1 BARRIERS TO MENTAL HEAL T H AND SUBSTANCE ABUSE SERVICE UTILIZATION AMONG HOMELESS ADULTS By MICHAEL D. BRUBAKER A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQ UIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009
2 2009 Michael D. Brubaker
3 To all of the women, men, and childre n who have lived without a home
4 ACKNOWLEDGMENTS I am truly g rateful to Dr. Michael Garrett, my c hair and advisor for offering me a creative space to develop and grow as a researcher and human being. I am also thankful to each of my committee members: Dr. Ellen Amatea, who encouraged me to critically engage the ex isting body of research ; Dr. Edil Torres Rivera, who introduced me to the ways of liberation psychology; and Dr. David Miller, who showed me the power of statistics and program evaluations as a way to serve those without permanent housing. I am thankful for the support of my family and friends, especially my parents, Dale and Barbara Brubaker. With good humor and constant encouragement, they have seen me through many challenges in both academics and life As a supervisor and friend, Dr. James Bass inspire d me to consider a career serving th ose without housing I am truly thankful for Dr. Brian Dew and Dr. West Olatunji for showing me how to make research meaningful for me and my community. I am incredible thankful for my two co -investigators on this resea rch, Niyama Ramlall and TaJuana Chisholm. These two women showed endless courage as they walked into the unknown and patience as they put up with me. Finally, I am most grateful for the support of the participants in this study, the women and men who are striving for a better life as they persevere without a home. Each having a different story, they welcomed us into their lives and trusted us to make a difference.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 8 LIST OF FIGURES ............................................................................................................................ 10 LIST OF TERMS ............................................................................................................................... 11 ABSTRACT ........................................................................................................................................ 13 CHAPTER 1 INTRODUCTION ....................................................................................................................... 15 Theoretical Framework ............................................................................................................... 20 An Emancipatory Communitarian Perspective .................................................................. 20 The Behavioral Model for Vulnerable Populations ........................................................... 24 Need for the Study ...................................................................................................................... 26 Purpose of the Study ................................................................................................................... 28 Research Questions ..................................................................................................................... 28 2 LITERATURE REVIEW ........................................................................................................... 30 Who are Those Who Become Homeless? ................................................................................. 32 Defining Homeless .............................................................................................................. 32 Demographics ...................................................................................................................... 33 The Chronic Homeless ........................................................................................................ 35 The Road to Homelessness ......................................................................................................... 38 Mental Health and Addictions ............................................................................................ 39 Interpersonal Violence and Trauma ................................................................................... 40 Natural and Human Inflicte d Disasters .............................................................................. 41 Factors Impacting Americas Youth and the Elderly ........................................................ 42 Specific Societal Factors ..................................................................................................... 42 Exiting Homelessness ................................................................................................................. 44 Barriers and Enablers to Substance Abuse and Mental Health Service Utilization ................ 46 Predisposing Factors ............................................................................................................ 48 Enabling Factors .................................................................................................................. 49 Need Factors ........................................................................................................................ 52 Applying the Behavioral Model for Vulnerable Populations to Homeless Individuals Seeking Mental Health and Substance Abuse Care .............................................................. 53 An Emancipatory Communitarianism Perspective on Mental Health and Substance Abuse Utilization ..................................................................................................................... 61 Summary ...................................................................................................................................... 63
6 3 METHODOLOGY ...................................................................................................................... 67 Research Hypotheses .................................................................................................................. 68 Research Design and Relevant Variables .................................................................................. 68 Population and Sample ............................................................................................................... 70 Procedure ..................................................................................................................................... 71 Instrumentation ............................................................................................................................ 73 Barriers to Treatment Instrument ........................................................................................ 73 Stages of Change Readiness and Treatment Eagerness Scale .......................................... 74 Data Analyses .............................................................................................................................. 76 4 RESULTS .................................................................................................................................... 78 Sample Description ..................................................................................................................... 78 Outcomes of Instrumentation ..................................................................................................... 81 Other Barrie rs Encountered ........................................................................................................ 83 Societal Barriers ................................................................................................................... 84 Programmatic Barriers ......................................................................................................... 84 Individual Barriers ............................................................................................................... 85 Summary of Other Identified Barriers ................................................................................ 86 Hypotheses Tests Results ........................................................................................................... 86 Hypothesis 1 Test Results ................................................................................................... 86 Hypothesis 2 Test Results ................................................................................................... 88 Hypothesis 3 Test Results ................................................................................................... 89 Hypothesis 4 Test Results ................................................................................................... 90 Hypothesis 5 Test Results ................................................................................................... 91 Post -Hoc Analyses ...................................................................................................................... 93 Hypothesis 6 Test Results ................................................................................................... 94 Hypothesis 7 Test Results ................................................................................................... 95 5 DIS CUSSION ............................................................................................................................ 107 Overview of the Study .............................................................................................................. 107 Discussion of Descriptive Data ................................................................................................ 107 Discussion of Barriers ............................................................................................................... 110 Discussion of Instrumentation .................................................................................................. 112 Discussion of Hypotheses ......................................................................................................... 113 Discussion of Hypothesis 1 ............................................................................................... 114 Discussion of Hypothesis 2 ............................................................................................... 116 Discussion of Hypothesis 3 ............................................................................................... 117 Discussion of Hypothesis 4 ............................................................................................... 118 Discussion of Hypothesis 5 ............................................................................................... 120 Di scussion of Post Hoc Analyses ............................................................................................. 123 Discussion of Hypothesis 6 ............................................................................................... 123 Discussion of Hypothesis 7 ............................................................................................... 124
7 Clinical Implications ................................................................................................................. 126 Community Implications .......................................................................................................... 129 Counselor Training Implications .............................................................................................. 131 Theoretical Implications ........................................................................................................... 133 Limitations ................................................................................................................................. 135 Study Design Limitations .................................................................................................. 135 Measurement Limitations .................................................................................................. 136 Recommendations for Future Studies ...................................................................................... 137 Conclusion ................................................................................................................................. 140 APPENDIX A Barriers to Treatment Instrument ............................................................................................. 142 B General Questionnaire .............................................................................................................. 146 C Personal Drug and Alcohol Use Questionnaire ....................................................................... 149 D Open -Ended Responses to Other Barriers Experienced ......................................................... 150 E Informed Consent ...................................................................................................................... 153 F Recruitment Script .................................................................................................................... 154 G Solicitation for In -Kind Donations ........................................................................................... 155 LIST OF REFERENCES ................................................................................................................. 156 BIOGRAPHICAL SKETCH ........................................................................................................... 166
8 LIST OF TABLES Table page 4 1 Demographics: gender and ethnicity..................................................................................... 98 4 2 Current place of residence ..................................................................................................... 98 4 3 Chronic criteria vari ables ....................................................................................................... 98 4 4 Time in Alachua County ........................................................................................................ 99 4 5 Mental health need variables ................................................................................................. 99 4 6 Number of persons willing to support .................................................................................. 99 4 7 Current mental health and substance abuse service utilization ......................................... 100 4 8 Attempts for substance abuse and mental health treatment services ................................ 100 4 9 Measurement reliabilities for BTI and SOCRATES (Modified/Combined) .................... 101 4 10 Ranking of barriers by mean for all cases .......................................................................... 101 4 11 Top barriers to treatment for persons seeking services ...................................................... 103 4 12 Bottom 5 barriers to treatment for persons seeking services ............................................. 103 4 13 Logistic regression results for barriers categories on mental health utilization ............... 103 4 14 Logistic regression results for barriers categories on substance abuse utilization ........... 103 4 15 Mean and standard deviations of barriers for chronic and n on -chronic groups ............... 104 4 16 Summary of logistic regression results for mental health service utilization ................... 104 4 17 Bivariate logis tic regression of significant variables on mental health service utilization .............................................................................................................................. 104 4 18 Summary of logistic regression results for substance abuse service utilization .............. 105 4 19 Bivariate logistic regression of significant variables on substance abuse service utilization .............................................................................................................................. 105 4 20 Summary of logistic regression results for mental health treatment attempts .................. 105 4 21 Bivariate logistic regression of significant variables on mental health treatment attempts ................................................................................................................................. 106
9 4 22 Summary of logistic regression results for substance abuse treatment attempts ............. 106 4 23 Bivariate logistic regression of significant variables on substance abuse treatment attempt s ................................................................................................................................. 106
10 LIST OF FIGURES Figure page 2 1 The Behavioral Model for Vulnerable Populations ............................................................. 66
11 LIST OF TERMS Barrier a personal, organizational, or environmental obstacle that imposes a cost to utilizing mental health or substance abuse services Chronically Homeless an unaccompanied individual with a physical, mental health, or substance abuse disability who has been co ntinuously homeless for the prior 12 months or has been homeless four or more times in the last three years. Enabling Factor material and interpersonal resources or lack thereof, that facilitate or inhibit ones u se of mental health or substance abuse ser vices Health Behavior a n action, either positive or negative, taken by an individual that will affect her or his physical or emotional well -being Homeless Individual a person lacking a regular nighttime residence including those living in emergency shelt er s welfare hotel s transitional housing, mental health or substance abuse treatment settings, or places not designed to provide sleeping accommodations. Individual Barrier a perceived obstacle that imposes a cost to utilizing mental health or substance a buse services that is primarily attributed to the person or the family of the one who would use these services. Mental Health Disability a diagnosed or undiagnosed impairment in ones cognitive or emotional ability severe enough to limit or prevent employm ent or routine functioning that is not primarily attributed to the prolonged use of alcohol or other chemical substances. Mental Health Services residential, outpatient, or private therapy or psychiatric care offered by recognized professionals including c ounselors, social workers, psychologists, nurses, medical doctors, or similar licensed practitioner that is not targeted to persons with psychological and social impairments related to prolonged use of alcohol and chemical substances. Natural Supports fami ly, friends, clergy, and acquaintances who would facilitate ones use of mental health or substance abuse services. Need Factor perceived or evaluated condition indicating a reaso n to seek or not seek mental health or substance abuse services Outcome Factor the evaluation of health and quality of care after having attempted to receive mental health or substance abuse services.
12 Perceived Barrier the subjective belief in the presence of obstacles that will impose a cost to using mental health or substance a buse services. Physical Disability a diagnosed or undiagnosed impairment in ones bodily ability that is severe enough to limit or prevent employment or routine function ing Predisposing Factor personal and familial characteristics that increase or decreas e the likelihood that one will have healthcare needs, whether one will perceive those needs as important, and whether enabling factors will be present to inhibit or facilitate mental health and substance abuse service utilization. Professional Supports nur ses, doctors, counselors, psychologists, social workers, and other healthcare staff who would facilitate ones use of mental health or substance abuse services. Programmatic Barrier a perceived obstacle that imposes a cost to utilizing mental health or sub stance abuse services that is primarily attributed to the program(s) where one would use these services or the staff or professionals who would provide them Societal Barrier a perceived obstacle that imposes a cost to utilizing mental health or substance abuse services that is primarily attributed to the society in which the one who would use these services lives. Substance Abuse Disability a diagnosed or undiagnosed impairment in ones cognitive or emotional ability that is severe enough to prevent employ ment or routine functioning primarily caused by the prolonged use of alcohol or chemical substances. Substance Abuse Services residential, outpatient, or private therapy or psychiatric care offered by recognized professionals including counselors, social w orkers, psychologists, nurses, medical doctors, or similar licensed practitioner that is targeted to persons with psychological and social impairments related to prolonged use of alcohol and chemical substances. Treatment Formal inpatient or outpatient sub stance abuse services provided in a structured format, often with psychoeducational classes counseling, and other various modes of therapy. Utilization A health behavior where one uses available healthcare services, including counseling and treatment serv ices.
13 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy BARRIERS TO MENTAL HEAL T H AND SUBSTANCE ABUSE SERVICE UTILIZATION AMONG HOMELESS ADULTS By Michael D. Brubaker August 2009 Chair: Michael T. Garrett Major: Mental Health Counseling M any men and women who live on the streets, in shelters, and other precarious circumstances experience high incidents of mental hea lth and substance abuse related problems. Despite the substantial need, a high percentage of these individuals are unable to get services This study explored the particular barriers and other contributing factors that impact mental health and substance abuse service utilization among pe ople without housing. A convenience s ample of 145 adult homeless men and women was recruited from emergency shelters, mental health and substance abuse programs, and public gathering sites to determine the perceived barrie rs to mental health and substance abuse services. Using logistic regression analys es it was determined that programmatic barriers negatively correlated with mental health service utilization when controlling for individual and societal barriers, and indi vidual barriers positively correlated with substance abuse service utilization when controlling for programmatic and societal barriers. Mixed analysis of variance procedures showed no interaction between category of barriers and chronic homelessness statu s. When controlling for all other variables, self reported need was the only variable to predict mental health service utilization Likewise, recognized need predicted substance abuse service utilization. Post hoc analyses revealed a significant predicti ve relationship of individual, programmatic, societal
14 barriers, and self identified need on mental health treatment attempts when controlling for all other variables. Additionally, individual barriers and recognized need predicted substance abuse treatmen t attempts. Based on the theory of Emancipatory Communitarian (Prilleltensky, 1997), it was concluded that there are many obstacles to obtaining mental health and substance abuse services and that each type plays a different role. Programmatic barriers d irect ly prevent full participation of individuals and families in counseling and treatment services. More subtly, individual barriers serve as a gateway to entry, as participants are encouraged to focus on these obstacles to getting care while in substanc e abuse programs. Societal barriers were the most highly recognized category, a signifier that counselors, prospective clients, and community members ne ed to advocate for larger social changes that could improve services. Implications and study limitatio ns are also discussed.
