visits, Interaction Style, and Provider Ratings by WIC/Healthy Start Participation (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider-.211.0821-.371-.050 6.61 .010 Secure-.004.0101-.023.016 .15 .698 Avoidant-.031.0101-.050-.011 9.05 .003 Anxious.051.0111.030.072 21.78 <.0001WIC/Healthy Start Participation-.230.1151-.456-.004 3.97 .046 Saturated Model Provider-.187.0841-.351-.022 4.96 .026 Secure-.030.0251-.080.019 1.44 .230 Avoidant-.079.0311-.140-.017 6.30 .012 Anxious-.059.0391-.136.018 2.24 .135 WIC/Healthy Start Participation-3.954.3291-6.560-1.349 8.85 .003 WIC/HS*Secure.0330.0281-.024.084 1.16 .281 WIC/HS*Anxious.116.0411.036.197 8.08 .005 WIC/HS*Avoidant.052.0331-.013.118 2.49 .115 Log Likelihood (Main Effect Model) 394.8837 Log Likelihood (Saturated Model) 400.4060 Difference in Log Likelihoods 5.5223 *2= 11.04

PAGE 288

265 Appendix U-24. Sick/Follow-up Visits, Interaction Style, a nd Provider Ratings by Mothers Employment Status (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider-.191.0821-.352-.030 5.38 .020 Secure-.002.0101-.022.017 .05 .826 Avoidant-.031.0101-.051-.011 9.12 .003 Anxious.053.0111.032.075 23.39 <.0001 Mother Worked > 30 Hours/Week.109.0971-.080.298 1.28 .257 Saturated Model Provider-.182.0831.344-.020 4.87 .027 Secure.002.0111.019.024 .05 .825 Avoidant-.041.0121.064-.017 11.23 .001 Anxious.066.0131.039.092 24.08 <.0001 Mother Worked > 30 Hours/Week.625.9821-1.3002.55 .40 .525 Work*Secure-.026.0261-.077.025 1.00 .316 Work *Anxious-.037.0241-.084.010 2.41 .121 Work *Avoidant.026.0231-.020.072 1.19 .275 Log Likelihood (Main Effect Model) 393.6316 Log Likelihood (Saturated Model) 395.7183 Difference in Log Likelihoods 2.0867 *2= 4.17

PAGE 289

266 Appendix U-25. Sick/Follow-up Visits, Interaction Style, a nd Provider Ratings by Mothers Feelings About Providers (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider-.200.0821-.359-.036 5.73 .017 Secure-.002.0101-.022.017 .05 .820 Avoidant-.029.0101-.049-.010 8.25 .004 Anxious.051.0111.030.073 22.31 <.001 Mothers Feelings -.007.0161-.038.024 .18 .667 Saturated Model Provider-.209.0841-.374-.043 6.12 .013 Secure.037.0251-.013.087 2.15 .142 Avoidant-.011.0231-.057.034 .24 .626 Anxious-.003.0231-.047.041 .02 .886 Mothers Feelings .091.1691-.239.422 .29 .588 Feeling*Secure-.071.0051-.016.002 2.50 .114 Feelings*Anxious.011.0041.003.019 6.83 .009 Feelings*Avoidant.004.0041-.012.003 1.28 .258 Log Likelihood (Main Effect Model) 393.0933 Log Likelihood (Saturated Model) 399.1433 Difference in Log Likelihoods 6.0500 *2= 12.10

PAGE 290

267 Appendix U-26. Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Mothers Age (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.192.0831.355-.030 5.39 .023 Secure-.002.0101.021.018 .04 .8471 Avoidant-.030.0101.049-.010 8.40 .004 Anxious.052.0111.031.074 22.53 <.0001 Mothers Age.005.0061-.008.017 .50 .481 Saturated Model Provider-.206.0831-.370-.043 6.12 .013 Secure-.05.0461-.144.034 1.45 .228 Avoidant-.073.0471-.164.019 2.41 .121 Anxious.005.0471-.088.097 .01 .921 Mothers Age-.107.0591-.221.008 3.32 .069 Age*Secure.002.0021-.001.005 1.50 .221 Age*Anxious.002.0021-.002.005 1.20 .273 Age*Avoidant.002.0021-.002.005 .94 .332 Log Likelihood (Main Effect Model) 393.2483 Log Likelihood (Saturated Model)395.4568 Difference in Log Likelihoods 2.2085 *2= 4.42

PAGE 291

268 Appendix U-27. Sick/Follow-up Visits, Interaction Style, and Provider Ratings by Mothers Bonding Score (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.201.0821-.362-.041 6.05 .014 Secure-.002.0101-.022.017 .05 .816 Avoidant-.029.0101-.049-.009 8.30 .004 Anxious.051.0111.030.072 22.01 <.0001 Mothers Bonding Score.001.0391-.076.077 .00 .989 Saturated Model Provider-.201.0831-.363-.038 5.84 .016 Secure-.003.0101-.023.016 .10 .752 Avoidant-.029.0101-.049-.009 8.00 .005 Anxious.052.0111.030.073 22.14 <.0001 Mothers Bonding Score.577.3741-.1571.311 2.38 .123 Bonding*Secure-.016.0111-.038.007 1.88 .170 Bonding*Anxious-.031.0111-.053-.010 8.19 .004 Bonding*Avoidant.012.0101-.007.031 1.43 .231 Log Likelihood (Main Effect Model) 393.0009 Log Likelihood (Saturated Model) 398.5198 Difference in Log Likelihoods 5.5189 *2= 11.04

PAGE 292

269 Appendix U-28. Sick/Follow-up Visits, Interaction Style, a nd Provider Office Ratings by Race/Ethnicity (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office-.036.0791-.191.120 .20 .652 Secure-.002.0101-.021.018 .03 .858 Avoidant-.026.0111-.046-.005 5.85 .016 Anxious.051.0111.029.073 21.25 <.0001Black-.142.0931-.325.041 2.30 .129 Hispanic.057.1261-.191.304 .20 .653 Saturated Model Provider Office-.033.0801-.190.124 .17 .683 Secure.013.0141-.013.040 .98 .323 Avoidant-.014.0151-.043.016 .79 .373 Anxious.049.0161.019.080 9.85 .002 Black1.057.8541-.6162.730 1.53 .215 Hispanic.1571.1001-.2002.312 .02 .887 Black*Secure-.030.0231-.074.014 1.77 .184 Black*Anxious-.016.0241-.063.030 .48 .488 Black*Avoidant-.012.0251-.061.037 .23 .634 Hispanic*Secure-.033.0301-.090.025 1.23 .267 Hispanic*Anxious.083.0441-.003.169 3.55 .060 Hispanic*Avoidant-.025.0261-.076.026 .95 .330 Log Likelihood (Main Effect Model) 391.4715 Log Likelihood (Saturated Model) 395.7190 Difference in Log Likelihoods 4.2475 *2= 8.50

PAGE 293

270 Appendix U-29. Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Birth Order (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider Office-.019.0791-.172.135 .06 .814 Secure-.002.0101-.021.018 .02 .903 Avoidant-.029.0101-.048-.009 7.94 .005 Anxious.048.0111.026.069 18.37 <.0001 Target Child Mothers First Child.114.0821-.047.274 1.93 .164 Saturated Model Provider Office.001.0791-.153.155 .00 .989 Secure-.012.0131-.037.014 .80 .372 Avoidant.004.0151-.025.034 .09 .766 Anxious.033.0161.002.064 4.27 .039 Target Child Mothers First Child.510.7551-.9691.989 .46 .499 First Child*Secure.025.0201-.015.065 1.48 .224 First Child*Anxious.032.0221-.021.076 2.02 .156 First Child*Avoidant-.061.0211-.101-.021 8.79 .003 Log Likelihood (Main Effect Model) 390.8215 Log Likelihood (Saturated Model) 396.8498 Difference in Log Likelihoods 6.0283 *2= 12.05

PAGE 294

271 Appendix U-30. Sick/Follow-up Visits, Interaction Style, a nd Provider Office Ratings by Childs Health Rating (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office-.003.0791-.158.152 .00 .968 Secure.003.0101-.017.022 .08 .779 Avoidant-.031.0101-.051-.011 9.45 .002 Anxious.044.0111.022.066 15.82 <.0001 Childs Overall Health Rating-.075.0251-.124-.026 8.99 .003 Saturated Model Provider Office.004.0811-.155.163 .00 .961 Secure-.091.0841-.255.074 1.16 .281 Avoidant-.002.0771-.153.150 .00 .981 Anxious-.043.0691-.178.092 .39 .534 Childs Overall Health Rating-.379.3171-.999.242 1.43 .232 Childs Health Rating*Secure.010.0091-.008.028 1.24 .266 Childs Health Rating*Anxious.010.0081-.005.024 1.59 .208 Childs Health Rating*Avoidant-.003.0081-.020.013 .14 .711 Log Likelihood (Main Effect Model) 394.0247 Log Likelihood (Saturated Model) 395.2903 Difference in Log Likelihoods 1.2656 *2= 2.53

PAGE 295

272 Appendix U-31. Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Mothers Health Rating (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office-.013.0781-.166.140 .03 .865 Secure-.004.0101-.023.016 .13 .719 Avoidant-.030.0101-.050-.010 8.80 .003 Anxious.054.0111.032.076 23.57 <.0001 Mothers Overall Health Rating.040.0241-.006.086 2.91 .088 Saturated Model Provider Office-.009.0791-.163.145 .01 .910 Secure-.052.0511-.152.048 1.05 .307 Avoidant-.050.0501-.148.048 .99 .319 Anxious.075.0501-.023.173 2.25 .134 Mothers Overall Health Rating-.108.2221-.543.326 .24 .625 Mothers Health Rating*Secure.006.0061-.006.018 .96 .327 Mothers Health Rating*Anxious-.003.0061-.014.009 .20 .651 Mothers Health Rating*Avoidant.003.0061-.009.015 .17 .680 Log Likelihood (Main Effect Model) 391.3367 Log Likelihood (Saturated Model) 391.9753 Difference in Log Likelihoods .6386 *2= 1.28

