USF Libraries
USF Digital Collections

Correlates of attitudes toward behavioral health services among older primary care patients

MISSING IMAGE

Material Information

Title:
Correlates of attitudes toward behavioral health services among older primary care patients
Physical Description:
Book
Language:
English
Creator:
Lynn, Nancy Bridger
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Stigma
Mental illness
Depression
Alcohol use
Elderly
Dissertations, Academic -- School of Aging Studies -- Doctoral -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Research suggests that more than ten percent of older adults experience behavioral health problems (including mental health problems and/or substance abuse). However, very few actually receive care from a behavioral health care provider or even a primary care provider. One major barrier to accessing and receiving care is the feeling of perceived stigma commonly associated with behavioral health problems. The present study examined the relationships among attitudinal variables, feelings of stigma, and behavioral health outcomes over time in an elderly population with the secondary analysis of data collected for a previously implemented research study, the Primary Care Research in Substance Abuse and Mental Health for Elderly (PRISM-E) study. The PRISM-E research project was a multisite, randomized, comparative trial examining two models of care for persons aged 65 and older with symptoms of depression, anxiety, and/or at-risk drinking.A total of 2,022 participants over the age of 65 were included in the database. Over half of the participants indicated that they had some feelings of stigma associated with mental health and substance abuse issues. An examination of the measure used in the PRISM-E study to measure stigma revealed the presence of two factors, or components, of stigma that we labeled Perceived Stigma and Comfort Level. Statistical analyses of the data demonstrated that feelings of stigma are not constant and can indeed change over time. However, in this sample, perceived stigma was not related to behavioral health outcomes, such as a reduction in symptoms of depression, anxiety, and/or at-risk drinking. Limitations of this study include the possibility that the sample may be biased due to the fact that all participants were under the care of a primary care provider and all agreed to enter treatment for their behavioral health problem.Implications of the findings are that it is possible to influence feelings of stigma and previous research has demonstrated that lower levels of feelings of stigma in older adults as well as other age groups may lead to improvements in accessing and engaging in behavioral health treatment.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2009.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Nancy Bridger Lynn.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 92 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 - 002029094
oclc - 436869742
usfldc doi - E14-SFE0002881
usfldc handle - e14.2881
System ID:
SFS0027198:00001


This item is only available as the following downloads:


Full Text
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 2200397Ka 4500
controlfield tag 001 002029094
005 20090916112950.0
007 cr bnu|||uuuuu
008 090916s2009 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002881
035
(OCoLC)436869742
040
FHM
c FHM
049
FHMM
090
HQ1061 (Online)
1 100
Lynn, Nancy Bridger.
0 245
Correlates of attitudes toward behavioral health services among older primary care patients
h [electronic resource] /
by Nancy Bridger Lynn.
260
[Tampa, Fla] :
b University of South Florida,
2009.
500
Title from PDF of title page.
Document formatted into pages; contains 92 pages.
Includes vita.
502
Dissertation (Ph.D.)--University of South Florida, 2009.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
520
ABSTRACT: Research suggests that more than ten percent of older adults experience behavioral health problems (including mental health problems and/or substance abuse). However, very few actually receive care from a behavioral health care provider or even a primary care provider. One major barrier to accessing and receiving care is the feeling of perceived stigma commonly associated with behavioral health problems. The present study examined the relationships among attitudinal variables, feelings of stigma, and behavioral health outcomes over time in an elderly population with the secondary analysis of data collected for a previously implemented research study, the Primary Care Research in Substance Abuse and Mental Health for Elderly (PRISM-E) study. The PRISM-E research project was a multisite, randomized, comparative trial examining two models of care for persons aged 65 and older with symptoms of depression, anxiety, and/or at-risk drinking.A total of 2,022 participants over the age of 65 were included in the database. Over half of the participants indicated that they had some feelings of stigma associated with mental health and substance abuse issues. An examination of the measure used in the PRISM-E study to measure stigma revealed the presence of two factors, or components, of stigma that we labeled Perceived Stigma and Comfort Level. Statistical analyses of the data demonstrated that feelings of stigma are not constant and can indeed change over time. However, in this sample, perceived stigma was not related to behavioral health outcomes, such as a reduction in symptoms of depression, anxiety, and/or at-risk drinking. Limitations of this study include the possibility that the sample may be biased due to the fact that all participants were under the care of a primary care provider and all agreed to enter treatment for their behavioral health problem.Implications of the findings are that it is possible to influence feelings of stigma and previous research has demonstrated that lower levels of feelings of stigma in older adults as well as other age groups may lead to improvements in accessing and engaging in behavioral health treatment.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
590
Advisor: Lawrence Schonfeld, Ph.D.
653
Stigma
Mental illness
Depression
Alcohol use
Elderly
690
Dissertations, Academic
z USF
x School of Aging Studies
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2881



PAGE 1

Correlates of Attitudes Toward Behavioral Health Services Among Older Primary Care Patients by Nancy Bridger Lynn A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy School of Aging Studies C ollege of Behavioral and Community Sciences University of South Florida Major Professor: Lawrence Schonfeld Ph.D. Marion Becker, Ph.D. Laurence Branch, Ph.D. Sandra Reynolds, Ph.D. Paul G. Stiles, Ph.D. Date of Approval: March, 2009 Keywords: Stigma, Mental Illness, Depression, Alcohol Use, Elderly Copyright 2009 Nancy Bridger Lynn

PAGE 2

Dedication I would like to dedicate this dissertation to my family. Without them, none of this would have been possible. To my wonderful h usband Jayson who has supported me in everything I have chosen to do, no matter what : I thank you for your encouragement in tough times, your willingness to read and re read drafts, and listen to countless numbers of practice presentations. To my daughter Teagan, who was born amid this dissertation madness: She makes me want to be the best person I can possibly be. One day I'll tell her how I wrote drafts while feeding her and I ran regression models while she napped. To my Mom, sister, brother in law, an d nephews: Thank you all for of your support. I couldn't have done it without you! This dissertation is also dedicated to the memory of my father. He passed away after a long battle with cancer soon after I passed my qualifying exams. I hope he knew that e verything I do in my life is an effort to be the kind of person he wanted me to be. He was a special man and I hope I've made him proud.

PAGE 3

Acknowledgements I would like to thank my Major Professor and Mentor, Dr. Larry Schonfeld, for his guidance an d support during this process. His ability to break seemingly impossible, complex ideas into understandable and manageable actions allowed me to forge ahead with the dissertation process even when I thought I was in over my head. His encouragement and und erstanding, through family tragedy and the birth of my daughter, made me realize that anything is possible if you try hard enough. I was very fortunate to have such a wonderfully supportive doctoral committee. I would like to thank Drs. Marion Becker (for her many, many years of support!), Larry Branch, Sandy Reynolds, and Paul Stiles for their constructive feedback and support. You are all fantastic teachers and we students are better researchers for having had the opportunity to learn from you. It was quite an effort to get the data for my dissertation. I would like to thank Eugenie Coakly of JSI Research and Training Institute, and Dr. Sue Levkoff of Harvard University, for their assistance in obtaining the entire PRISM E dataset. I would also like to thank Dr. Cynthia Zubritsky, University of Pennsylvania, for sharing information with me about the PRISM E study. Finally, I would like to acknowledge all of the Aging Studies faculty and staff for their guidance and support of me and my fellow students. Dr. Cathy McEvoy is a wonderful leader and provides invaluable support and opportunities for her students.

PAGE 4

i Table of Contents List of Tables ................................ ................................ ................................ ................... iv List of Figures ................................ ................................ ................................ .................. vi Abstract ................................ ................................ ................................ .......................... vii Chapter One Introduction ................................ ................................ ............................... 1 Prevalence of Mental Disorders in Older A dults ................................ .................. 3 Mood D isorders ................................ ................................ ........................ 4 Anxiety D isorder ................................ ................................ ....................... 6 Alcohol U se D isorders ................................ ................................ .............. 7 An Underserved Population ................................ ................................ .................. 8 Provider F actors ................................ ................................ ...................... 10 Service D elivery F actors ................................ ................................ ......... 11 Personal F actors ................................ ................................ ...................... 1 2 Stigma and B ehavioral H ealth S ervice U se ................................ ........................ 1 3 Measuring S tigma ................................ ................................ ................... 19 Facilitator s to Receiving S ervices ................................ ................................ ....... 20 The Impact of Behavioral Health Problems on Health Care Systems ................ 22 Conclusion ................................ ................................ ................................ .......... 24 Study Aims ................................ ................................ ................................ ......... 26 Hypothesis One ................................ ................................ ................................ ... 26 Hypothesis Two ................................ ................................ ................................ .. 27 Hypothesis Three ................................ ................................ ................................ 27 Chapter Two Method ................................ ................................ ................................ .... 28 Research Design ................................ ................................ ................................ 28 PRISM E ................................ ................................ ................................ ............. 28 Service M odels ................................ ................................ ................................ ... 29 Study S ample ................................ ................................ ................................ ...... 31 Measures ................................ ................................ ................................ ............. 32 Stigma ................................ ................................ ................................ ..... 33 Diagnosis/Outcome ................................ ................................ ................ 33 Measures of D epressive D isorders ................................ ......................... 34 Measures of A nxiety D isorders ................................ .............................. 36 Measures of A t R isk D rinking ................................ ................................ 37 Analysis Plan ................................ ................................ ................................ .... 38

PAGE 5

ii Chapter Three Results ................................ ................................ ................................ .. 41 Sample Characteristics ................................ ................................ ........................ 41 Lev el of Stigma ................................ ................................ ................................ ... 41 The Stigma Assessment Scale ................................ ................................ ............ 46 Research Questions ................................ ................................ ............................. 51 Research Question One : What D emographic and B ehavioral H ealth V ariables are A ssociated with L evel of S tigma at B aseline? ............................. 51 Embarrassed/Ashamed Item R egression M odel ................................ ..... 53 Perceived Stigma Component R egression M odel ................................ .. 54 Comfort Level Component R egression M odel ................................ ...... 55 Research Question Two : Does L evel of S tigma C hange O ver the T hree T ime P oints? Does C hang e V ary A ccording to E ngagement? ................. 56 Embarrassed/Ashamed Item ................................ ................................ ... 56 Perceived Stig ma Component ................................ ................................ 58 Comfort Level Component ................................ ................................ ..... 60 Research Question Three : What is the R elationship B etween C hange in S tigma and C hange in B ehavio ral H ealth V ariables O ver T ime? .................. 62 Embarrassed/Ashamed Item R egression M odel ................................ ..... 63 Perceived Stigma Component R egression M odel ................................ .. 64 Comfort Level Component R egression M odel ................................ ....... 65 Changes in CES D ................................ ................................ .................. 66 Changes in BA I ................................ ................................ ....................... 67 Changes in N umber of D rinks ................................ ................................ 67 Research Question Four : Assuming L evel of S tigma D oes C hange O ver T ime, W hat V ariables are R elated to C hange O ver T ime? ....................... 68 Chapter Four Discussion ................................ ................................ .............................. 71 Research Question One : What D emographic and B ehavioral H ealth V ariables are A ssociated with L evel of S tigma at B aseline? .................. 73 Research Question Two : Does L evel of S tigma C hange O ver the T hree T ime P oints? Does C han g e V ary A ccording to E ngagement? ................. 74 Hypothesis One ................................ ................................ ................................ ... 75 Research Question Three : What is the R elationship B etween C hange in S tigma and C hange in B ehavio ral H ealth V ariables O ver T ime? .................. 75 Hypothesis Two ................................ ................................ ................................ .. 7 7 Research Question Four : Assuming L evel of S tigm a D oes C hange O ver T ime, W hat V ariables are R elated to C hange O ver T ime? ....................... 7 7 Hypothesis Three ................................ ................................ ................................ 77 Conclusions ................................ ................................ ................................ ......... 78 Limitations ................................ ................................ ................................ .......... 79 Future Directions ................................ ................................ ................................ 81 List of References ................................ ................................ ................................ ........... 83

PAGE 6

iii Ab out the Author ................................ ................................ ................................ End Page

PAGE 7

iv List of Tables Table 1. Stigma Assessment Scal e ................................ ................................ ................ 34 Table 2. Sample Characteristics ................................ ................................ ...................... 4 2 Table 3 Frequency of Responses to the Stigma Assessment Scale ............................... 4 4 Table 4 Diagnosis ................................ ................................ ................................ .......... 47 Table 5 Inter Item Correlation Matrix of Stigma Assessment Measure Items .............. 50 Table 6 Factor Analysis Components ................................ ................................ ............ 51 Table 7 Correlations of Baseline Variables ................................ ................................ ... 53 Table 8 Summary of Simultaneous R egression Analysis for Variables Predicting Baseline Embarrassed/Ashamed Item ............................... 54 Table 9 Summary of Simultaneous R egression Analysis for Variables Predicting Perceived Stigm a Component ................................ ........... 55 Table 10 Summary of Simultaneous R egression Analysis for Variables Predicting Comfort Level Component ................................ ............... 56 Table 1 1 Change in Level of Stigma as Measured by Embarrassed/Ashamed Item ................................ ................................ ............... 57 Table 1 2 Change in Level of Stigma as Measured b y Embarrassed/Ashamed Item by Engagement ................................ ..................... 57 Table 1 3 Change in Level of Stigma as Measured by Embarrassed Ashamed Item by Service Model Type ................................ ......... 58 Table 1 4 Change in Level of Stigma as Measured by Perceived Stigma Component ................................ ................................ ............. 59 Table 1 5 Change in L evel of Stigma as Measured by Perceived Stigma Component by Engagement ................................ ................... 59

PAGE 8

v Table 1 6 Change in Level of Stigma as Measured by Perceived Stigma Component by Service Model Type ................................ ...... 60 Table 1 7 Change in Level of Stigma as Measured by Comfort Level Compon ent ................................ ................................ ................. 6 0 Table 1 8 Change in Level of Stigma as Measured by Comfort Level Component by Engagement ................................ ....................... 61 Table 1 9 Change in Level of Stigma as Measured by Comfort Level Component by Service Model Type ................................ .......... 62 Table 20 Correlations of Change in Stigma Items a nd Change in Behavioral Health Variables ................................ .............................. 63 Table 2 1 Summary of Simultaneous Regression Analysis for Behavioral Health Variables Predicting Change in Embarrassed/Ashamed Item ................ 64 Table 2 2 Summary of Simultaneous Regression Analysis for Behavioral Health Variables Predicting Change in the Perceived Stigma Component ........ 65 Table 2 3 Summary of Simultaneous Regression Analysis for Behavioral Health Variables Predicting Change in the Comfort Level Component ............ 65 Table 2 4 Summary of Simultaneous Regressi on Analysis for Variables Predicting Change in the CES D ................................ ........................ 66 Table 2 5 Summary of Simultaneous Regression Analysis for Variables Predicting Change in the BAI ................................ ............................. 67 Table 2 6 Summary of Simultaneous Regression Analysis for Variables Predicting Change in the Number of Drinks ................................ ...... 68 Table 2 7 Summary of Simultaneous Regression Analysis for Variables Predicting Change in Embarrassed/Ashamed Item ............................ 69 Table 2 8 Summary of Simultaneous Regression Analysis for Variables Predicting Change in the Perceived Stigm a Component .................... 70

PAGE 9

vi List of Figures Figure 1. Responses to Stigma Item 1 ................................ ................................ .. 4 3 Figure 2. Responses to Stigma Item 2 ................................ ................................ .. 46

PAGE 10

vii Correlates of Attitudes Toward Behavioral Health Services Among Older Primary Care Patients Nancy Lynn ABSTRACT Research s uggests that more than ten percent of older adults experienc e behavioral health problems ( including mental health problems and/or substance abuse) H owever, very few actually receive care from a behavioral health care provider or even a primary care provid er. One major barrier to accessing and receiving care is the feeling of perceived stigma commonly associated with behavioral health problems The present study examined the relationship s among attitudinal variables, feelings of stigma, and behavioral healt h outcomes over time in an elderly population with the secondary ana lysis of data collected for a previously implemented research study, the Primary Care Research in Substance Abuse and Mental Health for Elderly (PRISM E) study. The PRISM E research projec t was a multisite, randomized, comparative trial examining two models of care for persons aged 65 and older with symptoms of depression, anxiety, and /or at risk drinking. A total of 2,022 participants over the age of 65 were included in the database. Over half of the participants indicated that they had some feelings of stigma associated with mental health and substance abuse issues. An examination of the measure used in the PRISM E study to measure stigma revealed the presence of two factors, or components of stigma that we labeled Perceived Stigma and Comfort Level. Statistical analyses of the data demonstrated that feelings of stigma are not constant and can indeed

PAGE 11

viii change over time. However in this sample, perceived stigma was not related to behavioral h ealth outcomes such as a reduction in symptoms of depression, anxiety, and/or at risk drinking L imitation s of this study include the possibility that the sample may be biased due to the fact that all participants were under the care of a primary care pr ovider and all agreed to enter treatment for their behavioral health problem. I mplications of the findings are that it is possible to influence feelings of stigma and previous research h as demonstrated that lower levels of feelings of stigma in older adult s as well as other age groups may lead to improvements in accessing and engaging in behavioral health treatment.