15 CHAPTER 1 INTRODUCTION I feel like they dont care about me as an individual. And they have their own rigid structure or personal agenda that has nothing to do with any of their clients. They dont support what we are up to, theyre doing their thing. (Miller & Keys, 2001, p. 345) People who are homeless are not social inadequates. They are people without homes. (Dame Sheila McKechnie, 2004) Being homeless in America is a devastating experience, fraught with desperation, trauma, and social marginalization. For some, this fate may be the result of a natural disaster or a major health crisis that depletes ones savings and resources. Recent h urricanes in the Gulf Coast have exposed the vulnerability of the poor who live in th is reg ion casting people out of their homes (Brodie, Weltzien, Altman, Blendon, & Benson, 2006). These effects have been echoed in the Midwest as floods have devastated the homes of those living in low lying areas (Lydersen, 2008) Others find refuge on the s treets and shelters after surviving personal crises such as interpersonal violence, losing a job, or being overwhelmed by medical bills ( Crane et al., 2005; Tessler, Rosenheck & Gamache, 2001). The economic crisis that began in 2008, highlighted by home foreclosures, is an example of such dramatic cris e s that ha ve claimed many homes evicting many to the streets ( Erlenbusch, O Connor, Downing, & Phillips, 2008). Not everyone becomes homeless so abruptly. Some will experience a slower decline, falling dee per into poverty until a home is no longer affordable and the streets are the only option. In these cases, individuals and families may have suffer ed from the effects of substance abuse and mental health problems as well as other health and life concerns that gradually impact them (Tessler, Rosenheck & Gamache, 2001). Many of these life difficulties, such as trauma
16 and abuse have originated in childhood and slowly erode ones ability to manage lifes demands (Stein, Leslie, & Nyamathi; 2002 ). Each of th e aforementioned causes of homelessness has societal roots. Hurricane Katrina was a good example of societys role in contributing to homelessness. In New Orleans, those living in the low income areas were disproportionally affected by the flooding and b y the slow recovery efforts ( Connolly, 2005). During that time, outrage could be heard in news reports at the neglect of this vulnerable segment of society (Klein, 2005) Domestic violence, the lack of affordable housing, the absence of medical insurance mental health problems, and substance abuse all may be connected to social inequities (National Coalition for the Homeless, 2007c ). While the individual has a role in some of these events, societys contributions to these challenges often remain unmenti oned. A closer look at the demographics of those affected by homelessness reveals how society has unfairly favor ed certain segments of the population over others who live in this state of extreme poverty. Over the course of a year, there are an estimated 3.5 million homeless men, women and children (National Coalition of the Homeless, 2007b) The racial disparity is apparent as European Americans represent only 41% of the urban homeless, a figure matched nearly equally by African Americans (40%; Burt et a l., 1999). As African Americans only represent about 11% of the total population, their four -fold increase among those who are homeless reaffirms the impact of racism in American society ( Burt et al., 1999). This trend appears even more severely among Na tive Americans who are 8% of the homeless population and only 1% of the general population ( Burt et al., 1999). Men represent ing 51% of the urban homeless currently outnumber women (34%) and children (15%; U.S. Conference of Mayors 2007). T his trend ma y be reversing, however, as
17 evidenced by the fact that women and children are now the fastest growing segment of the homeless population (National Coalition for the Homeless, 2007b ). Some advocates have argued that veterans are also overrepresented among h omeless persons (National Coalition for the Homeless, 2007a), basing their estimates on older literature. More current urban studies reveal declining percentages of veterans among those who are homeless (U.S. Conference of Mayors), yet these figures may be increas ing with the recent conflicts in the Middle East resulting in high rates of mental and physical disabilities (Marshall & Souccar, 2005). Americas vulnerable populations are often the face s of homelessness. Mental health and substance abuse proble ms are pervasive among Americas homeless. U p to 39% report mental health problems within the last month, a figure that reaches 57% for lifetime prevalence (Burt et al., 1999). When alcohol or drug problems are added, these figures total 66% and 86%, resp ectively, for prior month and lifetime problems (Burt et al.). Yet b y self reports, mental health and substance abuse problems are not the primary reason for becoming homeless ( Tessler, Rosenheck & Gamache, 2001 ). Since general poverty contributes to c ompromised mental health and homelessness make persons susceptible to trauma on the streets (Lee & Schreck, 2005), it is apparent that many of these mental health conditions were not realized until after losing their residence. In addition to the immediat e traumas experienced while becoming homeless, there are many early life events that can compromise ones mental health and resistance to problematic substance use behaviors. For homeless women with a mental health diagnosis, domestic violence is the lead ing cause of homelessness ( Tessler, Rosenheck & Gamache, 2001). Living under the tyranny of a n abusive partner can leave physical and emotional wounds that will take years to heal. H omeless women are also more likely to experience childhood sexual abuse and
18 military sexual abuse than homeless men (Benda, 2006; Jainchill, Hawke, & Yagelka, 2000). Among the chronically homeless b oth men and women reported high rates of trauma and abuse (Christensen et al., 2005). All of these experiences can contribute to mental health and substance abuse problems ( Burnette, et al., 2008). Thus, b oth childhood and adult traumas must be addressed if mental health and reduced substance abuse is to be realized. Using social supports, including mental health and substance abuse services, are among the primary ways a homeless individual may achieve residential stability (Crane & Warnes, 2007; MacKnee & Mervyn, 2002; Thompson, Pollio, Eyrich, Bradbury, & North, 2004; Zlotnick, Tam, & Robertson 2003). Effective social suppo rts need to be offered from en gage ment through aftercare and termination. Moreover, d ignity is important i n these relationships as homeless individuals who are treated with more dignity report being more motivated to improve their lives, achieve self -suff iciency, help others, and leave the state of homelessness (Miller & Keys 2001). In turn, motivation provides the impetus to u tilize support services giving new opportunities for healing and life stability ( Thompson et al.) Counselors, internal motivati on, and material resources all combine to strengthen a homeless persons ability to make significant progress in her or his life While the need for substance abuse and mental health services are evident, there are many homeless people with such problems w ho do not obtain proper care. As with medical services, homeless individuals are more likely to utilize emergency mental health services instead of the longer -term outpatient services that could sustain their recovery (Folsom et al., 2005). In one study, the odds of utilizing emergency mental health services were seven times greater than those who were housed, while the odds of utilizing outpatient services were significantly less (.77)
19 than those who were housed ( Folsom et al., 2005). The lack of afforda ble housing and adequate shelter services may be one factor supporting this utilization pattern. Overall prior experiences with service providers affect mental health and substance abuse service utilization The Substance Abuse and Mental Health Service Administration has recognized the fact that many homeless people with severe mental illnesses have had poor and ineffective prior experiences with service providers (U.S. Department of Health and Human Services, 2008). T hose who are in the position to improve the lives of homeless people may actually serve as a barrier by providing ineffective and poor quality of care. Among the general population, poor interpersonal relationships with service providers are the most significant barrier to receiving medica l care ( Melnyk 1990). This relationship is mirrored a mong those who are homeless as poor relationships are one of the many named barriers to getting mental health and substance abuse services ( Bhui, Shananhan, & Harding, 2006; Thompson, McManus, Lantry, Windsor, & Flynn, 2006). Administratively, service providers can make programs too restrictive and cumbersome for prospective clients to enter ( Bhui, Shananhan, & Harding; Tsemberis, Moran, Shinn, Asmussen, & Shern, 2003). Additionally, those seeking ca re report facing long lines, high staff turnover, and a disjointed systems of care ( Bhui, Shananhan, & Harding; Wilde et al., 2004 ). Counselors have the ability to directly affect the most important factor, their relationship with the client. Additionall y, for those not in administrative positions, they may advocate for changes that may improve accessibility. This call for a dvocacy and social justice has re e merged at the forefront of the counseling literature in recent years, considered by some as the six th force of counseling ( Pack Brown, Thomas, & Seymour, 2008). In the ACA Code of Ethics (2005) counselors are called to examine potential barriers and obstacles that inhibit access and/or the growth and development
20 o f clients (Section A.6.a.). Homeles s individuals, already stigmatized by society ( Toro et al., 2007) face numerous barriers to substance abuse and mental health treatment. These barriers in combination with other factors that contribute to treatment utilization need to be examined to un derstand how counselors may intervene with and on the behalf of homeless individuals. Theoretical Framework To date, there have been many theoretical approaches with which to serve the homeless population. Among the more traditional ones the most studied theory is cognitive behavioral therapy ( CBT ; Foster, 2008; Milby et al., 2008; Thompson, McManus & Voss, 2006), a theory made readily measurable by its systematic procedures which fit more easily into a medical model of treatment (Wompold, 2001). Others focusing on intrapsychic matters of those who are homeless suggest using psychoanalysis and related techniques (Felix & Wine, 2001; Sklarew, Twemlow & Wilkinson, 2004). Both CBT and psychoanalysis offer contributions that may benefit homeless clients bu t fail to address the wider sociological problems which impede growth. Countering this limited focus, ecological approaches and empowerment strategies have been suggested to better understand the broader systems that impact homeless people and strengthen their abilities to overcome difficulties (Kuhlman, 1994; Wise, 2005). Furthermore, some have suggested postmodern approaches such as s olution focused b rief t herapy, narrative therapy, and feminist therapy to recognize the importance of those who are homel ess giving voice to their experiences and moving towards wellness (Hartman, Little & Ungar, 2008; Levy, 2004; Reid, Berman & Forchuk, 2005; Thompson, McManus & Voss). An Emancipatory Communitarian Perspective While each of these theories is helpful in their pursuits to better the lives of homeless persons, to date, none has suggested a theory which considers the systemic influences on homelessness, promotes client empowerment and communal responsibility for both the
21 counselor and client. Emancipatory Com munitarianism (EC; Prilleltensky, 1997) is such a theory to address these needs. It looks beyond individualistic values in psychotherapy and takes a strong stand for distributive justice for all persons. While recognizing client strengths, it takes a com munal approach to addressing barriers, insisting that all work to overcoming obstacles that oppress and demand mutual responsibility to achieving these ends. Prilleltenskys (1997) formulation of Emancipatory Communitarianism (EC) emerged as an aspiration al paradigm (p. 519) that would promote the development of good individual lives and good communities. He established this as an overarching category under which particular theories would emerge. Subsequent to its initial publication, this paradigm was suggested for use in the areas of vocational development (Blustein, McWhirter & Perry, 2005) d omestic violence (Chronister, 2006; Chronister, Wettersten, & Brown, 2004), multicultural counseling (Vera & Speight, 2003) and family therapy (Melito, 2003), al l issues relevant to the homeless population. EC is an appropriate fit for working with those who are homeless as it nurtures critical consciousness and demands responsible action, al l within a communal framework. At the root of this theory, Prilleltensky (1997) concerned himself with the lack of attention to underlying values and assumptions among the existing theories and how they were expressed in practice. He developed EC with these three areas in mind as outlined in the following paragraphs Values. EC pursues higher goals than to seek individual self -determination and autonomy as is found in many approaches; it also strives for distributive justice as well as the concern for other persons and communities (Prilleltensky, 1997; Prilleltens k y, 2000). It favors the oppressed to equalize the imbalances of power, and in doing so it respects the many facets of human diversity ( Prilleltensky, 1997; Prilleltensky 2000).
22 Assumptions. EC approaches ascribe to idea that oppression is at the root of many, but no t all mental health problems, and that solutions will come from critical consciousness invoking individual and societal change (Prilleltensky, 1997; Prilleltensy & Nelson, 2002). Clients and communities are invited to build upon existing strengths and re sources not simply focus ing on deficits or diagnoses (Prilleltensky & Nelson). Rooted in these collective assets, mental and societal health will come through the removal of oppression, practicing mutuality, and fulfilling social obligations (Prilleltensk y, 1997). Practices. Counselors using this approach will help clients define their problems in terms of interpersonal and societal oppression using interventions which seek to change individuals as well as social systems (Prilleltensky, 1997; Prilleltensky & Nelson, 2002). Personal agency is stressed with EC, a responsibility that must be assumed largely by the client less he or she becomes a victim of these social ills (Prilleltensky & Nelson). Furthermore, it acknowledges that importance of the counselo r -client relationship that can be oppressive if power is not shared and clients are not given the power to determine their own therapeutic processes (Martn Bar, 1994; Prilleltensky; Prilleltensky & Nelson). And like the post -modern approaches, EC counsel ors set aside the role of expert, understanding that knowledge and expertise is largely within the client and surrounding community (Prilleltensky, 2000; Prilleltensky, 2002). Lastly, EC advocates for change to occur proactively, emphasizing wellness and s trengt h s instead of waiting for further pathology (Prilleltensky, 2000). The underlying values and assumptions inform the practices of EC, which may readily be used to support homeless clients and communities in their pursuits to survive, and even thrive In the case of barriers, EC suggests that they are largely caused by society, yet without critical consciousness, the individual may tend to blame him or her self instead As the general
23 public tends to blame homeless individuals for their circumstances (Toro et al., 2007), it is expected that those more closely aligned with these values and beliefs, those curren tly in mental health or substance abuse treatment programs are more likely to share this view of self blame. It would also suggest that the si gnificant barriers will occur within the relationship between homeless individuals and service provider when power imbalances are exercised. When homeless individuals are treated as deficient, their well -being is compromised, and many will resist this cha racterization and therefore resist treatment. This theory also promotes the awareness of these barriers, as recognized by thos e who are homeless and th e people who support them. Such barriers would not be identified through indirect means; rather, instrum ents that measure barriers should ask direct questions related to the homeless experience By the assumptions and practices of EC, these named barriers are deemed important and relevant to ones ability to take responsibility and overcome them In relati on to demographic variables, barriers are expected to be different based on ethnicity, gender, age, sexual orientation, and veteran status, as each of these groups experience societal various forms of social oppression and biases. These values, assumptions and practices not only guide counseling practices but also research. Critical consciousness may be invoked through open acknowledgement of societal barriers and subsequent exploration of the connection of these factors to greater societal causes. Addit ionally, information may be shared with key stakeholders to advocate for changes in the treatment service system. In both research and practice, EC provides a guiding framework to understand barriers and resources from a social justice and community-based orientation. By combining EC with compatible utilization theory, counselors will have a unique perspective to support greater mental health and substance abuse service usage
24 The Behavioral Model for Vulnerable Populations While EC serves as a guiding fo rce to overcome barriers and encourage existing strengths, there is a need to conceptualize the many factors that contribute to utilization. The utilization literature provides such a framework where both barriers and resources are of significant concern. The concept of barriers was first recognized in this context by Rosenstock (1966) who defined them as the cost to taking action towards getting healthcare (Melnyk, 1988) If the subjective appraisal of costs to receiving care exceeded the benefit, the b arrier would prevent action. Rosenstock note d An individual may believe that a given action will be effective in reducing the threat of disease, but at the same time see that action itself as being inconvenient, expensive, unpleasant, painful or upsetti ng (p. 100). Shortly thereafter, Andersen (1968) looked at barriers and other factors as he proposed a more dynamic and comprehensive model to health care utilization Three categories were most important in this model predisposing factors (e.g. demogra phics) enabling factors (e.g. health insurance) and need factors (e.g. mental health or substance abuse problems) These underlying factors determine whether one will seek healthcare services when a need, either real or perceived, arises. In his model, enabling factors include both facilitative resources and the obstructive barriers which support or prevent one from seeking services. The Andersen (1968) model has been critiqued and revised over the years and later adapted for the homeless population. Ad dressing the original model, Melnyk (19 8 8) c riticized it for failing to recognize the subjective nature of health care barriers, which are not necessarily external objective realities for an individual. Later iterations of the Andersen model have recogniz ed perceived barriers within the enabling category, which includes both health promoting and hindering factors (Gelberg, Andersen, & Leake, 2000). Andersens (1968) model was subsequently modified and adapted for vulnerable populations, namely for those w ho are
25 homeless (Gelberg, Andersen, & Leake, 2000; Stein, Andersen, & Gelberg, 2007). This model added vulnerable domains to each of the three traditional categories. For example, ones history of homelessness would be a predisposing factor which may inc rease the likelihood that barriers such as hunger and other competing needs will be present. These conditions, in turn, can influence the perceived need for health care. Of the se three categories the enabling factors are the most mutable and are therefor e a target for programmatic and public policy changes (Andersen, 1995). As predisposing factors such as demographics and need factors are often difficult to alter, resources and barriers within the enabling category are the better target for impacting ones service utilization (Andersen). However, both predisposing and need factors are to be considered especially as they may impact enabling factors For instance, poor quality substance abuse and mental health services may be a result of staff biases ag ainst those of a different cultural background. Therefore, it is important to study all of these factors and their relationships to realize the source s of barriers and how they may be changed. The Behavioral Model for Vulnerable Populations ( Gelberg, Ande rsen, & Leake, 2000) is an appropriate framework to understand utilization along with Emancipatory Communitarianism (EC) for a number of reasons. First, it looks at barriers and resources to understand how one uses treatment services. It is not simply de ficit focused. Additionally, it begins to offer a meta perspective on the way that predisposing factors are related to enabling factors. Through the lens of EC, the relationships of these variables become more explicit, revealing societal biases and oppr ession that impact the lives of homeless individuals. Lastly, its acknowledgement of subjective needs, resources, and barriers respect the experiences of the individual and how perception is relevant to this equation.