PAGE 296

273 Appendix U-32. Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by WIC/Healthy Start Participation (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office-.035.0791-.190.120 .20 .656 Secure-.002.0101-.021.017 .04 .841 Avoidant-.031.0101-.051-.011 9.52 .002 Anxious.050.0111.029.072 20.96 <.0001WIC/Healthy Start Participation-.216.1161-.443.011 3.49 .0619 Saturated Model Provider Office-.041.0831-.203.122 .24 .624 Secure-.032.0261-.082.018 1.55 .214 Avoidant-.080.0321-.142-.018 6.46 .011 Anxious-.068.0391-.144.008 3.06 .080 WIC/Healthy Start Participation-4.2081.321-6.760-1.657 10.45 .001 WIC/HS*Secure.034.0281-.021.089 1.48 .224 WIC/HS*Anxious.126.0411.046.205 9.57 .002 WIC/HS*Avoidant.053.0331-.012.119 2.54 .111 Log Likelihood (Main Effect Model) 391.5158 Log Likelihood (Saturated Model) 397.9420 Difference in Log Likelihoods 6.4262 *2= 12.85

PAGE 297

274 Appendix U-33. Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Mothers Employment Status (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office-.022.0791-.176.132 .08 .781 Secure-.000.0101-.020.019 .00 .974 Avoidant-.032.0101-.052-.012 9.81 .002 Anxious.053.0111-.032.075 23.19 <.0001Mother Worked > 30 Hours/Week.138.0961-.050.325 2.06 .151 Saturated Model Provider Office.008.0821-.150.165 .01 .926 Secure.004.0111-.017.026 .16 .692 Avoidant-.042.0121-.066-.018 11.94 .001 Anxious.067.0141.041.093 24.7 <.0001Mother Worked > 30 Hours/Week.773.9811-1.1502.69 6 .62 .431 Work*Secure-.028.0261-.079.023 1.15 .284 Work *Anxious-.042.0241-.090.005 3.02 .082 Work *Avoidant.026.0241-.020.072 1.19 .275 Log Likelihood (Main Effect Model) 390.8629 Log Likelihood (Saturated Model) 393.1894 Difference in Log Likelihoods 2.3265 *2= 4.65

PAGE 298

275 Appendix U-34. Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Mothers Feelings About Provider Of fices (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office-.007.0811-.165.152 .01 .936 Secure-.004.0101-.020.019 .00 .966 Avoidant-.030.0101-.049-.001 8.52 .005 Anxious.051.0111.030.072 21.80 <.0001 Mothers Feelings -.012.0161-.044.020 .51 .474 Saturated Model Provider Office-.019.0841-.183.146 .05 .825 Secure.028 .0261-.022.079 1.23 .268 Avoidant-.011.0231-.057.035 .23 .631 Anxious-.001.0221-.054.034 .19 .662 Mothers Feelings .025.1681-.303.354 .02 .880 Feeling*Secure-.005.0051-.014.004 1.25 .264 Feelings*Anxious.012.0041.004.020 8.70 .003 Feelings*Avoidant-.005.0041-.012.003 1.42 .233 Log Likelihood (Main Effect Model) 390.1139 Log Likelihood (Saturated Model) 395.9973 Difference in Log Likelihoods 5.8783 *2= 11.76

PAGE 299

276 Appendix U-35. Sick/Follow-up Visits, Interaction Style, a nd Provider Office Ratings by Mothers Age (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office-.004.0801-.160.153 .00 .963 Secure.000.0101-.019.020 .00 .985 Avoidant-.030.0101-.050-.010 8.83 .003 Anxious.053.0111.031.075 22.45 <.0001Mothers Age.007.0061-.006.019 1.19 .275 Saturated Model Provider Office-.015.0801-.171.142 .03 .854 Secure-.052.0461-.142.037 1.32 .250 Avoidant-.071.0471-.163.021 2.29 .130 Anxious.013.0471-.079.105 .08 .784 Mothers Age-.097.0591-.212.018 2.76 .097 Age*Secure.002.0021-.001.005 1.48 .224 Age*Anxious.002.0021-.002.005 .85 .356 Age*Avoidant.002.0021-.002.005 .83 .362 Log Likelihood (Main Effect Model) 390.4475 Log Likelihood (Saturated Model) 392.2834 Difference in Log Likelihoods 1.8359 *2= 3.67

PAGE 300

277 Appendix U-36. Sick/Follow-up Visits, Interaction Style, and Provider Office Ratings by Mothers Bonding Score (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider Office-.019.0791-.173.135 .06 .811 Secure-.000.0101-.020.019 .00 .974 Avoidant-.030.0101-.050-.010 8.78 .003 Anxious.051.0111.029.072 21.26 <.0001 Mothers Bonding Score.005.0401-.072.083 .02 .891 Saturated Model Provider Office-.011.0801-.169.146 .02 .889 Secure-.001.0101-.020.019 .01 .926 Avoidant-.029.0101-.049-.009 8.20 .004 Anxious.051.0101.029.072 21.13 <.0001 Mothers Bonding Score.524.3741-.2081.26 1.97 .1604 Bonding*Secure-.012.0111-.034.010 1.19 .275 Bonding*Anxious-.034.0111-.055-.012 9.19 .002 Bonding*Avoidant.013.0101-.007.032 1.65 .199 Log Likelihood (Main Effect Model) 389.8659 Log Likelihood (Saturated Model) 395.4847 Difference in Log Likelihoods 5.6188 *2= 11.24

PAGE 301

278 Appendix U-37. Predicted Sick and Follow-Up Visits by Avoidant Interaction Style and Whether Target Child was Mother's First Controlling for Provider Office and Interaction Styles (Secure, Anxious, Avoidant)0 2 4 6 8 10 567891011121314151617181920212223242526272829303132333435 Avoidant Interaction ScorePredicted Visits First Child Not First Child

PAGE 302

279 Appendix U-38. Predicted Sick and Follow-Up Visits by Anxious Interaction Style and WIC/Healthy Start Participation Controlling for Provider Office and Interaction Styles (Secure, Anxious, Avoidant)0.0 0.2 0.4 0.6 0.8 12345678910111213141516171819202122232425 Anxious Interaction ScorePredicted Visits No WIC WIC

PAGE 303

280 Appendix U-39. Predicted Sick and Follow-Up Visits by Anxious Interaction Style and Feelings About Going to the Doctor Controlling for Provider Office and Interaction Styles (Secure, Anxious, Avoidant)0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 12345678910111213141516171819202122232425 Anxious Interaction ScorePredicted Visits 2 4 6 8 10

PAGE 304

281 Appendix U-40. Predicted Sick and Follow-Up Visits by Anxious Interaction Style and Bonding Issues Controlling for Provider Office and Interaction Styles (Secure, Anxious, Avoidant)0 5 10 15 20 25 12345678910111213141516171819202122232425 Anxious Interaction ScorePredicted Visits -1 0 1 2 3

PAGE 305

282 Appendix U-41. Emergency Department Visits, Interaction St yle, and Provider Ratings by Race/Ethnicity (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.023.2311-.475.429 .01.921 Secure-.018.0311-.079.043 .34.557 Avoidant.007.0331-.058.072 .04.835 Anxious.018.0331-.046.083 .31.576 Black.096.2801-.453.645 .12.732 Hispanic.015.4071-.783.813 .00.971 Saturated Model Provider.006.2321-.448.460 .00.979 Secure.020.0451-.069.109 .19.663 Avoidant.065.0511-.036.166 1.60.206 Anxious.017.0491-.079.113 .11.735 Black2.0152.5721-3.0267.056 .61.433 Hispanic6.303.5641-.68713.285 3.12.077 Black*Secure-.028.0681-.162.107 .16.687 Black*Anxious-.023.0671-.155.109 .12.732 Black*Avoidant-.041.0751-.188.106 .30.582 Hispanic*Secure-.166.1091-.379.047 2.33.127 Hispanic*Anxious.080.1361-.187.347 .034.558 Hispanic*Avoidant-.169.0901-.345.006 3.58.059 Log Likelihood (Main Effect Model) -107.9761 Log Likelihood (Saturated Model) -104.1519 Difference in Log Likelihoods 3.8242 *2=7.65

PAGE 306

283 Appendix U-42. Emergency Department Visits, Interaction St yle, and Provider Ratings by Birth Order (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.018.2331-.475.439 .01.939 Secure-.021.0311-.083.040 .45.501 Avoidant.012.0311-.049.074 .16.692 Anxious.013.0331-.052.077 .15.702 Target Child Mothers First.222.2511-.271.7142 .78.377 Saturated Model Provider.044.2391-.425.512 .03.856 Secure-.054.0421-.136.028 1.66.198 Avoidant-.059.0481-.152.025 1.52.217 Anxious-.040.0471-.131.052 .07.399 Target Child Mothers First-5.1002.4641-9.929-.270 4.28.039 First Child*Secure.058.0671-.073.189 .75.386 First Child*Anxious.079.0661-.050.208 1.45.228 First Child*Avoidant.117.0651-.011.245 3.24.072 Log Likelihood (Main Effect Model) -107.6504 Log Likelihood (Saturated Model) -104.3559 Difference in Log Likelihoods 3.2945 *2=6.59

PAGE 307

284 Appendix U-43. Emergency Department Visits, Interaction Styl e, and Provider Ratings by Childs Health Rating (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider.019.2341-.439.477 .01.936 Secure-.014.0321-.077.048 .21.650 Avoidant.008.0321-.055.071 .06.802 Anxious.005.0331-.061.070 .02.886 Childs Overall Health Rating-0.130.0701-.269.008 3.41.065 Saturated Model Provider-0.109.2411-.582.363 .21.650 Secure.692.2451.2111.173 7.94.005 Avoidant-.069.2281-.515.378 .09.764 Anxious.164.1581-.146.474 1.08.299 Childs Overall Health Rating1.642.8571-.0363.321 3.68.055 Childs Health Rating*Secure-.078.0271-.130-.025 8.34 .004 Childs Health Rating*Anxious-.017.0181-.052.018 .94.333 Childs Health Rating*Avoidant.008.0251-.041.056 .10.753 Log Likelihood (Main Effect Model) -106.5343 Log Likelihood (Saturated Model) -101.0785 Difference in Log Likelihoods 5.4558 *2=10.91