PAGE 12

1 Chapter One Introduction The proportion of older adults over the age of 65 has increased over the past century and is predicted to contin ue growing at high rates. Both the l engthening life expectancies and the aging of the large numbers of the Baby Boom generation contribute to this dramatic increase (USDHHS, 1999). As the number of older adults in this country grows, one would expect a pro portional amount of growth in the number of older adults seeking behavioral health services (mental health services and /or substance abuse treatment services). However, only a relatively small proportion of older adults with behavioral health disorders cur rently has any contact with service providers (Cole, 2002). Karlin and Duffy (2004) state that the "mental health needs of the nation's geriatric populations have been significantly neglected" (p. 509). Many barriers prevent those older adults in need fro m seeking and receiving health care (Fitzpatrick, Powe, Cooper, Ives, & Robbins, 2004) The Surg eon General's Report on Mental H ealth (USDHHS, 1999) identifies stigma as the most formidable barrier to the receipt of mental health services. I n 1963, Goffman defined stigma as an "attribute that is deeply discr editing" and reduces the bearer from a whole and usual person to a tainted, discounted one" ( as cited in Bambauer & Prigerson, 2006). Perceived or self stigma is the belief that most people will devalue and discriminate against

PAGE 13

2 individuals who use behavioral health services or who have a behavioral health disorder (Link, Cullen, Struening, Shrout, & Dohrenwend, 1989). Perceived stigma can be a problem for individuals of all ages with behavioral health p roblems. Research focused on older adults has found that o lder patients are often reluctant to discuss their behavioral health problems with a health care provider. As with individuals in other age groups, s ome older adults may reject the diagnosis of depr ession or another behavioral health disorder due to the stigma attached to it. A psychiatric diagnosis may spark concerns of a potential loss of independence and fears of being institutionaliz ed (Wagenaar, Mickus, Gaumer, & Colenda, 2002). Older males seem particularly affected by the stigma associated with a behavioral health diagnosis, which is cause for concern because older males have the highest rate of completed suicide (Reynolds & Kupfer, 1999). The problem of stigma not only affects the identificati on of behavioral problems in older adults, but treatment adherence as well. Greater perceived stigma toward individuals with behavioral health problems is associated with a greater likelihood of treatment discontinuation (Sirey et al., 2001). This study w ill explore the relationship s among attitudinal variables including feelings of stigma and behavioral health outcomes over time by analyzing data collected for a previously implemented research study, the Primary Care Research in Substance Abuse and Menta l Health for Elderly (PRISM E) study This research project was a multisite, randomized, comparative trial examining two models of care for persons aged 65 and older with symptoms of depression, anxiety, and at risk drinking. PRISM E was the largest random ized study of behavioral health care among the elderly to date. The overall goal of the study was to compare the effects of integrated behavioral health care

PAGE 14

3 models and enhanced referral care models on engagement, participation, patient clinical outcomes, and cost effectiveness for the target conditions of depression, anxiety, and at risk drinking in older adults (Levkoff et al., 2004). The PRISM E dataset will be explored in this dissertation in an effort to examine the relationship s among perceptions of s tigma and behavioral health outcomes in older adults. Prevalence of Mental Disorders in Older Adults Evidence suggests that behavioral health problems are relatively common among the elderly population. However, f ew, if any, large scale epidemiological st udies have focused specifically on the prevalence of behavioral health issues in this population. To understand the prevalence of behavioral health problems among older adult populations, we must examine the d ata that are available from epidemiological stu dies examining the behavioral health issues in populations across the lifespan. Some of these studies provide much needed information about the rates of these disorders among an elderly population. The National Institute of Mental Health (NIMH) Epidemiolog ic Catchment Area (ECA) Program was developed over two decades ago to determine the prevalence and incidence of specific disorders in community and institutional samples. This seminal study provided much needed data on the prevalence of mental disorders in all ages, including those 65 and older. Using the Diagnostic Interview Schedule (DIS) as the diagnostic tool, results indicated that the one month prevalence of any disorder among non institutionalized adults 65 and over was 12.3%, lower than rates seen i n their younger counterparts. The most prevalent diagnostic category in this age group was anxiety disorder s (5.5%; Regier et al., 1988); this category includes disorders such as phobias, obsessive compulsive disorder, and posttraumatic stress disorder. Ph obic

PAGE 15

4 disorders were the most common anxiety disorder, with a 1 month prevalence rate of 4.8%. A one month prevalence rate of 0.7% was found for major depression. Some gender differences were found as well, with older women experiencing mental health disord ers at higher rates than older men (13.6% vs. 10.5%). Alcohol and drug disorders, however, were more common in men than in women. A one month prevalence rate for alcohol disorders was estimated to be 0.9% in this population Since the ECA was completed, a number of epidemiological studies have been conducted and the results provide more information about the extent of behavioral health problems in this population. Many of these studies have focused on specific disorders, such as anxiety, mood, or substanc e use disorders. Data from several of the studies are presented below. However, Hybels and Blazer (2004) caution that these epidemiological trends for older adults must be interpreted with the understanding of the importance of cohort effects. Cohorts born during the same general time experienced the same historical events. These events shape their behavior and attitudes throughout their lives, resulting in cohort effects. For example, cohorts growing up during the Prohibition Era in the 1920s may have deve loped negative values about alcohol use that have biased the views and practices of older adults currently in their 9 0s and older (Atkinson, Ganzini, & Bernstein, 1992). For this reason, the prevalence rates for behavioral disorders may not remain stable o ver time. In fact, as the Baby Boom generation ages, the prevalence of some disorders, specifically alcohol and substance use disorders, may increase (Hybels & Blazer, 2004). Mood d isorders Depression and depressive symptoms are associated with reduced qu ality of life, decreased functioning and independence, premature death, and

PAGE 16

5 suicide (Lebowitz et al., 1997). In a recent study, Chachamovich, Fleck, Laidlaw, and Power (2008) found that even relatively minor levels of depression (subsyndromal levels) are a ssociated with decreases in several domains of quality of life among older adults. In a four year prospective study of HMO enrollees, the percentage of older adults with depressive symptoms increased from 14% at baseline to 18% at the four year follow up ( Unutzer et al., 1997), a significant increase suggesting that depressive symptoms increase as we age Older adults also experience other mood disorders such as dysthymia, bipolar disorder, and hypomania, but these disorders generally occur at a much lower rate than major depression. Using criteria from the American Psychiatric Association's Diagnostic and Statistical Manual for Mental Disorders (DSM; APA, 1994), Gurland and colleagues (Gurland, Cross, & Katz, 1996) found one year prevalence rates of major d epression to be about 5% in older adults. Depression may be more evident among subsamples of the general population. Among a sample of frail, low income, community dwelling seniors, Rogers and Barusch (2000) found 29% exhibiting depressive symptomatology. The prevalence of diagnosable depression among nursing home residents is also substantial (Smyer & Qualls, 2004) and has been estimated to be as high as 22 % (Burrows, Satlin, Salzman, Nobel, & Lipsitz, 1995). While t he prevalence of major depression declin es with age, depressive symptoms increase over time. This trend is evident when comparing DSM based measures for diagnosis to symptom based assessments. However, diagnosing older adults with depression can be more difficult than younger adults. Older adult s with depression often report different symptoms than those commonly seen in younger adults. Older adults are less likely to report feelings of dysphoria, which is often a classic symptom of depression

PAGE 17

6 in younger adults (USDHHS, 1999). Older adults with d epressive symptoms may complain of physical symptoms such as unexplained pain, headache, fatigue, loss of appetite, or gastrointestinal symptoms (Charney et al., 2003; Lebowitz et al, 1997). Because symptoms reported by older adults are often somatic in na ture, depression may be mistakenly diagnosed as physical illnesses, such as atherosclerosis, Alzheimer's disease, or simply normal aging (USDHHS, 1999) possibly leading to lower estimates of behavioral health problems in this population Anxiety d isorder s The category of anxiety disorders include s generalized anxiety disorder, phobic disorders, panic disorder, obsessive compulsive disorder, and post traumatic stress disorder. In a community based study, one year prevalence rates indicate that about 11.4% of adults 55 or older meet criteria for an anxiety disorder (Flint, 1994). Phobic disorders are the most common type of anxiety disorder found in older adults (Regier et al., 1988). In a study of Canadian community dwelling older adults, Bland, Newman, an d Orn (1988) found a prevalence rate of 3.0% for phobic disorders. Generalized anxiety disorder (GAD) was not assessed in the ECA study, but one ECA site did include assessments for GAD. Blazer, Hughes, and George reported that 2.2% of individuals over the age of 65 met criteria for GAD in the year prior to the assessment interview ( as cited in Hybels & Blazer, 2004). Similar to trends in depression, symptoms of anxiety may be more common than diagnosable anxiety disorders as age increases. Worrying, which usually does not meet the criteria for an anxiety disorder in itself, can be troublesome for older adults. In a study of Swedish community dwelling older adults aged 78 and older, Forsell and Winblad

PAGE 18

7 (1998) found that nearly one quarter of the sample expe rienced feelings of excessive worry and anxiety. Alcohol u se d isorders One challenge for researchers in this field is that DSM diagnostic criteria for alcohol use disorders were initially developed for use with young and middle aged adults, not older adul ts. The DSM includes criteria su ch as problems at work and problems with family which may not apply to older adults since many are retired from the workplace and many are widowed and do not live with or have regular contact with family More recently, seve ral screening and diagnostic tools have been developed and validated for use with older adults, making it possible to gather more accurate epidemiological data. Liberto, Oslin, and Ruskin (1992) examined the prevalence of heavy drinking ( defined here as 1 2 to 21 alcoholic drinks per week) in older adults and found it to be between 3% and 9%. Data from the ECA study using one month prevalence rates of alcohol abuse and dependence were much lower (0.9%; Regier et al., 1988). Gender differences in drinking pa tters are also evident. In an analysis of three nationally representative studies, Breslow, Faden, and Smothers (2003) discovered prevalence rates of moderate drinking ranged from 27% to 38% for males and 21% to 32% for females. For heavier drinking, preva lence rates were around 9% for males and 2% for females. Often, mental illnesses and substance abuse disorders co occur in the same individual. A study by Holroyd and Duryee (199 7 ) found that 9% of older clients receiving treatment for a mental health dis order through an outpatient psychiatric facility also had a diagnosable dependence on alcohol. Recent estimates of persons with mental illness and alcohol abuse or dependence are greater than 20%. Alcohol use disorders can

PAGE 19

8 cause or exacerbate the symptoms of many mental disorders, including personality disorders (Westreich, 2005). P revalence rates for at risk alcohol use are expected to rise as the Baby Boom generation ages, since this group has heavier drinking patterns (and substance abuse rates) than th e current group of older adults (Patterson & Jeste, 1999). Research has shown that although alcohol consumption generally decrease with age, problem drinkers often continue the same drinking patterns as they age (Schonfeld et al., 2000). Additionally, the effects of alcohol on the individual often changes with aging leading to an increase in the dangers of drinking. Although a recent study found potential health benefits to moderate alcohol use among women (two or less drinks a day), other variables must be taken into account when interpreting these findings (Balsa, Homer, Fleming, & French, 2008). Alcohol use in conjunction with certain medications, changes in body composition, and the presence of comorbid medical conditions could all potentially magnify th e effects of the alcohol, in turn possibly leading to increased risk of accidents and falls. An Underserved Population Recent research has estimated that as many as one in four older adults has a significant mental disorder (Bartels, 2003) H owever this number may be underestimated due to an underidentification of older adults with behavioral health disorders. Although their numbers are growing, o lder adults are less likely than younger adults to receive mental health treatment of any type (Swartz et al., 1998). According to the Surgeon General (USDHHS, 1999), less than half of those older adults with a mental disorder will receive treatment in their lifetime When examining data from the nationally

PAGE 20

9 representative Healthcare for Communities (HCC) household survey, Klap and colleagues (Klap, Unroe, & Unutzer, 2003) found results similar to the ECA study in that only half of older adults with a probable mental disorder had received any type of treatment Further older adults were significantly less likely to receive specialty mental health treatment (i.e. outpatient specialty mental health services, emergency room, or counseling visits) than their younger counterparts Stated in another context Cole and Yaffe write that "among 27 elderly per thousand who hav e moderate severe depression, 22 consult a family physician who detects the disorder in 5.3; of these, only 2.8 are referred to a psychiatric service, let alone a geriatric psychiatry service" (p. 3; as cited in Cole, 2002). If an older adult is identifie d in a primary care setting as having a behavioral health disorder, they are more likely to receive a psychotropic medication without a recordable diagnosis than younger individuals (Larson, Lyons, Hohmann, & Beardsley, 1991). In addition, having a behavio ral health disorder is related to a greater risk of the prescri ption of inappropriate medications in older adults, especially benzodiazepines (Stuck et al., 1994). These findings underscore the importance of encouraging and empowering older adults to acces s and engage in specialized behavioral health treatment Older adults residing in nursing homes are at even greater risk of being underserved. In an examination of Medicare claims data, Shea and colleagues (Shea, Russo, & Smyer, 2000) found that 80% of nur sing facility residents with some indication of a mental illness do not receive mental health visits from a psychiatrist, clinical psychologist, or licensed clinical social worker. In addition, nearly three quarters of all nursing facility residents had no visits with any health care provider, including a primary

PAGE 21

10 care physician. When a health care provider did visit, the visits were infrequent, with less than 10% of residents with a mental illness receiving a monthly visit. Nursing facility residents living in rural areas have been identified as an especially underserved population, with even lower rates of seeing any provider, much less a behavioral health specialist. The factors contributing to older adults with behavioral health needs being underidentifie d by the health care system are complex. Factors can be categorized into three main groups: provider factors, service delivery factors and personal factors These factors will be discussed below. Provider f actors Although most older adults report that they would use some type of a psychological service if needed, few actually do (Arean, Alvidrez, Barrera, Robinson, & Hicks, 2002). In reality, older adults are less likely than middle aged adults to consult with a specialty mental health care provider or any health care provider about their depression (Crabb & Hunsley, 2006). If they do seek treatment, older adults prefer to receive care from their primary care physicians ( Dupree, Watson, & Schneider, 2005; Gallo, Rabins, & Illife, 1997 ; Mickus & Colenda, 2000 ). Primary care offers the potential advantages of proximity, affordability, convenience, and coordination of care for behavioral and physical disorders, given that comorbidity is typical. Unfortunately, this coordination rarely happens. According to t he President's New Freedom Commission on Mental Health (2003), mental health disorders "often go undiagnosed, untreated, or under treated in primary care (p. 59). Some physicians harbor their own beliefs that behavioral health disorders simply do not occur in older adults or that depression is a normal part of the aging process. Others may feel that it is more important to treat co

PAGE 22

11 morbid physical health illnesses than mental health illnesses (Unutzer et al, 2001). In reality, many primary care physicians r eceive little or no training in mental health issues or geriatrics therefore may be unqualified to appropriately identify and treat behavioral health problems in their patents. Often, primary care physicians do not screen patients, including older adults, for behavioral health problems. Klap and colleagues (2003) reported that only 15% of older adults in the HCC dataset reported being asked by their primary care physicians about symptoms of anxiety, the most common behavioral health disorder in older adults Only 1% of older adults indicated that their physician had referred them to a mental health specialist in the past year for evaluation or treatment. The consequences of this inaction can be very serious; Conwell, Rotenberg, and Caine (1990) conducted an analysis of suicides and found that 75% of the older adult suicide completers had visited their primary care physician in the month preceding their deaths. Service d elivery f actors. The behavioral health system in the United States encompasses a variety of providers including psychologists, psychiatrists, mental health counselors, and social workers. Similarly, services are offered in varied settings such as mental health centers and private practice, and less frequently in nursing homes, primary care, crim inal justice, and aging services. The service system is fragmented and difficult to navigate (Bartels, 2003). In many rural areas, behavioral health services may not be available Limited access to transportation in all areas prevents many older from regul arly accessing needed services (Arcury et al., 2005) The current health care system lacks preparedness for the projected influx of future older adults in need of behavioral health services. Borson and colleagues (Borson,

PAGE 23

12 Bartels, Colenda, Gottlieb, & Mey ers, 2001) label the behavioral health service system as vertically and horizontally fragmented. Vertical fragmentation refers to situations when multiple service systems are involved in providing care, such as specialty mental health, primary medical care aging services, and substance abuse. Often, these multiple service sectors do not collaborate on care leading to a fragmentation of services. This is often the case for those older adults with comorbid conditions, meaning the coexistence of two or more c onditions, often mental health and substance abuse. Borson and colleagues also use the term vertically fragmented to describe the lack of or disruption of continuity of care over time experienced by many older adults. Personal factors. When compared with younger adults, older adults often exhibit atypical symptoms of mental disorders. Older adults with behavioral health issues often present with somatic complaints, such as fatigue and pain making it difficult to make an appropriate diagnosis In some case s, older adults may exhibit symptoms of a mental disorder that do not meet the full DSM criteria for that disorder (USDHHS, 1999). Although the person's symptoms do not meet criteria, it is possible even probable, that their quality of life is being negati vely affected by the symptoms (Hybels & Blazer, 2004). Subsyndromal symptoms of depression have been associated with significant disability reduced quality of life, and suicidal ideation ( Chachamovich et al., 2008; Chopra et al., 2005). Older adults ofte n have comorbid physical health problems along with their behavioral health problems. In some cases, the somatic complaints will mimic or mask the behavioral health problem, making diagnosis difficult (USDHHS, 1999). Charney and colleagues (2003) state tha t depression should be considered of hospitalized older adults

PAGE 24

13 with myocardial infarction, congestive heart failure, stroke, hip fracture, cancer, or alcohol abuse if they exhibit delayed recovery, poor compliance with rehabilitation programs, or treatment refusal. Other conditions which mimic or mask psychiatric symptoms include thyroid dysfunction, mitral valve prolapse, epilepsy, multiple sclerosis, hypoglycemia, and brain tumors (Segal, Coolidge, & Hersen, 1998). When depression co occurs with physical illnesses, treatment adherence is reduced and chances for recovery from the physical illness are lessened (Goldman, Niels e n, & Champion, 1999). Due to an increase in chronic medical conditions in older age, older adults consume a disproportionate amount of prescribed and over the counter medications. These medications may mask symptoms or be the underlying cause of the symptoms (Hybels & Blazer, 2004). With increased medication use, older adults are exposed to a greater chance of adverse reactions and harmf ul drug interactions (Segal et al., 1998). In some cases, older adults themselves may feel that illnesses such as depression and anxiety are part of the natural course of growing older. Some older adults mistakenly believe that their symptoms are a part of their somatic illnesses. Klap and colleagues (2003) found that older adults with a probable mental health disorder were less likely than younger or middle aged adults to perceive a need for mental health services (28% vs. 49% or 43%, respectively). Others due to the stigmatization of behavioral health problems, may deny or minimize symptoms or rationalize them in an effort to avoid being labeled as having a mental illness (Goldman et al., 1999). Stigma and B ehavioral H ealth S ervice U se The stigma surroun ding behavioral health care has been widely studied in younger populations. Less is known about the impact of stigma on the health care

PAGE 25

14 utilization of older adults T he topic of stigma and the negative outcomes associated with it has surfaced in the past d ecade The Surgeon General's Report on mental health states that stigma is the most formidable obstacle to further progress in the field of mental illness (USDHHS, 1999). The President's New Freedom Commission (2003) has recommended a national campaign to reduce or eliminate the stigma surrounding behavioral health care, especially in suicide prevention. In 2003, the World Health Organization (WHO) and the Old Age Psychiatry section of the World Psychiatric Association (WPA) jointly distributed a technical consensus paper on the problem of stigma and discrimination against older adults with mental disorders and calling for further research in the area (Graham et al., 2003). In addition, the U.S. Substance Abuse and Mental Health Services Administration (SAMH SA) has developed an on line resource center dedicated to combating stigma and discrimination associated with mental illness and substance abuse issues ( www.stopstigma.samhsa.gov). Corrigan (2004) defines two types of stigma that overlap and interact: publ ic stigma and self stigma. Public stigma is "what a na•ve public does to the stigmatized group" and self stigma is "what members of a stigmatized group may do to themselves if they internalize the public stigma" (p. 616). Barney, Griffiths, Jorm, and Chris tensen (2006) describe another type of stigma similar to self stigma perceived stigma, as the belief that other people hold stigmatizing ideas; that others will respond negatively to them if they seek help. Link and colleagues ( Link, Struening, Neese Todd Asmussen, & Phelan, 200 2 ) posit that people develop conceptions of mental illness early on in life and form expectations as to how they and the rest of society would react to a person with mental illness. When a person goes on to develop a mental illness they may struggle with

PAGE 26

15 reconciling their beliefs about people with mental illness and fear how society will react to them. This may lead to persons with mental illnesses feeling set apart and different from others and ashamed. Possible coping mechanisms include secrecy, distancing, and withdrawal from others. Perceived stigma has been closely associated with self esteem feelings of shame, and depressive symptoms. This association indicates that feelings of stigma and ways of coping with feelings of stigm a may be able to change over time (Link, Struening, Neese Todd, Asmussen, & Phelan, 2001; 2002). Stigma influences behavioral health services utilization in a number of ways. Older patients are often reluctant to discuss their behavioral health problems wi th a health care provider due to the stigma surrounding such problems Concerns about having a mental health diagnosis substantially decrease the likelihood of service use (Bambauer & Prigerson, 2006). Barney and colleagues (2006) surveyed adults and found that s elf stigma and perceived stigma are common and both types decrease the likelihood of seeking help from any professional source, including general practitioners, counselors, psychologists, and psychiatrists. Respondents reported greater embarrassment associated with visiting mental health professionals, especially psychiatrists. One fifth of respondents expected to receive negative responses when they did seek help, especially from general practitioners. In another study, Cooper, Corrigan, and Watson (2003) found that individuals were less likely to consider seeking care in the future if they perceived people with mental illness as responsible for their disorders, reacted to them angrily because of this attribution, and withheld pity and helping behavi ors. As would be expected, those who felt favorably toward seeking help from a professional are more likely to do so (Komiti, Judd, & Jackson, 2006).