26 Need for the Study Currently research on mental health and substance abuse service utilization among those who are homeless is minimal. Within the utilization and barriers literature, most studies have identified separate contributing factors which affect service usage with only a few stu dies that have explored the relationships between variables. In these studies, researchers have identified that the predisposing factors including gender, age, education, and parental status are related to utilization ( Bonin, Fournier, & Blais, 2007; Lemm ing & Calsyn, 2004; McCarthy et al ., 2007; Sosin & Bruni, 2000). Enabling factors are more researched and include staff -client relationships, administration of program services, fragmented services, and overly restrictive admissions criteria ( Bhui, Shanan han, & Harding, 2006 ; Calloway & Morrissey, 1998; Lemming & Calsyn; Rothbard, Min, Kuno, & Wong, 2004; Sosin & Bruni; Thompson, McManus, Lantry, Windsor, & Flynn; Tsemberis, Gulcur, & Nakae, 2004; Wilde et al., 2004) There has been some attention to the larger societal factors such as health insurance and programmatic funding resources ( North, Pollio, Perron, Eyrich & Spitnzgel, 2005; Sosin & Bruni ; Kertesz et al., 2006 ). Additionally, some studies have focused on the role of natural support systems inc luding family, friends, and other non-clinical sources of care ( Kertesz et al.; Lam & Rosenheck, 1999 ; Lemming & Calsyn ). Need factors are generally assumed, especially among those who are already in a treatment setting ( Kertesz et al.; McCarthy et al. ). One study added a measure of the Stages of Readiness to Change to determine the degree to which study participants perceived a need for treatment (Kertesz et al. ). Again, the factors that are most easily changed are the enabling ones. However, none of t hese studies have produced a comprehensive list of perceived barriers to understand which obstacles are the most salient for various homeless populations. Women may experience different barriers than men, African Americans different than European American s and
27 Hispanics, the elderly different than younger populations, and so on. By understanding how predisposing factors are linked to perceived barriers, there can be a more targeted focus on the important challenges identified by those who need mental heal th and/or substance abuse treatment, but do not have a home. Furthermore, there needs to be information about whether chronically homeless persons experience barriers differently than those whose homeless experience is short term, or transitional, in natu re. Another problem is the notion that barriers and other enabling factors may frequently be determined by indirect means and assumptions. For instance, one study determined that living with children indicated a need for childcare, an assumed barrier (Ker tesz, et al.). By directly asking the degree to which the lack of childcare was a barrier, the participant is given fuller voice to express this need, and she or he may prioritize this barrier in contrast to others. One of the biggest problems in the exis ting research is that the quantitative studies have all recruited homeless individuals from programs instead of from shelters and street locations. By following these procedures, the researchers are neglecting the group of people who are of greatest inter est, those who are not getting help. A study with more extensive recruiting, looking at both those in treatment and those outside of it will substantially contribute to the knowledge base that may better serve potential program participants. Lastly, most of the studies lack evidence on how the results were utilized to benefit the population aside from submitting a peer reviewed publication. A few of the studies have tracked program changes that have been made by clinicians and administrators ( Gamache, Ros enheck, & Tessler 2000; Lam & Rosenheck, 1999; Lemming & Calsyn 2004; Sosin & Bruni 2000), but none of them have showed evidence of sharing their results with the population they are researching. Also, there is no evidence in sharing this information w ith public policy makers as
28 an effort to advocate for the needs of those who are homeless By the premises espoused by EC, researchers and clinicians alike are called to a higher level of responsibility which would include advocacy. The current research in this study include s this type of community involvement. Purpose of the Study The main purpose of th is study wa s to identify the most salient barriers that prevent mental health and substance abuse service utilization among homeless men and women. By un derstanding the se barriers and their relationship to other enabling predisposing, and need factors, counselors may better understand their role in inviting new clients into counseling relationships and how to effectively serve the ones who have already s o ught their care. Furthermore, counselors can gain a broader perspective of the contextual forces that impact service utilization. Subsequently, counselors will be better able to advocate for individual, programmatic and societal changes that can support homeless individuals as they attempt to get assistance. Research Questions 1 What are the most significant perceived barriers that prevent homeless person from obtaining mental health or substance abuse services? 2 How do societal, programmatic, and individu al barriers contribute to mental health services utilization? 3 How do societal, programmatic, and individual barriers contribute to substance abuse services utilization? 4 Does being chronically homeless versus non-chronically homeless impact each of the barr iers categories?
29 5 What is the relationship between predisposing factors (gender, ethnicity, age, homeless duration), enabling factors (perceived barriers, natural supports, professional supports), and need factors (perceived needs, and mental health rating) on mental health service utilization? 6 What is the relationship between predisposing factors (gender, ethnicity, age, homeless duration), enabling factors (perceived barriers, natural supports, professional supports), and need factors (recognized need to c hange substance use, and actions to change substance use) on substance abuse service utilization?
30 CHAPTER 2 LITERATURE REVIEW There are many reasons why a man, woman, or child may become homeless. Losing a job, becoming bankrupt from medical bills, suff ering the effects of a natural or human-made catastrophe are all possibilities that may lead someone to the streets without safe and secure housing (Brodie, Weltzien, Altman, Blendon, & Benson, 2006; Crane et al., 2005). Others may fall into homelessness over a steady decline, already living on the lower margins of poverty They may have mental health, substance abuse, or relationship problems that they have endured over many years and finally reach the tipping point where the y can no longer pay the rent and no one is willing or able to help (Tessler, Rosenheck, & Gamache, 2001). While there are many causes of homelessness, society often neglects to pay attention to the underlying systemic factors that contribute to homelessness focusing instead on the pe rsonal decisions and vulnerabilities of this population (Toro et al., 2007). In fact, there are many societal causes including low wages, a lack of affordable housing, and a declining economy, among other reasons that contribute to a person or family losi ng their home (National Coalition for the Homeless, 2007c ). Furthermore, as certain populations are more likely to become homeless such as African Americans, Native Americans, those with physical and mental disabilities, and veterans (Burt et al., 1999), one must question the degree to which other factors such as racism, ableism, and variable support for our military personnel reflect the way society systematically oppresses some and fails to support those in greatest need. Once someone becomes homeless, t he road to recovery can be tough as many of the same burdens exist that contributed to ones circumstances. In addition, with few resources available, often distributed through bureaucratic means, it can be difficult to access the needed resources to impr ove ones situation. Like the many factors that contributed to becoming homeless, there are
31 many personal and societal barriers that prevent making life gains. Besides the lack of funding for social programs, there are also poorly designed and implemente d services for those who are homeless (Bhui, Shananhan, & Harding, 2006). Furthermore there are personal and family beliefs about the efficacy of these services among other psychological factors, often shaped by life experiences, which can prevent someone from getting the help they need (Gelberg, Andersen, & Leake, 2000). These are but a few factors that may deter a homeless individual from seeking mental health or substance abuse counseling services. These barriers to care are an issue for counselors who serve marginalized populations such as those who are homeless With greater attention to social justice issues, counselors are called upon to advocate for client needs. As homeless individuals and families are in great need of mental health and substanc e abuse services (Burt et al., 1999; Tessler, Rosenheck, & Gamache, 2001), counselors are remised to find more effective ways to provide services. In fact the ACA Code of Ethics (2005) calls for counselors to examine potential barriers and obstacles that inhibit access and/or the growth and development of clients (Section A.6.a.). Where these injustices occur, counselors are called to advocate. This chapter will show that there remains a dearth of information on the barriers and other factors that inhibit mental health and substance abuse service utilization by homeless persons Of the few studies that exist, there are significant limitations and an often narrow scope that prevents counselors from understanding how to best reach potential clients who do not receive needed care. In order to best understand these utilization factors, this chapter will offer a description of the homeless population, the factors which contribute to this condition, and how many people exit homelessness. This information w ill offer the backdrop for identif ying the key
32 barriers and enabling factors that affect if and how a homeless individual uses these critical counseling services. Who are Those Who Become Homeless? On any given year, there are up to 3.5 million homeless pe rsons in the United States who will take refuge in shelters, under bridges and overpasses, in cars, tents, abandoned buildings, or a myriad of places (Urban Institute, 2001). Others who have lost their homes may be temporarily living with family members or friends, or they may seek shelter in hotels or other such places as long as they have enough income (Burt, 2008). Even though the definition of homelessness may seem apparent, this is not the case as the ways for defining this population has been debated The following sections will discuss the leading definition of homelessness, provide demographics about this group, and will discuss one particular subpopulation, the chronically homeless, that has gained much attention in recent years. Defining Homele ss There are many interpretations to the word homeless depending on its use. For those living on the street, being homeless can carry many negative connotations as evidenced by American biases against this population (Toro et al., 2007). As a marginali zed population, homeless persons may avoid the label homeless as an adaptive way to blend in among those who are housed (Harter et al., 2005). Some rightly claim their dwelling in a car or in the woods as their home. In such cases, having home is not d efined by the structure as much as it is by the sense of normalcy and stability (Rivlin & Moore, 2001). On the other end, a dvocates with the intent of securing funding and services for those without a house or apartment may identify the homeless to incl ude persons who are marginally housed, living with relatives, or other temporary circumstances due to extreme poverty and traumatic life circumstances (Burt, 2004).
33 For legal purposes, the Federal government more narrowly define d a homeless person through the McKinney -Vento Homeless Assistance Act of 1987. By this account the term homeless includes a person who is: 1. an individual who lacks a fixed, regular, and adequate nighttime residence; and 2. an individual who has a primary nighttime residence that is a. a supervised publicly or privately operated shelter designed to provide temporary living accommodations (including welfare hotels, congregate shelters, and transitional housing for the mentally ill); b. an institution that provides a tempora ry residence for individuals intended to be institutionalized; or c. a public or private place not designed for, or ordinarily used as, a regular sleeping accommodation for human beings. (42 USC Sec. 11302) This definition does not recognize those who do uble up with another family or friends for shelter (National Coalition for the Homeless, 2007c ). Persons living in these situations may fail to qualify for some programs reserved for those who meet the federal criteria. Although not exclusively, most rec ent studies on those who are homeless have used this federal definition (e.g., Kertesz et al. 2006) This is important to remember as there is a portion of the population who live in extreme poverty, having lost their home yet are not included in the current research. Furthermore, as noted above, some may not even qualify for services if they can not be covered by the federal definition. Demographics Ethnicity and race vary by region of the country and between urban and rural settings. In a large study of homeless persons living in either urban or rural settings 41% were White/nonHispanic, 40% Black/nonHispanic, 11% Hispanic, 8% Native American, and 1% other (Burt et
34 al., 1999). Ethnic minorities, particularly African American s and Native Americans are strongly overrepresented as they are 11% and 1% of the general population, respectively (Burt et al.). When looking at the urban setting alone, 23 major cities reported an aggregate racial estimate among those who are homeless at 50% White, 45.7% Black, 2.5% Native American, and 1.6% Asian; of the total 12.8% report to be Hispanic/Latino (U.S. Conference of Mayors, 2007). When comparing gender, single men comprise the largest portion of urban homeless adults at 67.5% (U.S. Conference of Mayors), even though women represent the largest portion of those who live in poverty ( DeNavasWalt, Proctor, & Smith, 2007) Some attribute this difference to the fact that women, and particularly women with children, are more likely to get social support than men (Morri s, 1997). But this trend may be reversing as evidenced by the fact that women and children are now the fastest growing segment of those who are homeless (National Coalition for the Homeless, 2007b ). Some advocates have argued that veterans are also overrep resented among homeless persons (National Coalition for the Homeless, 2007a), basing their estimates on older literature. Current urban studies reveal declining percentages of veterans among those who are homeless (U.S. Conference of Mayors), yet these fig ures may soon increase with the recent conflicts in the Middle East resulting in high rates of mental and physical disabilities (Marshall & Souccar, 2005). Being homeless does not necessarily equate to being jobless. In fact, 13% of the urban homeless rep ort working either part or full time (U.S. Conference of Mayors, 2007). Many others are disabled, some of whom receive or await disability income (Rosen, McMahon, Lin, & Rosenheck, 2006). Another faulty stereotype is the assumption that homeless persons l ive in urban settings. There are many homeless persons who live in rural settings, taking refuge in their
35 cars, campers, or doubled up in the homes of family or friends (Rollinson, 2007). These persons are more likely to be single working mothers of Europe an -American ethnicity (Rollinson). The Chronic Homeless In recent years there has been increased attention to differentiating the chronic homeless from the transitionally homeless, those who experience brief homelessness and do not return to the streets. Public attention peaked in 2002 when President Bush announced that one of his top objectives was to eliminate chronic homelessness in the next decade (U.S. Department of Housing and Urban Development, 2004). The U.S. Conference of Mayors followed suit and by 2007 nearly 300 cities nationwide committed to having a plan to end chronic homelessness (U.S. Department of Housing and Urban Development, 2007). Although the term chronic homeless appears earlier in the literature, the study that identified the cha racteristics of the chronically homeless was written by Kuhn and Culhane (1998). The authors performed a cluster analysis with the data from public shelter in New York and Philadelphia and determined that there were three categories of homeless people bas ed on their shelter usage as follows: the transitionally homeless, the episodically homeless, and the chronically homeless. Although limited to two urban settings, this study represented a large number of persons (n = 73,263) over a period of two to three years (Philadelphia and New York samples, respectively). The authors first analyzed the New York sample and then used the Philadelphia sample to confirm their findings. The transitionally homeless were those persons who entered the shelter system for a s hort period of time and typically stayed for only one time during the three -year period studied. These persons would become homeless because of various crises and would not have any options for staying with family or friends. During the study in New York this cluster, which comprised
36 81% of the total, used the shelter on average 1.4 times for a total number of 58 days per person (Kuhn & Culhane, 1998) The episodically homeless were those who experienced more homeless episodes than either of the two ot her clusters, but these episodes were interspersed with periods of stable housing. These persons tended to be younger but frequently had medical, mental health, or substance abuse problems. These factors contributed to the fluctuating patterns of housing, as one experienced periods where these debilitating conditions become more or less acute. The sample revealed that this cluster stayed in shelters, on average, 4.9 different times in the 3 year period for an average total of 264 days person. This group represented 9.1% of the citys total sample (Kuhn & Culhane, 1998) Chronically homeless persons were the most severe group of the three in terms of total days homeless and represented a similar percentage of the total homeless population as the episodic cluster (9.8% in the New York sample). They are more likely to be older, African American and often have disabilities and substance abuse problems. The New York and Philadelphia samples contradicted each other in terms of gender representation. In New York there were no significantly differences in gender between the chronic cluster and the transitional group; however, in Philadelphia, there were significantly more women in the chronic cluster (28.9%) when compared to the transitional cluster (16.9%). The authors attributed these differences to social services delivery systems in each of the cities. Instead of using shelters as emergency housing, the chronically homeless use these services as if they were long-term housing options. The sample showed that this cluster experienced 2.3 homeless episodes for an average total of 638 days during the three year period (Kuhn & Culhane, 1998).