PAGE 308

285 Appendix U-44. Emergency Department Visits, Interaction Style, and Provider Ratings by Mothers Health Rating (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider-.024.2301-.474.427 .01 .917 Secure.022.0311-.083.040 .47 .494 Avoidant.010.0311-.051.071 .10 .751 Anxious.021.0331-.045.086 .39 .531 Mothers Overall Health Rating.029.0721-.112.170 .17 .683 Saturated Model Provider-.048.2341-.507.411 .04 .839 Secure-.002.1571-.309.305 .00 .990 Avoidant.098.1531-.202.398 .41 .521 Anxious-.011.1451-.296.274 .01 .938 Mothers Overall Health Rating.298.6911-1.0561.653 .19 .666 Mothers Health Rating*Secure-.003.0191-.040.035 .02 .891 Mothers Health Rating*Anxious.004.0181-.031.038 .04 .832 Mothers Health Rating*Avoidant -.011.0191-.048.026 .35 .556 Log Likelihood (Main Effect Model) -107.9545 Log Likelihood (Saturated Model) -107.7798 Difference in Log Likelihoods .1747 *2= .35

PAGE 309

286 Appendix U-45. Emergency Department Visits, Interaction St yle, and Provider Rati ngs by WIC/Healthy Start Participation (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.015.2311-.468.438 .00 .948 Secure-.017.0311-.078.044 .30 .584 Avoidant.012.0311-.050.073 .13 .714 Anxious.019.0321-.045.082 .33 .568 WIC/Healthy Start .299.4311-.5451.143 .48 .487 Saturated Model Provider-.051.2341-.510.409 .05 .829 Secure.045.1031-.157.247 .19 .660 Avoidant.052.1051-.154.259 .25 .619 Anxious.151.1771-.196.499 .73 .933 WIC/Healthy Start 5.2574.6311-3.81214.334 1.29 .256 WIC/HS*Secure-.070.1091-.283.143 .41 .521 WIC/HS*Anxious-.137.1801-.491.216 .58 .447 WIC/HS*Avoidant-.045.1101-.262.181 .17 .681 Log Likelihood (Main Effect Model) -107.7775 Log Likelihood (Saturated Model) -107.1475 Difference in Log Likelihoods .6300 *2= 1.26

PAGE 310

287 Appendix U-46. Emergency Department Visits, Interaction Style, and Provider Ratings by Mothers Employment Status (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.014.2321-.469.441 .00 .952 Secure-.019.0311-.080.042 .38 .536 Avoidant.009.0321-.054.071 .07 .785 Anxious.020.0331-.045.085 .36 .546 Mother Worked > 30 Hours .088.2971-.495.670 .09 .768 Saturated Model Provider.022.2361-.441.485 .01 .926 Secure.006.0341-.061.074 .03 .885 Avoidant.027.0381-.048.101 .49 .483 Anxious.031.0411-.049.110 .58 .448 Mother Worked > 30 Hours 6.9143.181-.68813.141 4.74 .030 Work*Secure-.152.0861-.320.015 3.17 .075 Work *Anxious-.067.0701-.205.070 .092 .338 Work *Avoidant-.105.0781-.257.047 1.84 .175 Log Likelihood (Main Effect Model) -107.9957 Log Likelihood (Saturated Model) -105.5292 Difference in Log Likelihoods 2.4665 *2= 4.93

PAGE 311

288 Appendix U-47. Emergency Department Visits, Interaction Style, and Provider Ratings by Mothers Feelings About Providers (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model .54 .462 Provider-.069.2301-.519.381 .09 .763 Secure-.020.0311-.081.042 .40 .530 Avoidant.009.0311-.523.070 .07 .787 Anxious.015.0331-.050.080 .21 .649 Mothers Feelings .063.0481-.032.157 1.68 .195 Saturated Model Provider.026.2431-.449.502 .01 .913 Secure-.087.0801-.244.070 1.18 .278 Avoidant.008.0771-.143.159 .01 .914 Anxious-.116.0621-.236.005 3.53 .064 Mothers Feelings -.570.4921-1.534.394 1.34 .246 Feeling*Secure.014.0131-.012.040 1.09 .296 Feelings*Anxious.026.0111.005.048 5.79 .016 Feelings*Avoidant-.003.0121-.026.020 .06 .808 Log Likelihood (Main Effect Model) -107.2049 Log Likelihood (Saturated Model) -104.2769 Difference in Log Likelihoods 2.9280 *2= 5.86

PAGE 312

289 Appendix U-48. Emergency Department Visits, Interaction St yle, and Provider Ratings by Mothers Age (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider-.095.2311-.547.357 .17 .680 Secure-.021.0311-.083.040 .46 .495 Avoidant.012.0311-.049.072 .15 .703 Anxious.008.0331-.056.072 .07 .799 Mothers Age-.046.0221-.089-.004 4.52 .034 Saturated Model Provider-.097.2341-.556.362 .17 .679 Secure.347.1381.077.617 6.36 .012 Avoidant.038.1531-.262.337 .06 .806 Anxious.157.1421-.122.436 1.22 .269 Mothers Age.372.1961-.013.756 3.59 .058 Age*Secure-.014.0051-.024-.004 7.91 .005 Age*Anxious-.006.0061-.017.005 1.30 .254 Age*Avoidant-.001.0061-.013.011 .03 .857 Log Likelihood (Main Effect Model) -105.5969 Log Likelihood (Saturated Model) -101.0117 Difference in Log Likelihoods 4.5852 *2= 9.17

PAGE 313

290 Appendix U-49. Emergency Department Visits, Interaction Style, and Provider Ratings by Mothers Bonding Score (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider.009.2321-.445.464 .00 .969 Secure-.015.0321-.080.044 .33 .564 Avoidant.005.0311-.055.066 .03 .864 Anxious.011.0331-.053.075 .12 .734 Mothers Bonding Score.227.1071-.018.436 4.52 .034 Saturated Model Provider.017.2371-.448.481 .00 .442 Secure-.018.0321-.081.044 .33 .945 Avoidant.022.0331-.044.087 .43 .564 Anxious.009.0341-.057.076 .08 .513 Mothers Bonding Score1.0541.121-1.143.244 .89 .782 Bonding*Secure.001.0351-.068.069 .00 .345 Bonding*Anxious.018.0291-.040.075 .36 .982 Bonding*Avoidant-.046.0331-.110.018 1.96 .551 Log Likelihood (Main Effect Model) -105.9726 Log Likelihood (Saturated Model) -104.8557 Difference in Log Likelihoods1.1169 *2= 2.23

PAGE 314

291 Appendix U-50. Emergency Department Visits, Interacti on Style, and Provider Office Ratings by Race/Ethnicity (Adj usted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider Office.210.2291-.238.658 .84 .359 Secure-.017.0311-.078.044 .30 .583 Avoidant.009.0331-.057.074 .07 .797 Anxious.019.0331-.047.084 .23 .571 Black.104.2811-.448.656 .14 .712 Hispanic-.013.4061-.810.783 .00 .974 Saturated Model Provider Office.256.2311-.197.709 1.23 .268 Secure.020.0451-.069.109 .21 .650 Avoidant.072.0521-.031.174 1.89 .170 Anxious.015.0491-.081.110 .09 .764 Black2.0692.5461-2.9217.058 .66 .416 Hispanic6.4333.5541-.52213.400 3.28 .070 Black*Secure-.024.0691-.158.111 .12 .733 Black*Anxious-.018.0691-.152.117 .07 .798 Black*Avoidant-.050.0751-.197.097 .45 .505 Hispanic*Secure-.173.1091-.386.041 2.52 .113 Hispanic*Anxious.088.1381-.183.358 .40 .525 Hispanic*Avoidant-.175.0891-.350-.000 3.85 .050 Log Likelihood (Main Effect Model) -107.5734 Log Likelihood (Saturated Model) -103.5616 Difference in Log Likelihoods 4.0118 *2= 4.02

PAGE 315

292 Appendix U-51. Emergency Department Visits, Interaction St yle, and Provider Offi ce Ratings by Birth Order (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider Office .210.2321-.244.664.82 .3638 Secure -.020.0311-.081.041.41 .5198 Avoidant .014.0311-.047.076.21 .6491 Anxious .013.0331-.052.079.16 .6924 Target Child Mothers First .226.2511-.266.718.81 .3679 Saturated Model Provider Office .185.2281-.261.632.66 .416 Secure -.053.0421-.135.0291.59 .208 Avoidant -.053.0481-.147.0411.24 .265 Anxious -.040.0471-.133.053.71 .399 Target Child Mothers First -4.9652.4391-9.745-.1854.14 .042 First Child*Secure .057.0661-.070.187.74 .390 First Child*Anxious .082.0661-.048.2121.53 .217 First Child*Avoidant .111.0661-.017.2402.87 .090 Log Likelihood (Main Effect Model) -107.2526 Log Likelihood (Saturated Model) -104.0507 Difference in Log Likelihoods 3.2019*2= 6.40

PAGE 316

293 Appendix U-52. Emergency Department Visits, Interaction St yle, and Provider Office Ratings by Childs Health Rating (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider Office.252.2341-.206.710 1.16 .281 Secure-.014.0321-.076.048 .20 .652 Avoidant.011.0321-.052.073 .11 .744 Anxious.006.0341-.061.072 .03 .870 Childs Health Rating-.139.0711-.278-.000 3.84 .050 Saturated Model Provider Office.153.2431-.0324.630 .39 .530 Secure.636.2361.1741.10 7.28 .007 Avoidant-.077.2291-.525.371 .11 .738 Anxious.193.1661-.132.518 1.36 .244 Childs Health Rating.1535.8411-.1133.184 3.33 .068 Childs Health*Secure-.071.0261-.122-.021 7.64 .006 Childs Health*Anxious-.021.0191-.057.016 1.24 .266 Childs Health*Avoidant.009.0251-.040.058 .13 .722 Log Likelihood (Main Effect Model) -105.9743 Log Likelihood (Saturated Model) -100.9903 Difference in Log Likelihoods 4.9840 *2= 9.97

PAGE 317

294 Appendix U-53. Emergency Department Visits, Interacti on Style, and Provider Office Ratings by Mothers Health Rating (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office .209.2291-.239.658.84 .360 Secure -.021.0311-.082.041.44 .510 Avoidant .012.0311-.050.073.14 .704 Anxious .022.0341-.045.088.42 .517 Mothers Overall Health Rating .033.0731-.110.175.20 .654 Saturated Model Provider Office .197.2291-.252.646.74 .390 Secure -.008.1581-.319302.00 .959 Avoidant .082.1511-.213.377.30 .586 Anxious -.010.1501-.303.284.00 .949 Mothers Health Rating .226.6871-1.1211.573.11 .742 Mothers Health*Secure -.002.0191-.040.036.01 .931 Mothers Health*Anxious .004.0181-.032.039.04 .839 Mothers Health*Avoidant -.009.0191-.045.028.23 .635 Log Likelihood (Main Effect Model) -107.5542 Log Likelihood (Saturated Model) -107.4409 Difference in Log Likelihoods .1133*2= .23