PAGE 27

16 Older adults living in rural communities are also affected by the stigma associated with behavioral healt h problems, perhaps more so than those living in suburban and rural areas In a survey of rural residents, respondents often felt that people in their communities would gossip about a person with mental illness and would be wary of a person who had been ho spitalized for a mental illness (Komiti et al., 2006) P aradoxically, respondents also felt that their neighbors were more supportive and caring of a person with a behavioral health illness than those living in urban communities. Many individuals living in rural communities preferred informal assistance to formal assistance and viewed formal mental health care as a last resort, only to be used after trying to work out their problems themselves. A study of at risk drinkers living in rural communities found t hat respondents fel t a lack of privacy when seeking assistance from primary care providers. This is concerning because a lack of specialty behavioral health providers in many rural locations leaves primary care the only option for receiving needed behavior al health care (Fortney et al., 2004). Ethnicity and race also play an important role in service utilization and the influence of stigma. Although there is wide variation among ethnicities, American Indian/Alaskan Natives have higher rates of mental healt h problems and reported unmet needs when compared to Whites. African America n Asian, Mexican, Central and South America n and other Hispanic Latino groups generally have among the lowe st rates of mental health problems in comparison to other ethnic groups (Harris, Edlund, & Larson, 2005). M any factors play a part in the relationship between health care utilization and ethnicity, such as education, finances, health insurance, and mistrust of the service system. Some evidence suggests that non Caucasians are less likely to think that the

PAGE 28

17 medical system is a useful source of mental health treatment when compared to Caucasians (Corrigan, 2004). However, very little research has been conducted examining the influence of stigma on the behavioral health decisions of ethnic minorities, especially elderly minorities, in the United States. While some older adults may be willing to seek services, many are unaware of where to find behavioral health service providers. In a survey of older and younger adults, older adults described themselves as less knowledgeable about mental health care and appropriate treatment than younger adults. M ost older adults surveyed wished they had a better understanding of when to see a mental health professional and felt that adequate access to mental health care was very important (Robb, Haley, Becker, Polivka, & Chwa, 2003). However, the acknowledgement of the importance of psychological treatments such as medication and psychotherapy does not predict use of mental health services. Among ber eaved older adults, receptivity to professional assistance for mental health issues does not necessarily influence the rate of actually accessing these services. Oftentimes, while older adults accept the usefulness of these services in general, their perso nal need for professional assistance may be perceived as a threat to their ability to live independently (Bambauer & Prigerson, 2006). Older adults have been found to be more likely to report feeling less receptive to professional mental health services an d were more concerned with possible family reactions than those in other age groups (Leaf, Bruce, Tischler, & Holzer, 1987). The stigma associated with behavioral health problems not only affects a person's decision to seek help, it may influence an indivi dual's decision to continue treatment. A review of 34 studies found that more than 40% of people receiving antipsychotic

PAGE 29

18 medication failed to adhere to prescribed regimens (Cramer & Rosenbeck, 1998). Similar attrition is evident in psychosocial treatments as well (Corrigan, 2004). In a study by Sirey and colleagues (2001), perceived stigma was found to predict early treatment discontinuation in older adults. Younger patients reported greater levels of perceived stigma but it did not influence their treatmen t participation decisions as heavily as it did for older adults. In 2005, the Substance Abuse and Mental Health Services Administration (SAMHSA) issued the report Mentally Healthy Aging: A Report on Overcoming Stigma for Older Americans (USDHHS, 2005). Written by leading researchers in the field, the report recommends strategies to overcoming the barriers to reduce stigma associated with mental illness. The report suggests (1) empowering and educating older Americans with mental illnesses, and (2) educat ing the public on mental health and aging. The 2003 consensus paper written by WHO and the WPA make a number of suggestions for reducing stigma and discrimination against older adults with mental illnesses including creating supportive environments and app ropriate health and social care systems, place the mental health of older adults on the public agenda, and encourage more research. The authors of this report suggest that government policy makers as well as non governmental organizations have major parts to play in reducing stigma against people with mental illnesses (Graham et al., 2003). T he SAMHSA and WHO/WPO reports mostly focus on changing public stigma not on how to improve self or perceived stigma. Link and colle a g u es ( 2002) developed an interventi on targeting the coping strategies of people with mental illnesses. The authors hypothesized that by modifying coping strategies of dealing with perceived stigma, self

PAGE 30

19 esteem and depressive symptoms would also improve. Results of the ir study did not suppor t this hypothesis. H owever, this is one of the only studies to attempt to change perceptions of stigma within a person with mental illness rather than changing the beliefs of society at large. Measuring stigma Several measures of perceived stigma exist in the mental health literature. In a review of the literature, Link and colleagues (Link, Yang, Phelan, & Coll ins, 2004) categorize measures applicable to patients and consumers of mental health services as falling into the following categories: measures ex amining mental health consumers' experience of stigma, measures fo r consumers associated with modified labeling theory, measures of rejection and perceptions of rejection, measures of coping orientations, and measures of stig ma related feelings. Link and c olleagues caution against using self report measures of stigma due to measures of neuroticism being related to the perception of being stigmatized. For example, "a person who is unemployed, isolated, or beset by low self esteem may seek to explain his or h er disadvantaged status by invoking stigma. In such a scenario, levels of measured stigma do not cause bad outcomes but are instead consequences of those outcomes" (p. 525). The research conducted in this area by Link and colleagues demonstrates the comple xities and variation s within the stigma construct and the challenges inherent in its measurement. Stigma is a complex construct; many instruments exist which attempt to measure its various facets. Many measures of perceived stigma attempt to quantify the e xperience of the person with mental illness. One such measure is the Stigma Coping Scale (Link, Struening, Cullen, Shrout, & Dohrenwend, 1989), which assesses beliefs about devaluation of and discrimination toward individuals with mental illnesses as repor ted by

PAGE 31

20 a person with mental illness (Sirey et al., 2001). In a review of measuring mental illness stigma, Link and colleagues (2004) discuss several measures of stigma appropriate for patients/consumers, including his own work in developing measures of cop ing techniques (secrecy, withdrawal, and distancing), and stigma related feelings (misunderstood and different/ashamed) Included in the review is Wahl' s (1999) Mental Health Consumer Experience of Stigma, which includes questions about stigma experience a nd discrimination. The Self Stigma of Seeking Help Scale (Vogel, Wade, & Haake, 2006) consists of ten items such as "I would feel inadequate if I went to a therapist for psychological help." T he newly developed instrument, Depression Self Stigma Scale (Kan ter, Rusch, & Brondino, 2008) measures general self stigma, secrecy, public stigma, treatment, and stigmatizing experiences. Although many measures of perceived stigma exist in the literature base, few if any, have been developed specifically for use wi th older adults (with the exception of the HIV/AIDS field) One measure developed for older adults focuses on bereaved older adults, not older adults in general (Bambauer & Prigerson, 2006). In light of demonstrated findings that older adults in need of be havioral health services are acutely affected by perceived stigma, measures developed specifically for this population would be a huge step toward understanding how to combat the problem of perceived stigma for older adults Facilitators to Receiving Serv ices Much research has been conducted to identify barriers to re ceiving treatment. I n addition, researchers have attempted to identify which personal factors s ignificantly predict seeking treatment for behavioral health problems. Being female, not being

PAGE 32

21 ma rried, and having higher levels of education are all predictors of seeking and receiving treatment (Crabb & Hunsley, 2006). Additionally, having a driver's license and access to transportation are also associated with greater numbers of health care visits (Arcury et al., 2005). Recently, increasing numbers of "carve in" or co located behavioral health services are being offered to patients. "Carve in" services integrate behavioral health care into the system of general health care services (USDHHS, 1999). A s Speer and Schneider (2003) state, when a primary care physician is able to take a patient down the hall to meet with a counselor, the physician is "implicitly vouching for the mental health provider, providing personal reassurance and support to the pati ent, and demonstrating to the patient the mutually respectful relationship between the two providers" (p. 95). Recent studies have found less perceived stigma, improved communication, and improved convenience for patients receiving care from behavioral hea lth professionals co located within the offices of primary care providers (Williams, Shore, & Foy, 2006). Models that integrate mental health treatment into primary care have had some success engaging and treating older adults. Because older adults tend to seek mental health treatment in primary care (Kaplan Adamek, & Calderon 1999), Katon and colleagues introduced a structured depression treatment program into the primary care setting. Patients participating in the program displayed better medication adh erence, better satisfaction with care, and a greater decrease in severity of major depression (Katon, Von Korff, Lin, Bush, & Ormel, 199 7 ). Another study examining integrated service delivery found a declining trend in institutionalization, lower rates of caregiver

PAGE 33

22 burden, and lower rates of deterioration among the group receiving integrated care (Tourigny, Durand, Bonin, Hebert, & Rochette, 2004). In the PRISM E study, Bartels and colleagues (2004) co located integrated mental health and substance abuse se rvices in a primary care setting. When compared with an enhanced referral model, patients using the integrated model had a greater rate of treatment engagement, greater mean number of visits, and a greater rate of return for subsequent visits. The integrat ed model was also associated with a greater proportion of first visits occurring within two weeks from randomization compared with the referral group. This integrated model was also particularly effective at engaging individuals with at risk alcohol use an d those with active suicidal ideation. In another study of integrated services, clinicians reported preferring the integrated care rather than the referral system for older adults with psychiatric disturbances (Gallo et al., 2004). The Impact of Behavioral Health Problems on Health Care Systems As evidenced by the epidemiological studies discussed above, many older adults have behavioral health problems or are at risk for their development. The impact of these behavioral health problems are far reaching and are especially evident in the health care industry. Historically, the mental health service delivery system has been fragmented and ill equipped to accommodate the growing and various needs of older adults living in the community (Bartels, 2002; 2003). Wi th the numbers of older adults growing at such a high rate, the entire health care system may need reform to be able to meet the needs of the elderly population in the U.S. Mental disorders in older adults have been associated in research literature with h igher health care costs (Katon, Lin, Russo, & Unutzer, 2003; Luber et al., 2001). In a

PAGE 34

23 four year prospective study of HMO Medicare enrollees, Unutzer and colleagues (1997) discovered that participants with significant depressive symptoms had higher health care costs (about 50% higher) than their non depressed counterparts throughout the four years of the study, even after adjusting for age, gender, and chronic illness. The depressed group exhibited higher costs in all categories of medical care (i.e., inpat ient/outpatient visits, laboratory tests, and emergency room visits). Only about 1% of the health care costs were from the specialty mental health or substance abuse sectors, indicating that if behavioral health care was being delivered, it was happening i n primary care settings. This underutilization of appropriate behavioral health services contributes to the economic burden. Older adults who do not receive services or receive inappropriate services eventually need more expensive intensive treatment in th e future. If behavioral health problems are identified early and managed properly with evidence based treatment approaches, the economic burden of these disorders would be reduced (Wang, Simon, Kessler, 2003). Older adults living in long term care faciliti es have a high need for behavioral health services. In a study of Florida residents of assisted living facilities (ALFs), Becker, Stiles, and Schonfeld (2002) discovered that the proportion of residents receiving mental health care was greater than those o lder adults residing in the community. The implication of this is that community dwelling older adults are underrepresented in the receipt of mental health treatment and /or that ALF residents exhibit a great need for mental health services. The ALF residen ts had greater mental health service expenditures, in some cases more than double, than community dwelling older adults. The authors posit that "given the reported negative outcomes for the ALF population with mental health

PAGE 35

24 needs, the effectiveness of the mental health services provided remains questionable" (p. 96). Untreated behavioral health problems have been associated with overutilization of medical services and increased health care costs (Speer & Schneider, 2003). When a diagnosis of a behavioral di sorder such as depression is missed by service providers, the search for physical explanations of symptoms causes unnecessary increases in medical utilization rates and expenditures (Goldman et al., 1999). Conclusion Research has consistently shown that older adults are in great need of behavioral health services, but very few are identified as having such needs and fewer still actually receive appropriate treatment. Anxiety, depression, and alcohol use disorders are not uncommon in older adults, although these disorders are not a part of normal aging. Successful treatment modalities do exist that have been specifically designed for use with older adults. The reasons for this underidentification are many and varied. Factors associated with providers inclu de erroneous beliefs that symptoms are a normal part of aging, a lack of training in gerontology and behavioral health, a tendency to focus on somatic complaints, and problems with diagnostic criteria. The fragmentation of the health care service system ma kes it difficult for many people, especially older adults, to navigate and access appropriate services. Often, it may be difficult for providers to identify symptoms of a behavioral health problem Older adults may display atypical symptoms or may emphasiz e somatic complaints instead of behavioral health problems. Co occurring disorders and certain medications also make the diagnosis of a behavioral health problem

PAGE 36

25 difficult for the health care provider. Older adults may also be reluctant to seek out service s due to limited transportation, financial barriers, or a whole host of other personal factors. The stigma associated with mental illnesses is also a major obstacle to seeking and receiving appropriate care. The consequences of stigma are far reaching. St igma has been related to poor quality of life outcomes, family disintegration, and lack of community participation and contribution to name a few (Ferrante, March 2007). By identifying stigma as a major barrier to receiving health care, researchers and po licymakers are one step closer to eliminating the problem. Public education campaigns and empowering older adults will enable more older adults to seek out appropriate services. Interventions aimed at reducing feelings of stigma within the individual with behavioral health problems is also an important step for researchers. The costs associated with not receiving treatment are high for this population. Untreated behavioral health problems lead to complications with the treatment of physical health problems and make recovery difficult to attain. When identified early, behavioral health problems in older adults can be treated appropriately and cost effectively. This calls for improvements in the health care service system, including improved training for provi ders in working with older adults to prevent and treat behavioral health problems. All of these factors make identification and treatment provision to older adults with behavioral health problems difficult. There is a pressing need to remedy the situation as the Baby Boom generation ages and the numbers of older adults living in this country grow However, researchers in the field of aging studies are working toward remedying

PAGE 37

26 these issues and ensuring that those people in need of services are identified and served in a timely and appropriate manner. Study Aims The aims of this study are to 1) examine the levels of perceived stigma associated with behavioral health problems and mental health services within a sample of older primary care patients 2) determi ne whether stigma is a static or dynamic variable, and 3) identify which mental health factors are affected by perceived stigma. To address these aims, several research questions were generated and tested through analyses of the data collected through the PRISM E Study. PRISM E is a multi site, longitudinal (three points in time), randomized trial comparing two behavioral health models of care for older adults with depression, anxiety, and/or at risk drinking. For this study, the research questions and hyp otheses are: 1. Which demographic and behavioral health variables are associated with level of stigma at baseline? 2. Does level of stigma change over the three time points? Does change vary according to engagement? Hypothesis One Respondents randomly assigned to the integrated care model (co located services) are hypothesized to demonstrate greater reductions in perceived stigma than those in the enhanced referral model. Previous research has demonstrated that older adults receiving care in co located treatmen t models display better treatment adherence, satisfaction with care, and a decrease in symptoms (Bartels et al., 2004; Gallo et al., 2004; Katon et al., 1997) than those in usual care models. Co located practices also lend an added credibility to the behav ioral health specialist, perhaps leading to improved

PAGE 38

27 engagement in services (Speer & Schneider, 2003), improved communications, and less stigma (Williams et al., 2006). 3. What is the relationship between change in stigma and change in behavioral health varia bles over time? Hypothesis Two It is predicted that as feelings of stigma improve over time, behavioral health outcomes will also improve over time. Research has demonstrated strong associations between self esteem, depressive symptoms, and perceived sti gma. Improvements in self esteem and depression have been linked to decreased feelings of perceived stigma, indicating that perceived stigma is not a stable character trait and is able to change (Link et al, 2001; 2002). Although the current research study will not be able to determine causality, it is believed that these variables will have a positive relationship. 4. Assuming level of stigma does change over time, which variables are related to change over time? Hypothesis Three Based on previous literatur e ( Rusch, Angermeyer, & Corrigan, 2005; Sirey et al., 2001), it is predicted that respondents with l ower levels of feelings of stigma at baseline will demonstrate greater improvement s in behavioral health outcomes over time. The reasoning behind this hypot hesis is that if a person feels no or little stigma at the onset of treatment, he or she will be more likely to become engaged in treatment and attend the treatment sessions.

PAGE 39

28 Chapter Two Method Research Design The research design for the curre nt study is a secondary analysis of data collected for the Primary Care Research in Substance Abuse and Mental Health for Elderly (PRISM E) study. This dataset allowed for an examination of stigma and how it is related to behavi oral health variables within a large sample of older adults from around the United States The PRISM E study will be described in detail below. PRISM E The Primary Care Research in Substance Abuse and Mental Heal th for Elderly (PRISM E) study wa s a multi site, randomized, comparative trial examining two models of care for persons aged 65 and older with symptoms of depression, anxiety, an d at risk drinking. This study is the largest randomized study of behavioral health care among the elderly to date. The overall goal of the study wa s to compare the effects of integrated behavioral health care models and enhanced referral care models on engagement, participation, patient clinical outcomes, and cost effectiveness for the target conditions of depression, anxiety, and at risk drinking in o lder adults (Levkoff et al., 2004). Patients randomized to the integrated care model received behavioral health services in their primary care clinic from a mental health provider while patients randomized to the enhanced referral model received a referra l from their primary care provider to a ttend a

PAGE 40

29 specialty behavioral health clinic geographically separate from their primary care provider (Chen et al., 2006). P rimary care patients were enrolled in the study at ten sites across the United States. Each sit e contained at least one primary care clinic for recruitment, and integrated care and referral models for treatment within a specific geographic region. A total of 34 primary care clinics/practices and 22 mental health/substance abuse specialty agencies pa rticipated in the trial. The study sites included five V eterans A ffairs (VA) medical centers, three community mental health centers, and two hospital networks (Levkoff et al., 2004). Service m odels Participants were randomly assigned to receive treatment for depression, anxiety, and/or at risk drinking thorough one of two models: the integrated care model or the enhanced referral model. The integrated care models were sites where the behavioral health care providers were co located within primary care. Bec ause older adults tend to seek mental health treatment in primary care (Kaplan et al. 1999), integrating behavioral health and primary care may reduce the stigma and other barriers associated with specialty mental health care. Integrated models have been found to improve medication adherence, increase patient satisfaction with care, and decrease psychological symptoms (Katon et al. 1 997 ). I ntegrated service delivery has also been related to declining trend s in institutionalization, lower rates of caregive r burden, and lower rates of deterioration (Tourigny et al. 2004). In the PRISM E study, integrated care model sites met four criteria: 1) sites had to have their integrated model i n operation for a minimum of six months before the start of the study ; 2) mental health and substance abuse services must have been co located in a

PAGE 41

30 primary medical care clinic with no distinction between the two in terms of signage, staff, or clinic names ; 3) mental health and substance abuse services must have been provided by certified specialists in collaboration with primary care providers; and 4) primary care providers were to remain involved in the patients' care by documenting ongoing communication with the mental health/substance abuse services staff (Levkoff et al., 200 4). Appointments with the mental health and substance abuse provider were required to be scheduled within 2 to 4 weeks following the primary care provider visit (Bartels et al., 2004). The enhanced referral model sites provided mental health and substance abuse services in a specialty setting that was physically separate and designated as a mental health/substance abuse clinic. The referral model at each site was required to include three elements: 1) all patients with an identifiable mental health or at ri sk drinking problems received an appointment with the specialty mental health agency designated by the primary care clinic ; 2) specialty services were provided in a physically separate location and designated as a mental health/substance abuse clinic; and 3) mental health and substance abuse services were provided by licensed clinicians. Enhanced referral sites were required to meet certain criteria to ensure it was well functioning, such as facilitating transportation for patients to the referral clinic, e nsur ing patients received an appointment with the specialist within 4 weeks of randomization notify ing the primary care clinic if the patient failed to attend their first visit, and ensuring a process for emergency or urgent consults (Levkoff et al., 2004 ). Adherence to the required criteria wa s assessed during site visits by the coordinating center staff. Since many of the sites had well functioning treatment models