37 Since this report, the department of Housing and Urban Development collapsed the episodically homeless and chronical ly homeless into one category and recognized them both as chronically homeless (Kertesz, Larson, Horton, Winter, Saitz & Samet, 2005). By their definition a person who is chronically homeless is an unaccompanied adult with a disabling condition who is 1) currently homeless for one year or longer, or 2) having four or more episodes of homelessness in the last three years ( National Coalition for the Homeless, 2008) Parents with children and those without a disabling condition (physical or mental health) d o not qualify as chronically homeless. As for the general definition of homelessness, the chronic homeless definition has guided funding initiatives for HUD and other federal agencies. Part of the reason for focusing on the chronically homeless relates to the social cost of this subpopulation. Kuhn and Culhane (1998) note that chronically homeless persons represent approximately 10% of the total homeless population but use almost 50% of the total shelter days. Adding the episodic cluster to this group only raises these costs. Culhane would later add that the cost of shelter, medical, counseling and other social services reaches approximately $60,000 to $70,000 per year for each homeless individual (Mitka, 2006). With women the cost of chronic homelessne ss includes the personal and financial problems of lower birth weights and higher preterm births (Stein, Lu & Gelberg, 2000). The resulting infant mortality and higher medical costs have not been studied, but are presumably serious. Chronic homeless ness is also a concern for individuals and their family members. Those who live without housing for extended periods of time tend to have higher rates of substance abuse, mental illness and medical disabilities (Kuhn & Culhane, 1998). Although these conditio ns may predate their homeless condition, the lack of housing may contribute to these debilitating conditions which impact whole families (Evans, 2004; Murali & Oyebode, 2004).
38 Furthermore, chronically homeless individuals have increased mortality rates ev en when adjust ed for age differences. Barrow, Herman, Cordova, and Stuening (1999) found that extended periods of homelessness in men increased the odds of mortality to 2.2 to 1 over men with stable housing. With the high personal and societal costs, home lessness is an important focus for society and those who have the ability to impact their lives. The chronically homeless are of particular interest to counselors and other service providers as they have some of the greatest mental health and substance abu se needs, factors that may be contributing to their condition. And if programmatic, personal, and societal barriers are preventing them from receiving these services, counselors have the ability to advocate and make a significant impact. A discussion about the causes of homelessness will better enlighten the awareness of these barriers and how they persist. The Road to Homelessness Much of the research attention to homeless men, women, and children has been largely focused on the causes of homelessness. There have been many reasons identified, both from the personal and societal levels but by different groups On one side, a dvocates suggest that declining wages, reduced public assistance, a lack of affordable housing and affordable health care are all l arger structural reasons that persons become homeless (National Coalition for the Homeless, 2007c ). Alternatively, research ers often ignore these factors in favor of issues more readily identifiable among homeless individuals such as addictions, domestic violence, mental health problems, and trauma (Mallett, Rosenthal, & Keys, 2005; Tessler, Rosenheck, & Gamache, 2001). Researcher biases appear to reflect American societys biases, implicating the individual, rather than society, as the primary cause (Tor o et al., 2007) In essence, these studies are focusing on the symptoms of individual and larger societal influences ; without considering all of these factors, both the client and counselor are likely to fall into a blame the victim or
39 homelessness as a disease mentality which can hinder client progress (Cosgrove, 2006; Prilleltensky & Nelson, 2002; Steinhaus, Harley & Rogers, 2004). The following sections identify both individual and societal factors including mental health, addictions, and domestic v iolence; natural and human inflicted disasters; factors that lead youth and the elderly to homelessness; and societal factors. Mental Health and Addictions For many counselors, a common reason for seeing a homeless client would be to assist with a mental health problem. Among this population, up to 39% report mental health problems within the last month, a figure that reaches 57% for lifetime prevalence (Burt et al., 1999). When alcohol or drug problems are added, these figures total 66% and 86%, respectively, for prior month and lifetime problems (Burt et al.) Among those who identify as homeless and having a mental illness, the majority do not identify their mental health as the reason for becoming homeless. In one large study (n = 7,224), only 20% of those diagnosed with mental health problems reported mental illness as the primary cause of being homeless, and an additional 13% attributed alcohol or drug problems to their situation, still far from half (Tessler, Rosenheck & Gamache, 2001). If homelessne ss is not the primary reason for becoming homeless among those identified with a mental health problem, their depression, anxiety, and other mental difficulties may be the result of being homeless Researchers have noted how living in poverty exacerbates mental illnesses through stress, poor living environments, increased criminal victimization, and the lack of resources to face these difficulties (Evans, 2004; Murali & Oyebode, 2004). This would suggest that homelessness, as an extreme form of poverty, ca n be considered both a cause and effect of compromised mental health.
40 This circular effect is even more evident among the chronically homeless. Kertesz et al. (2005) studied patients in a detoxification unit and found that mental health scores improved le ss over time among the chronically homeless than among the transitionally homeless and even less than those who were housed. Not surprisingly, the chronically homeless were more likely to have greater medical severity, alcohol use severity, greater levels of depression at baseline and less support from family and friends (Kertesz et al.) A lack of support from family and friends has correlated with being homeless over extended periods (Wong & Piliavin, 2001). All of these factors could have contributed to slower mental health improvements and kept them at risk for remaining homeless. Interpersonal Violence and Trauma Among women, interpersonal conflict, including domestic violence, exceeds either mental health or drug abuse as the primary cause of homeless ness (Tessler, Rosenheck, & Gamache, 2001). A high proportion of these women will find safety at a domestic violence shelter, but many others will find other more traditional shelters for women and children (Stainbrook & Hornick, 2006). These women have h igh posttraumatic stress disorder (PTSD) rates, have often experienced violence in the last th re e months prior to entering the shelter, and are now experiencing the trauma of homelessness (Stainbrook & Hornick). Domestic violence is but one form of trauma that homeless women may experience. These women are more likely to experience sexual abuse during their lifetime than men (58% versus 14% in men; Jainchill, Hawke, & Yagelka, 2000). Furthermore, homeless female military veterans are more likely to be sex ually abused while serving their country (Benda, 2006). These traumas are found to increase the likelihood of an early onset of homelessness (Jainchill et al.) Childhood abuse increases the likelihood that women will become chronically homeless, but it ap pears to be less the case for men. Women who have survived these early life traumas are
41 more likely to experience physical abuse as an adult and have lower self -esteem, which in turn contributes to their risk of becoming chronically homeless (Stein, Lesli e, and Nyamathi, 2002). Early studies have disputed the correlation between childhood sexual, physical, or verbal abuse and becoming chronically homeless in men (Sumerlin, 1999). However, chronically homeless men with co -occurring disorders have been sho wn to report high rates of lifetime physical or sexual abuse (68.6%; Christensen et al., 2005). These high incidents of trauma certainly have an impact on the lives of both homeless men and women. Natural and Human Inflicted Disasters Another form of trauma can directly lead to homelessness, natural and human inflicted disasters. Two major U.S. disasters which began the 21st century brought to the public attention the way that whole communities may become homeless and overwhelmed by disaster. In 2002, mont hs after the terrorist attacks in New York City, local shelters became so overwhelmed with newly homeless persons that they began to consider evicting the longer term patrons of these services (Steinhauer, 2002). Then, in 2005, after Hurricanes Katrina and Rita hit the Gulf Coast, thousands of people became temporarily homeless seeking shelter in nearby cities ( Brodie, Weltzien, Altman, Blendon, & Benson, 2006). Upon reaching safe shelter, these homeless persons began to share stories of survival not only f rom the natural elements, but also from predators who took advantage of their disadvantaged situation, often through violence (Brodie et al.). This vulnerability is experienced by many other homeless persons who live with little to no protection on the str eets. In New Orleans, it became apparent that poverty increased ones vulnerability to becoming homeless as the more impoverished and largely African American 9th Ward was subject to greater flooding than other areas of the city; it wa s also the least lik ely to be rebuilt (Connolly, 2005). Later disasters, such as the Midwestern floods of 2008, have hit low income neighbors
42 disproportionately high as they tend to live in low lying areas, often without flood insurance (Lydersen, 2008). W hile many experience the trauma of these events, it is the poor and socially marginalized who more likely bear t he burden of becoming homeless. Factors Impacting Americas Youth and the Elderly Beyond disasters, there are some factors more likely to cause youth and elderly in dividuals to become homeless. Among adolescents many leave the home because of family conflict, child sexual abuse, other traumas, drug and alcohol use, psychological difficulties, or criminal activity (Mallett, Rosenthal, & Keys, 2005; Martijn & Sharpe, 2006). Some of these adolescents are seeking the safety of the streets over their homes and others are put out because of their behaviors or their identification as a lesbian, gay, bisexual, or transgender individual (Martijn & Sharp; Rew & Horner, 2003 ). The individual decisions of these families and their children should not be ignored; however, it is essential to understand the role of social policies, economic difficulties, and societal biases that have contributed to a young persons feeling unsafe a nd unsupported in their own family. Unlike Americas youth, the elderly become homeless for seemingly quite different reasons. Most have never been homeless before and simply find mounting rent or mortgage payments, medical care, and other costs too overwhelming (Crane et al., 2005). Other underlying factors leading to homelessness among the elderly include depression and other mental health problems, which are often untreated as many seniors do not seek help for these problems (Crane et al.). Again, personal choices and individual vulnerabilities play a role, yet so do social and economic policies, rarely spoken contributors to this crisis of Americas youth and the elderly. Specific Societal Factors Very few studies look directly at the societal factors th at lead a person to being homeless. One such study in Toronto, Canada, used qualitative means to inquire about the societal causes
43 and discovered that homeless persons cited the following reasons for their condition: a lack of supportive counseling service s, limited employment that offered an acceptable wage, a lack of safe and drug -free affordable housing, and a lack of incentive to change (Morrell Bellai, Goering, and Boydell, 2000). Interestingly, when those who are homeless identified societal factors, they did not identify the ways that social policies, norms, and biases all contribute to these causes; rather, they tended to identify those factors that would most directly impact their condition. It is unclear the degree to which homeless persons and counselors in the United States view homelessness as a result of societal factors. Perhaps more attention will be paid to these broader societal factors in light of the recent U.S. economic and housing crisis. By the end of 2007, foreclosure rates surpassed t hose of 2006 by 75% (Barris, 2008), and while many of those who lost their primary residence have been able to rent apartments or stay with family and friends, others have turned to emergency shelters along side other homeless persons (Erlenbusch, OConnor Downing & Phillips, 2008). As the debate continues over who is to blame high -risk homeowners, greedy lenders, or economic policy makers those living on the streets may be paying the ultimate price. Becoming homeless is not a welcoming experience. Co unselors who are social justice minded will be concerned with both the societal and personal factors that contribute to one being homeless. They may play a direct role in the mental health and substance abuse services provided to this population and will advocate for changes in the social systems that perpetuate the inequities that lead some to the streets and others not. Likewise, counselors who share these broader perspectives will be interested in those factors that facilitate and inhibit progress for these men, women, and children to exit homelessness.