PAGE 318

295 Appendix U-54. Emergency Department Visits, Interacti on Style, and Provider Office Ratings by WIC/Healthy Start Particip ation (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.213.2281-.234.660 .87 .351 Secure-.016.0311-.077.046 .25 .617 Avoidant.013.0311-.048.075 .17 .676 Anxious.020.0331-.045.084 .35 .553 WIC/Healthy Start .319.4301-.5261.164 .55 .459 Saturated Model Provider Office.217.2331-.240.673 .87 .354 Secure.029.1031-.174.231 .08 .781 Avoidant.069.1051-.138.275 .43 .514 Anxious.144.1771-.202.490 .66 .415 WIC/Healthy Start 5.0814.5441-3.82513.988 1.25 .264 WIC/HS*Secure-.050.1091-.263.163 .21 .645 WIC/HS*Anxious-.128.1801-.481.224 .51 .476 WIC/HS*Avoidant-.061.1101-.277.155 .31 .579 Log Likelihood (Main Effect Model) -107.3575 Log Likelihood (Saturated Model) -106.7519 Difference in Log Likelihoods .6056 *2= 1.21

PAGE 319

296 Appendix U-55. Emergency Department Visits, Interacti on Style, and Provider Office Ratings by Mothers Employment Stat us (Adjusted) (N=126) Poisson Regression B SE dfWald CI Limits Wald Sig. Main Effects Model Provider Office.207.2301-.244.658 .81 .368 Secure-.018.0311-.079.42 .35 .552 Avoidant.011.0321-.052.073 .11 .740 Anxious.021.0341-.045.087 .38 .536 Mother Worked > 30 Hours.096.2941-.480.673 .11 .744 Saturated Model Provider Office.255.2391-.212.723 1.15 .285 Secure.009.0351-.059.077 .07 .796 Avoidant.026.0381-.048.100 .47 .493 Anxious.035.0411-.045.116 .73 .392 Mother Worked > 30 Hours7.0403.1641-.83813.242 4.95 .026 Work*Secure-.160.0861-.328.008 3.50 .062 Work *Anxious-.078.0711-.218.061 1.22 .270 Work *Avoidant-.097.0791-.250.057 1.51 .219 Log Likelihood (Main Effect Model) -107.6035 Log Likelihood (Saturated Model) -104.9773 Difference in Log Likelihoods 2.6262 *2= 5.25

PAGE 320

297 Appendix U-56. Emergency Department Visits, Interacti on Style, and Provider Office Ratings by Mothers Feelings About Provide r Offices (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.132.2341-.328.591.32 .574 Secure-.018.0311-.079.043.32 .570 Avoidant.010.0311-.052.071.09 .758 Anxious.016.0341-.050.082.23 .631 Mothers Feelings .053.0501-.045.1501.13 .288 Saturated Model Provider Office.257.2521.237.7501.04 .308 Secure-.106 .0821.266.0551.66 .198 Avoidant.010.0771.141.162.02 .893 Anxious-.118.0611.238.0023.69 .055 Mothers Feelings -.673.4981-1.648.3031.83 .177 Feeling*Secure.017.0141-.001.0441.64 .200 Feelings*Anxious.027.0111.006.0496.23 .013 Feelings*Avoidant-.003.0121-.026.020.06 .806 Log Likelihood (Main Effect Model) -107.0947 Log Likelihood (Saturated Model) -103.7700 Difference in Log Likelihoods 3.3247*2= 6.65

PAGE 321

298 Appendix U-57. Emergency Department Visits, Interacti on Style, and Provider Office Ratings by Mothers Age (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.127.2301-.323.577.30 .581 Secure-.020.0311-.081.042.40 .527 Avoidant.013.0311-.048.074.18 .672 Anxious.010.0331-.055.074.08 .771 Mothers Age-.044.0221-.087-.0013.96 .047 Saturated Model Provider Office.216.2351-.245.676.84 .359 Secure.359.1391.086.6316.66 .001 Avoidant.044.1541-.259.346.08 .778 Anxious.179.1451-.105.4641.53 .216 Mothers Age.401.1991-.011.7914.07 .044 Age*Secure-.015.0051-.025-.0058.19 .004 Age*Anxious-.001.0061-.018.0041.63 .202 Age*Avoidant-.001.0061-.014.011.05 .830 Log Likelihood (Main Effect Model) -105.5343 Log Likelihood (Saturated Model) -100.6893 Difference in Log Likelihoods 4.8450*2= 9.69

PAGE 322

299 Appendix U-58. Emergency Department Visits, Interacti on Style, and Provider Office Ratings by Mothers Bonding Score (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.273.2311-.180.726 1.39 .238 Secure-.016.0311-.078.045 .27 .605 Avoidant.008.0311-.053.069 .07 .797 Anxious.011.0331-.054.076 .11 .743 Mothers Bonding Score.247.1091.033.461 5.13 .024 Saturated Model Provider Office.287.2351-.173.747 1.50 .221 Secure-.017.0321-.080.045 .30 .586 Avoidant.026.0331-.040.091 .59 .443 Anxious.010.0341-.058.077 .08 .774 Mothers Bonding Score1.011.1251-1.2043.21 .79 .373 Bonding*Secure.006.0351-.063.074 .03 .869 Bonding*Anxious.015.0301-.042.073 .27 .602 Bonding*Avoidant-.046.0331-.109.018 1.97 .161 Log Likelihood (Main Effect Model) -105.3010 Log Likelihood (Saturated Model) -104.1379 Difference in Log Likelihoods 1.1721 *2= 3.44

PAGE 323

300 Appendix U-59. Predicted Emer g enc y Department Visits b y Secure Interaction St y le and Child Health Ratin g Controllin g for Provider Office and Interaction Styles (Secure, Anxious, Avoidant)0 10 20 121314151617181920212223242526 Secure Interaction ScorePredicted Visits 2 4 6 8 10

PAGE 324

301 Appendix U-60. 0 100 200 300 400 500 15182124273033 Anxious Interaction ScorePredicted Visits 18 24 30 36 45Predicted Emergency Department Visits by Secure Interaction Style and Mother's Age Controlling for Provider Office and Interaction St y les

PAGE 325

302 Appendix U-61. Immunizations, Interaction Style, and Provi der Ratings by Race/Ethnicity (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider.009.0491-.088.106 .03 .856 Secure.002.0071-.011.015 .11 .739 Avoidant-.000.0071-.014.013 .00 .949 Anxious.002.0071-.012.016 .07 .788 Black-.001.0601-.117.116 .00 .991 Hispanic.006.0831-.157.168 .00 .945 Saturated Model Provider.013.0501-.086.110 .07 .795 Secure.000.0101-.018.020 .00 .973 Avoidant.008.0101-.012.029 .64 .422 Anxious.001.0101-.020.021 .00 .958 Black.337.5601-.7611.435 .36 .547 Hispanic.146.7031-1.2331.524 .04 .836 Black*Secure.006.0151-.024.035 .14 .711 Black*Anxious.007.0151-.022.035 .19 .660 Black*Avoidant-.023.0161-.054.008 2.07 .150 Hispanic*Secure.003.0181-.033.038 .02 .891 Hispanic*Anxious-.004.0271-.056.048 .02 .877 Hispanic*Avoidant-.006.0191-.043.030 .11 .735 Log Likelihood (Main Effect Model) 2196.3978 Log Likelihood (Saturated Model) 2197.7108 Difference in Log Likelihoods 1.313 *2= 2.62

PAGE 326

303 Appendix U-62. Immunizations, Interaction Style, and Pr ovider Ratings by Birth Order (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider.011.0491-.086.107 .05 .828 Secure.002.0071-.011.015 .10 .755 Avoidant.000.0071-.013.013 .00 .987 Anxious.001.0071-.013.015 .02 .899 Target Child Mothers First.040.0531-.064.145 .57 .452 Saturated Model Provider.009.0501-.0891.07 .03 .853 Secure.003.0081-.014.019 .12 .726 Avoidant.001.0101-.018.020 .01 .918 Anxious.001.0101-.018.021 .01 .907 Target Child Mothers First.135.5071-.8581.128 .07 .790 First Child*Secure-.002.0141-.029.025 .03 .872 First Child*Anxious-.000.0141-.028.027 .00 .985 First Child*Avoidant-.002.011-.028.025 .02 .898 Log Likelihood (Main Effect Model) 2196.6772 Log Likelihood (Saturated Model) 2196.6957 Difference in Log Likelihoods .0185 *2= .04

PAGE 327

304 Appendix U-63. Immunizations, Interaction Style, and Pr ovider Ratings by Childs Health Rating (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider .001.0491-.087.106 .04 .845 Secure .002.0071-.011.015 .12 .731 Avoidant -.000.0071-.013.013 .00 .946 Anxious .002.0071-.012.015 .06 .809 Childs Overall Health Rating -.002.0191-.039.035 .01 .926 Saturated Model Provider .007.0501-.091.106 .02 .888 Secure .003.0601-.116.121 .00 .967 Avoidant .022.0551-.086.130 .17 .985 Anxious -.031.0481-.124.062 .44 .509 Childs Health Rating -.001.2251-.441.439 .00 .995 Childs Health*Secure -.000.0061-.013.013 .00 .989 Childs Health*Anxious .004.0051-.006.014 .50 .479 Childs Health*Avoidant -.003.0061-.014.009 .17 .676 Log Likelihood (Main Effect Model) 2196.3991 Log Likelihood (Saturated Model) 2196.6614 Difference in Log Likelihoods .2623 *2= .52

PAGE 328

305 Appendix U-64. Immunizations, Interaction Style, and Pr ovider Ratings by Mothers Health Rating (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider .010.0491-.086.106.04 .8380 Secure .003.0071-.011.016.15 .697 Avoidant -.000.0071-.013.013.00 .958 Anxious .001.0071-.012.015.04 .844 Mothers Health Rating -.005.0151-.034.024.11 .745 Saturated Model Provider .013.0501-.084.111.07 .790 Secure -.003.0331-.067.061.01 .929 Avoidant -.013.0321-.076.050.17 .679 Anxious .007.0291-.051.065.05 .819 Mothers Health Rating -.050.1461-.337.236.12 .731 Mothers Health *Secure .001.0041-.007.009.03 .857 Mothers Health *Anxious -.001.0041-.008.006.03 .855 Mothers Health*Avoidant .002.0041-.006.009.17 .680 Log Likelihood (Main Effects Model) 2196.4477 Log Likelihood (Intera ction Model) 2196.5377 Difference in Log Likelihoods .0900*2= .18