PAGE 42

31 in place prior to the start of the study, investigators decided against requiring specifi c clinical interventions, with the exception of at risk drinking in the integrated model. Basic information was collected on each behavioral health treatment visit, including the type and amount of intervention each participant received. Participants rando mized to the integrated model who exhibited at risk drinking received B rief A lcohol I ntervention, a manualized intervention developed specifically for older adults by Barry, Oslin, and Blow (2001 ; Levkoff et al., 2004). Study Sample Any p erson aged 65 or older who had a primary care appointment at a participating clinic between March 2000 and March 200 2 was screened for a mental healt h disorder or at risk drinking ( N = 24, 154 ) Another 776 patients were referred to the study by their primary care provider. Out of the 24,930 initially screened 1,102 were ineligible to continue in the screening process due to cognitive impairment or incomplete data VA sites also disqualified any females from participating in the study. Another 17,398 screened negative for a ny of the three required behavioral health conditions (depression, anxiety, and/or at risk drinking). Of the 6,430 eligible participants who screened positive for signs of mental health and/or alcohol misuse problems (about 25% of those screened), 3,225 pa tients were ineligible because they were currently receiving mental health/substance abuse treatment or refused to complete the next step of the screening process which was the baseline assessment Following the first screening, 3,205 patients co mpleted t he baseline assessment. Of those, 1,183 were excluded due to incomplete information, no target diagnosis, identified as being hypomanic, psychotic, and/or were receiving mental health or substance abuse treatment. The final study group

PAGE 43

32 comprised 2,022 part icipants who met study criteria, consented to participate in the study, and were randomized to one of the two study conditions (Bartels et al., 2004). Patients at eight sites were randomized to treatment models through a permuted blocks design, stratified by site, major diagnostic category, and age group (65 to 74 years, and 75 years or older). Blocks of size six were used for younger participants, and size four for older participants. Two VA sites already had well functioning randomization procedures in pl ace prior to the onset of the study. These two sites assigned patients to the models based on their Social Security number. Analyses conducted by the investigators showed this method to be unbiased and comparable to the randomization procedures used at the other eight sites (Levkoff et al., 2004). Measures The screening interview included measures such as the General Health Questionnaire, suicidal ideation questions modified from the PRIME MD, and questions on quantity/frequency of alcohol consumption. The baseline interview diagnosed depression, anxiety, and at risk drinking using the following instruments: Mini International Neuropsychiatric Interview, Center for Epidemiological Studies Depression Scale (CES D), Beck Anxiety Inventory, an alcohol frequenc y/quantity scale, and a detailed medication review. Additional assessments included demographic data, the Paykel Suicide Scale, the Short Michigan Alcohol Screening Test Geriatric Version (S MAST G), and the Medical Outcomes Study 36 Item Short Form, and a stigma assessment scale, along with other attitudinal questions. In addition to completing a screening interview to identify potential participants, study staff completed interviews with participants at baseline (conducted within six weeks of the initial screening) 3

PAGE 44

33 months post enrollment (90 days after baseline assessment) and again at 6 months post enrollment (180 days after baseline assessment; Herr & Zubritsky, 2005 December ) Stigma The PRISM E protocol includes the SAMHSA Mental Health and Alco hol Abuse Stigma Assessment. The assessment contains seven items and was administered at baseline, three, and six months post enrollment The items are presented in Table 1. Chen and colleagues (2006) analyzed the first item from the Stigma Assessment Scal e in a multiple regression procedure predicting satisfaction with mental health services provided in the PRISM E study. In their study, this one stigma item was found to be marginally associated with satisfaction, with those participants indicating higher levels of stigma feeling less satisfied with their behavioral health services. Unfortunately, no psychometric information on the Stigma A ssessment Scale was included in this article. This is the only published article based on PRISM E that uses the S tigma A ssessment Scale therefore it is an area of PRISM E that has been relatively unexplored until this point. Diagnosis /Outcome Participants were assessed at baseline, 3 months, and 6 months for depression, anxiety, and at risk alcohol use. A number of meas ures were used in order to capture this information from participants and will be described below. Several of the measures are from the Mini International Neuropsychiatric Interview (MINI ; Sheehan et al., 1998) a short, structured interview designed for d iagnosing disorders based on the Diagnostic and Statistical Manual Disorders IV (DSM IV; American Psychiatric Association, 1994)

PAGE 45

34 Table 1 Stigma Assessment Scale SAMHSA Mental Health and Alcohol Use Stigma Assessment "Some people consider a mental h ealth or an alcohol problem a mark of shame, and others do not. We are trying to find out what older persons feel about this issue. The next few questions ask how you would react if you had to deal with such a problem." 1. Would you be embarrassed or ash amed if you had a mental health problem? Not at all Not Very Somewhat Very Extremely Don't Know/Refused 2. Would you be embarrassed or ashamed if you had an alcohol abuse problem? Not at all Not Very Somewhat Very Extremely Don't Know/Refused NOTE: If stigma exists ("Not Very" or above) for only mental health or only alcohol, ask remaining questions accordingly. 3. Do you think people around you would think differently of you if you received mental health or alcohol abuse treatment? Not at all Not Very Somewhat Very Extremely Don't Know/Refused 4. Would it be difficult for you to start mental health or alcohol abuse treatment if other people knew that you were going to be in treatment? Not at all Not Very Somewhat Very Extremely Don't Know/Refus ed 5. How comfortable would you be talking about your mental health or alcohol abuse problems with your primary care doctor? Not at all Not Very Somewhat Very Extremely Don't Know/Refused 6. How comfortable would you be talking about your mental health or alcohol abuse problems with a counselor or mental health professional? Not at all Not Very Somewhat Very Extremely Don't Know/Refused 7. Would it be difficult for you to obtain treatment for a mental health or alcohol abuse problem in a setting that was clearly identified as a mental health clinic or alcohol treatment center? Not at all Not Very Somewhat Very Extremely Don't Know/Refused Measures of depressive disorders Depression was assessed using four scales: MINI Major Depression Scale, MIN I Dysth y mia Scale, MINI Depression History Scale, and the Center for Epidemiological Studies Depression Scale (CES D). The MINI Major

PAGE 46

35 Depression Scale was asked of participants who indicated they had felt consistently depressed for the past two weeks and/o r felt less interested in most things over the past two weeks. The MINI Major Depression Scale consists of seven items assessing respondents' appetite, sleep habits, movement problems, energy level, feelings of worthlessness and guilt, ability to concentra te, and feelings of suicide that occurred over the past two weeks. The responses were summed to create a range from 0 to 7, with higher scores indicating a greater severity of depression. Those who scored a 3 or higher (along with meeting other criteria) w ere identified as having major depression. The MINI Dysthymia Scale was asked of those participants who had feelings of depression but did not meet the criteria for Major Depression. Those who indicated that they had felt sad most of the time for the last two years and that period was not relieved for two months or more were administered the MINI Dysthymia Scale. The MINI Dysthymia Scale consists of six items assessing appetite, sleep habits, energy level, self confidence, ability to concentrate, and feelin gs of hopelessness during the past two years. The responses were summed to create a range from 0 to 6, with higher scores indicating a greater severity of dysthymia. Those who scored a 2 or higher were identified as having dysthymia. The MINI Depression Hi story Scale was asked of those participants who had feelings of depression but did not meet the criteria for major depression or dysthymia. The MINI Depression History Scale was asked of those participants who indicated that they had felt depressed for a p eriod of two weeks or more during their lifetime. The MINI Depression History Scale consisted of seven questions assess ing the participants' appetite, sleep habits, movement problems, energy level, feelings of worthlessness,

PAGE 47

36 ability to concentrate, and sui cidal feelings. The responses were summed to create a range from 0 to 7, with higher scores indicating a greater severity of depressive symptoms. Those who scored a 3 or higher were identified as having a positive h istory of depression The MINI Depression History Scale was only assessed at Baseline (Time 1). The CES D (Radloff, 1977) was asked of a participant s identified as having major depression, dysthymia, minor depression, a history of depression, or who was taking anti depressant/anti anxiety medicat ions. The CES D is a twenty item self report inventory designed to assess an individual's depressive sym ptoms over the past seven days. A score is 16 or more indicates depression. The author reports high estimates of internal consistency (Radloff, 1977). M easures of anxiety disorders Anxiety was assessed using three scales: MINI Panic Disorder Scale, MINI Generalized Anxiety Scale, and the Beck Anxiety Inventory (BAI). The MINI Panic Disorder Scale was asked of those participants who indicated during their lifetime they had had spells of feeling anxious occurring unexpectedly and lasting more than 10 minutes. The MINI Panic Disorder Scale contained 13 items assessing respondents' feelings during those spells including physical and emotional symptoms. Respon ses were summed to create a total score that ranged from 0 to 13, with higher scores indicating a greater severity of symptoms of panic disorder. Those who scored a 4 or higher were identified as having a panic disorder. The MINI Generalized Anxiety Disor der Scale was asked of those who indicated that they had been excessively worried or anxious most days over the past six months which interfered with functioning and did not meet criteria for panic disorder. The MINI Generalized Anxiety Disorder Scale cons isted of six items assessing feelings of

PAGE 48

37 restlessness, tenseness, tiredness, irritability, difficulty sleeping, and difficulty concentrating during the past six months. Responses were summed to create a total score that ranged from 0 to 6, with higher scor es indicating a greater severity of symptoms of generalized anxiety disorder. Those who scored a 3 or higher were identified as having a generalized anxiety disorder. T he Beck Anxiety Inventory (BAI; Beck, Epstein, Brown, & Steer, 1988) was asked of parti cipants who scored positive for panic disorder, generalized anxiety disorder or currently taking anti depressant or anti anxiety medication The BAI is a self report questionnaire consisting of 21 items assessing anxiety symptoms over the past week. Respo nses were summed to create a total score ranging from 0 to 62, with higher scores indicating greater severity of symptoms. The BAI has been reported to have good reliability and validity (Carmin, Pollard, & Gillock, 1999). Measures of at risk drinking Par ticipants who indicated having at least one alcoholic drink during the past year were screened for at risk alcohol use. Participants who had an alcoholic beverage in the past week were asked to indicate the total number of drinks ingested during the past w eek. Respondents were also asked to indicate the number of times they had four or more drinks in the same day during the last three months. At risk drinking was defined as drinking more than 14 drinks per week for men and 12 drinks per week for women or ha ving four or more binges (defined as four or more drinks in a day) during the last three months (Levkoff et al., 2004). The Short Michigan Alcohol ism Screening Test Geriatric Version (SMAST G ; Blow, Gillespie, Barry, Mudd, & Hill, 1998 ) was administered to those participants who were positive for at risk drinking. The SMAST G consisted of ten items assessing a

PAGE 49

38 respondents drinking behavior, for example, the frequency of underestimating drinking, skipping meals when drinking, memory loss associated with d rinking. The responses to the SMAST G were summed to create a score ranging from 0 to 10, with scores of two or higher indicating a need for further assessment for at risk drinking behavior. The S MAST G has been demonstrated to have a specificity of 97% a nd a sensitivity of 85% when compared with DSM 3 R diagnosis of alcohol abuse or dependence (Blow et al. 1 998). Analysis Plan Analysis of the PRISM E data will begin with descriptive statistics in order to gain a more detailed understanding of the chara cteristics of the sample and the levels of perceived stigma of participants. A detailed examination of the Stigma Assessment Scale will take place. This examination will focus on the content and structure of the instrument and will include inter item corre lation, item total correlation, reliability analysis, and finally, a factor analysis. Research Question One will attempt to identify those demographic and behavioral health variables that are associated with baseline levels of stigma. Attempts to answer this question include the development of a correlation table examining the relationship between baseline stigma variable and selected mental health variables, for example scores from the CES D, S MAST G, and MINI depression scale. The correlation table wil l provide descriptive information about the relationships between the behavioral health variables and level of stigma. Due to skip patterns of the PRISM E research protocol, not all participants answered all of the behavioral health questions. T he followi ng measures have the highest

PAGE 50

39 response rate and will be included in a multiple regression analysis to examine the relationship between level of stigma at baseline and behavioral health variables: CES D, BAI and the frequency of alcohol use. Demographic var iables (e.g., age, gender, race/ethnicity) will also be included in multiple regression analysis. When appropriate, demographic variables will be dichotomized. Research Question Two seeks to identify if level of stigma change s over the three time points. T he second part of the question will examine if engagement has an impact on change in perceived stigma over time. The analysis plan for this question include s conducting a repeated measure analysis of variance (ANOVA) model to identify any significant chang es in the scores of the Stigma Assessment Scale over time. If there is a significant change, the next step will include an ANOVA mixed design which will examine both change over time and differences by group (classified as engaged/not engaged ) PRISM E pro tocol defines engaged as attending at least one behavioral health treatment session. Hypothesis One will also seek to be supported through this analysis plan. Hypothesis One predicts that the participants assigned to the integrated care model (which utiliz ed a co location service delivery model) will demonstrate greater improvements in feelings of stigma compared to their peers assigned to the enhance referral model of care. Research Question Three attempts to clarify the re lationship between change in stig ma and change in behavioral health variables over time. A s with Research Question One the analysis for this question will begin with the dev elopment of a correlation table Change in level of stigma will be captured by calculating the change in stigma fro m Time 1 to Time 3, a 6 month interval. Changes in the mental health variables will be calculated

PAGE 51

40 in the same way. The correlation table will pair the change in stigma variable with selected mental health variables, for example change scores from the CES D S MAST G, MINI depression scale, etc. The next step will include conducting a multiple regression analysis to examine the relationship between change in level of stigma and in the main behavioral health measures : CES D, BAI and the frequency of alcohol use. These analyses will also seek to confirm Hypothesis Two which states that as feelings of stigma improve over time, behavioral health variables will also improve over time. The fourth Research Question is an attempt to understand which variables are r elated to the change in stigma levels over time. To answer this question, multiple regression will be performed to predict change in level of stigma from Time 1 to Time 3 by past behavioral health service use, number of PRISM E treatment sessions attended, satisfaction with PRISM E services, cultural competence of PRISM E services, diagnosis, and the model of assignment. The variable categorizing a site as VA or non VA will be included to control for any influence by site. These analyses will also seek to c onfirm Hypothesis Three which predicts that respondents with lower levels of feeling of stigma at baseline will demonstrate greater improvements in behavioral health outcomes over time compared to those with greater levels of stigma.

PAGE 52

41 Chapter T hree Results Sample Characteristics The final sample included in the analyses contained 2,022 participants. The majority of participants were male, White, married/partnered, and averaged 74 years of age. Participants generally lived with others, had a high sc hool education or less, and rated their physical health as fair. Nearly two thirds of the sample was enrolled into the study through V A Medical Centers, where only males were eligible for participation, explaining why roughly three quarters of the sample w ere males (see Table 2 ). The sample was evenly divided between the two treatment models with 9 82 participants randomly assigned to the integrated care model and 1,0 10 randomized to the enhanced referral care model (30 cases are missing this information) L evel of Stigma Respondents were asked to indicate their feelings of stigma associated with mental health or alcohol abuse problems at the baseline interview. Item 1 of the Stigma Assessment Scale asks "Would you be embarrassed or ashamed if you had a ment al health problem?" and respondents were asked to respond on a five point Likert type scale ranging from N ot at A ll to E xtremely Figure 1 illustrates the responses to this question. While nearly half of respondents indicated that they would not be emb arrassed or ashamed if they had a mental health problem (46.3%), 53.7% of respondents indicated

PAGE 53

42 Table 2 Sample Characteristics ( N = 2,022) Total Sample ( N = 2,022) No Mental Health Stigma Indicated ( n = 896) Mental Health Stigma Indicated ( n = 1,039) Age (Years) Mean (SD) Range 73.5 (6.2) 65 103 73.3 (6.1) 65 94 73.5 (6.2) 65 103 Gender Male n (%) 1,461 (72.3%) 706 (78.8%) 1 705 (67.9%) 1 Race/Ethnicity White n (%) Black n (%) Hispanic n (%) Asian n (%) Other n (%) 1,0 65 (52.7%) 510 (25.2%) 302 (14.9%) 112 (5.5%) 53 (2.6%) 426 (47.5%) 1 296 (33.0%) 1 115 (12.8%) 47 (5.2%) 12 (1.6%) 607 (58.4%) 1 187 (18.0%) 1 163 (15.7%) 63 (6.1%) 36 (3.5%) Marital Status Married/Partnered Widowed Divorced Never married Separated 972 (48.1%) 500 (24.7%) 310 (15.3%) 123 (6.1%) 108 (5.3%) 400 (44.6%) 1 225 (25.1%) 147 (16.4%) 62 (6.9%) 58 (6.5%) 539 (51.9%) 1 245 (23.6%) 149 (14.3%) 58 (5.6%) 44 (4.2%) Living Arrangement Live with others Live alone 1,308 (64.7%) 693 (34.3%) 569 (63.5%) 316 (35.3%) 687 (66.1%) 343 (33.0%) Highest Level of School Completed Less than 8 th grade Less than 12 th grade High school graduate/GED Some college/Trade school College graduate Graduate school 428 (21.2%) 447 (22.1%) 451 (22.3%) 380 (18. 8%) 180 (8.9%) 126 (6.2%) 177 (19.8%) 215 (24.0%) 196 (21.9%) 173 (19.3%) 77 (8.6%) 55 (6.1%) 222 (21.4%) 216 (20.8%) 233 (22.4%) 198 (19.1%) 98 (9.4%) 68 (6.5%) Site Type VA Site 1,220 (60.3%) 620 (69.2) 1 566 (54.5%) 1 Self Rated Physical Health Excellent Very good Good Fair Poor 52 (2.6%) 201 (9.9%) 527 (26.1%) 795 (39.3%) 414 (20.5%) 34 (3.8%) 88 (9.8%) 242 (27.0%) 353 (39.4%) 175 (19.5%) 38 (3.7%) 113 (10.9%) 304 (29.3%) 367 (35.3%) 214 (20.6%) 1 Significantly different at p < .05

PAGE 54

43 Fi gure 1. Would you be embarrassed or ashamed if you had a mental health problem? ( n = 1,935) that they did have some level of stigma associated with mental health problems. One fifth (19.7%) of respondents indicated that they would feel "Very" or "Extreme ly" embarrassed or ashamed. Table 3 presents the frequencies of responses on the Stigma Assessment Scale. Nearly half of participants indicated they believed others would think differently of them if they sought behavioral health treatment (43.2% responded "Somewhat", "Very", or "Extremely" to Item 3). Most participants indicated that it would not be very difficult for them to start treatment if others knew (58.3% responded "Not at All" or N ot Very" to Item 4). The majority of participants responded that t hey would feel comfortable talking about their behavioral health problems with their primary care physician or a behavioral health professional (60.0% and 55.1% respectively, answered "Very" or "Extremely" to Item s 5 and 6). Just over half of participants responded that it would not be difficult for