44 Exiting Homelessness The research addressing the factors that facilitate ones exit from homelessness are far fewer than those that cite the causes (Karabanow, 2008). The few available do offer some ins ights into the key reasons a person can exit this state of living They indicate that personal, social, and societal resources are all important in this process. The importance of family and friends as well as service providers top the lists as reasons fo r exiting homelessness (Crane & Warnes, 2007; MacKnee & Mervyn, 2002; Thompson, Pollio, Eyrich, Bradbury, & North, 2004; Zlotnick, Tam, & Robertson, 2003). Counselors play a role b oth as they make direct contact and facilitate contact with family members. Relationships are critical. Outreach programs can effectively recruit homeless individuals to engage treatment, but this alone will not lead to stable housing (Bradford, Gaynes, Kim, Kaufman, & Weinberger, 2005). Effective social supports need to be engaged throughout the recovery period. Dignity plays an important role in these relationships, as homeless individuals who are treated with more dignity report being more motivated to improve their lives, achieve self -sufficiency, help others, and leave th e state of homelessness (Miller, 2001). All of these may play a role in how one utilizes support services, which are critical in helping someone change their circumstances (Thompson et al.) There are also many internal processes that occur which promote a transition towards stable housing. Taking responsibility for ones future is reported as a key factor in exiting homelessness (MacKnee & Mervyn, 2002; Thompson et al., 2004). This is not to say that a homeless person needs to assume the blame for their circumstances, when in fact there are many contributing factors. Those who exit homelessness take an active role in their life decisions. As one homeless individual explained, If you are on the streets, you just got to decide if you want to stay out there, or you want to help yourself to get off the streets
45 (Thompson et al., p. 427). Other internal changes that promote a move to stable housing include discovering self -esteem, accomplishing mainstream lifestyle goals such as a high school diploma, and changing perceptions often as a result of bottoming out (MacKnee & Mervyn). Societal factors also play a role in ones exit from homelessness. Those who take advantage of government subsidized housing and have consistent entitlements are more likely to gain residential stability (Zlotnick, Tam, & Robertson, 2003). Government resources may also decrease motivation if the individual fails to acknowledge their own worth and potential (MacKnee & Mervyn, 2002). By offering affordable housing, society has a n opportunity to provide both dignity and a stable home from which other life changes may be based. Homeless youth and the elderly have similar factors that inspire their exit from homelessness, but some factors unique to their developmental levels. Among homeless youth, there are often multiple attempts to leave the streets, inspired by negative, even traumatic events. These crises cause a young woman or man to seek the help of family and friends which in turn, lead s to getting help from their schools or other professional caregivers (Karabanow, 2008). The elderly homeless also benefit from contact with relatives and their professional supports, but they are also more likely to maintain newly found housing if they adopted more activities in their lives (Crane & Warnes, 2007). Even though the research is still developing about the ways that homeless individuals and families find stable housing, there are key resources that, when made available, can positively impact this process. Many of the changes oc cur internally, yet professional counselors often serve as the midwife of these realizations and can facilitate renewed connections with family and other community resources. And when there are barriers to accessing these resources,
46 counselors have the op portunity to address them so that there may be greater opportunities to achieve stable housing and a more secure lifestyle. Barriers and Enablers to Substance Abuse and Mental Health Service Utilization As recognized in the literature, there are many facto rs which may either facilitate or inhibit ones successful transition out of homelessness. As many who live without stable housing experience high rates of substance abuse and mental health problems, treatment services are certainly needed (Burt et al., 1 999; Tessler, Rosenheck, & Gamache, 2001). Despite this high need, homeless individuals are more likely to use short term residential services instead of utilizing longer term treatment options that may substantially improve their circumstances. In one l arge San Diego study (n = 10,340), the likelihood of homeless patients with serious mental illnesses using crisis residential services were 7.3 times that of similar patients who were housed (Folsom et al., 2005). In contrast, the odds of using outpatient day treatment, or case management services were significantly lower (0.77, 0.56, and 0.59, respectively) than the housed patients (Folsom et al.) As housing is a primary need, it is not surprising that homeless persons would opt to receive services tha t included a residential component. This trend holds true for medical services as well, where excessive emergency room visits have been associated with homeless severity (Stein, Andersen, & Gelberg, 2007). This would suggest similar reasons for such utilization patterns. The utilization literature expresses great concern over the barriers to healthcare and treatment services (Andersen, 1968; Melnyk, 1988; Rosenstock, 1966). Melnyk notes that the concept of barriers to health care was originally introduce d by Rosenstock who defined this as the cost to taking action towards getting healthcare. Perceived costs and benefits work against one another and ultimately determine whether one will utilize services. He notes An individual may believe that a given action will be effective in reducing the threat of disease, but at the same
47 time see that action itself as being inconvenient, expensive, unpleasant, painful or upsetting (p. 100). Andersen proposed a more dynamic and comprehensive model to health care u tilization, looking at the predisposing factors (e.g. demographics) which influence the enabling resources (e.g. health insurance). These underlying factors contribute to whether one will seek healthcare services when a need either real or perceived, ar ises. In his model, enabling factors include both facilitative resources and name the obstructive barriers which prevent one from seeking services. The Andersen (1968) model has been critiqued and revised over the years and later adapted for the homeless p opulation. Addressing the original model, Melnyk (198 8) c riticized it for failing to recognize the subjective nature of health care barriers, which are not necessarily external objective realities for an individual. Later iterations of the Andersen model have recognized perceived barriers within the enabling category, which includes both health promoting and hindering factors (Gelberg, Andersen, & Leake, 2000). Andersens (1968) model was subsequently modified and adapted for vulnerable populations, name ly those who are homeless (Gelberg, Andersen, & Leake, 2000; Stein, Andersen, & Gelberg, 2007). This model added vulnerable domains to each of the three traditional categories (Figure 2 1). For example, ones history of homelessness would be a predisposi ng factor which may increase the likelihood that other competing factors such hunger and other unmet needs will be present. These conditions, in turn, can influence the perceived need for health care. Of the three categories named in the Behavioral Model for Vulnerable Populations (Gelberg, Andersen, & Leake, 2000), the enabling factors are the most mutable and are therefore a target for programmatic and public policy changes (Andersen, 1995). As predisposing factors such as demographics and need factors a re often difficult to alter, the primary focus is directed
4 8 on those resources and barriers within the enabling category (Andersen). However, both predisposing and need factors need to be considered as they are related to the enabling factors For instanc e, poor quality substance abuse and mental health services may be related to staff biases against those of a different cultural background. Therefore, it is important to study all of these factors and their relationships to inform the source of barriers a nd how they may be changed. The following sections will outline the current research on each of the models primary categories: predisposing factors, enabling factors, and need factors, as they relate to the homeless population. Subsequently, there will be a review of the existing studies that use the factors of this model to examine these factors and suggest ions for the next research steps. Predisposing Factors There are many individual characteristics that predispose someone to be more or less likely to use substance abuse and mental health services. These factors may not be the cause of less service utilization, however, they can be indicators of the interactions of individual, family, and societal issues that lead one to get care. It has been shown t hat certain demographic characteristics among those who are homeless promote service utilization. Being female, younger, having more education, not living with children and being married are all demographic characteristics that predict higher service util ization (Bonin, Fournier, & Blais, 2007; Lemming & Calsyn, 2004; McCarthy et al., 2007; Sosin & Bruni, 2000). Among homeless veterans, being African American predicts higher service utilization ( Gamache, Rosenheck, & Tessler 2000). This finding contrast s with the general homeless population, where being African American actually predicts lower service utilization ( Lemming & Calsyn 2004). Homeless women are also more likely to use outpatient services than men (Kertesz et al., 2006). Gender and child ca re needs may be indicative of larger societal influences as well as programmatic deficiencies that need to be addressed.
49 Enabling Factors There has been more attention to the area of enabling factors, and in particular barriers. In her operationalizing of healthcare barriers, Melnyk (1990) found that those factors related to the provider -consumer relationship were the most salient among those she surveyed. This finding parallels the common factors theory, which also recognizes the primacy of the client -co unselor relationship (Wompold, 2001). Difficulties with homeless service providers can erode trust in this relationship, an essential component to counseling (Bhui, Shananhan, & Harding, 2006; Thompson, McManus, Lantry, Windsor, & Flynn, 2006). Studies h ave shown the importance of counselors and other service providers in facilitating improved service utilization among those who are homeless (Lemming & Calsyn, 2004; Rothbard, Min, Kuno, & Wong, 2004; Sosin & Bruni, 2000; Wilde et al., 2004). Homeless youth also report that they are attracted to service providers who are caring, motivating, respectful, and pet -friendly (Thompson, McManus, Lantry, Windsor, & Flynn) Sometimes the relationship issues can be related to programmatic and even societal issues. Homeless persons have complained of delayed treatment and high staff turnover, both factors that may be related to low funding or poor management (Bhui, Shananhan, & Harding, 2006; Wilde et al., 2004). Homeless clients may begin to feel like industrializ ed products instead of human beings. The problems can occur between agencies as well. With a fragmented system, consumers may find it difficult to negotiate ; therefore t hey therefore rely on emergency services, if any services at all (Swick, 2005). An in itiative in the 1990s called Access to Community Care and Effective Services and Supports (ACCESS) was developed to improve linkages between service providers and facilitate a more effective use of resources among those who are homeless (Rothbard, Min, Kun o, & Wong, 2004). These linkages were shown to increase under this
50 system, decreasing the number of days in inpatient psychiatric care and increasing outpatient utilization (Calloway & Morrissey, 1998; Rothbard, Min, Kuno, & Wong). Furthermore it was sh own that the ACCESS program facilitated homeless individuals use of outpatient services within 30 days of discharge from an inpatient setting (Rothbard, Min, Kuno, & Wong). Another shift being explored among service providers is the movement away from a bstinence requirements to receive housing as well as mental health care. When housing is provided separately in a housing first model, clients are able to access treatment and other social services while living in a residence without overcoming the oft en difficult abstinence hurdle (Tsemberis, Moran, Shinn, Asmussen, & Shern, 2003). The results of one such program, called Pathways, revealed that those given housing without abstinence requirements utilized addiction treatment services at significantly lower rates when compared to a control group; however, substance use rates did not differ (Tsemberis, Gulcur, & Nakae, 2004). In essence, although the more restrictive program policies facilitated higher utilization, these individuals coerced by housing did not necessarily change their substance usage over the long term. The housing first model, however, did provide an active multidisciplinary team of nurses, social workers, psychiatrists, substance abuse and vocational counselors for everyone who entere d the program (Tsemberis, Gulcur, Nakae). When the abstinence barriers were removed, mental health and other social services were still available. Housing also enables someone to have physical security as well as a permanent address to receive mail, suc h as medical notices and other important documents (Swick, 2005). All of these may support greater utilization of outpatient services which require making appointments and future planning. Another similar program entitled Choices Unlimited provided me ntal health and other social services without abstinence requirements and found higher substance abuse and mental health service utilization
51 than comparable drop in centers that required abstinence (Tsemberis et al., 2003). Even though abstinence may be p erceived as facilitative in getting those who are homeless into treatment, it can actually serve as a barrier. Organizational structures may also impact the utilization of certain programs. Among substance abuse programs, it has been found that agencies w ith fewer types of services, higher ratios of professional staff, and diverse funding sources are more likely to be used by those who are homeless ( North, Pollio, Perron, Eyrich, & Spitnzgel, 2005 ). Interestingly, mental health programs with fewer funding sources that are larger in size, and also have higher ratios of professional staff are utilized more frequently (North, Pollio, Perron, Eyrich & Spitnzgel). More research is needed to determine why some of these factors are different between mental healt h and substance abuse services. The only consistent factor is the ratio of professional staff; higher ratios would imply better trained service professionals to meet the needs of the clients. While socioeconomic status is considered a predisposing fact or, income itself is an enabling one in the Behavioral Model for Vulnerable Populations (Gelberg, Andersen, & Leake, 2000). Likewise, insurance and other benefits such as Medicaid and Medicare are often viewed as enabling factors. Conversely, the lack of income or benefits would be a barrier to care. This appears intuitive, but research has shown contradictory results. In some cases, having access to income and welfare benefits is associated with decreased service utilization (Sosin & Bruni, 2000). It is hypothesized that when clients have other resources, the restrictive nature and time commitment of treatment is too costly (Sosin & Bruni). In some cases, o ther priorities are greater than mental health or substance abuse treatment (Wilde et al., 2004) Furthermore, those who are working while homeless may be less inclined to take time off in order to make an appointment or leave their position to enter a treatment program (Sosin & Bruni). It is not the
52 income that is a barrier; rather, it is the less accommodating nature of available programs. In other cases, health insurance (Medicaid) has correlated with increased utilization (Kertesz et al., 2006). However, with a closer look, it appears that service utilization may increase Medicaid coverage as hospital case workers get their patients enrolled after admission. Subsequent utilization after enrollment is an area that needs further study Research has shown how important a clients social system is to utilization, as well. Family members, friends, employers, clergy and other spiritual caregivers can all play an important part in helping a homeless individual get substance abuse or mental health care. Alternatively, they may also serve to impede ones seeking help. Some have reported the stigma th ey have experienced being homeless and having a mental illness (Bhui, Shananhan, & Harding, 2006). By disclosing the need for psychological and housing services, they become associated with two highly stigmatized groups in American society, the mentally ill and the homeless. Therefore, having a large social support system is not enough to encourage utilization; it is the quality of these supports (Lam & Rosenheck, 1999). When homeless individuals are able to overcome these social biases and receive p ositive supports, they are more likely to utilize treatment services as well as mutual -help support groups (Kertesz et al., 2006; Lemming & Calsyn, 2004). Need Factors In the Behavioral Model for Vulnerable Populations (Gelberg, Andersen, & Leake, 2000), b oth perceived health needs and evaluated health needs are important. Evaluated health needs are those diagnosed by a medical professional and shared with the patient. Perceived health needs are from the perspective of the patient. Evaluated health needs can influence perceived needs, and are important to consider, yet perceived needs will ultimately determine whether one seeks services in the majority of non-mandated cases. In most studies, the evaluated health
53 needs are assumed by a clients presence i n a substance abuse or mental health program where they are likely to receive a diagnosis (Kertesz et al., 2006; McCarthy et al., 2007). Perceived needs are less frequently evaluated among studies of those who are homeless One study used the Stages of C hange Readiness and Treatment Eagerness Scale (SOCRATES; Miller & Tonigan, 1996) to measure perceived needs This scale wa s used to determine how ready a client is to make changes in their alcohol and drug usage. None of the studies found for this chapte r inquire about attempts, even if unsuccessful, to receive services. Applying the Behavioral Model for Vulnerable Populations to Homeless Individuals Seeking Mental Health and Substance Abuse Care Even with the number of correlative factors identified, the re are only a few studies which have looked at the relationships between predisposing, enabling, and need factors and how they contribute to mental health and substance abuse service utilization among those who are homeless (Lemming & Calsyn, 2004). The ea rliest of these focused on social supports and the relationship to individual factors (Lam & Rosenheck, 1999). A second one used a broad social situational perspective more similar to Rosenstocks (1966) cost -benefit utilization model (Sosin & Bruni, 2000). The four later studies framed their analysis with the original Andersen (1968) model (Gamache, Rosenheck, & Tessler, 2000; Lemming & Calsyn, 2004; McCarthy et al., 2007), or the related Behavioral Model for Vulnerable Populations (Kertesz et al., 2006). Each is reviewed below. With a focus on the correlation between social supports and utilization, Lam & Rosenheck (1999) used data from the ACCESS program, which sought to better integrate social services to those who are homeless Only those who rec eived case management services were analyzed for the study (n = 1,393). Those excluded from the study (n = 4,403) did not participate because of a lack of interest, being ineligible, or because there were not enough case management positions
54 open. In the first phase of the analysis, the researchers found that young people, women, African Americans, those with less time homeless, those in better health, and drug abusers tended to have larger social supports. In the second phase, they controlled for the pr edisposing factors of age, race, education, duration of homeless episode, childhood abuse, quality of life, and childhood stability, and found that social supports did not impact psychiatric service utilization. They found perceived availability of resour ces from supports to be the only factor predict ing substance abuse service usage when these other factors were controlled. Neither the number of persons in ones network nor the amount of contact with these persons predicted utilization of services. A num ber of issues limit the conclusions of this study. First, there is a substantial amount of people who do not participate in this study. By its own report, there were some excluded because of a lack of case managers to meet the need, a noted barrier to care. Additionally, those who refused the intervention were not surveyed to determine if they were substantially different from the treatment cohort. It is likely that some refused the intervention for reasons related to perceived barriers to treatment (e. g. perceived lack of efficacy). It would also improve the study to analyze the quality of the support systems available to those who are homeless If the network is supportive of mental health or substance abuse services, this could make a substantial di fferent in comparison to a network that is not supportive. A problem in their analysis was that they did not look at the relationship between client characteristics and utilization, using social supports as a mediating variable. This type of regression a nalysis may have shown greater significance between the variables of interest. Furthermore, the study could have followed numerous other enabling factors t o see their relative impact. In a later study, Sosin and Bruni (2000) questioned the relative influe nce of personal verses environmental factors on service utilization by analyzing the difference between treatment
55 acceptors (n = 96) and treatment rejecters (n = 181) in a Chicago -based intervention. As clients left an intensive in-patient setting, a sing le recruiter offered them a case -management program which included substance abuse counseling, support resources, and advocacy. Participants were administered a number of instruments including the Addiction Severity Index (ASI), the Personal History Form, and the General Health Questionnaire (GHQ), measures which the study cites as reliabil e with this population. Both the ASI and the GHQ may be viewed as need variables used to determine mental health and substance abuse problems. In addition, the study i nquired about military experience, childhood out -of -home placement, marital status, former foster care placement, and general disaffiliative beliefs, all of which may be considered predisposing factors. Enabling factors included adding recreational servic es to treatment, having the recruiting case manager explain the efficacy of the program, having other sources of care which are less demanding (barrier), welfare (in this study considered a barrier), and access to an automobile. In addition, the study inq uired about conflict with family and friends and whether they had received recent employment help. In logistical regression analyses, the study found that none of the predisposing factors contributed to treatment rejection and that many of the enabling fac tors did. The greatest enabling factor was having a case manager fully explain the services. Offering recreation services acted as a barrier, reducing utilization, as did having conflict with their social network. Access to income and other less demandi ng treatment appeared to divert these clients from these addiction treatment services, and having access to an automobile had no statistically significant effect. The authors note that their findings question the assumptions that personal attributes are a significant reason for service rejection.