PAGE 329

306 Appendix U-65. Immunizations, Interaction Style, and Provider Ratings by WIC/Healthy Start Participation (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider .010.0491-.086.106 .04 .838 Secure .002.0071-.011.015 .13 .720 Avoidant -.000.0071-.013.013 .00 .963 Anxious .002.0071-.012.05 .08 .782 WIC/Healthy Start .024.0831-.139.187 .08 .772 Saturated Model Provider .011.0501-.087.108 .04 .833 Secure .002.0191-.035.039 .01 .928 Avoidant -.003.1971-.042.035 .03 .863 Anxious .003.0301-.056.062 .01 .916 WIC/Healthy Start -.060.8651-1.7551.636 .00 .945 WIC/HS*Secure .001.2021-.039.040 .00 .967 WIC/HS*Anxious -.002.0311-.062.059 .00 .962 WIC/HS*Avoidant .004.0211-.038.045 .03 .866 Log Likelihood (Main Effect Model) 2196.4370 Log Likelihood (Saturated Model) 2196.4523 Difference in Log Likelihoods .3947 *2= .79

PAGE 330

307 Appendix U-66. Immunizations, Interaction St yle, and Provider Ratings by Mothers Employment Status (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider.002.0491-.095.099 .00 .975 Secure.002.0071-.011.015 .11 .740 Avoidant.001.0071-.012.014 .02 .879 Anxious-.000.0071-.014.014 .00 .993 Mother Worked > 30 Hours-.082.0651-.209.045 1.60 .207 Saturated Model Provider-.001.0501-.098.097 .00 .987 Secure.003.0071-.012.017 .13 .719 Avoidant.000.0081-.016.016 .00 .998 Anxious-.002.0081-.018.015 .03 .858 Mother Worked > 30 Hours-.293.6561-1.578.992 .20 .655 Work*Secure-.002.0171-.035.032 .01 .931 Work *Anxious007.0161-.024.038 .18 .670 Work *Avoidant.006.0161-.025.037 .15 .702 Log Likelihood (Main Effect Model) 2197.2033 Log Likelihood (Saturated Model) 2197.4437 Difference in Log Likelihoods .2404 *2= .48

PAGE 331

308 Appendix U-67. Immunizations, Interaction Style, and Provi der Ratings by Mothers Feelings About Providers (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider.010.0501-.087.107 .04 .838 Secure.002.0071-.011.015 .11 .739 Avoidant-.000.0071-.013.013 .00 .947 Anxious.002.0071-.012.015 .08 .783 Mothers Feelings -.001.0101-.022.019 .02 .891 Saturated Model Provider-.006.0511-.106.094 .02 .899 Secure.018.0171-.015.052 1.12 .289 Avoidant-.009.0161-.039.022 .30 .585 Anxious.014.0151-.015.043 .86 .354 Mothers Feelings .071.1101-.145.287 .41 .520 Feeling*Secure-.003.0031-.009.003 1.09 .296 Feelings*Anxious-.003.0031-.008.003 .97 .324 Feelings*Avoidant.002.0031-.003.007 .39 .534 Log Likelihood (Main Effect Model) 2196.4041 Log Likelihood (Saturated Model) 2197.5797 Difference in Log Likelihoods 1.1756 *2= 3.51

PAGE 332

309 Appendix U-68. Immunizations, Interaction Style, and Provi der Ratings by Mothers Age (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider.003.0451-.095.100 .00 .958 Secure.002.0071-.011.015 .09 .768 Avoidant-.000.0071-.013.013 .00 .963 Anxious.001.0071-.013.015 .03 .873 Mothers Age-.003.0041-.011.005 .53 .465 Saturated Model Provider.002.0501-.097.100 .00 .975 Secure-.001.0301-.060.058 .00 .973 Avoidant-.018.0301-.078.040 .37 .544 Anxious.005.0291-.052.063 .03 .858 Mothers Age-.020.0391-.096.056 .27 .604 Age*Secure.000.0011-.002.002 .01 .915 Age*Anxious-.000.0011-.002.002 .02 .895 Age*Avoidant.001.0011-.002.003 .38 .539 Log Likelihood (Main Effect Model) 2196.6622 Log Likelihood (Saturated Model) 2196.8605 Difference in Log Likelihoods .1983 *2= .40

PAGE 333

310 Appendix U-69. Immunizations, Interaction Style, and Pr ovider Ratings by Mothers Bonding Score (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider.008.0501-.088.105 .03 .860 Secure.002.0071-.011.015 .11 .743 Avoidant-.000.0071-.013.013 .00 .962 Anxious.002.0071-.012.015 .08 .480 Mothers Bonding Score-.001.0271-.060.044 .09 .764 Saturated Model Provider.007.0501-.090.104 .02 .886 Secure.002.1171-.011.015 .11 .745 Avoidant-.001.0071-.014.013 .01 .925 Anxious.002.0071-.012.015 .06 .804 Mothers Bonding Score-.078.2661-.599.444 .09 .770 Bonding*Secure.001.0081-.015.016 .01 .933 Bonding*Anxious.003.0071-.012.017 .12 .728 Bonding*Avoidant.001.0071-.013.014 .01 .940 Log Likelihood (Main Effect Model) 2196.4399 Log Likelihood (Saturated Model) 2196.5275 Difference in Log Likelihoods .0876 *2= .18

PAGE 334

311 Appendix U-70. Immunizations, Interaction Style, and Provider Office Ratings by Race/Ethnicity (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.015.0511-.086.115 .08 .776 Secure.002.0071-.011.015 .11 .742 Avoidant-.000.0071-.014.013 .00 .968 Anxious.002.0071-.012.016 .08 .781 Black.001.0601-.116.118 .00 .992 Hispanic.005.0831-.158.168 .00 .951 Saturated Model Provider Office.025.0521-.076.127 .24 .628 Secure.000.0101-.019.019 .00 .976 Avoidant.009.0111-.012.030 .72 .395 Anxious.000.0101-.020.021 .00 .968 Black.340.5601-.7571.437 .37 .543 Hispanic.157.7041-1.2241.537 .05 .824 Black*Secure.006.0151-.023.035 .15 .696 Black*Anxious.007.0151-.022.036 .23 .634 Black*Avoidant-.024.0161-.055.008 2.17 .141 Hispanic*Secure.002.0181-.034.038 .01 .908 Hispanic*Anxious-.004.0271-.056.048 .02 .880 Hispanic*Avoidant-.007.0191-.043.030 .12 .725 Log Likelihood (Main Effect Model) 2196.4219 Log Likelihood (Saturated Model) 2197.7942 Difference in Log Likelihoods 1.3723 *2= 2.75

PAGE 335

312 Appendix U-71. Immunizations, Interaction Styl e, and Provider Office Ratings by Birth Order (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.016.0511-.084.116 .10 .757 Secure.002.0071-.011.015 .09 .762 Avoidant.000.0071-.013.013 .02 .963 Anxious.001.0071-.013.015 .02 .890 Target Child Mothers First.040.0531-.064.144 .56 .453 Saturated Model Provider Office.016.0511-.084.117 .10 .750 Secure.003.0081-.013.019 .13 .717 Avoidant.002.0101-.018.020 .02 .879 Anxious.001.0101-.018.020 .01 .905 Target Child Mothers First.154.5011-.8281.136 .09 .759 First Child*Secure-.003.0141-.030.024 .04 .848 First Child*Anxious-.000.0141-.028.027 .00 .994 First Child*Avoidant-.002.0131-.029.024 .03 .867 Log Likelihood (Main Effect Model) 2196.7016 Log Likelihood (Saturated Model) 2196.7293 Difference in Log Likelihoods .0277 *2= .06

PAGE 336

313 Appendix U-72. Immunizations, Interaction Style, and Provi der Office Ratings by Childs Health Rating (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.015.0511-.0846.115 .09 .765 Secure.022.0071-.011.015 .11 .736 Avoidant-.000.0071-.013.023 .00 .971 Anxious.002.0071-.012.015 .06 .801 Childs Overall Health Rating -.002.0191-.039.035 .01 .924 Saturated Model Provider Office.010.0531-.093.114 .04 .847 Secure.002.0611-.118.121 .00 .981 Avoidant.022.0551-.087.130 .16 .693 Anxious-.031.0481-.124.063 .41 .520 Childs Health Rating-.004.2261-.447.438 .00 .985 Childs Health*Secure.000.0071-.013.013 .00 .998 Childs Health*Anxious.004.0051-.007.014 .48 .490 Childs Health*Avoidant-.002.0061-.014.009 .16 .687 Log Likelihood (Main Effect Model) 2196.4246 Log Likelihood (Saturated Model) 2196.6702 Difference in Log Likelihoods .2456 *2= .49

PAGE 337

314 Appendix U-73. Immunizations, Interaction Styl e, and Provider Office Ratings by Mothers Health Rating (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.015.0511-.085.114 .08 .775 Secure.003.0071-.011.016 .14 .705 Avoidant-.000.0071-.013.013 .00 .982 Anxious.002.0071-.012.015 .04 .834 Mothers Overall Health Rating -.005.0151-.034.025 .10 .758 Saturated Model Provider Office.016.0511-.084.116 .10 .751 Secure-.003.0331-.067.061 .01 .926 Avoidant-.013.0321-.074.050 .16 .694 Anxious.007.0301-.051.065 .05 .818 Mothers Health Rating-.048.1451-.333.236 .11 .739 Mothers Health*Secure.001.0041-.007.009 .03 .857 Mothers Health*Anxious-.001.0041-.008.006 .03 .857 Mothers Health*Avoidant.002.0041-.006.009 .16 .690 Log Likelihood (Main Effect Model) 2196.4675 Log Likelihood (Saturated Model) 2196.5526 Difference in Log Likelihoods .0851 *2= .17