PAGE 55

44 them to obtain treatment from a setting clearly identified as a behavioral health treatment center (52.7% answered "Not at All" or Not Very to Item 7). Table 3 also illustrates the amount of data missing from the Stigma Assessment Scale. Item s 1 and 2 are missing 4.3% and 6.5% respectively, while Item s 3 through 7 are missing between 16% and 19%. It is not clear from the PRISM E published literature why one fifth of the responses are missing for these items. Ta ble 3 Frequency of Responses to the Stigma Assessment Scale Response Options n (%) Item Not at All Not Very Some what Very Extremely Missing 1. Would you be embarrassed or ashamed if you had a mental health problem? 896 (44.3) 199 (9.8) 459 (22.7) 197 (9.7) 184 (9.1) 87 (4.3) 2. Would you be embarrassed or ashamed if you had an alcohol or substance abuse problem? 681 (33.7) 129 (6.4) 364 (18.0) 367 (18.2) 349 (17.3) 132 (6.5) 3. Do you think people around you would think differently of you if you received mental health or alcohol abuse treatment? 566 (28.0) 190 (9.4) 512 (25.3) 190 (9.4) 172 (8.5) 392 (19.4) 4. Would it be difficult for you to start mental health or alcohol abuse treatment if other people knew that you were going to be in tr eatment? 970 (48.0) 209 (10.3) 296 (14.6) 116 (5.7) 72 (3.6) 359 (17.8) 5. How comfortable would you be talking about your mental health or alcohol abuse problems with your primary care doctor? 112 (5.5) 99 (4.9) 250 (12.8) 741 (36.6) 473 (23.4) 339 (16 .8) 6. How comfortable would you be talking about your mental health or alcohol abuse problems with a counselor or mental health professional? 115 (5.7) 113 (5.6) 312 (15.4) 712 (35.2) 402 (19.9) 368 (18.2) 7. Would it be difficult for you to obtain trea tment for a mental health or alcohol abuse problem in a setting that was clearly identified as a mental health clinic or alcohol treatment center? 793 (39.2) 273 (13.5) 327 (16.2) 173 (8.6) 97 (4.8) 359 (17.8)

PAGE 56

45 When the sample is divided and examined by b aseline level of stigma associated with mental health problems, several statistically significant differences between the two groups are identified (see Table 2 ). No significant differences were seen for age ( t (1,933) = 0.92, p = .34). Women were more lik ely to indicate feelings of stigma associated with mental health problems ( X 2 (1, 1,908) = 28.22, p = .00) than their male counterparts Each racial/ethnic group offered as an option was coded as a separate variable, allowing participants to choose more th an one category. To examine if any differences in feelings of stigma existed among the groups, each racial/ethnic group was included in a chi square analysis, with mental health stigma indicated or not indicated and racial/ethnic group indicated or not ind icated. Regarding racial/ethnic group differences, Whites were more likely than not to indicate feelings of stigma ( X 2 (1, 1,932) = 23.09, p = .00) and Blacks were less likely to indicate those feelings ( X 2 (1, 1,932) = 57.95, p = .00). No statistically si gnificant differences were seen in the other racial/ethnic groups. Significant differences in feelings of stigma were evident by marital status ( X 2 (4, 1,927) = 12.95, p = .01), with those who were married likely to indicate feelings of stigma, while those who were separated, divorced, widowed, and never married were less likely to indicate feelings of stigma. No significant differences were seen when examined by living arrangement, level of schooling completed, diagnosis, or general health. Significant dif ferences were seen for type of site, with participants from non VA sites more likely to indicate feelings of stigma related to mental health problems ( X 2 (1, 1,935) = 43.94, p = .00) than were participants from VA sites Item 2 of the Stigma Assessment Sca le asks "Would you be embarrassed or ashamed if you had a n alcohol abuse problem?" and respondents were asked to respond

PAGE 57

46 on a five point Likert type scale ranging from "N ot at A ll to "E xtremely Figure 2 illustrates the responses to this item Over two thirds of respondents (64.0%) indicated that they had some level of stigma associated with alcohol abuse problems, with over one third (37.9%) responding "Very" or "Extremely" embarrassed or ashamed. Figure 2. Would you be embarrassed or ashamed if you had an alcohol abuse problem? ( n = 1,890) Just over half of the sample was identified as having either a depressive or anxiety disorder and another 23% was identified as having both a depressive and an anxiety disorder. Twenty percent of the sample was id entified as engaging in at risk drinking behavior and 5% had a dual diagnosis of at risk drinking along with a depressive and/or anxiety disorder (Table 4 ). The Stigma Assessment Scale The Stigma Assessment Scale is a seven item, self report measure of pe rceived stigma (presented in Table 1). When examining the content and scaling, it appears that there may be some structural problems with the measure. The

PAGE 58

47 questions in the Stigma Assessment Scale are assessing the respondents' amount of embar rassment or comfort level; however, the response items are in fact measuring frequency. For example, the first question asks "Would you be ashamed or embarrassed if you had a mental health problem?" which structurally would elicit a dichotomous (yes/no) r esponse. The question might have been worded better to ask "To what degree would you be ashamed". The original scale ranges from "Not at all" to "Extremely" with "Somewhat" as a midpoint. According to Bass, Cascio, and O'Connor (1974), these scale items a re measuring frequency not amount and are not of approximately equal intervals. M ore appropriate response categories to the question of "To what degree would you be ashamed" would be "None" or "Hardly Any" "Some", "Quite a bit", "An extreme amount" and "All." In addition, these suggested scale items have been measured to be relatively equal in their distance from each other, leading to a more precise measure. The first two items of the Stigma Assessment Scale (Table 1) ask a person's level of embarrassme nt if he or she had either a mental health problem or an alcohol abuse problem. For the remaining analyses, the inclusion of either Item 1 or Item 2 will be based on the individual's diagnostic category. For example, if a person was identified as Table 4 Diagnosis Depressive Disorder Only 986 (48.8%) Anxiety Disorder Only 62 (3.1%) Anxiety Disorder and Depressive Disorder 457 (22.6%) At Risk Drinking Only 408 (20.2%) At Risk Drinking and Depressive and/or Anxiety Disorder 100 (4.9%) Missing 9 (0.4%)

PAGE 59

48 having a mental health disorder (depressive or anxiety), Item 1 was used for the analyses. If an individual was identified as having an at risk drinking classification, Item 2 was used. If an individual was identified as having a mental health and alcohol use disor der (dual diagnosis), the item with the highest level of stigma (either Item 1 or Item 2) was used in subsequent analyses. This procedure resulted in a new variable which will be referred to as The Embarrassed/Ashamed Item Internal consistency is the degr ee to which all the items in a scale hang together', or are associated and are measuring the same construct (Pallant, 2001). In order to assess the internal consistency of the Stigma Assessment Scale, several procedures were utilized. First, an inter item correlation was calculated where each pair of the six items is correlated and then averaged to obtain an overall correlation for the items in the measure. Pearson P roduct Moment C orrelations between the items ranged from a high of .429 between Embarrassed /Ashamed Item and Item 3 ("Would you be embarrassed or ashamed" and "Do you think people around you would think differently", respectively) and a low of .006 between items 3 and 5 ("Do you think people around you would think differently" and "How comfo rtable would you be talkingprimary care doctor?", respectively). The average inter item correlation was 234, small according to Cohen's guidelines (1988). In addition, a total score for the six items was computed and used as a seventh variable in the in ter item analysis to compute the average item total correlation. Correlations with the new s ummed i tem ranged from a high of .643 with Item 3 ("Do you think people around you would think differently") and a low of .321 with Item 5 ("How comfortable would you be talkingprimary care doctor?"). Table 5 displays the

PAGE 60

49 item correlation matrix including the new summed item. The average item total correlation was medium at 482 (Cohen, 1988) The small average inter item and medium item total correlations, along w ith the variability in the inter item correlations themselves indicate that the Stigma Assessment Scale is measuring more than one construct. If a scale is measuring more than one construct, or is multidimensional, it should not be combined to create a sin gle scale. To further explore this scale and its dimensionality, a reliability analysis for the scale was conducted using SPSS software, which provides more information on the internal consistency of the scale. The Cronbach's alpha coefficient for the rel iability of the Stigma Assessment Scale was .437, which is well below the recommended level of .7 further indicating that the scale is multidimensional and measuring more than one construct (Pallant, 2001). Finally, an exploratory factor analysis was con ducted to determine the number of constructs that the stigma assessment is measuring. The sample size exceeded conventionally accepted minimum level for conducting a factor analysis (Nunnally, 1978; Tabachnick & Fidell, 1996) and inter item correlations we re at an acceptable level, with a Kaiser Meyer Olkin of Sampling Adequacy value of .617, meeting the recomm ended value of .6 (Kaiser, 1970; 1974) and the Bartlett's Test of Sphericity (Bartlett, 1954) reached statistical significance, meaning that it is ap propriate to conduct a factor analysis with these data. The factor analysis revealed the presence of two components with eigenvalues exceeding 1, explaining 36.3% and 25.6% of the variance respectively. The screeplot also indicated a break after the second component. Two components were extracted for further examination using a Varimax rotation. Component 1 included two

PAGE 61

50 variables, Component 2 included two variables, and two variables overlapped components (Table 6 ). The two factor solution explained a total of 61.9% of the variance, with Component 1 contributing 31.0% and Component 2 contributing 30.8%. Items 1 2, 3, 4, and 7 comprise Component 1 and Items 5 and 6 comprise Component 2. According to Kline (1994), items are selected for a component based on it s highest loading. Items 4 and 7 load on both components, but both items load higher on Component 1 (Kline, 1994) therefore were grouped with Component 1 Component 1 ( which we labeled Perceived Stigma) includes items measuring level of embarrassment and the perception of a negative response from others regarding seeking behavioral health treatment Component 2 ( which we labeled Comfort Level) includes items measuring the respondents' level of comfort in talking with a service provider about their own mental heath or alcohol abuse problems. Table 5 Inter Item Correlation Matrix of Stigma Assessment Measure Items ( n ) Q1 2. Q3. Q4. Q5. Q 6. Q7. Embarrassed/Ashamed Item (Q1 2) 1.00 (1,844) .429* (1,531) .269* (1,561) .033 (1,575) .023 (1,550) .214* (1,561) Q3. .429* (1,531) 1.00 (1,630) .322* (1,604) .006 (1,605) .045 (1,583) .149* (1,596) Q4. .269* (1,561) .322* (1,604) 1.00 (1,663) .241* (1,640) .223* (1,617) .478* (1,633) Q5. .033 (1,575) .006 (1,605) .241* (1,640) 1.00 (1,683) .638* (1,641) .215* (1,643) Q6. .023 (1,550) .045 (1,583) .223* (1,617) .638* (1,641) 1.00 (1,654) .228* (1,619) Q7. .214* (1,561) .149* (1,596) .478* (1,633) .215* (1,643) .228* (1,619) 1.00 (1,663) Sum of Q1 2 through Q7 .628* (1,844) .643* (1,630) .513* (1,663) .321* (1,683) .348* (1,654) .441* (1,663) *Correlation is significant at p < .01

PAGE 62

51 Based on the results of these analyses, the Stigma Assessment Scale measured more than one dimension of stigma and collapsing all items into one overall score would no t be a valid measure for the purposes of this dissertation. The remaining research questions will utilize the Embarrassed/Ashamed Item, the Perceived Stigma Component, and the Comfort Level Component. Perceived Stigma and Comfort Level scores were computed using the mean of responses to the Embarrassed/Ashamed Item, 3, 4, and 7, and Items 5 and 6, respectively. Means were used instead of a summative score due to missing data on some of the items. Table 6 Factor Analysis Components Abbreviated Items Perce ived Stigma ( Component 1 ) Comfort Level ( Component 2 ) Item 3. Do you think that others will think differently of you if you received MH/SA treatment? .765 Embarrassed/Ashamed Item Would you be embarrassed or ashamed if you had a MH/SA problem? .744 I tem 4. Would it be difficult to start MH/SA treatment if others knew? .667 .384 Item 7. Would it be difficult to get treatment from a place identified as a MH/SA clinic? .527 .426 Item 6. How comfortable would you be talking with a MH/SA professional? .871 Item 5. How comfortable would you be talking about your MH/SA problems to your PCP? .859 Research Questions Research Question One : Which demographic and behavioral health variables are associated with level of stigma at baseline? The analysis for this question began with the development of a correlation matrix pairing the baseline stigma variable s ( Embarrassed/Ashamed Item ; Perceived Stigma, and Comfort Level ) with selected baseline behavioral health variables, for example scores

PAGE 63

52 from the CES D, S MAST G, MINI depression scale. The correlation table provides descriptive information about the relationships between baseline behavioral health variables and baseline level of stigma. For the correlation table a diagnostic hierarchy was assumed of the i nterview protocol of the PRISM E study For example, if a participant was identified as having major depression, questions about dysthymia and depressive symptoms were not asked during the baseline interview. For the purposes of this analysis, if a person was identified as having major depression, their data were recoded as positive for dysthymia and depressive symptoms. The same is true for those participants identified as having panic disorder their missing data were recoded as positive for generalized anxiety disorder. This recoding method is both logical and has the effect of increas ing the size of the sample available for creating the correlation table The correlation s are presented in Table 7 and display some variability in the degree of the relatio nships between the baseline behavioral health variables and the baseline stigma variable s ( Embarrassed/Ashamed Item ; Perceived Stigma, and Comfort Level ). C orrelations range d from a low of .00 1 ( Perceived Stigma and Number of A lcoholic D rinks in the L ast W eek) to a high of .16 9 ( Perceived Stigma and CES D) and are considered small according to Cohen's guidelines (1988). It should be noted that higher scores on the Embarrassed/Ashamed Item and the Perceived Stigma Component indicate more feelings of stigm a, higher scores on the Comfort Level Component indicate greater feelings of comfortableness. Because not all participants answered all of the behavioral health questions (due to skip patterns), scores from the CES D, BAI and the frequency of alcohol use were included in a complete (not adjusted for other variables) multiple regression analysis to

PAGE 64

53 examine the relationship between level of stigma at baseline and behavioral health variables. These three items were asked of the most participants therefore de creasing the incidence of missing responses. Demographic variables (age, gender, race/ethnicity) were also included in the multiple regression analysis. Race/ethnicity variables were coded as categorical variables (White/Non White and Hispanic/non Hispanic ). Embarrassed/Ashamed Item r egression m odel. The first model pr edicted the baseline score of Embarrassed/Ashamed Item using the variables age, gender ( 0 = male; 1 = female), race (0 = Non White; 1 = White), ethnicity (0 = Non Hispanic; 1 = Hispanic), Table 7 Correlations of Baseline Variables Baseline Behavioral Health Variables Correlation with Baseline Embarrassed/ Ashamed Item r ( n ) Correlation with Baseline Perceived Stigma r ( n ) Correlation with Baseline Comfort Level r ( n ) MINI Major Depre ssion Scale .117** (1,250) .084** (1,339) .078** (1,168) MINI Dysthymia Scale .034 (1,094) .021 (1,173) .010 (1,018) MINI Depression History Scale .048 (1,384) .031 (1,480) .080** (1,285) CES D .162** (1,435) .169** (1,532) .042 (1,337) MINI Panic D isorder Scale .043 (192) .080 (198) .083 (173) MINI Generalized Anxiety Scale .072 (500) .056 (509) .080 (442) Beck Anxiety Inventory (BAI) .030 (728) .079* (757) .038 (662) Number of alcoholic drinks in the last week .003 (726) .001 (802) .129** (73 5) Number of alcoholic binges in the last three months .019 (961) .044 (1,050) .007 (943) S MAST G .085 (472) .010 (547) .120** (490) Significant at p < .05 ** Significant at p < .01

PAGE 65

54 b aseline scores of the CES D, BAI and frequency of alcohol use. Sim ultaneous complete multiple regression was used, meaning that all independent variables were entered into the equation at once. Results indicate that m ulticollinearity was not a problem in this analysis as evidenced by the small strength correlations betwe en the independent variables and the acceptable tolerance levels. The regression model had an R 2 value of .06 2 meaning the independent variables accounted for only 6. 2 % of the variance in the dependent variable (Baseline Embarrassed/Ashamed Item ) and was not significant (see Table 8 ) These results indicate that age, race/ethnicity, gender, baseline level s of depression, anxiety, and frequency of drinking were not significantly associated with baseline stigma levels. Table 8 Summary of Simultaneous Regr ession Analysis for Variables Predicting Baseline Embarrassed/Ashamed Item Variable B SE B p Gender (0 = male; 1 = female) .137 .215 .043 .525 Age .003 .015 .013 .846 Race (0 = Non White; 1 = White) .410 .205 .146 .047 Ethnicity (0 = Non Hispanic; 1 = Hispanic) .652 .292 .166 .026 CES D .025 .010 .199 .010 BAI .009 .009 .081 .283 # of Alcoholic Drinks in the Last Week .003 .008 .027 .677 Perceived Stigma Component r egression m odel. The next model predicted the baseline score of the Perceived S tigma Component using the variables age, gender ( 0 = male; 1 = female), race (0 = Non White; 1 = White), ethnicity (0 = Non Hispanic; 1 = Hispanic), b aseline scores of the CES D, BAI and frequency of alcohol use. S imultaneous complete multiple regression was used, meaning that all independent variables were entered into the equation at once. Results indicate that multicollinearity was not a problem in this analysis as evidenced by the small strength correlations

PAGE 66

55 between the independent variables and the ac ceptable tolerance levels. The regression model had an R 2 value of .046, meaning the independent variables accounted for only 4.6% of the variance in the dependent variable (Baseline Perceived Stigma Component), and was not significant (Table 9 ) These res ults indicate that age, race/ethnicity, gender, baseline levels of depression, anxiety, and frequency of drinking were not significantly associated with baseline perceived stigma levels. Table 9 Summary of Simultaneous Regression Analysis for Variables P redicting Baseline Perceived Stigma Component Variable B SE B p Gender (0 = male; 1 = female) .183 .147 .085 .214 Age .007 .010 .045 .497 Race (0 = Non White; 1 = White) .183 .140 .096 .192 Ethnicity (0 = Non Hispanic; 1 = Hispanic) .130 .199 .049 .516 CES D .015 .007 .180 .021 BAI .000 .006 .004 .960 # of Alcoholic Drinks in the Last Week .002 .005 .021 .751 Comfort Level Component r egression m odel. The next model predicted the baseline score of the Perceived Stigma Component using the variables age, gender ( 0 = male; 1 = female), race (0 = Non White; 1 = White), ethnicity (0 = Non Hispanic; 1 = Hispanic), baseline scores of the CES D, BA I and frequency of alcohol use. Simultaneous complete multiple regression was used, meaning that all independent variables were entered into the equation at once. Results indicate that multicollinearity was not a problem in this analysis as evidenced by the small strength correlations between the independent variables and the acceptable tolerance levels. The regression model had an R 2 value of .049, meaning the independent variables accounted for only 4.9% of the variance in the dependent variable (Baseline Comfort Level Component), and was not significant (see Table 10 ) These results indicate that age, race/ethnicity, gender,

PAGE 67

56 baseline levels of depression, anxiety, and fr equency of drinking were not significantly associated with baseline Comfort Level. Table 10 Summary of Simultaneous Regression Analysis for Variables Predicting Baseline Comfort Level Component Variable B SE B p Gender (0 = male; 1 = female) .153 .1 57 .067 .329 Age .000 .011 .002 .974 Race (0 = Non White; 1 = White) .221 .149 .109 .140 Ethnicity (0 = Non Hispanic; 1 = Hispanic) .288 .212 .101 .176 CES D .003 .007 .035 .654 BAI .006 .006 .074 .328 # of Alcoholic Drinks in the Last Week .00 9 .006 .104 .117 Research Question Two : Does level of stigma change over the three time points? Does change vary according to engagement? The analysis plan to test this research question included a repeated measures analysis of variance (ANOVA) model t o analyze the total group and a n ANOVA mixed design dividing the sample by engaged/not engaged. The PRISM E protocol defines engaged as attending at least one behavioral health treatment session. Nearly two thirds of the sample ( 59. 8%) was classified as en gaged. Embarrassed/Ashamed Item The means and standard deviations of the change in level of stigma for the total sample over time (baseline, 3 months, and 6 months post enrollment) as measured by Stigma Assessment Scale Item 1 2 are presented in Table 1 1 Results from the repeated measures ANOVA examining the change in level of stigma for the total sample over time (baseline, 3 months, and 6 months post enrollment) as measured by Embarrassed/Ashamed Item indicates that there is a significant effect for ti me, Wilks' Lambda = .939, F (2, 1,258) = 41.177, p < .01. These results indicate that the level of stigma as measured by Embarrassed/Ashamed Item significantly improved over