56 One major factor that was not explored was the perceived relationship between the recruiting case manager and the prospective clients. As this relationship is an essential component to service utilization (Melnyk, 1990), it could have added much to their conclusions. When the recruiter provided more information about the programs efficacy, prospective clients may have experienced higher levels of care and concern from her. By only having one case manager to recruit, it is uncertain how much of the rejection was associated with her delivery and bias for or against certain components of the program. Furthermore, as they only measured rejection rates for two phases of the same program, it is difficult to ascertain what other parts of the program were unappealing to these prospective clients. An evaluation of barriers and other enabling factors across various settings would provide a better understanding of the full experience of clients in seeking services. Furthermore, by only interviewing those who have already received treatment, the study is missing a number of people who never entered treatment in the first place. Two of the studies which used Andersens (1968) utilization theory focused on veterans. The fir st of these, conducted by Gamache, Rosenheck, and Tessler (2000), examined data from the ACCESS program noted earlier in this chapter. From this large data set, they analyzed a subset (n = 698) and found that for mental health services, serving in Vietnam was the only predisposing factor to predict utilization. Statistically significant enabling characteristics included VA disability benefits and participating in an ACCESS site with VA services nearby. Need factors showed that addiction severity predicte d mental health service usage. For substance abuse services, the Vietnam experience as well as being African American predicted utilization. Being in an ACCESS program site with a nearby VA medical center also predicted utilization, but the other enablin g factors (VA service -connected disability status, income, and
57 VA discharge status) did not. Once again high addiction severity predicted substance abuse service utilization, as expected. Of equal importance in this study is what it did not find. Among t he predisposing factors, marital status, gender, age, and years of education did not predict service utilization among those who are homeless (Gamache, Rosenheck, & Tessler, 2000). As all of the participants in the ACCESS program had a mental health diagn osis, psychological needs are assumed. However, a PTSD diagnosis did not determine the extent of mental health or substance abuse service utilization. As in the previous ACCESS -based study, there are many limitations in the sampling methods in how client s were excluded. And in this analysis, the authors do not concern themselves with utilization rates of non -VA services. Some who are not using VA services may receive care from other locations to meet their needs. It would be helpful to understand overa ll utilization for veterans as well. The second study to track veterans was conducted by McCarthy et al. (2007), who directly used VA administrative data to understand utilization patterns among those diagnosed with a bi polar or schizophrenia disorder. O ne of their primary concerns was to see which factors of the Andersen model (1968) most likely contributed to a 12-month gap in mental health services. Predisposing factors included age, gender, marital status, race/ethnicity, VA service connection, and h omeless status. While this study was not restricted to homeless people this population represented 9.8% of the total sample, a significantly higher percentage than the percentage of homeless (1%) found in the general population. Enabling factors include d geographic accessibility ( measured by distance to VA service centers ) and availability ( measured by VA and non -VA inpatient beds in the county of residence ) As all participants received mental health
58 services at the baseline measure, the need factors a re considered from an evaluated perspective, not the perceived needs from the veterans perspective. The study found that those who were homeless were significantly more likely to have a 12 month gap in services in comparison to the total sample. In fact, 15.35% of the service gap cohort was homeless, where only 8.33% of those without a gap in services were homeless. The remainder of the study did not focus on those who are homeless but its findings may still be relevant. The authors did find that both geographic accessibility and availability significantly contributed to utilization rates. Given the limited means of transportation among many homeless, distance may be an even greater barrier. The study also found that the other predisposing factors, be ing male, European American, married, younger, and having a VA service connected disability all contributed to increased service utilization. Without further analysis, it is uncertain whether these differences remain with the homeless population alone. On ce again, a major limitation to this study is the fact that all of the participants had already received mental health care from the VA system. Those who never received services but have mental health needs were excluded. It also narrowly focused on two enabling factors, which is informative, however, there are many other enabling factors that may rank higher in importance. The authors report that these factors were statistically significant, although the effect sizes were small. In another study using A CCESS data (n = 3,855), Lemming and Calsyn (2004) used Andersens (1968) behavioral model to predict case management, and a composite of total services including psychiatric care, substance abuse services, housing, medical, and other service utilization. This longitudinal design followed homeless participants from baseline to a 12 -month follow up, measuring all three categories of variables. For this study, predisposing variables
59 included demographics (gender, ethnicity, and education) as well as diagnost ic variables (schizophrenia, depression, etc.) Enabling factors focused on the following three variables: income, natural support, and social support from a professional (SFP). Natural support was operationalized by a composite score of perceived suppor t from significant others, frequency of contact with those supports, and satisfaction from those supports. As such, they were able to identify the quality and frequency of these perceived supports. Need variables were based on mental health and substance abuse problem symptoms. Using hierarchical multiple regression analyses, they identified that enabling variables were the largest predictor of both case management and total services. Case management services were predicted only by SFP and a diagnosis of schizophrenia. More predisposing variables predicted total service usage (the total of case management, substance abuse, housing, and medical), but in all they only accounted for 2% of the variance in utilization. At baseline, African Americans, those w ith less education, and those without a diagnosis of major depression or bipolar disorder utilized total services less frequently. Enabling variables predicted 8% of the overall variance in total service utilization and included SFP as well as natural sup port. In terms of need variables, those who perceived greater overall needs, had depressive symptoms, and less stable housing were more likely to utilize total services. However, the need variables only accounted for less than 1% of the variance in total service utilization. Contrary to their prediction, those with greater severity in their substance abuse problems utilized services less often. With housing, medical services, and other social services included in the composite dependent variable, it is d ifficult to know how these factors contributed to substance abuse and mental health services alone. The stigma of receiving mental health or substance abuse services
60 are different than the other services, and should be considered separately to obtain grea ter detail. The limits of the ACCESS program are already addressed in this chapter. Both the analysis and measure appear to be appropriate for the study questions. As in other studies, there are many other enabling variables that are not accounted for a nd would contribute much to the study. The final study in this review was conducted by Kertesz et al. (2006) who followed clients from discharge from detoxification programs for a two -year period to determine what type of substance abuse services, if any, they utilized. They used the Behavioral Model for Vulnerable Populations (Gelberg, Andersen, & Leake, 2000) to compare the three categories of contributing factors and analyzed them using logistical regression models for each of the four potential outcome s including: residential inpatient treatment, outpatient care, mutual help groups, and a composite any treatment category. Predisposing categories included gender, race, and age, as well as a measure of chronic homelessness. Enabling factors included h ealth insurance (including Medicaid), being hospitalized (an event where one might receive a referral for treatment), type of social network, and the need for childcare (barrier). They also included cognitive status and other measures of health status, fa ctors that may best be categorized among the predisposing factors. The evaluated need factors were assumed as all participants received care from a detoxification unit. The perceived needs were measured indirectly through an instrument which measures one s stage of readiness to change. The results indicated that those seeking inpatient treatment were more likely to have more drug and alcohol consequences, be more highly motivated to change, be either transitionally or chronically homeless, be younger, have physical health problems, have health insurance, and not to live with their children. Those to utilize outpatient services were similar in that they also had higher odds of having drug and alcohol consequences, being female, having health insurance,
61 and having had an overnight hospitalization. Those using outpatient services were less likely to be transitionally homeless or live in a substance abusing environment. Those using mutual support groups were predicted by drug and alcohol consequences, higher motivation, being European American, having health insurance, not living with children, and living in an abstinence -supporting environment. Causality should be interpreted carefully as each of these substance abuse service environments may contribute to the independent variables. For instance, outpatient services are recommended for persons living in supportive non -substance abusing environments (American Society of Addiction Medicine, 2001). The screening for such persons would increase these characteri stics in this category rather than such persons self selecting these types of services. Yet, as one of the more comprehensive designs, this study is revealing in terms of understanding the factors which contribute to differing levels of care. Like the ot hers, it remains limited to those who have already entered a level of care at baseline and it measures only a sma ll number of enabling factors. An Emancipatory Communitarianism Perspective on Mental Health and Substance Abuse Utilization At the core of An dersens (1968) utilization theory was the concern for equal distribution of health services. He proposed an ideal system of healthcare where social class, race, and income have less impact (Andersen, p. 59) on service utilization, and need based charac teristics have a larger impact. In effect, he was seeking a model that would detail the type of factors that prejudice the American healthcare system from providing adequate care to all individuals. The Behavioral Model for Vulnerable Populations (Gelber g, Andersen, & Leake, 2000) is an extension of these objectives. To these initial efforts a new paradigm is needed to better understand the relationships between factors in the model and promote increased social justice with and on behalf of those who ar e homeless.
62 Emancipatory Communitarianism (EC) was formulated by Prilleltensky (1997) as an aspirational paradigm (p.519) that would promote the development of good individual lives and good communities. He noted the lack of attention to values by other psychotherapeutic paradigms as well as their lack of mutual responsibility within the community context. For this reason, he conceptualized EC which strives for distributive justice and seeks to liberate those who are oppressed under the imbalances of po wer while respecting human diversity (Prilleltensky, 1997, Prilleltensy, 2000). It invites a strengths -based perspective and minimizes the emphasis on diagnostic criteria (Prilleltensky & Nelson, 2002). Similarly, it recognized that even the counselors a nd researchers who are trying to help, may in fact be oppressing those who are already marginalized by not sharing in the decisions of their therapeutic processes (Martn Bar, 1994; Prilleltensky; Prilleltensky & Nelson). In the spirit of liberation psyc hology, EC would suggest the importance of critical consciousness to see the deeper contextual roots that are influencing the quality of life experienced by the oppressed (Freire, 1970/2006; MartnBar). All of these principles have implications when cons idering the ways that homeless individuals utilize mental health and substance abuse services. In the case of barriers, EC suggests that they are largely caused by society, yet without critical consciousness, the individual may tend to blame him or her sel f. As the general public tends to blame homeless individuals for their circumstances (Toro et al., 2007), it is expected that those currently in mental health or substance abuse treatment programs are more likely to share this view of self -blame. It woul d also suggest that the significant barriers will occur within the relationship between homeless individuals and service provider when power imbalances are exercised. When homeless individuals are treated as deficient, their well being is compromised, and many will resist this characterization, and therefore resist treatment.
63 This theory also promotes the awareness of these barriers, as recognized by those who are homeless and those who support them. Such barriers would not be identified through obtuse me ans; rather, individuals w ould be asked directly what primary barriers they experience, and their subjective experiences are deemed important and relevant. In relation to demographic variables, barriers are expected to be different based on ethnicity, gen der, age, sexual orientation, and veteran status, as each of these groups experience societal various forms of social oppression and biases. These values, assumptions, and practices not only guide counseling practices but also research. Critical conscious ness may be invoked through open acknowledgement of societal barriers and subsequent exploration of the connection of these factors to greater societal causes. Additionally, information may be shared with key stakeholders to advocate for changes in the tr eatment service system. In both research and practice, EC provides a guiding framework to better understand the barriers and resources which affect mental health and substance abuse service utilization by homeless individuals. Summary The road to becoming homeless is fraught with many negative events as a result of societal attitudes and priorities, individual characteristics, and natural fate. With a combination of social supports, professional care, and motivation -driven effort, many homeless individual s and families are able to take the steps necessary to exit this compromised situation. With societys emphasis on personal initiative, barriers to care are a less considered reason for someone remaining on the streets or in homeless institutions. Yet, a s enabling factors are essential and the most accessible factor to increasing substance abuse and mental health utilization rates, they are a necessary focus that could benefit the homeless population.