PAGE 338

315 Appendix U-74. Immunizations, Interaction Style, and Pr ovider Office Ratings by WIC/Healthy Start Participation (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.016.0511-.086.116 .10 .749 Secure.002.0071-.011.015 .12 .724 Avoidant-.000.0071-.013.013 .00 .990 Anxious.002.0071-.012.015 .08 .774 WIC/Healthy Start .026.0831-.138.189 .09 .759 Saturated Model Provider Office.016.0531-.088.120 .09 .761 Secure.001.0191-.036.038 .00 .952 Avoidant-.002.0201-.041.038 .01 .922 Anxious.004.0301-.055.062 .02 .092 WIC/Healthy Start -.031.8561-1.7091.65 .00 .971 WIC/HS*Secure.001.0201-.039.041 .00 .946 WIC/HS*Anxious-.002.0311-.062.058 .00 .952 WIC/HS*Avoidant.002.0211-.040.044 .01 .918 Log Likelihood (Main Effect Model) 2196.4672 Log Likelihood (Saturated Model) 2196.4763 Difference in Log Likelihoods .0091 *2= .02

PAGE 339

316 Appendix U-75. Immunizations, Interaction Style, and Provi der Office Ratings by Mothers Employment Status (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.015.0511-.084.114 .09 .769 Secure.002.0071-.011.015 .12 .733 Avoidant.001.0071-.012.015 .03 .856 Anxious-.000.0071-.014.014 .00 .994 Mother Worked > 30 Hours-.082.0641-.208.044 1.63 .202 Saturated Model Provider Office.016.0521-.085.117 .10 .753 Secure.003.0071-.011.017 .15 .695 Avoidant.000.0081-.015.016 .00 .998 Anxious-.001.0081-.018.015 .03 .874 Mother Worked > 30 Hours-.276.6551-1.5591.007 .18 .674 Work*Secure-.002.0171-.036.032 .02 .897 Work *Anxious.006.0161-.025.037 .15 .697 Work *Avoidant.006.0161-.024.037 .16 .689 Log Likelihood (Main Effect Model) 2197.2459 Log Likelihood (Saturated Model) 2197.4928 Difference in Log Likelihoods .2469 *2= .49

PAGE 340

317 Appendix U-76. Immunizations, Interaction Style, and Provi der Office Ratings by Mothers Feelings About Provider Offices (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office.017.0521-.085.119 .11 .744 Secure.002.0071-.011.015 .11 .745 Avoidant-.000.0071-.013.013 .00 .974 Anxious.002.0071-.012.016 .08 .771 Mothers Feelings -.002.0111-.023.019 .03 .857 Saturated Model Provider Office.002.0541-.103.108 .00 .968 Secure.018 .0171-.016.051 1.04 .307 Avoidant-.009.0161-.039.022 .30 .582 Anxious.014.0151-.016.043 .85 .358 Mothers Feelings .067.1101-.150.283 .37 .545 Feeling*Secure-.003.0031-.009.003 1.01 .315 Feelings*Anxious-.003.0031-.008.003 .96 .328 Feelings*Avoidant.002.0031-.003.007 .40 .529 Log Likelihood (Main Effect Model) 2196.4362 Log Likelihood (Saturated Model) 2197.5726 Difference in Log Likelihoods 1.1364 *2= 2.27

PAGE 341

318 Appendix U-77. Immunizations, Interaction Style, and Provider Office Ratings by Mothers Age (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office .008.0521-.093.011 .03 .873 Secure .002.0071-.011.015 .09 .766 Avoidant -.000.0071-.013.013 .00 .976 Anxious .001.0071-.012.015 .03 .869 Mothers Age -.003.0041-.011.005 .51 .477 Saturated Model Provider Office .007.0521-.095.108 .02 .894 Secure -.000.0301-.060.058 .00 .978 Avoidant -.018.0301-.078.041 .36 .549 Anxious .005.0291-.052.063 .03 .854 Mothers Age -.020.0391-.096.057 .26 .613 Age*Secure .000.0011-.002.002 .01 .921 Age*Anxious -.000.0011-.002.002 .02 .891 Age*Avoidant .001.0011-.002.003 .37 .542 Log Likelihood (Main Effect Model) 2196.6736 Log Likelihood (Saturated Model) 2196.8688 Difference in Log Likelihoods .1952 *2= .39

PAGE 342

319 Appendix U-78. Immunizations, Interaction Style, and Pr ovider Office Ratings by Mothers Bonding Score (Adjusted) (N=126) Poisson Regression B SE df Wald CI Limits Wald Sig. Main Effects Model Provider Office .013.0511-.087.114.07 .799 Secure .002.0071-.011.015.10 .747 Avoidant -.000.0071-.013.013.00 .982 Anxious .002.0071-.012.015.08 .774 Mothers Bonding Score -.001.0271-.060.045.08 .782 Saturated Model Provider Office .012.0521-.089.114.06 .811 Secure .002.0071-.011.015.10 .746 Avoidant -.001.0071-.014.013.00 .944 Anxious .002.0071-.012.015.07 .798 Mothers Bonding Score -.080.2671-.603.442.09 .763 Bonding*Secure .001.0081-.015.017.01 .920 Bonding*Anxious .003.0071-.012.017.12 .734 Bonding*Avoidant .001.0071-.013.014.01 .934 Log Likelihood (Main Effect Model) 2196.4586 Log Likelihood (Saturated Model) 2196.5456 Difference in Log Likelihoods .0870*2= .17

PAGE 343

320 Appendix V Well Child Care Data Recording Issues The current study indicates most infa nts received 6 of the recommended eight visits (American Academy of Pe diatrics, 2002). The differences were primarily in the use or recording of th e visits under one month. Some of this variability is probably due to a lack of a comprehensive medical record. For example, infants may have still been in th e hospital at the time of the 2-4 day well child care visit. Another explanation may be that wo men go to one visit with a pediatric provider, decide they do not want that one and change before the next visit. In both instances, if the researcher does not know to ask or if mother does not remember about those first visits, this information does always get recorded. Even if there is knowledge of the additional information in other medical records, it may not be accessible. More than one woman was excluded from the study due to the inability to obtain all medical records for a child. For example, when a mother changes providers, little of the record is transferred, usually just the immunizations received. The initial prov ider closes the file and often archives the records in an off-site location wher e retrieval is inconvenient, making the already difficult task of obtaining the reco rd nearly impossible. Additionally, if researchers are not aware that the medical record is incomplete at the providers office for which the original release of medical information was obtained, researchers must go back to the mother and obtain another signed release form for the provider who has the additional record s. Given the high rates of mobility in this population, locating the mother for the second contact can also be difficult.

PAGE 344

321 Appendix W. Issues Regarding the Measurement of We ll Child Care Visits and Immunizations The distribution of both the number of well child care visi ts and immunizations is truncated, or restricted, to 8 and 13 respect ively. As a result, the assumptions of the Poisson regression analysis may not have b een met. A more appropriate calculation would have been an analysis referred to as a truncated Poisson Regression. However, currently no software is available that comput es these analyses. Because of the lack of any statistically significant fi ndings associated with these two types of visits and the truncation issues, additional analyses were conducted. Two dichotomous variables were calculated including five or more well ch ild care visits, and twelve or more immunizations. Nine separate main effects models were calculated including provider rating, interaction scores, and one of the ni ne potentially moderating factors. Using a dichotomized well child care visit, Black wome n (p<.05) were more likely to take their child to at least 5 well child care visi ts than White or Hispanic women. Using a dichotomized immunization measure, women who worked more than 30 hours per week were more likely to have received all or most immunizations. In regard to immunizations, a number of other methods have also been used to measure timeliness of immunizations. Fo r example, Glauber (2003) developed a composite measure obtained thr ough averaging a score for each vaccination event. Others have defined delayed initiati on of immunizations as being wh en the first vaccination was not provided until the child was older than 90 days (Gaudino, 2005). This 90 day delay in

PAGE 345

322 initiation was explored briefly in this st udy. Given the 90 day definition, ten percent of participants delayed immunizati ons. Those delays were positiv ely correlated (p<.05) with child health and children in the family. Si milar findings were found when using the dichotomized immunization variable of 12 or more shots (p<.05). The lack of association with the research questions and the correlations with delayed initiation of immuniza tions indicate that more exploration of immunization patterns is needed. One difficulty with this endeavor is the rapidly changing array of vaccinations being received by children. To minimize the number of shots an infant receives, manufactures have been working to combine more vaccines into one shot such as combining the Hib with either the DTaP or HepB vaccine. According to the CDC the potential for combining even more vaccines, such as one shot for DTaP, Hib, HepB, IPV, and Hep A, is being considered (Centers for Disease Control, 2005a). Currently, the seven illnesses for which infants are vaccina ted were generally combined into between three and five shots. Therefore, rather than looking specifically at the number of vaccines received, researchers should explore tre nds in when vaccinations are received.

PAGE 346

323 Appendix X. Measurement of Maternal Feelings About Doctors and Maternal Bonding Maternal Feelings About Doctors Mothers feelings about going to the doc tor was a significant interaction term in models exploring associations with sick and follow-up visits. Women were asked a general question regarding whet her they really hated or en joyed going to the doctor. The original intent behind this question was to tr y to capture whether the mother had a fear about going to the doctor that would keep her from taking her child. Similar to the maternal bonding measure, no instrument was id entified in the litera ture that could be administered within the scope of the full study. As a result, th e measure used in this study was brought before experts in the field for its face validity but had no further testing regarding its reliability or validity. The significant (p<.01) correlation (r=.235) between feelings about doctors and provider office i ndicates the two measures are both capturing a motivating construct that is significant to unde rstanding one aspect of health care utilization patterns. Further expl oration of this issue is need. Maternal Bonding When developing this study, no validated measure to capture maternal bonding that was also short enough to be incorporated into the study was identified. As a result, the bonding measure used was aggregated from information regarding factors known to be associated with bonding. Although various efforts were made to ensure the instruments validity and reliability, it was not exposed to rigorous testing prior to its use. The significant results of this study indicate that there is a need for the development of an abbreviated measure of maternal bonding that has been adequately tested.