PAGE 68

57 time. Post hoc t tests indicate that the Embarrassed/Ashamed Items differ significa ntly from each other from Baseline to 3 months, from 3 months to 6 months, and from Baseline to 6 months. The means and standard deviations of the change in level of stigma (baseline, 3 months, and 6 months post enrollment) as measured by Embarrassed/Ashamed Item by engagement are presented in the Table 1 2 Results from the ANOVA mixed design examining the change in level of stigma by engagement indicates that there is not a significant interaction, Wilks' Lambda = .99 9 F (2, 1, 257 ) = 0.333 p = 71 6. These re sults indicate that the level of stigma as measured by Embarrassed/Ashamed Item did not significantly change over time as a function of engagement. Table 1 2 Change in Level of Stigma as Measured by Embarrassed/Ashamed Item by Engagement Time of Interview Engaged ( n = 817 ) M ( SD ) Not Engaged ( n = 443) M ( SD ) F ( p ) Baseline 2.39 (1.42) 2.44 (1.39) 3 Months Post Enrollment 2.17 (1.36) 2.23 (1.37) 6 Months Post Enrollment 2.04 (1.32) 2.04 (1.32) .333 (.716) In order to further investigate the significant change in level of stigma over time, service model type (Enhanced Referral or Integrated Model) was included in an ANOVA mixed design equation. The means and standard deviations of the change in level of Table 1 1 Change in Level of Stigma as Measured by Embarrassed/Ashamed Item T ime of Interview Total Sample ( N = 1,260) M ( SD ) F ( p ) Baseline 2.41 (1.41)* 3 Months Post Enrollment 2.19 (1.36)* 6 Months Post Enrollment 2.04 (1.29)* 41.17 7 (.000) *p < .01

PAGE 69

58 stigma (baseline, 3 months, and 6 months post enrollment) as measur ed by Embarrassed/Ashamed Item by type of treatment model are presented in Table 1 3 Results from the ANOVA mixed design examining the change in level of stigma by model indicates that there is not a significant interaction, Wilks' Lambda = .99 9 F (2, 1, 23 9 ) = 0.347 p = 707 These results indicate that the level of stigma as measured by Embarrassed/Ashamed Item did not significantly change over time as a function of type of service model. Perceived Stigma Component The means and standard deviations of the change in level of stigma for the total sample over time (baseline, 3 months, and 6 months post enrollment) as measured by the Perceived Stigma Component are presented in Table 1 4 Resul ts from the repeated measures ANOVA examining the change in the Perceived Stigma C omponent for the total sample over time (baseline, 3 months, and 6 months post enrollment) indicates that there is a significant effect for time, Wilks' Lambda = .952, F (2, 1 ,341) = 34.043, p < .01. These results indicate that the level of stigma as measured by the Perceived Stigma Component did significantly improve over time. Post hoc t tests indicate that the Perceived Stigma Component Items differ significantly from each o ther from Baseline to 3 months, from 3 months to 6 months, and from Baseline to 6 months. Table 13 Level of Stigma as Measured by Embarrassed/Ashamed Item b y Service Model Type Time of Interview Enhanced Referral Model ( n = 613) M ( SD ) Integrated Care Model ( n = 443) M ( SD ) F ( p ) Baseline 2.46 (1.42) 2.36 (1.41) 3 Months Post Enrollment 2.21 (1.37) 2.17 (1.36) 6 Months Post Enrollment 2.07 (1.32) 2.00 (1.28) .3 47 (.7 07 )

PAGE 70

59 Table 1 4 Change in Level of Stigma as Measured by the Perceived Stigma Component Time of Interview Total Sample ( N = 1,343) M ( SD ) F ( p ) Baseline 2.09 (0.95)* 3 Months Post Enrollment 1.97 (0.89)* 6 Months Post Enrollment 1.83 (0.88)* 34.043 (.000) *p < .01 The means and standard deviations of the change in level of stigma (baseline, 3 months, and 6 months post enrollment) as measured by the Perceived Stigma Component by engagement are presented in Table 1 5 Results from the ANOVA mixed design examining the change in level of stigma by engagement indicates that there is not a significant interaction, Wilks' Lambda = .997, F (2, 1,340) = 1.945, p = .143. These results indic ate that the level of stigma as measured by the Perceived Stigma Component did not significantly change over time as a function of engagement. In order to further investigate the significant change in lev el of stigma over time, model type was included in an ANOVA mixed design equation. The means and standard deviations of the change in level of stigma (baseline, 3 months, and 6 months post enrollment) as measured by the Perceived Stigma Component by type of treatment model are presented in Table 1 6 Results from the ANOVA mixed design examining the change in level of stigma by model indicates that there is not a significant interaction, Wilks' Table 1 5 Change in Level of Stigma as Measured by the Perceived Stigma Component by Engagement Time of Interview Engaged ( n = 8 19) M ( SD ) Not Engaged ( n = 524) M ( SD ) F ( p ) Baseline 2.05 (0.95) 2.15 (0.95) 3 Months Post Enrollment 1.98 (0.90) 1.95 (0.87) 6 Months Post Enrollment 1.82 (0.88) 1.84 (0.88) 1.945 (.143)

PAGE 71

60 Lambda = 1.000 F (2, 1, 321 ) = 0.303 p = 739 These results i ndicate that the level of stigma as measured by the Perceived Stigma Component did not significantly change over time as a function of type of model. Comfort Level Component. The means and standard deviations of the change in level of stigma for the total sample over time (baseline, 3 months, and 6 months post enrollment) as measured by the Comfort Level Component are presented in Table 1 7 Results from the repeated measures AN OVA examining the change in the Comfort Level Component for the total sample over time (baseline, 3 months, and 6 months post enrollment) indicates that there is not a significant effect for time, Wilks' Lambda = .995, F (2, 825) = 2.035, p = .131. These re sults indicate that the level of stigma as measured by the Comfort Level Component did not significantly improve over time Table 1 6 Change in Level of Stigma as Measured by the Perceived Stigma Component by Service Model Type Time of Interview Enhanced Referral Model ( n = 659) M ( SD ) Integrated Model ( n = 665) M ( SD ) F ( p ) Baseline 2.11 (0.99) 2.07 (0.91) 3 Months Post Enrollment 1.96 (0.89) 1.99 (0.90) 6 Months Post Enrollment 1.83 (0.89) 1.84 (0.88) 0.303 (.739) Table 1 7 Change in Level of Stigma as Measured by the Comfort Level Component Time of Interview Total Sample ( N = 827) M ( SD ) F ( p ) Baseline 3.77 (1.03) 3 Months Post Enrollment 3.80 (0.97) 6 Months Post Enrollment 3.85 (0.92) 2.035 (.131)

PAGE 72

61 The means and standard deviations of the change in level of stigma (baseline, 3 months, and 6 months post enrollment) as measured by the Comfort Level Component by engagement are presented in Table 18. Results from the ANOVA mixed design examining the change in level of stigma by engagement indicates that there is not a significant interaction, Wilks' Lambda = .998, F (2, 824) = 1.022, p = .360. These res ults indicate that the level of stigma as measured by the Comfort Level Component did not significantly change over time as a function of engagement. In order to further investigate the significant change in level of stigma over time, model type was included in an ANOVA mixed design equation. The means and standard deviations of the change in level of stigma (baseline, 3 months, and 6 months post enrollment) as measured by the Comfort Level Component by type of treatment model are presented in Table 1 9 Results from the ANOVA mixed design examining the change in level of stigma by model indicates that there is not a significant interaction, Wilks' Lambda = .998 F (2, 809 ) = 0. 840 p = 432 These results indic ate that the level of stigma as measured by the Comfort Level Component did not significantly change over time as a function of type of model. To further examine the relationship between stigma and engagement, a set of three Pearson correlations were cond ucted to determine if baseline level of stigma (as Table 1 8 Change in Level of Stigma as Measured by the Comfort Level Component by Engagement Time of Interview Engaged ( n = 503) M ( SD ) Not Engaged ( n = 324) M ( SD ) F ( p ) Baseline 3.84 (1.00) 3.67 (1.06) 3 Months Post Enrollment 3.82 (0.94) 3.77 (1.01) 6 Months Post Enrollment 3.86 (0.91) 3.84 (0.94) 1.022 (.360)

PAGE 73

62 measured by the Embarrassed/Ashamed Item, the Perceived Stigma Component, and the Comfort Level Component) affected the number of treatment sessions attended. These correlations were all small, according t o Cohen's guidelines (1988); all were less than .008. Table 1 9 Change in Level of Stigma as Measured by the Comfort Level Component by Service Model Type Time of Interview Enhanced Referral Model ( n = 423) M ( SD ) Integrated Model ( n = 389) M ( SD ) F ( p ) Baseline 3.77 (1. 77) 3.77 (0.99) 3 Months Post Enrollment 3.80 (0.96) 3.81 (0.98) 6 Months Post Enrollment 3.88 (0.92) 3.81 (0.92) 0.840 (.432) Research Question Three : What is the relationship between change in stigma and change in behavioral health variables over time? As with Question 2, the analysis for this question began with the development of a correlation table Change in level of stigma was captured by calculating the change in Embarrassed/Ashamed Item the Perceived Stigma Component, and the Comfort Level Component from baseline to 6 month s Changes in the behavioral health variables were captured by calculating the changes in the variables from baseline to 6 months The correlation table pair s the change in the stigma variable s with selected behavioral he alth variables, for example change scores from the CES D, S MAST G, MINI depression scale, etc. It should be noted that higher scores on the Embarrassed/Ashamed Item and the Perceived Stigma Component indicate greater feelings of stigma and higher scores o n the Comfort Level Component indicate greater feelings of comfortableness. The correlation table displayed in Table 20 and illustrates the variability in the strengths of the relationships between the change s in the behavioral health variables and

PAGE 74

63 the cha nge in the stigma variable s ( Embarrassed/Ashamed Item Perceived Stigma Component, and Comfort Level Component ). Variables were removed if fewer than 50 participants had a score for that scale. The correlations range from a low of .0 0 1 ( the Perceived Stigm a Component and CES D ) to a high of .174 ( Comfort Level Component and the MINI Major Depression Scale ) A ll of the correlations are considered small according to Cohen's guidelines ( < .290; 1988). Because not all participants answered all of the behavioral health questions (due to skip patterns), the following variables were included in a simultaneous complete multiple regression analysis to examine the relationship between change in level of stigma and change in behavioral health variables from baseline to 6 months: CES D, BAI and the frequency of alcohol use. These items were asked of the most participants therefore decreasing the incidence of missing responses. Table 20 Correlation s of Change in Stigma Items and Change in Behavioral Health Variables B ehavioral Health Variables Change in Embarrassed/ Ashamed Item r ( n ) Change in Perceived Stigma r ( n ) Change in Comfort Level r ( n ) MINI Major Depression Scale .118* (735) .027 (651) .174* (434) CES D .179* (991) .001 (852) .159* (547) MINI Generaliz ed Anxiety Scale .081 (104) .067 (89) .042 (65) Beck Anxiety Inventory (BAI ) .057 (345) .035 (279) .057 (184) Number of alcoholic drinks in the last week .014 (555) .042 (472) .028 (333) Number of alcoholic binges in the last three months .027 (604) .049 (511) .017 (349) S MAST G .022 (333) .001 (325) .057 (227) Significant at p < .01 Embarrassed/Ashamed Item r egression m odel. The first model predicted the change in Embarrassed/Ashamed Item from baseline to 6 months using the change scores of the CES D, BA I and frequency of alcohol use (change from baseline to 6 months ).

PAGE 75

64 Simultaneous complete multiple regression was used, meaning that all independent variables were entered into the equation at once. Results indicate that multicollinearity was not a problem in this analysis as evidenced by the small strength correlations between the independent variables and the acceptable tolerance levels The regression model had an R 2 value of .0 33 meaning it accounted for only 3.3 % of the variance in the de pendent variable, and was not significant (Table 2 1 ) These results indicate that change s in the level s of depression, anxiety, and frequency of drinking variables were not significantly associated with changes in the Embarrassed/Ashamed Item Table 2 1 S ummary of Simultaneous Regression Analysis for Behavioral Health Variables Predicting Change in Embarrassed/Ashamed Item Variable B SE B p CES D .024 .014 .198 .094 BAI .004 .013 .039 .744 # of Alcoholic Drinks in the Last Week .000 .014 .003 .979 Perceived Stigma Component r egression m odel. The second model predicted the change in the Perceived Stigma Component from baseline to 6 months using the change scores of the CES D, BAI and frequency of alcohol use (change from baseline to 6 months ). Si multaneous complete multiple regression was used, meaning that all independent variables were entered into the equation at once. Results indicate that multicollinearity was not a problem in this analysis as evidenced by the small strength correlations betw een the independent variables and the acceptable tolerance levels The regression model had an R 2 value of 003 meaning it accounted for only 0 3 % of the variance in the dependent variable, and was not significant (see Table 2 2 ) These results indicate th at change s in the level s of depression, anxiety, and frequency of drinking variables were not significantly associated with changes in Perceived Stigma levels

PAGE 76

65 Table 2 2 Summary of Simultaneous Regression Analysis for Behavioral He a l th Variables Predictin g Change in the Perceived Stigma Component Variable B SE B p CES D .003 .013 .025 .833 BAI .004 .012 .041 .729 # of Alcoholic Drinks in the Last Week .005 .013 .039 .713 Comfort Level Component r egression m odel. The third model predicted the ch ange in the Comfort Level Component from baseline to 6 months using the change scores of the CES D, BA I and frequency of alcohol use (change from baseline to 6 months ). Simultaneous complete multiple regression was used, meaning that all independent varia bles were entered into the equation at once. Results indicate that multicollinearity was not a problem in this analysis as evidenced by the small strength correlations between the independent variables and the acceptable tolerance levels The regression mo del had an R 2 value of .0 28 meaning it accounted for only 2 8 % of the variance in the dependent variable, and was not significant (Table 2 3 ) These results indicate that change s in the level s of depression, anxiety, and frequency of drinking variables wer e not significantly associated with changes in stigma levels. Table 2 3 Summary of Simultaneous Regression Analysis for Behavioral Health Variables Predicting Change in the Comfort Level Component Variable B SE B p CES D .020 .014 .174 .141 BAI .002 .012 .020 .864 # of Alcoholic Drinks in the Last Week .006 .014 .045 .668 In order to further examine the relationship between feelings of stigma and change in behavioral health outcomes over time, a second s et of complete regression models were completed. The following variables were included in a simultaneous

PAGE 77

66 multiple regression analysis to examine the relationship between baseline level of stigma and change in behavioral health variables from baseline to 6 months: CES D, BAI and the frequency of alcohol use. These items were asked of the most participants therefore decreasing the incidence of missing responses. Because the Embarrassed/Ashamed Item contains items that overlap with the other two stigma items, it was not included in the models. Two stigma items, the baseline Perceived Stigma Component and the baseline Comfort Level Component, were used to predict the behavioral health items. Change in CES D The first model predicted the change in CES D scores from baseline to 6 months using the baseline scores of the Percei ved Stigma and Comfort Level Components. Simultaneous complete multiple regression was used, meaning that all independent variables were entered into the equation at once. Results indicate that multicollinearity was not a problem in this analysis as eviden ced by the small strength correlations between the independent variables and the acceptable tolerance levels The regression model had an R 2 value of .0 13 meaning it accounted for only 1.3 % of the variance in the dependent variable, and was statistically significant ( p = .002; see Table 2 4 ) These results indicate that lower Comfort Level scores at baseline (indicating less comfort) would predict a worsening of depressive symptoms over time as measured by the CES D. However, these results should be interpr eted with caution because the associations may not be strong enough to be considered meaningful Table 2 4 Summary of Simultaneous Regression Analysis for Variables P redicting Change in the CES D Variable B SE B p Baseline Perceived Stigma Component .819 .412 .065 .047 Baseline Comfort Level Component 1.227 .386 .104 .002

PAGE 78

67 Change in BAI The second model predicted the change in BAI scores from baseline to 6 months using the baseline scores of the Perceived Stigma and Comfort Leve l Components Simultaneous complete multiple regression was used, meaning that all independent variables were entered into the equation at once. Results indicate that multicollinearity was not a problem in this analysis as evidenced by the small strength c orrelations between the independent variables and the acceptable tolerance levels The regression model had an R 2 value of 00 0 meaning it accounted for none of the variance in the dependent variable, and was not significant (Table 2 5 ) These results indi cate that baseline level s of stigma were not significantly associated with changes in BAI scores. Table 2 5 Summary of Simultaneous Regression Analysis for Variables Predicting Change in the BAI Variable B SE B p Baseline Perceived Stigma Component .0 61 .804 .004 .940 Baseline Comfort Level Component .280 .753 .021 .710 Change s in n umber of d rinks The third model predicted the change in the number of alcoholic drinks in the past week from baseline to 6 months using the change scores of the Percei ved Stigma and Comfort Level Components Simultaneous complete multiple regression was used, meaning that all independent variables were entered into the equation at once. Results indicate that multicollinearity was not a problem in this analysis as eviden ced by the small strength correlations between the independent variables and the acceptable tolerance levels The regression model had an R 2 value of 001 meaning it accounted for only 0.1 % of the variance in the dependent variable, and was not significan t (Table 2 6 ) These results indicate that baseline level s of stigma were not significantly associated with changes in the number of alcoholic drinks consumed

PAGE 79

68 Table 2 6 Summary of Simultaneous Regression Analysis for Variables Predicting Change in the Nu mber of Drinks Variable B SE B p Baseline Perceived Stigma Component .412 .470 .037 .381 Baseline Comfort Level Component .006 .441 .001 .990 Research Question Four : Assuming level of stigma does change over time, what variables are related to chang e over time? As the analysis for Research Q uestion Three demonstrated, level of stigma as measured by Embarrassed/Ashamed Item and the Perceived Stigma Component did indeed improve over time for the total sample. To answer this research question complete simultaneous multiple regression was performed to predict change in level of stigma (Time 1 Time 3) by past behavioral health service use, number of PRISM E treatment sessions attended, satisfaction with PRISM E services, cultural competence of PRISM E s ervices, diagnosis, and the model of assignment. Also included is a variable to control for any influence by site (VA/Non VA). Simultaneous complete multiple regression was used for the first model, predicting change in the Embarrassed/Ashamed Item meani ng that all independent variables were entered into the equation at once. Results of the analysis show that multicollinearity was not problematic as evidenced by the small strength correlations between the independent variables. The regression model had an R 2 value of .0 2 4, meaning it accounted for only 2 .4% of the variance in the dependent variable, and was not significant (Table 2 7 ) These results indicate that past behavioral health service use, number of PRISM E treatment sessions attended, satisfaction with PRISM E services,