64 Of the current research about barriers to these types of services, there has been some notable progress. To date, there have been many barriers identified by those who are homeless While the list is not exhaustive, the foundation has been established. There has also been a theoretical model on utilization developed for and applied to the homeless population. This has inspired new research on each of the categories of population factors, predisposing, enabling, and need, which contribute to service usage. M uch more is needed. Of the current research studi es none ha s taken a comprehensive perspective on the many barriers that exist and their relative impact on the homeless population. Many barriers contribute to ongoing homelessness through lower service utilization but it is important to determine which ones those who are homeless believe to be the most salient in their lives. Similarly, many researchers have attempted to objectify these factors by measuring indirect indicators rather than direct questions A survey of barriers to treatment would simplify this research and would empower research participants to more directly express their needs. These studies have also lacked a full representation of homeless individuals, instead recruiting from existing programs and emergency detoxification services. Those who rarely, if ever, utilize these services have been excluded. A broader recruiting strategy would include these types of participants as well as those in shelters and living outside of an institutional setting. The population not currently using services are likely the population of greater interest, assuming that many of these services could provide them needed assistance. On a related note, the chronic population needs to be included in such a study, with measures consistent with federal standa rds If the chronic homeless population experiences greater need for services, there should be greater utilization. But they also may be experiencing greater barriers to care which keeps them from using mental health and substance abuse services
65 that may help them exit homelessness. The potential phenomenon needs to be explored in order to determine the roles of the individual, programs and society in inhibiting the use of these services Finally, there is no evidence of the distribution of this informa tion with the homeless population. This information may reach service providers, other researchers, and other s who read professional literature, but it is unlikely to be shared directly with homeless individuals and families. Such information could empow er this marginalized population to see the perspective of other homeless people and the possible contributing factors. Additionally, there is no evidence of sharing this information with public policy makers who could support funding and other initiatives to improve the accessibility and effectiveness of substance abuse and mental health services. With greater information, individuals and institutions alike could be given the opportunity to change, so that those in need of services could obtain them and u se them for their benefit. Recognizing the limitations of current research reveals the need for an emphasis on the fundamental reason for Andersens (1968) utilization model, distributive justice. Marrying the Behavioral Model for Vulnerable Populations ( Gelberg, Andersen, & Leake, 2000) with Emancipatory Communitarianism inspires counseling research and practices that may better serve those who are homeless with respect and a higher level of advocacy. This study has strive n to achieve such goals as detai led in the following chapters.
66 Figure 2 1. The Behavioral Model for Vulnerable Populations. Gelberg, L, Andersen, R. M. & Leake, B. D. (2000). The behavioral model for vulnerable populations: Application to medial care and outcomes for homeless people. Health Services Research, 34, 12731302. Figure 1, p. 1278. Reprinted with permission from Wiley-Blackwell.
67 CHAPTER 3 METHODOLOGY As the utilization literature shows, there is little research on the role of perceived barriers in preventing homeless a dults from seeking care for mental health or substance abuse problems Furthermore, there is no information on the relationship between these subjective barriers, the predis posing factors and need factors. The purpose of this study was to examine the ro le of perceived barriers to mental health and substance abuse services in relation to the predisposing (gender, ethnicity, age, duration of homelessness), enabling (natural and professional supports), and need factors that impact mental health and substanc e abuse service utilization. Subjective barriers to mental health and substance abuse treatment were measured with the Barriers to Treatment Instrument ( BTI; Miller, Hodgkins, Estlund, & Brubaker, 2008) a 54 item instrument used to determine the individua l, programmatic, and societal barriers, as perceived by ch ronically homeless individuals. Treatment status, homeless status, predisposing variables (gender, ethnicity, age, duration of homelessness), other enabling variables (natural and professional supp orts), and service utilization were measured through face valid questions on a general questionnaire. Mental health needs were measured by a single item response indicating a need for services as well as a four point Likert scale rating ones current ment al health condition. Substance abuse n eed variables w ere measured by two factors of the modified version of the Stages of Change Readiness and Treatment Eagerness Scale (SOCRATES; Burrow Sanchez & Lundberg, 2007). This chapter will discuss the methodology that was used in this study to examine the identified research questions. Within are the proposed null hypotheses, research design and relevant variables, population and sample, sampling procedure, instrumentation, and data analyses.
68 Research Hypotheses The following null hypotheses w ere tested in this study: H1 There is no significant relationship between mental health service utilization and category of perceived barriers. H2 The re is no significant relationship between substance abuse service utilizati on and category of perceived barriers. H3 The re is no significant relationship between homeless status (chronic versus non-chronic) and category (individual, programmatic, or societal) of perceived barriers. H4 There is no variance in mental health service utilization in the last 12 months explained by predisposing factors (gender, ethnicity, age, and duration of homelessness), enabling factors (societal barriers, programmatic barriers, individual barriers, natural supports, professional supports), and need factors (self identified need and mental health rating). H 5 There is no variance in substance abuse service utilization in the last 12 months explained by predisposing factors (gender, ethnicity, age, and duration of homelessness), enabling factors (societal barriers, programmatic barriers, individual barriers, natural supports, professional supports), and need factors (recognizing the need to change and taking steps to change). Research Design and Relevant Variables A research design using comparative an d correlation methods w as employed for this study. The variables for each hypothesis are discussed below. Hypothesis 1. T he dependent variable, mental health service utilization w as measured on a dichotomous basis from a direct report of whether or not the participant was currently receiving m ental health treatment s ervices. This measure reflect ed those who ha d overcome mental health treatment barriers. The category of perceived barrier s wa s the independent variable and w as
69 determined by using the thre e subscale s, Individual, Programmatic, and Social/Financial as measured on the BTI (See Appendix A ) Hypothesis 2. The dependent variable, substance abuse service utilization, w as measured on a dichotomous basis from a direct report of whether or not the participant was currently receiving substance abuse treatment services. The category of perceived barriers wa s the independent variable and w as determined by using the three subscale s, Individual, Programmatic, and Social/Financial measured on the BTI. Hypothesis 3 The dependent variable in this hypothesis wa s the category of perceived barrier, which wa s measured as stated above. The independent variable, homeless status was a signifier as to whether a person wa s chronically homeless or not as define d by the Department of Housing and Urban Development (HUD; National Coalition for the Homeless, 2008) Five questions that use the HUD criteria w ere asked in the General Questionnaire to determine homeless status (See Appendix B) Hypothesis 4 The depen dent variable in this hypothesis wa s mental health service utilization which w as obtained by a face valid measure on the General Questionnaire which ask ed if the individual wa s currently receiving mental health counseling or treatment services. There were eleven in dependent variables. The predisposing factors of gender, ethnicity, age, and homelessness status w ere asked from the General Questionnaire. The enabling factors include d individual, programmatic, and societal barriers, factor s measured on the B TI Enabling factors also include d natural and professional support and were measured by the number of people in each category who m the participant believe d would support their receiving mental health services. The need factor s for this hypothesis w ere m easured by two face valid item s on the General Questionnaire asking if the participant believe d that she or he need ed mental health
70 services as well as a question a sking participant s to rate their current mental health ( Andersen 1973). Hypot hesis 5 The dependent variable in this hypothesis wa s substance abuse service utilization which w as obtained by a face -valid self report measure on the General Questionnaire asking if the individual wa s currently receiving substance abuse counseling or treatment servi ces. There were eleven in dependent variables in this hypothesis. The predisposing and enabling factors were the same as Hypothesis 3 and w ere measured in like manner. There are t wo measure s of need for substance abuse services. Recognized need w as meas ured by the AMREC factor of the SOCRATES (Modified/Combined; Burrow -Sanchez & Lundberg, 2007), an instrument related to the stage of readiness for change of both alc ohol and drug use (See Appendix C) A second need variable taking steps to change drinking and/or drug use, w as measured by the Taking Steps factor of the SOCRATES (Modified/Combined) Population and Sample The population to which this study w as generalized wa s the adult homeless population in the United States. To approximate this population, c onvenience sampling methods w ere used, seeking participants through agencies that serve homeless men and women in Gainesville, Florida. According to the Gainesville/Alachua County Office on Homelessness and the Alachua County Coalition for the Homeless and Hungry (200 8 ), there were 1,365 homeless men, women, and children in Alachua County between January 2 4 and 2 5 2008 Of these, 968 were adults living in program s jail, hospitals detoxification units, and noninstitutional settings (streets, cars, t ents, etc.) Among the adult population, 72.9% were male, 52% European American/White, 42. 3 % African American /Black 2.6% Hispanic/Latino, and 1.8 % Native American. Additionally, 3 3 2 % were U.S. Military Veterans and 55. 2 % indicated a mental
71 health and/or substance abuse disability. (Gainesville/Alachua County Office on Homelessness and the Alachua County Coalition f or the Homeless and Hungry, 2008) Procedure Recruiting for the study occur red at both the agency and participant level. Agencies were ident ified from the 2007 Community Resource Guide distributed by the Alachua County Coalition for the Homeless and Hungry ( ACCHH) Furthermore, the principal investigator of this study met with the Executive Director of the ACCHH to find additional contacts a nd garner his support for the project. In addition, two of the investigators attended the ACCHH monthly meeting to solicit agency participation. Of the 1 6 agencies identified in the Community Resource Guide 14 were contacted and 9 decided to participate All agencies expressed interest when contacted. Stated reasons for not participations included significant administrative hurdles and being too small to significantly contribute. One small agency decided to participate after the survey collection was completed. The agencies represented a mix of services offered to the community including emergency shelter, domestic violence shelter and services, residential substance abuse treatment, veteran based housing, pregnancy protection and parenting, faith -base d food and ministry, and outpatient medical and mental health services. Two programs offered services in public gathering spaces as well as in homeless camps. The investigators accompanied outreach workers in these c ases to survey persons. One participa nt took the survey in the public library The two co investigators were v olunteer students, recruited from the Department of Counselor Education at the University of Florida. The lead investigator was a middle aged European American male in the Ph.D. prog ram. Both of the co investigators were younger masters level students, one of African descent and one of Indian descent. These students w ere
72 trained in multicultural awareness through the department and specifically about homelessness and survey implemen tation procedures from the principal investigator All investigators read the publication Homeless conditions in Gainesville: Results from the January 2008, Point -in Time Census & Survey of Alachua County residents without housing (Gainesville/Alachua Cou nty Office on Homelessness & Alachua County Coalition for the Homeless and Hungry, 2008) that described the demographic composition of this population, their self reported causes of homelessness, and other information that will familiarize them with the study participants. They also read Prilleltenskys (1997) article Values, A ssumption s, and P ractices: Assessing the M oral Implications of P sy chological Discourse and Action to gain an understanding of the theoretical foundation of the study. Finally, all inve stigators were required to study the University of Florida Institutional Board (UF -IRB) required readings including the IRB Researcher Responsibilities, Code of Federal Regulations, Belmont Report, and take a HIPPA test. Subsequent to reading th is material, the research team collected surveys for the 2009 Point -in Time Census and Survey to prepare for their own data collection. All participants were approved to collect data by the U F IRB The investigation team wore name tags identifyi ng them wi th the University of Florida and the approved UF IRB number. Upon greeting persons who had not previously been identified as homeless, the investigators introduced themselves to individuals and asked if they knew anyone who may want to take the survey. T he approved UF IRB statement is presented in Appendix F. Each participant w as given an informed consent which provided for a person to withdrawal from the study at any time (See Appendix E ). The informed consent w as read aloud to all participants unless they asked the investigator not to do so. The survey packet include d the General Questionnaire, the Barriers to Treatment Instrument, and the Stages of Change Readiness and
73 Treatment Eagerness Scale (SOCRATES Modified/Combined). Survey administrators con tinued to read the instruments to those participants who are not able to read or write. Participants w ere offered a small in kind donation of socks, toiletries, or other items that did not exceed $ 5 in value. Consistent with the Emancipatory Communitaria n perspective, a substantial portion of the in kind gifts were donated by students taking classes in the Department of Counselor Education who responded to a letter soliciting their support (See Appendix G) All survey results w ere kept secure and report ed only in collective form. The UF IRB approved the study prior to collecting any data. Instrumentation There were two instruments used in addition to the General Questionnaire for this study. The y were the Barriers to Treatment Instrument (BTI) and the Stages of Change Readiness and Treatment Eagerness Scale (SOCRATES Modified/Combined), both of which will be discussed in the following sections. Barriers to Treatment Instrument Subjective barriers to mental health and substance abuse treatment w ere meas ured with the Barriers to Treatment Instrument (Miller, Hodgkins, Estlund, & Brubaker, 2008) a 55 item survey used to determine the individual, programmatic, and societal barriers, as perceived by chronically homeless individuals. Four items on the surve y measure the number of attempts to receive substance abuse and mental health treatment within the last 12 months and over ones lifetime. It also measures 49 barriers identified within the literature each of which are written in personalized statement f orm (e.g. Services were not available where I was living .) and rated on a five -point Likert scale from strongly agree to strongly disagree. A confirmatory item asks the participant to identify whether they identify the greatest barriers to be individ ual,
74 programmatic, or societal. The final item asks participants to indicate the various types of services they have received, or attempted to receive over the last 12 months. The Content validity for the BTI was established by an expert panel which consi sted of profession al s serving those who are homeless from the disciplines of psychiatry, mental health counseling, and research evaluation. Internal reliability was measured in a study of 52 chronically homeless men and women, where the three subscales In dividual P rogrammatic and S ocietal/Financial barriers produced Cronbachs alphas of .87, .74, and .8 6 respectively. Stages of Change Readiness and Trea tment Eagerness Scale The Stages of Change Readiness and Treatment Eagerness Scale (SOCRATES) is curre ntly one of the most cited instruments to measure motivation to change alcohol use (Burrow -Sanchez & Lundberg, 2007). It was originally created by Miller and Tonigan (1996) to improve upon Prochaska and DiClemente (1992) instrument which measured ones ge neral stage of change a transtheoretical concept SOCRATES is a self report instrument that was designed to focus upon problem drinking behaviors specifically ; a later form was designed for problem drug use. A number of versions have evolved since the f irst iteration including a 39 item long form and a 19 item short form (Version 8) the la t ter of which Miller and Tonigan recommend for research purposes. Each item is a statement related to ones perception of his or her own drug use, consequences, and d esires to change. Respondents indicate their level of agreement on a 5 -point Likert scale. In a large sample (n = 1,672) of men and women presenting for substance abuse treatment as part of Project MATCH (Matching Alcoholism Treatments to Client Hetergen eity), Miller and Tonigan (1996) found three factors which emerged, Ambivalence Recognition and Taking Steps The Cronbach alphas for each of these subscale s on the short
75 form were .60, .85, and .83, respectively. The authors of the instrument note its validity with the general population as well as those in treatment settings. Burrow Sanchez and Lundberg (2007) combined Version 8 with Version 8D so that motivation to change alcohol and drug use could be measured together. They compared the three facto r model (Miller & Tonigan, 1996) with a two factor model which combined Ambivalence with Recognition (AMREC ) and deleted four of the 19 questions with lower factor loadings (Maisto et al., 1999) The instrument was administered to homeless adults (n = 338), the majority of whom were waiting for substance abuse treatment services The two factor model provided a better fit than the original three and produced Cronbach alphas of .79 and .78 for the AMREC and Tak ing Steps factor s, respectively. By seeing Am bivalence and Recognition as opposite ends of the same factor the SOCRATES (Modified/Combined) was simplified and the lower internal consistency reflected in the earlier used Ambivalence scale was eliminated. These two factor s are appropriate to measure n eeds in this study. Higher levels of AMREC denotes higher levels of recognition of the problems of ones own alcohol and/or drug use as well as self identification of being an alcoholic and/or drug addict. For the Behavioral Model, p roblem identification is a significant indicator of need ( Andersen & Newman, 1973). This factor is expected to be particularly effective in measuring the need for substance abuse services among those not currently getting care. Higher levels on the Taking Steps factor denotes increased levels of taking steps of any kind to change drug and/or alcohol use. Those currently in treatment settings may no longer have significant problems indicated in the AMREC factor but are expected to score higher in the Taking Steps factor show ing the need for continued substance abuse services.