PAGE 347

324 Appendix Y. National Comparisons of Ratings of Experiences Of Care Responses regarding experiences of care indicate mothers generally had more positive ratings than those found in the nationa l CAHPS data set. This disparity may be due to the studys selection crit eria restricting the childs age and excluding ill children. Also, Health issues change ra pidly in infants, making delays in care more significant. Upon contacting providers, most mothers reported being told to bring the child to the office that day or to take the child to the emergency department. Longer appointment delays are more common for older children. This increased frequenc y of visits during infancy also enhances development of bonds between the mother and the provider. The variations in provider ra tings locally versus nationa lly could also be due, in part, to Medicaid implementation differences across states. Study part icipants reported no significant difficulties in changing providers. Both mothers and providers reported that women could easily change providers with the longest wait pe riod being one month. One issue related to the operation of the provider office is the ability to make well-child care appointments in a timely manner. Most (73.4%) women reported always getting appointments as soon as they wante d. These rates are higher than those found in the National CAHPS Benchmarking Database (53% Always). However, the number of well child care visits are more freque nt for children 12 months and younger when compared to older children. As a result, most providers schedule the next well-child care

PAGE 348

325 visit upon discharge from the previous visi t. This proactive approach maximizes the opportunities to schedule appointments that are convenient for the mother. Sick and follow-up visits are scheduled in much shorter time intervals making it more difficult for the clinic to offer times th at meet the needs of mothers. Even so, the ratings of the timeliness of sick visits in this study were even higher tan both well child care visits of study participants as well as sick/follow-up visit sc heduling data reported from the Nation CAHPS Benchmarking Da tabase (61% Always, 24% usually, 14 sometimes or never). Another issue that influences the rating of the provider o ffice is the rapport that develops between a mother and health care sta ff. In the case of mothers in the study, most (78.9%) reported staff always treated them w ith courtesy and respect compared to the national CAHPS benchmarking database ratin gs of 72%. However, even though mothers reported providers showed respect for what mo ther had to say, they were less likely to report that providers listened to them car efully. Sixty-two perc ent of mothers said providers always listened to them carefully while nationally, 67% of respondents reported providers listened to them carefully (Age ncy for Health Care Research and Quality, 2004b). Another study found that Hispanics also reported lower ratings on the question regarding whether physicians lis tened to them carefully (M errill & Allen, 2003). This involvement in the childs health care is im portant. For example, research has indicated that women who perceived that they had some level of control over their health status were more likely to be compliant with medical recommendations (Tinsley, Trupin, Owens & Boyum, 1993).

PAGE 349

326 Health Care providers were rated higher than the National CAPHS Benchmarking Database in regard to explai ning things in a way mothers could understand (81.3% versus 69% always). In at least two instances involving mothers in this study, the level of trust was affected by inaccurate diagnoses and the lack of follow-through regarding the results of laboratory tests. Furtherm ore, mothers specifically indicated that they really appreciated when the providers office called the home to follow-up with the child after being seen for an illness.

PAGE 350

327 Appendix Z. Additional Study Findings, Limitations, and Strengths Recruitment Personalized approaches elicited most of the interviews. However, these approaches can be resource intensive and requi re committed individuals in direct contact with potential participants. To prevent referr al biases, a formalized sampling structure is needed. Furthermore, the pace of recruitment efforts and resulting sc heduling issues also influenced the successful completion of the interview. Interviews scheduled within 48 hours of initial contact were more likely to result in completed interviews than those scheduled at a later date. Delays beyond 48 hours usually did not re sult in a successful interview. Another recruitment issue was the lack of inclusion of Hispanic mothers in the study. Attempts were made with in the studys methodology to en sure adequate racial and ethnic representation of infants in this study, ta rgeting one third of th e study participants who were non-Hispanic White, one third w ho were non-Hispanic Black, and one third who were Hispanic. However, only twelve percent of the final study population was Hispanic. Discussions with both clinic staff a nd Healthy Start care c oordinators indicated that the majority of their Hispanic Medicaid clients spoke little to no English excluding them from the study. Informants believed that most Hispanics were reluctant to ask for assistance unless they were in extreme need su ch as new immigrants to the United States who have limited employment opportunities. As a result, bilingual approaches are needed to better include these women in future research.

PAGE 351

328 Inclusion of Hispanic indi viduals in research such as the current study are needed because some of the greatest health care disparities are among this population. More specifically, Hispanic individua ls have reported more issues with obtaining care, going without care, having a usual source of care, and being convinced that family members could receive needed care (Phillips, Maye r, & Aday, 2000). Hispanic respondents also reported more often that their providers did no t provide them with needed information or listen to them carefully. Disappearing Racial Differences The body of public health literature contains a number of studies describing racial and ethnic differences in health care utiliza tion. For example, one study regarding racial and ethnic disparities in the use of pregnancy-related health care among women receiving Medicaid-funded prenatal care found that minority women were less likely to receive services that the woman initiates, discre tionary services, and services potentially requiring specialized follow-up care, whereas, th ey were more likely to receive screening for diseases related to high-risk behaviors (Gavin, Adams, Hartman, Benedict & Chireau, 2004, p113). Another study found that for women of African American descent, those who had less than a high school education, were not married, had multiple children, were not participating in WIC even though they were eligible and t hose below 50% of the federal poverty level were less likely to ta ke their children to receive adequate immunizations (Luman et al., 2003). Consistent with the literatur e, the current study identifie d a number of significant bivariate correlations of Bl ack, non-Hispanic women with being in a single household,

PAGE 352

329 having less education, younger age, more avoi dant interaction st yle, less bonding, lower child health status, and mothers who worked at least thirty hours per week. However, when race was included in a more comp lex model containing other independent variables, confounders, and/or moderators, these differences became non-significant. This finding regarding the ch anging level of association be tween race and health care utilization factors is important to emphasi ze because many studies only report bivariate associations leading to the conclusion that Black race, a non-changeable factor, is the primary contributing factor. In reality, the inter-relationships am ong factors is more complex and many of these other factors are able to be modified to some extent. Improvements to health indicators have begun to plateau in recent years. For further progress to be made, the there is a need to build upon the current l iterature and explore these multi-faceted associations, focusi ng on factors that can be modified. For example, one factor that may impact experiences of care a nd utilization rates is the racial concordance between Black, nonHispanic women and their pediatric health care provider. Racial concordance has been shown to improve satisfaction with care (Laveist & Nuru-Jeter, 2002). Study partic ipants who were Black, non-Hispanic were half (24%) as likely to report seeing a pediatric health care provider that was the same race/ethnicity as White, non-Hispanic a nd Hispanic mothers in the study (49.4%). Focusing specifically on the associations betw een racial concordance and pediatric health care may provide addition insight into utilization patterns.

PAGE 353

330 Need for Electronic Medical Record The barriers faced during the data collect ion process for this study highlighted the need for a continuous medical record for study participants to ensure that all necessary information is present. This can be achiev ed through a variety of methods. For example, future studies could narrow their inclusion criteria to mothers who have never changed their childs health care provi der office. However, this would greatly limit the population from which to sample and make recruitment ev en more difficult. Similar to the prenatal passport, mothers could maintain their ow n infant health care passport as well. A solution to address the issue of incomplete medical records gaining more interest is the use of an electronic medical record. The scope of access to this medical record is currently limited. Many of the providers offices visite d in this study were in the process of putting their patient files in the computer. Efforts to expand access to multiple sites owned by the same provider also are under way. The interest has gotten so great that one retailer had more than 100 different electr onic medical record software applications in their inventory that were devel oped for this purpose (Capterra, 2005). Within Florida, the Department of Hea lth provides a significant portion of the childhood immunizations for its residents. To share the immunization histories of children in a confidential manner across multiple providers, the Florida Department of Health has recently implemented a webbased, password-protected state-wide immunization registry (Florida Department of Health, 2005c). Although moving in the right di rection, the existence of a more global system is still in need. For at least a decade, discussi ons regarding a national medical record have occurred (Kohane, Greenspun, Fackler, Ci mino & Szolovits, 1996). More recently, the

PAGE 354

331 January, 2004, presidential stat e of the union address outlined a plan that would provide most Americans with electronic medical reco rds within the next ten years (Bush, 2004). The benefits of an electronic medical record system are many in cluding avoiding some dangerous medical mistakes, to be able to reduce costs, such as through duplication of services, and improve patient care by having all pertinent information at hand. One example of the progress has been ma de regarding electronic medical records is the AHLTA internet-based electronic medi cal system developed with off-the-shelf software for the United States Military (Gilmore, 2005). This system has been implemented for about 60% of the military, incl uding in combat situations. It is projected that by the end of 2007, this sy stem will serve over nine m illion service members, their families, and retirees globally. Martial Status Reliability There were racial differences in marital ra tes which may be due, in part to the age differences among the racial and ethnic groups However, marital st atus was a variable that has questionable reliability. Medicaid income guidelines often prompt women to falsely claim to be single even if they may ha ve a live-in partner. For example, one of the women interviewed indicated she was singl e even though a man came to the door and asked if the womans husband was home. Th e woman replied by saying yes and woke up a man sleeping on her couch. Conversely, the informed consent process clearly stated that information provided would not be shared with anyone. As a result of this process and experiences, it is believed that some women reported bei ng married or have a live-in partner who may not have reported that relationship to Medicaid.

PAGE 355

332 Family Planning and Baby Spacing At least 9.2% of women did not adequate ly space their conception experiences by getting pregnant within a few m onths of delivering the infant being targeted in this study. In a somewhat extreme case, one 31 year old mother in the study who had a total of ten children experienced 4 consecutive pregnancies resulting in 5 infants under the age of 26 months. Although many subsequent pregnancie s result in healthy infants, becoming pregnant again too soon significantly increases the risk of fetal death and inferior growth and development (physically and intellectually)(Senanayake, 1982). Furthermore, nearly two-thirds (60.3 %) of mothers had not intended on getting pregnant at the time they did. Similarly, st atewide rates indicate that 64% of women receiving prenatal Medicaid services reporte d that the current pr egnancy was unintended (Florida Department of Health, 2004). Women were not asked whether they had experienced an abortion, spontaneous or therap eutic, after the birth of the child involved in the study. As a result, unintended pre gnancies and inadequate spacing between pregnancies is even greater issu e than is currently being measur ed through birth histories. Clinic Environment The clinic environments varied greatly in the various clinics visi ted as part of this study. Some clinic waiting areas lacked any co lor and dcor (childre ns toys, television, reading material), making them uncomfortable to sit in. This is especially true given how long some women had to wait before being s een by the provider. In comparison, some clinics had beautifully decora ted waiting rooms that addressed the needs of children and