PAGE 80

69 cultural competence of PRISM E services, diagnosis, and the model of assignment were not significantly associated with changes in stigma levels. Simultaneous complete multiple regression was used in the second model, predicting change in the Perceived Stigma Component, meaning that all independent variables were entered into the equation at once. Results of the analysis show that multicolli nearity was not problematic as evidenced by the small strength correlations between the independent variables. The regression model had an R 2 value of .051 meaning it accounted for only 5.1 % of the variance in the dependent variable, and was not significa nt (see Table 2 8 ) These results indicate that past behavioral health service use, number of PRISM E treatment sessions attended, satisfaction with PRISM E Table 2 7 Summary of Simultaneous Regression Analysis for Variables Predicting Change in Embarrassed/Ashamed Item Variable B SE B p Past treatment for depression (0 = No, 1 = Yes) .255 .219 .084 .246 Outpatient MH visits in the past 3 months (0 = No, 1 = Yes) .053 .738 .005 .943 Outpatient SA visits in the past 3 months (0 = No, 1 = Yes) .690 1.642 .028 .675 Number of PRISM E sessions attended .004 .028 .011 .876 Satisfaction with PRISM E services .279 .157 .125 .076 Cultural competence of PRISM E services .040 .129 .022 .758 Depression diagnosis (0 = No, 1 = Yes) .537 1. 774 .158 .762 Anxiety diagnosis (0 = No, 1 = Yes) .680 1.667 .206 .684 At risk drinking diagnosis (0 = No, 1 = Yes) .529 1.758 .163 .764 Depression and At Risk Drinking diagnosis (0 = No, 1 = Yes) .474 1.683 .081 .778 Anxiety and At Risk Drinking diagnosis (0 = No, 1 = Yes) .195 .841 .022 .817 Depression and Anxiety diagnosis (0 = No, 1 = Yes) .540 1.656 .156 .745 Assigned Treatment Model (0 = Integrated, 1 = Referral) .008 .196 .003 .966 VA site (0 = Non VA Site, 1 = VA Site) .154 .212 .052 .467

PAGE 81

70 services, cultural competence of PRISM E services, diagnosis, and the model of assignment were not s ignificantly associated with changes in stigma levels. Regression analysis was not performed to examine change over time for the Comfort Level Component because previous analyses showed that it did not significantly change over time. Table 2 8 Summary o f Simultaneous Regression Analysis for Variables Predicting Change in Perceived Stigma Component Variable B SE B p Past treatment for depression (0 = No, 1 = Yes) .214 .195 .078 .272 Outpatient MH visits in the past 3 months (0 = No, 1 = Yes) .027 .656 .003 .967 Outpatient SA visits in the past 3 months (0 = No, 1 = Yes) .108 1.460 .005 .941 Number of PRI SM E sessions attended .001 .025 .002 .976 Satisfaction with PRISM E services .168 .139 .083 .230 Cultural competence of PRISM E services .229 .115 .140 .047 Depression diagnosis (0 = No, 1 = Yes) .896 1.578 .292 .570 Anxiety diagnosis (0 = No, 1 = Yes) .777 1.482 .262 .600 At risk drinking diagnosis (0 = No, 1 = Yes) .859 1.563 .293 .583 Depression and At Risk Drinking diagnosis (0 = No, 1 = Yes) .724 1.496 .138 .629 Anxiety and At Risk Drinking diagnosis (0 = No, 1 = Yes) .361 .748 .045 .630 Depression and Anxiety diagnosis (0 = No, 1 = Yes) .528 1.472 .169 .720 Assigned Treatment Model (0 = Integrated, 1 = Referral) .071 .175 .027 .686 VA site (0 = Non VA Site, 1 = VA Site) .144 .188 .054 .446

PAGE 82

71 Chapter Four Discussion T he primary goals of this study were to investigate the usability of the PRISM E Stigma Assessment instrument and to examine the relationship between feelings of stigma on behavioral health variables among a sample of older adults with depressive disorders, anxiety disorders, and at risk drinking behaviors. Secondary goals were to examine if feelings of stigma changed over time and if so, to identify which variables influenced that change. These goals were accomplished by conducting a secondary analysis of d ata collected through the Primary Care Research in Substance Abuse and Mental Health for the Elderly (PRISM E) study. Over half of the sample indicated that they would be embarrassed or ashamed if they had a mental health problem and over two thirds respo nded that they would be embarrassed if they had an alcohol abuse problem. These findings are disconcerting because the sample is comprised only of older adults with behavioral health disorders. In general, females were more likely than males to indicate th at they had feelings of stigma associated with mental health issues as were White Non Hispanics when compared to other racial/ethnic groups. Married respondents, as well as respondents from non VA sites were more likely to indicate having feelings of stigm a associated with behavioral health problems.

PAGE 83

72 Several different methods were used to determine the reliability of the Stigma Assessment Scale indicate that several problems exist with the instrument. An examination of the scaling of the measure indicates flaws with the response categories in that the categories are measuring frequency when the categories should be measuring amount and do not represent approximately equal intervals Further investigations using inter item and inter total correlations, SPSS reliability analysis, and an exploratory factor analysis revealed that the Stigma Assessment Scale is multidimensional, which means that it is assessing more than one construct of stigma. Therefore results indicate that the items of the scale should not b e aggregated to form one total score. The exploratory factor analysis revealed the presence of two factors: one measuring what we call Perceived Stigma and another measuring what we call Comfort Level, referring to the respondents' level of comfort in disc ussing behavioral health issues with a professional healthcare provider. It is unclear why the authors of the PRISM E study chose to develop and use this particular instrument when there are several other validated measures of perceived stigma available (K anter et al., 2008; Link et al., 1989; Vogel et al., 2006; Wahl, 1999). While the reliability of the instrument was questionable, these findings were used to create the three measures of stigma used in the remaining analyses: (a) Embarrassed/Ashamed Item, comprised of Item 1 or 2, depending on the diagnosis of the respondent, (b) Perceived Stigma, comprised of the Embarrassed/Ashamed Item, and Items 3, 4, and 7, and (c) the Comfort Level Component, comprised of Items 5 and 6.

PAGE 84

73 Research Question One : What demographic and behavioral health variables are associated with level of stigma at baseline? Results from a series of correlations pairing each of the three baseline stigma measures (Embarrassed/Ashamed, Perceived Stigma, and Comfort Level) and the behavi oral health measures indicate small correlations. Most notable, two measures of depressive symptoms (MINI Major Depression and the CES D) had two of the higher correlations found with the Embarrassed/Ashamed Item (.117 and .162, respectively). These findin gs indicate that respondents exhibiting depressive symptoms at baseline also had higher levels of feelings of embarrassment and shame about behavioral health disorders. In addition, scores on the CES D were correlated with the Perceived Stigma Component at r = .169, indicating that those respondents with higher levels of depressive symptoms also reported feeling higher levels of perceived stigma at baseline. Two measures of alcohol use (Number of alcoholic drinks consumed in the last week and S MAST G) had higher levels of correlations with the Comfort Level Component ( .129 and .120 respectively). T hese findings seemingly contradict each other, indicat ing that as the risk for alcohol misuse increases as measured by the S MAST G the respondents level of co mfort in talking with a professional health care provider increases as well However, as the number of alcoholic drinks consumed in the last week increases, the respondents level of comfort decreases. Although this finding could be interpreted to mean that as the number of alcoholic dr inks in the last week decreases ( indicating the respondent may have a mental health problem instead of an at risk alcohol use diagnosis ) their level of comfort increases

PAGE 85

74 A series of three regression models were constructed to further examine the relationship between baseline levels of stigma and baseline behavioral health variables. However, all three models yielded non significant results, indicating that age, race/ethnicity, gender, baseline levels of depression, anxiety, and frequency of drinking were not significantly associated with baseline stigma levels. R esearch Question Two : Does level of stigma change over the three time points? Does change vary according to engagement? The three stigma measures (Embarrassed/Asham ed, Perceived Stigma, and Comfort Level) were examined using repeated measures analysis of variance (ANOVA) techniques to determine if they changed significantly over time. Two of the three measures (Embarrassed/Ashamed and Perceived Stigma ) showed statist ically significant improvement over time. The third measure (Comfort Level) showed a trend toward improvement over time, but this trend was not statistically significant. The variables of engaged/not engaged (as defined as attending one or more treatment sessions) and type of service delivery model (integrated/enhanced referral model) were added to the analyses, resulting in a series of mixed design ANOVAs. There were no significant differences in the change in level of stigma over time between the engaged and not engaged groups, or the enhanced referral and integrated model groups. The findings indicate that while some facets of stigma significantly changed over time, those changes were not influenced by engagement or model type. To further examine the is sue of level of stigma at baseline and engagement, a Pearson correlation was conducted pairing the three baseline stigma measures and with the number of treatment sessions attended. Very small correlations (all below .008)

PAGE 86

75 indicate that there was no associ ation between baseline feelings of stigma and number of treatment sessions attended. Hypothesis One Previous literature has suggested that co locating primary care and behavioral health providers has been successful in improving patient treatment adherence satisfaction with care, and psychological symptoms (Bartels et al., 2004; Gallo et al., 2004; Katon et al., 1997). Other researchers have found that patients may be more comfortable seeing a behavioral health specialist located within their primary care physicians office who can work closely with their own physician (Speer & Schneider, 2003). However, this hypothesis was not supported. Results showed that feelings of stigma decreased for groups assigned to both models of treatment. Research Question Three : What is the relationship between change in stigma and change in behavioral health variables over time? In order to explore this question, change scores were calculated for the three stigma items (Embarrassed/Ashamed, Perceived Stigma, and Comfort Level ) and the behavioral health variables from Baseline to 6 months after baseline. Results from a series of correlations pairing the stigma item change scores with the behavioral health variables change scores indicate that most (with two exceptions) correlat ions were small according to Cohen's guidelines (1988). Of note, two measures of depressive symptoms MINI Major Depression Scale and CES D) had two higher correlations with both the Embarrassed/Ashamed Item, at .118 and .179 respectively. These findings in dicate that as feelings of depression improved, so did feelings as stigma as measured by the Embarrassed/ Ashamed Item. Interestingly, these same two measures of depression also

PAGE 87

76 had two of the higher correlations with Comfort Level at .174 and .159, respec tively. In this case, due to the directionality of the scaling, these correlations indicate that as depressive symptoms improve over time, feelings of comfortableness in speaking with a professional decrease. These results indicate that change s in the lev el s of depression, anxiety, and frequency of drinking variables were not significantly associated with changes in the Embarrassed/Ashamed Item A series of three regression models were conducted to further examine the relationship between change in levels of stigma and change in behavioral health variables over time. However, all three models yielded non significant results, indicating that change s in the level s of depression, anxiety, and frequency of drinking variables were not significantly associated wi th changes in stigma levels. A second series of three regression models were conducted to further examine the relationship between baseline levels of stigma and change in behavioral health variables over time. However, two of the three models yielded non s ignificant results, indicating that change s in the level s of depression, anxiety, and frequency of drinking variables were not signif icantly associated with baseline stigma levels. The model predicting change in CES D scores over time by baseline stigma le vel s was statistically significant; however the model only accounted for 1.3% of the variance. Results indicate that baseline level of stigma, particularly Comfort Level, can predict a small amount of the variance associated with depressive symptoms. As ba seline comfortableness decreases, depressive symptoms worsen over time.

PAGE 88

77 Hypothesis Two It was predicted that as feelings of stigma improve over time, behavioral health outcomes would also improve over time. Although the current research study can not de termine causality, it was believed that those variables would have a positive relationship. However, the results did not support this hypothesis. There was no relationship found between change in stigma levels over time and change in behavioral health vari ables over time. Research Question Four : Assuming level of stigma does change over time, what variables are related to change over time? As was demonstrated in Research Question Three, level of stigma did indeed change over time in this sample. Regression models determined that past behavioral health service use, number of PRISM E treatment sessions attended, respondent s atisfaction with PRISM E services, respondent rating of the cultural competence of PRISM E services, diagnosis, and the model of assignme nt were not significantly associated with changes in stigma levels. Hypothesis Three Based on previous literature (Rusch et al. 2005; Sirey et al., 2001), it was predicted that respondents with lower levels of feelings of stigma at baseline would demons trate greater improvements in behavioral health outcomes over time. Some support was found for this hypothesis. A significant relationship was found between baseline level of Comfort Level and depressive symptoms as measured by the CES D. Higher levels of comfortableness at baseline predicted improvements in depressive

PAGE 89

78 symptoms. However, the remaining two analyses examining this relationship were not significant Conclusions Research has demonstrated that certain attitudinal characteristics have an impact o n a person's responsiveness to behavioral health treatment. Along with receiving q uality treatment, remaining in treatment and adherence to treatment regimens are important factor s in improving behavioral health problems. Those who have a negative attitude toward mental health services are more likely to drop out of treatment than those with a more posi t ive attitude ( Bambauer & Prigerson, 2006; Edlund et al., 2002). Conversely, those patients who report being satisfied with their mental health care are more likely to attend more treatment sessions than those who are less satisfied (Chen et al., 2006 ; Komiti et al. 2006 ). One important barrier to the receipt of needed behavioral health treatment is perceived stigma. Perceived stigma has been linked with tre atment discontinuation in ol der adults (Sirey et al., 2001), as well as lower levels of quality of life among older adults (Depla, de Graaf, van Weeghel, & Heeren, 2005). This dissertation sought to explore the relationships between attitudinal characteris tics such as perceived stigma and behavioral health outcomes. Most research in this field has focused on perceived stigma as a barrier to accessing services or as a barrier to treatment adherence Few studies have examined the impact of perceived stigma on behavioral health outcomes. Although the effects of stigma are far reaching, the majority of the results from this study do not support a link between feelings of stigma and behavioral health outcomes. Ther e was limited support

PAGE 90

79 for a link between feeling comfortable discussing problems with a provider and improvement in depressive symptoms ; however, the relationship may not be strong enough to be considered meaningful. Explorations of associations between perceived stigma and other attitudinal variables, such as satisfaction with care and cultural competence ratings, did not reveal any meaningful relationships. In addition, no meaningful relationships were found between feelings of stigma and demographic characteristics, such as age, race/ethnicity, gender and military status (VA/Non VA). Feelings of stigma did change over time, indicating that it is a transient state of being versus a permanent character trait. Few research studies have examined the persistence of perceived stigma in a person over time. The research in this dissertation illustrates that it is possible for a person's level of perceived stigma to change ove r a relatively short time frame. Interventions aimed at reducing a persons' feelings of perceived stigma may be successful in motivating people to seek help and engage in treatment, therefore reducing the costs associated with untreated behavioral health problems and improving the quality of life of many older adults. Limitations One main consideration for the lack of significant results found in this secondary analysis concerns the measurement item. Examinations of the Stigma Assessment Scale revealed several problems related to its structure, questioning its reliability and the identification of two factors, indicating that the instrumen t was measuring at least two different aspects of stigma (Perceived Stigma and Comfort Level). The inclusion of a validated measure of stigma may have potentially been more sensitive in finding

PAGE 91

80 associations between feelings of stigma and behavioral health, attitudinal, and demographic characteristics. However, this illustrates the major problem with conducting secondary data analysis: the research design (including the measures) does not always dovetail precisely with the research questions of the second ary study. In addition, the measurement of perceived stigma is a complex task. As pointed out by Link and colleagues (2004), the various conceptualizations of stigma ha ve led to a diverse lot of instruments from which researchers can choose. As with other at titudinal measures, it can be difficult to establish the validity of such instruments. Other limitations for the current research are related to the design of the PRISM E study as it relates to the current analyses The goal of the original study was to examine two types of service delivery methods. In general, no restrictions were placed on the types of treatments offered to participants. While data were collected regarding the general type of treatment (e.g., psychotherapy, medication management), who p rovide d that therapy (e.g., psychiatrist, psychologist), and the number of treatment sessions attended, the details about treatment and fidelity to a treatment model were not measure d Therefore the quality of the behavioral health treatments provided is u nknown. In addition, the current study focused on the 2,022 respondents who agreed to receive behavioral health treatment and participate in the PRISM E study. This group of participants may differ in a number of ways from those individuals who declined t o participate in the study. Namely, those who declined may have greater feelings of stigma which have a greater influence over their behavioral health treatment choices than those who agreed to participate in the study. If this occurred, the sample would b e biased toward those whose feelings of stigma have less of an impact on their behavioral health

PAGE 92

81 care utilization. Stigma may be more of a barrier to treatment for those who declined to participate in the study than for those who a greed to participate. Fur ther, respondents' history of behavioral health treatment was not captured beyond the three months prior to the onset of the study. Previous experience with the behavioral health system can influence feelings of perceived stigma as well as willingness to e ngage in treatment and response to treatment Future Directions Perceived stigma has been in the crosshairs of recent reports from the Surgeon General (1999), the World Health Organization and the World Psychiatry Organization (2003), and SAMHSA (2005). T hese reports, as well as the plethora of research articles linking perceived stigma as a barrier to receiving behavioral health treatment, all highlight the complexity and far reaching effects of negative stereotypes about people with behavioral health pro blems. There is hope for the future. With the spotlight trained on the detrimental effects of the stigmatization of behavioral illnesses, one can hope that these issue s become a matter of national importance. The development and implementation of anti sti gma campaigns including such strategies as protest, education, and contact ( Rusch et al., 2005) will go a long way in reducing one of the barriers to receiving behavioral health care. WHO and WPA's report on reducing stigma aimed specifically at older adu lts with mental illnesses provides recommendations for the general public, the media, the corporate sector, and academia (Graham et al., 2003). SAMHSA provides strategies to overcome the barriers of stigma as well as the internet based Resource Center to A ddress Discrimination and Stigma Associated with Mental Illness

PAGE 93

82 ( www.stopstigma.samhsa. gov). The Resource Center includes the guide "Developing a Stigma Reduction Initiative" (USDHHS, 2006). Interventions designed to lessen the feelings of perceived stigm a in an individual with behavioral health challenges will play an important part in ensuring older adults get treatment they need. Improving self esteem and self efficacy has also been linked to treatment access and adherence (Link et al., 2002) The co lo cation of behavioral health services with physical health services has been linked in the literature to lessening stigma, improving communication, and improving symptoms (Katon et al., 1997) It is important for researchers to focus on these area s as thes e initiatives may help an underserved population receive needed services. Future research endeavors to understand the complex relationship between feelings of stigma and behavioral health outcomes will add to the research base and help to reduce the impac t of perceived stigma as a barrier to seeking and receiving behavioral health treatment. By raising awareness and educating the general public as well as policymakers, it is possible to break through the barrier of stigma and ensure that all older adults r eceive the care that they need.

PAGE 94

83 List of References American Psychiatric Association (APA) (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Arcury, T. A., Gesler, W. M., Preisser, J. S., Sherman, J., Spencer, J., & Perin, J. (2005). The effects of geography and spatial behavior on health care utilization among the residents of a rural region. Health Services Research, 40 (1), 135 156. Arean, P. A., Alvidrez, J., Barrera, A., Robinson, G. S., & Hicks, S (2002). Would older medical patients use psychological services? The Gerontologist, 42 392 398. Atkinson, R. M., Ganzini, L., & Bernstein, M. J. (1992). Alcohol and substance use disorders in the elderly. In J. E. Birren, R. B. Sloane, & G. D. Cohen (E ds.), Handbook of mental health and aging (2 nd ed.). San Diego: Academic Press. Balsa, A. I., Homer, J. F., Fleming, M. F., & French, M. T. (2008). Alcohol consumption and health among elders. The Gerontologist, 48 622 636. Bambauer, K. Z., & Prigerson, H. G. (2006). The Stigma Receptivity Scale and its association with mental health service use among bereaved older adults. Journal of Nervous and Mental Disease, 194 (2), 139 141. Barney, L J ., Griffiths, K M ., Jorm, A F ., & Christensen, H (2006). Sti gma about depression and its impact on help seeking intentions. Australian and New Zealand Journal of Psychiatry 40 (1) 51 54. Barry, K. L., Oslin, D., W., & Blow, F. C. (2001). Prevention and management of alcohol problems in older adults New York: Spr inger. Bartels, S. J. (2002). Quality, costs, and effectiveness of services for older adults with mental disorders: A selective overview of recent advances in geriatric mental health services research. Current Opinion in Psychiatry, 15 411 416. Bartels, S. J. (2003). Improving the system of care for older adults with mental illness in the United States. American Journal of Geriatric Psychiatry, 11 486 497.

PAGE 95

84 Bartels, S. J., Coakley, E. H., Zubritsky, C., Ware, J. H., Miles, K. M., Arean, P. A., et al. ( 2004). Improving access to geriatric mental health services: A randomized trial comparing treatment engagement with integrated versus enhanced referral care for depression, anxiety, and at risk alcohol use. American Journal of Psychiatry, 161 (8), 1455 1462 Bartlett, M. S. (1954). A note on the multiplying factors for various chi square approximations. Journal of the Royal Statistical Society, 16 (Series B), 296 298. Bass, B. M., Cascio, W. F., & O'Connor, E. J. (1974). Magnitude estimations of expressions of frequency and amount. Journal of Applied Psychology, 59 313 320. Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56 893 897. Bec ker, M., Stiles, P., & Schonfeld, L. (2002). Mental health service use and cost of care for older adults in assisted living facilities: Implications for public policy. Journal of Behavioral Health Services & Research, 29 (1), 91 8. Bland, R. C., Newman, S. C., & Orn, H. (1988). Prevalence of psychiatric disorders in the elderly in Edmonton. Acta Psychiatrica Scandinavica, 338 (Suppl.), 57 63. Blow, F. C., Gillespie, B.W., Barry, K. L., Mudd, S. A., & Hill, E. M. (1998). Brief screening for alcohol problems in elderly populations using the Short Michigan Alcohol Screening Test Geriatric Version (SMAST G). Alcoholism: Clinical and Experimental Research, 22 (Suppl), 131A. Borson, S., Bartels, S. J., Colenda, C. C., Gottlieb, G. L., & Meyers, B. (2001). Geriat ric mental health services research: Strategic plan for an aging population. American Journal of Geriatric Psychiatry, 9 (3), 191 204. Breslow, R. A., Faden, R. A., & Smothers, B. (2003). Alcohol consumption by elderly Americans. Journal of Studies on Alco hol, 64 884 892. Burrows, A. B., Satlin, A., Salzman, C., Nobel, K., & Lipsitz, L. A. (1995). Depression in a long term care facility: Clinical features and discordance between nursing assessment and patient interviews. Journal of the American Geriatrics Society, 43 1118 1122. Carmin, C. N., Pollard, C.A., & Gillock, K. L. (1999). Assessment of anxiety disorders in the elderly. In P. Lichtenberg (Ed.), Handbook of assessment in clinical gerontology (pp. 59 90). New York: John Wiley & Sons.

PAGE 96

85 Chachamovich E., Fleck, M., Laidlaw, K., & Power, M. (2008). Impact of major depression and subsyndromal symptoms on quality of life and attitudes toward aging in an international sample of older adults. The Gerontologist, 48 593 602. Charney, D., Reynolds, C. F., Lewis, L., Lebowitz, B. D., Sunderland, T., Alexopoulos, G. S., et al. (2003). Depression and bipolar support alliance consensus statement on the unmet needs in diagnosis and treatment of mood disorders in late life. Archives of General Psychiatry, 60 (7), 664 672. Chen, H., Coakley, E. H., Cheal, K., Maxwell, J., Costantino, G., Krahn, D. D., et al. (2006). Satisfaction with mental health services in older primary care patients. American Journal of Geriatric Psychiatry, 14 (4), 371 379. Chopra, M. P., Zubr itsky, C., Knott, K., Ten Have, T., Hadley, T., Coyne, J. C., et al. (2005). Importance of subsyndromal symptoms of depression in elderly patients. American Journal of Geriatric Psychiatry, 13 597 606. Cohen, J. W. (1988). Statistical power analysis of t he behavioral sciences (2 nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. Cole, M. G. (2002). Public health models of mental health care for elderly populations. International Psychogeriatrics, 14 (1), 3 6. Conwell, Y., Rotenberg, M., & Caine, E. D. ( 1990). Completed suicides at age 50 and over. Journal of the American Geriatric Society, 38 640 644. Cooper, A E ., Corrigan, P W ., & Watson, A C. (2003). Mental illness stigma and care seeking. Journal of Nervous and Mental Disease 191 (5), 339 341. Corrigan, P. (2004). How stigma interferes with mental health care. American Psychologist, 59 (7), 614 625. Crabb, R., & Hunsley, J. (2006). Utilization of mental health care services among older adults with depression. Journal of Clinical Psychology, 62 299 312. Cramer, J. A., & Rosenbeck, R. (1998). Compliance with medication regimens for psychiatric and medical disorders. Psychiatric Services, 49 196 210. Depla, M., de Graaf, R., van Weeghel, J., & Heeren, T. J. (2005). The role of stigma in the qual ity of life of older adults with severe mental illness. International Journal of Geriatric Psychiatry, 20 146 153. Dupree, L. W., Watson, M., & Schneider, M. G. (2005). Preferences for mental healthcare: A comparison of older African Americans and older Caucasians. Journal of Applied Gerontology, 24 (3), 196 210.

PAGE 97

86 Edlund, M. J., Wang, P. S., Berglund, P. A., Katz, S. J., Lin, E., & Kessler, R. C. (2002). Dropping out of mental health treatment: Patterns and predictors among epidemiological survey respondent s in the United States and Ontario. American Journal of Psychiatry, 159 845 851. Ferrante, S. (March, 2007). De stigmatizing older adult mental health and substance abuse care. Paper presented at the annual conference of the Florida Coalition on Aging, J acksonville, FL. Fitzpatrick, A. L., Powe, N. R., Cooper, L. S., Ives, D. G., & Robbins, J. A. (2004). Barriers to health care access among the elderly and who perceives them. American Journal of Public Health, 94 1788 1794. Flint, A. J., (1994). Epidem iology and comorbidity of anxiety disorders in the elderly. American Journal of Psychiatry, 151 640 649. Forsell, Y., & Winblad, B. (1998). Incidence of major depression in a very elderly population. International Journal of Geriatric Psychiatry, 14 368 372. Fortney, J., Mukherjee, S., Curran, G., Fortney, S., Han, X., & Booth, B. M. (2004). Factors associated with perceived stigma for alcohol use and treatment among at risk drinkers. Journal of Behavioral Health Services & Research, 31 418 429. Gallo J. J., Rabins, P. V., & Illife, S. (1997). The "research magnificent" in later life: Psychiatric epidemiology and the primary care of older adults. International Journal of Psychiatry in Medicine, 27 185 204. Gallo, J. J., Zubritsky, C., Maxwell, J., N azar, M., Bogner, H. R., Quijano, L. M., et al. (2004). Primary care clinicians evaluate integrated and referral models of behavioral health care for older adults: Results from a multisite effectiveness trial (PRISM E). Annals of Family Medicine, 2 (4), 305 309. Goldman, L. S., Nielsen, N. H., & Champion, H. C. (1999). Awareness, diagnosis, and treatment of depression. Journal of General Internal Medicine, 14 569 580. Graham, N., Lindesay, J., Katona, C., Bertolote, J. M., Camus, V., Copeland, J. R. M., e t al. (2003). Reducing stigma and discrimination against older people with mental disorders: A technical consensus statement. International Journal of Geriatric Psychiatry, 18 670 678. Gurland, B. J., Cross, P. S., & Katz, S. (1996). Epidemiological pers pectives on opportunities for treatment of depression. American Journal of Geriatric Psychiatry, 4 (Suppl. 1), S7 S13.

PAGE 98

87 Harris, K. M., Edlund, M. J., & Larson, S. (2005).Racial differences in the mental health problems and use of mental health care. Medical Care, 43 775 784. Herr, B. M., & Zubritsky, C. (2005, December). PRISM E: Mental health services and primary health care for older adults. Paper presented at the annual conference of the American Public Health Association, Philadelphia, PA. Holroyd, S. & Duryee, J. J. (1997). Substance use disorders in a geriatric psychiatry outpatient clinic: Prevalence and epidemiologic characteristics. Journal of Nervous and Mental Disease, 185 (10), 627 632. Hybels, C. F., & Blazer, D. G. (2004). Epidemiology of me ntal disorders in older adults. In B. L. Levin, J. Petrila, & K. D. Hennessy (Eds.), Mental health services: A public health perspective (pp. 211 231). New York: Oxford. Kaiser, H. (1970). A second generation Little Jiffy. Psychometrika, 35 401 415. Kai ser, H. (1974). An index of factorial simplicity. Psychometrika, 39 31 36. Kaplan, M. S., Adamek, M. E., Calderon, A. (1999). Managing depressed and suicidal geriatric patients: Differences among primary care physicians. Gerontologist, 39 417 425. Kant er, J. W., Rusch, L. C., & Brondino, M. J. (2008). Depression self stigma: A new measure and preliminary findings. The Journal of Nervous and Mental Disease, 196 663 670. Katon, W Lin, E., Russo, J., & Unutzer, J. (2003). Increased medical costs of a population based sample of depressed elderly patients. Archives of General Psychiatry, 60 897 903. Katon, W., Von Korff, M., Lin, E., Bush, T., & Ormel, J. (1997). Population based care of depression: Effective disease management strategies to decrease p revalence. General Hospital Psychiatry, 19, 169 178. Karlin, B. E., & Duffy, M. (2004). Geriatric mental health policy: Impact on service delivery and directions for effecting change. Professional Psychology: Research and Practice, 35 (5), 509 519. Klap, R., Unroe, K. T., & Unutzer, J. (2003). Caring for mental illness in the United States: A focus on older adults. American Journal of Geriatric Psychiatry, 11 517 524. Kline, P. (1994). An easy guide to factor analysis New York: Routledge.

PAGE 99

88 Komiti, A., J udd, F., & Jackson, H. (2006). The influence of stigma and attitudes on seeking help from a GP for mental health problems. Social Psychiatry and Psychiatric Epidemiology, 41 738 745. Larson, D. B., Lyons, J. S., Hohmann, A. A., Beardsley, R. S. ( 1991). P sychotropics prescribed to the U.S. elderly in the early and mid 1980s: Prescribing patterns of primary care practitioners, psychiatrists, and other physicians. International Journal of Geriatric Psychiatry, 6 63 70. Leaf, P. J., Bruce, M. L., Tischler, G. L., & Holzer, C. E. (1987). The relationship between demographic factors and attitudes toward mental health services. Journal of Community Psychology, 15 275 284. Lebowitz, B. D., Pearson, J. L., Schneider, L. S., Reynolds, C. F., Alexpoulos, G. S., B ruce, M. L., et al. (1997). Diagnosis and treatment of depression in late life: Consensus statement update. Journal of the American Medical Association, 278 1186 1190. Levkoff, S. E., Chen, H., Coakley, E., McDonel Herr, E. C., Oslin, D. W., Katz, I., et al. (2004). Design and sample characteristics of the PRISM E multisite randomized trial to improve behavioral health care for the elderly. Journal of Aging and Health 16 3 27. Liberto, J. G., Oslin, D. W., & Ruskin, P. E. (1992). Alcoholism in older per sons: A review of the literature. Hospital & Community Psychology, 43 975 984. Link, B. G., Cullen, F. T., Struening, E., Shrout, P. E., & Dohrenwend, B. P. (1989). A modified labeling theory approach to mental disorders: An empirical assessment. America n Sociological Review, 54 400 423. Link, B. G., Struening, E., Cullen, F. T., Shrout, P. E., & Dohrenwend, B. P. (1989). A modified labeling theory approach to mental disorders: An empirical assessment. American Sociological Review, 54 400 423. Link, B G., Struening, E. L., Neese Todd, S., Asmussen, S., & Phelan, J. C. (2001). The consequences of stigma for the self esteem of people with mental illnesses. Psychiatric Services, 52 1621 1626. Link, B. G., Struening, E. L., Neese Todd, S., Asmussen, S., & Phelan, J. C. (2002). On describing and seeking to change the experience of stigma. Psychiatric Rehabilitation Skills, 6 (2), 201 231. Link, B. G., Yang, L. H., Phelan, J. C., & Collins, P. Y. (2004). Measuring mental illness stigma. Schizophrenia Bulle tin, 30 (3), 511 541.

PAGE 100

89 Luber, M. P., Meyers, B. S., Williams Russo, P. G., Hollenberg, J. P., DiDomenico, T. N., Charlson, M. E., et al. (2001). Depression and service utilization in elderly primary care patients. American Journal of Geriatric Psychiatry, 9 169 176. Mickus, M., & Colenda, C. (2000). Knowledge of mental health benefits and provider preference. Psychiatric Services, 51 199 203. Nunnally, J. O. (1978). Psychometric theory New York: McGraw Hill. Pallant, J. (2001). SPSS survival manual. Ph iladelphia: Open University Press. Patterson, T. L., & Jeste, D. V. (1999). The potential impact of the baby boom generation on substance abuse among elderly persons. Psychiatric Services, 50 (9), 1184 1188. President's New Freedom Commission on Mental He alth (2003). Achieving the Promise: Transforming Mental Health Care in America. Final Report ( DHHS Pub lication No. SMA 03 3832 ) Rockville, MD : U.S. Department of Health and Human Services Radloff, L. S. (1977). The CES D scale: A self report depression scale for research in the general population. Applied Psychological Measurement, 1 385 401. Regier, D. A., Boyd, J. H., Burke, J. D., Rae, D. S., Myers, J. K., Kramer, M., et al. (1988). One month prevalence of mental disorders in the United States. Arc hives of General Psychiatry, 45 977 986. Reynolds, C. F., & Kupfer, D. J. (1999). Depression and aging: A look to the future. Psychiatric Services, 50 1167 1172. Robb, C., Haley, W. E., Becker, M. A., Polivka, L. A., & Chwa, H. J. (2003). Attitudes tow ards mental health care in younger and older adults: Similarities and differences. Aging & Mental Health, 7 (2), 142 152. Rogers, A., & Barusch, A. (2000). Mental health service utilization among frail, low income elders: Perceptions of home service provi ders and elders in the community. Journal of Gerontological Social Work, 34 (2), 23 28. Rusch, N., Angermeyer, M. C., & Corrigan, P. W. (2005), Mental illness stigma: Concepts, consequences, and initiatives to reduce stigma. European Psychiatry, 20 529 53 9.

PAGE 101

90 Schonfeld, L., Dupree, L. W., Dickson Fuhrmann, E., Royer, C. M., McDermott, C. H., Rosansky, J. S., et al. (2000). Cognitive behavioral treatment of older veterans with substance abuse problems. Journal of Geriatric Psychiatry & Neurology, 13 124 129 Segal, D. L., Coolidge, F. L., & Hersen, M. (1998). Psychological testing of older people. In I. H. Nordhus, G. R. VandenBos, S. Berg, & P. Fromholt (Eds.), Clinical geropsychology (pp. 231 257). Washington DC: American Psychological Association. Shea D. G., Russo, P. A., & Smyer, M. A. (2000). Use of mental health services by persons with a mental illness in nursing facilities: Initial impacts of OBRA87. Journal of Aging and Health, 12 560 578. Sheehan, D. Lecrubier, Y., Sheehan, K., Amorim, P., Ja navs, J., Weiller, E., et al. (1998). The Mini International Neuropsychiatric Interview (M.I.N.I): The development and validation of a structured diagnostic psychiatric interview for DSM IV and ICD 10. Journal of Clinical Psychiatry, 59 (Suppl 20), 22 33, q uiz 34 57. Sirey, J. A., Bruce, M. L., Alexopoulos, G. S., Perlick, D. A., Raue, P., Friedman, S. J., et al. (2001). Perceived stigma as a predictor of treatment discontinuation in young and older outpatients with depression. American Journal of Psychiatr y, 158 479 481. Smyer, M. A., & Qualls, S. H. (2004). (Eds.). Aging and mental health. Malden, MA: Blackwell. Speer, D. C., & Schneider, M. G. (2003). Mental health needs of older adults and primary care: Opportunity for interdisciplinary geriatric tea m practice. Clinical Psychology: Science and Practice, 10 (1), 85 101. Stuck, A. E., Beers, M. H., Steiner, A., Aronow, H. U., Rubenstein, L. Z., & Beck, J. C. (1994). Inappropriate medication use in community residing older persons. Archives of Internal M edicine, 154 2195 2200. Swartz, M. S., Wagner, H. R., Swanson, J. W., Burns, B. J., George, L. K., & Padgett, D. K. (1998). Administrative update: Utilization of services. I. Comparing use of public and private mental health services: the enduring barri ers of race and age. Community Mental Health Journal, 34 (2), 133 44. Tabachnick, B. G., & Fidell, L. S. (1996). Using multivariate statistics (3 rd ed.). Mahweh, NJ: Lawrence Erlbaum.

PAGE 102

91 Tourigny, A., Durand, P. J., Bonin, L., Hebert, R., & Rochette, L. (200 4). Quasi experimental study of the effectiveness of an integrated service delivery network for the frail elderly. Canadian Journal on Aging, 23 (3), 231 246. Unutzer, J., Katon, W., Williams, J. W., Callahan, C. M., Harpole, L., Hunkeler, E. M., et al. (2 001). Improving primary care for depression in late life: The design of a multicenter randomized trial. Medical Care, 39 785 799. Unutzer, J., Patrick, D. L., Simon, G., Grembowski, D., Walker, E., Rutter, C., et al. (1997). Depressive symptoms and the c ost of health services in HMO patients aged 65 years and older: A 4 year prospective study. Journal of the American Medical Association, 27 1618 1623. U.S. Department of Health and Human Services ( USDHHS ). (1999). Mental h ealth: A r eport of the Surgeon G eneral. Rockville, MD: Author, Substance Abuse and Mental Health Services Administration, Center for Mental Health Services, National Institutes of Health, National Institute of Mental Health. U.S. Department of Health and Human Services ( USDHHS ). (2005). Mentally healthy aging: A report on overcoming stigma for older Americans. Rockville, MD: Author, Substance Abuse and Mental Health Services Administration, Center for Mental Health Services. U.S. Department of Health and Human Services ( USDHHS ). (2006). Developing a s tigma r eduction initiative ( SAMHSA Pub. No. SMA 4176). Rockville, MD: Center for Mental Health Services, Substance Abuse and Mental Health Services Administration. Vogel, D. L., Wade, N. G., Haake, S. (2006). Measuring the self stigma assoc iated with seeking psychological help. Journal of Counseling, 53 325 337. Wagenaar, D. B., Mickus, M. A., Gaumer, K. A., & Colenda, C. C. (2002). Late life depression and mental health services in primary care. Journal of Geriatric Psychiatry & Neurology 15 134 140. Wahl, O. F. (1999). Mental health consumers' experience of stigma. Schizophrenia Bulletin, 25 467 478. Wang, P. S., Simon, G., & Kessler, R. C. (2003). The economic burden of depression and the cost effectiveness of treatment. Internatio nal Journal of Methods in Psychiatric Research, 12 (1), 22 33. Westreich, L. M. (2005). Alcohol and mental illness. Primary Psychiatry, 12 (1), 41 46.

PAGE 103

92 Williams, J., Shore, S. E., & Foy, J. M. (2006). Co location of mental health professionals in primary ca re settings: Three North Carolina models. Clinical Pediatrics, 45 537 543.

PAGE 104

About the Author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aurence G. Branch Doctoral Student Research Award. &