PAGE 356

333 mothers including toys, reading material, brea st feeding rooms and separations between sick and well children. Given that the clinic waiting area is the begi nning of a health care experience, it should be inviting, respect ful, and functional for its users. Study Limitations The first issue to be addressed is the thre at to internal validity related to the study not being an experimental design. As a result, the data used in this research cannot be used to make causal inferences about the da ta but rather focuses on correlations among study variables. There are also limitations to the study th at restrict the generalizability of the findings. An issue common to most studies is the selection bias that results from the voluntary nature of recruiting participants. Th ere may be significan t differences between those who choose to particip ate in the study and those who refuse participation. Randomization of the days, times and locations of the clinics attempted to minimize some of those biases among women recruited in the clinics, or at least equalize their effect across individuals whose lifestyle s lead them to attend the clinics at different times. The restrictive inclusion and exclusion crit eria as well as the low response rates, also added selection bias to the study with pa rticipants being health ier and generally more satisfied with their health care services th an national comparison groups. Even so, there were significant difference in re spondents ratings of provide rs and utilization of care. Comparisons were made between study part icipants and the state-wide Medicaid population to illustrate some of the differences. Study participants were somewhat older than the Medicaid population of mothers havi ng 11.1% of prenatal Medicaid participants

PAGE 357

334 being age 19 and younger (versus 23%), 31% were age 20-24 (versus 36%), 43.7% were age 25 to 34 (versus 33%), and 14.3% (ver sus 8.4%) were age 35 and over (Florida Department of Health, 2004). This lag time between the rates of prenatal Medicaid participants and infant Medi caid participants of approxi mately two years accounts for some of the age differences. This lag also allows more time to develop more long-term relationships (married and livein partners), and develop more maternal health issues. Additionally, women who volunt eered for the study expressed increased interest when they found out the study was part of a graduation requirement. Many of the mothers were in the process of earning higher leve ls of education and wanted to support the research of another student. Mothers in th e study achieved an aver age of 12.4 years of education, with 34 (25.8%) havi ng less than a high school diploma or GED, 49 (37.1%) having a high school diploma or GED, and 32 (24.2%) having at least a Bachelors Degree [Table 3]. Statewide, 37% of new mo thers had less than a high school education, 45% had a high school educa tion, and 19% had more than a high school education (Florida Department of Health, 2004). Another restriction is the age of the res pondent. In most cases, in order to collect information from individuals under 18 years of age, the protection of human subjects requires parental consent. Given the natu re and structure of the study design, this approval process was not feasible. Therefor e, the study excluded mothers who were under 18 years of age at the time of the interview. However, the inclusion criteria did not exclude an excessive number of women due to age (272, 1.1%). The 28-month time span between becoming pregnant and having a child 18 months of age allo ws for the inclusion of women who became pregnant at 15 y ears of age. There were approximately 272

PAGE 358

335 births (1.1% of all births) born to youth under 16 1/4 years of age in Pinellas and Hillsborough Counties in 2004 (Florida Department of Health, 2005). The low number of Hispanic mothers who participated in the study limited the ability to detect smaller differences in responses by ethnic category. Furthermore, excluding other racial and ethnic groups from the study also limits th e generalizability of the data. Since the number of births to indivi duals in these other ra cial and ethnic groups are so small (122, .05% of all births) in th e Tampa Bay area, the information lost was minimal (Florida Department of Health, 2005) However, a broader perspective of the issue illustrates a need for more studies. Th ese studies should use research designs more suited to the population characteristics includi ng ensuring that data collection instruments are adequately validated fo r the population of interest. The study also excluded infants with poten tially chronic illnesses such as those spending time in the neonatal intensive care unit prior to going home from the hospital. These special needs children generally requ ire more intensive health care than is normally provided and the nature of those visi ts is different. This increased utilization creates a bias in at least two ways. First, th e number and pattern of he alth care visits for chronically ill children would be different from an infant w ith a more typical level of health. Furthermore, the mothers reported experiences of care may be influenced by the different level of services pr ovided. This influence on percepti ons is also the reason that only women attending a health care visit or follow-up to a sick visit were interviewed. The study focused only on mothers with ch ildren receiving Medicaid, restricting the generalizability of the results to only a pproximately half of all infants born in the target area (Agency for Health Care Administ ration, 2004). The benefit of this restriction

PAGE 359

336 is that the variability of individual characteri stics is minimized. Furthermore, this concentration of women and infants who were at increased risk for underutilization of health care provided a greater opportunity to identify the underlying fa ctors that influence utilization within this high-risk population. From a design perspective, the study was limited by the specific items on the survey instruments. The questions included in these instruments may not adequately capture the true complexity of the factors dr iving health care utiliz ation. One issue is the potential ceiling effect for the number of immunizations being lim ited to a maximum of thirteen shots. Although the Po isson regression analysis is specifically designed for such count data, the limited possible number of vi sits and shots may not allow sufficient dispersion to detect st atistical difference. Another issue is due to much of the data relying on self-repo rted recall regarding health care experiences. Carefu l assessment of the information to be collected was made to determine the most appropriate method for ga thering the data. It is for this reason that specific details regarding the number of health care visits and types of immunizations were abstracted from the medical record. Th e mother was also asked about health care services so that visits to other health care providers may be included in the data. Finally, for the purpose of this study a comp romise was made between the need to capture and analyze more information regard ing factors that infl uence health care utilization than has been repor ted in the literature and th e feasibility of collecting a sample size large enough to include a dditional variables in the analyses.

PAGE 360

337 Strengths of the Study One of the greatest strengths of this e xploratory study is in its design. A face-toface interview format was chosen to allow for collection information women believed pertinent to the discussion that was not a ddressed in the questions within interview protocol. This allowed for a better understa nding of the context fr om which the mothers in the study operate. Furthermore, conducting fa ce-to-face interviews reduces, and in the case of this study eliminated, issu es involved with missing data. Additionally, the inclusion and exclusion cr iteria were specifically designed to establish the most homogeneous population whil e also retaining the largest sample size possible. This included Medicaid mothers with children having no chronic illnesses requiring excessive health care services. The re sult was a set of criteria that focused the study on a population that is at increased risk for poor heal th outcomes and inadequate preventive health care utilizat ion rates. Subsequently, alt hough the findings from this study can only be generalized to a proportion of the pediat ric Medicaid population, the study findings may provide the a dditional insight needed for in terventions to continue in their progress towards eliminating health disparities among vulnerable populations. Another strength of this study is the novel approach of exploring the role of attachment theory in the expression of hea lth behaviors. Progress in improving health outcomes and disparities in t hose outcomes among different sub-populations has slowed in recent years. To continue moving forward, there is a need to explore the more complex nature of human behavior w ithin a health care setting. Fo cusing on theories, such as attachment theory, provides a foundation for understanding the dynamics among the multitude of motivators that drive human behavior. The literature in the area of

PAGE 361

338 interaction styles and health care utilization is limited. As a result, the breadth of the questions for this explorat ory study and the novelty of the topic provide guidance regarding a number of issu es for future research. Additionally, to enhance the validity and reliability of the instrumentation used in this study, pre-existing questions were used whenever possible. For example, the Relationships Scales Questionnaire was used to measure interaction style. The Consumer Assessment of Health Plans questions were used to assess the mothers reported experiences of care questions with her pediatri c health care provider. Finally, most of the demographic questions were obt ained from other sources such as the US Census Bureau (2004).

PAGE 362

339 About the Author In 1988, Wendy received her Bachelors Degree in Psychology from Trinity University in San Antonio, Texas. She earned a Masters De gree in Educational Measurement and Research at the University of South Florida in 2001 and a Masters Degree in Epidemiology at Tulane University in 1997. She taught four years in a motor behavior laboratory, introductory biostatistic s at Tulane, and Contem porary Health at the University of South Florida. In 2000, Mrs. Shellhorn was recogni zed nationally by the Coalition for Excellence in MCH Epidemiology as the Outstanding Young Professional. In 2005, she was inducted into the Phi Kappa Phi and Omicron Delta Kappa National Leadership Honor Societies. She served on the Mammography Voucher Program Board and volunteered for the Komen Breast Can cer Foundation. Finall y, she co-authored multiple published articles, submitted two a dditional articles for publication, presented numerous papers at national conferences and pa rticipated in the National Fetal and Infant Mortality Review process.


xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001789990
003 fts
005 20070612092655.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 070612s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001478
035
(OCoLC)143607781
040
FHM
c FHM
1 100
Shellhorn, Wendy Lauran Struchen.
0 245
Maternal interaction style, reported experiences of care, and pediatric health care utilization
h [electronic resource] /
by Wendy Lauran Struchen Shellhorn.
260
[Tampa, Fla] :
b University of South Florida,
2006.
520
ABSTRACT: U.S. immunization and well child-care rates are below desired levels with lower income individuals being at higher risk for receiving inadequate care. To enhance the understanding of motivating factors to health care utilization, this study explored relationships between a mother's interaction style (secure, anxious, avoidant), her reported experiences with pediatric health care and her child's utilization of pediatric health care. Participants included 126 US-born, English-speaking women with an infant 12 to18 months of age. Linear regression analyses found no bivariate associations between maternal interaction style and reported experiences of care. Poisson regression analyses measured associations of maternal interaction style, reported experiences of care, and moderating variables with health care visits and immunizations received. Main effect models found no associations between maternal interaction style and reported experiences of care. Significant associations were identified between provider ratings and sick visits. There were no associations between provider office ratings and utilization rates. When interaction style and provider/provider office ratings were included in the model, high provider ratings (P < .05) and high anxious interaction scores (p < .0001) were associated with more sick visits while higher avoidant interaction style scores (p < .01) were associated with decreased use of sick visits. Multivariate modeling identified provider rating (p < .05) and anxious interaction score (p < .01) as main effects, child's health rating as a confounder, as well as target child being mother's first, WIC/Healthy Start participation, maternal bonding and feelings about going to the doctor acting as moderators to associations between interaction style and sick/follow-up visits. Secure interaction style scores were associated with increased use of emergency department visits, controlling for the confounding effects of maternal bonding andmoderating effects of child's health status and maternal age. Findings indicate that, in some cases, maternal interaction style is associated with how and when mothers access health care for their children. The confounders and moderators identified also highlight the need for more understanding regarding what motivates individuals. Finally, there were racial and ethnic differences including higher rates of avoidant interaction styles in Black, non-Hispanic mothers. Predicting health care utilization patterns will help better target the specific needs of mothers and ultimately improve health outcomes.
502
Dissertation (Ph.D.)--University of South Florida, 2006.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 338 pages.
Includes vita.
590
Adviser: Karen "Kay" Perrin, Ph.D.
653
Attachment.
Health.
Utilization.
Bonding.
Experiences of care.
Infant health.
Immunization.
Well child care.
690
Dissertations, Academic
z USF
x Public Health
Doctoral.
049
FHMM
090
RA425 (ONLINE)
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1478