76 Data Analyses There are 1 7 variables of interest in the f ive hypotheses of this study : 2 measures of service utilization for substance abuse and mental health services; 4 predisposing factors including gender, ethnicity, age, and homeless duration; 6 measures of enabling factors including total perceived barriers perceived societal barriers perceived programmatic barriers, and perceived individual barriers, as measured by the BTI as well as n atural and professional supports ; 2 mental health service need variables including perceived mental health needs and mental health rating; 2 substance abuse need variables including recognition of the need to change substance use and activity to change substance use as measured by the SOCRATES (Modified/Combined); and 1 measure of chronic homeless status, measured by a combination of eight questions to determine if a participant meets the Housing and Urban Development (HUD) criteria for chronic homelessness The firs t step was to provide descriptive statistics for the sample and a ranking of the barriers. Frequency distributions were calculated for all categorical data and mean and standard deviations for continuous data. In addition, Cronbachs alpha w as calculated for the subscales of the BTI and the factors of the SOCRATES (Modified/Combined) as well as the total BTI score to measure reliability of these instruments. Then each of the hypotheses w as tested. The first hypothesis test ed the relationship between me ntal health service utilization and category of perceived barriers and w as analyzed using logistic regression procedures. Mental health service utilization was the dependent variable and societal, programmatic, and individual barriers were the independent variable s The second hypothesis test ed the relationship between substance abuse service utilization and category of perceived barriers and, like Hypothesis 1, it w as analyzed using logistic
77 regression procedures. Substance abuse service utilization wa s the dependent variable and societal, programmatic, and individual barriers were the independent variables. The third hypothesis tested the relationship between category of perceived barrier and homeless status. It w as analyzed using a mixed analysis of va riance procedure. The independent variable homeless status (chronic versus non -chronic) wa s the between group effect and the within group effect wa s the source of perceived barriers (societal, programmatic, and individual) The dependent variable wa s the mean perceived barriers experienced. The fourth hypothesis test ed the relationship between predisposing, enabling, need factors and mental health service utilization. It w as analyzed using logistic regression procedures. Mental health service utilizati on wa s the dependent variable. The independent variables were gender, ethnicity, age, homeless duration societal barriers, programmatic barriers, individual barriers, natural supports, professional supports, perceived mental health needs, and mental heal th rating. The fifth hypothesis test ed the relationship between predisposing, enabling, need factors and substance abuse service utilization. It w as analyzed using logistic regression procedures. Substance abuse service utilization wa s the dependent var iable. The independent variables were gender, ethnicity, age, homeless duration societal barriers, programmatic barriers, individual barriers, natural supports, professional supports, recogni zed need to change substance use, and activity to change substa nce use. For all analyses, the Type I error rate w as set at 0.05 (Dooley, 2001)
78 CHAPTER 4 RESULTS Using the methodology described previously, each of the five hypotheses was tested for statistical significance. This chapter will provide the outcomes of this analysis beginning with a description of the sample demographics and other categorical variables. Subsequently, the outcomes of the instruments used in the study will be reported, including mean score s reliabilities and correlations of the scales w ithin each measure. Finally, the results of the hypotheses will be reported in addition to related post hoc analyses. Sample Description The 145 survey participants represent ed a diverse sample of individuals without housing Of the total, there were hig her percentages of men (66.9%) and ethnicity was largely comprised of African Americans (46.2%) and European Americans (41.4%). The mean age for the sample was 42 with a standard deviation of 11.9 years. Those identifying as Latina or Latino appear to be underrepresented in this sample, yet other demographic variables appear to mirror the national statistics. The full description of these demographic are shown in Table 4 1. The majority of persons in the sample lived in the woods or on the street. Thos e who lived in apartments or houses were in a program that paid for these residences, and thus they are still considered homeless by HUD definitions. Of those indicating other as their residence, open-ended responses included hotel, church, and mot herhood community. Those living in a hotel setting were persons in an emergency housing program offered by the local county housing authority. Table 4 2 provides a full description of the current places of residence. Participants were asked a series of questions to determine if they fit the criteria of chronic homelessness as defined by HUD. To fit this category the individual needed to live unaccompanied (no spouse or children), have a disabling condition, and have either one or more
79 year s of continuo us homelessness or four or more times homeless in the last three years. Of the total sample, 32.4% fit the criteria for being chronically homeless. Table 4 -3 shows the details of each of these criteria. It is possible that these numbers may actually und er represent the chronic population in this sample as some participants may not have considered themselves homeless anymore, now living in an agency setting. Also of interest in this study was the duration of homelessness. The men and women in this sample indicated a mean of 30. 7 months of continuous homelessness w ith a standard deviation of 40. 8 months. Many people who participated in the survey had been homeless for a long period of time. Furthermore, the survey asked how long the participant had lived in the local county. This was a concern as perceived barriers over the last year could have been experienced in other places outside of the sample area As noted in Table 4 4, the majority had been in the county for over a year, with only a little over 20% bein g here for six months or less. Mental health needs were high among the sample, with slightly over half indicating the need for mental health counseling or treatment services (See Table 4 5) Considering the stigma of admitting the need for these se rvices in American culture ( Overton & Medina, 2008), this number may be lower than the actual need. Similarly, w hen person s rated their mental health, nearly half stated that it was either fair or poor. This outcome confirm s the former measure of sel fadmitted need for services. T he correlation coefficient between the two scales (n = 144) reached .579, indicating a negative relationship between the rating and self -identified need. This relationship was statistically significant at the .01 level. Na tural and professional supports for substance abuse counseling or treatment services were relatively similar to the suppo rts for mental health services. To obtain these services,
80 a lmost a third of the participants had no family, friends, or professional s upport s Noteworthy is the fact that two of the original response options were combined for the tally For all four questions related to supports, there were two options that overlapped, 5 6 and 6 or more. T he two were combined to create the categor y 5 or more. Table 4 6 shows the full results. Although the majority of persons in the sample were not receiving mental health or substance abuse counseling or treatment services, over a third were receiving one or both (see Table 4 7) Of the 51 wom en and men receiving these services, 27.5% were residing in a residential treatment setting, 29.4% lived in an apartment or other form of transitional housing, 27.5% stayed in the woods or on the streets, 5.9% were in a shelter, and 16.1% lived in other se ttings as described previously. These results would indicate that a large number of persons were receiving outpatient services from community providers. Those attempting to get mental health or substance abuse treatment services over the last 12 months r eached 4 7 6 % of the total sample with the lifetime rates at 57.2% (see Table 4 8) A s many attempting to get services are not receiving them, there appears to be a significant unmet need, which is consistent with local and national reports (Gainesville/A lachua County Office on Homelessness & Alachua County Coalition for the Homeless and Hungry, 2008; U.S. Conference of Mayors, 2007). Overall, the sample was a diverse representation of the total population of homeless persons, some of whom have had a long history of homelessness as well as mental health and substance abuse service attempts. Gender, ethnicity, age, living conditions, homeless severity, social supports, and mental health needs are all suggested as potential factors that contribute to service utilization. The fact that many persons are attempting but not receiving services supports the hypothesis that some factors are preventing these individuals from using these counseling
81 and treatment services. Two other factors, perceived barriers and ne ed for substance abuse services were measured with validated instruments, which are described in the next section. Outcomes of Instrumentation Perceived barriers w ere measured with the Barriers to Treatment Instrument (BTI), and substance abuse counseling and treatment service needs were measured with two factor s of the Stages of Change Readines s and Treatment Eagerness Scale Version 8, Modified, drug and alcohol versions combined (SOCRATES Modified/Combined). Each of the scales and factors on both instr uments proved to show high reliability with this sample. All three subscales of the BTI Social/ Financial, Programmatic and Individual produced Cronbachs alpha s in excess of .90. Furthermore, the reliability of the AMREC factor of the SOCRATES (Modi fied/Combined) exceeded .90 and the Taking Steps factor nearly reached this threshold (see Table 4 9) Some study participants would, on occasion, not fill out items on the instruments This may have been a result of there not being an option to indicate Not Applicable or simple omissions due to user oversight. In order not to disqualify many of the BTIs small numbers of incomplete items were allowed as follows: Social/Financial no more than 1 of 7 items missing; Programmatic no more than 2 of 17 items missing; Individual no more than 3 of 25 items missing The missing items were replaced using the person mean substitution approach, a method that has shown to provide good representations of missing data and reliabilities on Likert scales where t he number of missing items is 20% or less (Downey & King, 1998). The item on the BTI with the highest nonresponse rate (5.6%) was Most providers would not accept Medicare/Medicaid All others fell within a 5% non response rate (Czaja & Blair, 2005).
82 Because of the small number of items on the SOCRATES (Modified/Combined) factor s with missing items w ere eliminate d from the analysis and subsequent tests of the hypotheses. To determine their intra measure relationships correlation coefficients were cal culated for each scale or factor in comparison to the other within the same instrument. On the BTI, the Social/Financial subscale positively correlated with both the Programmatic and Individual subscale s at the rates of .48 and .38, respectively. Additio nally, the Programmatic and Individual subscale s correlated positively at the rate of .78. All three correlations were significant at the .001 level. When the AMREC and the Taking Steps factors of the SOCRATES (Modified/Combined) were analyzed, they also showed a positive correlation of .55 to one another. This was also significant at the .0 0 1 level. Table 4 10 shows the ranking of individual items on the BTI by mean for all cases. Missing items were excluded individually. Although the lower mean scores represent those statements that reflect the most salient barriers that the homeless community are encountering, their absolute ranking should be viewed with caution. An analysis was not performed to determine whether each item statistically differed from one another. The rankings are provided to offer general trends seen in the community. For those attempting to receive mental health or substance abuse treatment services in the last 12 months (n = 67) m any of the top barriers reflected th ose of the ove rall population ; however, many of the mean scores dropped to lower levels Table 4 11 lists all of the items that have a mean score which falls below 3.00. Many of the top barriers relate to insurance coverage and meeting required co -payments. One item that advanced, when compared to the general population related to not having a safe place to stay to begin treatment. Other needs, the lack of program information, access, waiting lists, and a general sense of providers not caring about
83 those who are hom eless were all top barriers identified by those seeking treatment services. Similarly, the bottom barriers as ranked by their mean scores were comparable between those seeking treatment services and the whole sample population (see Table 4 12) Two items in particular moved lower on the list of barriers programs being to religious and participants not wanting to attend a program where one had to quit using all drugs and alcohol. Although there may be some who feel more strongly about the s e barrier s on average, they were two of the least encountered. A further analysis of the open-ended barriers question in the following section offers insights to other barriers that are impacting the homeless community and need to be considered for future studies. Othe r Barriers Encountered As the BTI was developed for chronically homeless individuals with substance abuse and mental health diagnoses, an open response item was added to the instrument asking for a description of additional barriers that the participant ha d experienced while trying to obtain addiction and/or mental health treatment services over the last 12 months. Of the total, 84 participants responded on this item, 19 of whom indicated that the questions did not apply to them or that they did not encounter any other barriers. An additional 5 responders indicated that they did not have a problem with substances or their mental health. Of the remaining responses, many barriers were named, many of which were noted in the primary survey. The individuals m ay have been emphasizing these points or recognizing them in their own words to reflect slightly different meanings. All of the responses are noted in Appendix D. The following provides a summary of these barriers, categorized by s ocietal, programmatic, and i ndividual barriers to remain consistent with the previous analysis.
84 Societal Barriers Many responses indicated the barriers which appear to be largely attributable to public policy, the local community, or society in general. The lack of funding for programs was mentioned by multiple individuals as was the lack of housing. Not getting benefits including insurance, social security disability insurance, Medicaid, and unemployment were all identified as barriers experienced in the last 12 months. Some indicated that they received benefits from the Veterans Administration and this had helped them. One noted that there needs to be more done for families. This can also be a programmatic issue, but the general nature of the comment is the reason for bein g placed in the societal category. One spoke of broken promises, which could be a societal, programmatic, or individual. She wrote: They need to help more than what they are doing. Told us for years they were trying to help make easier and havent done it yet. This statement could also relate to the continued delay of a One -Stop Center which has been approved by local officials, but not implemented due to resistance of community members who do not want the center placed in their neighborhoods (The Sam e Old Song, 2008). Programmatic B arriers Three major sub -categories were identified, each related to programmatic sources. The first of these was communication. Respondents cited the difficulty in getting information about programs, not having personal contacts that could facilitate getting services, and then maintaining contact with providers. Some others mentioned barriers related to the quality of care they were receiving. One comment stated, Pass the buck to other agencies! a seeming reference t o getting passed from agency to agency without getting help. Others noted bad experiences, concerns over staff competence, structure, and the limited time available that one could stay in a program. Lastly, responses included difficulties in the admissions process which
85 included not having a severe enough addiction and not having enough beds open. Again, some of these could be attributable to societal and individual barriers, but the language of the comments primarily identif ied the program as the primar y source of these barriers. Individual Barriers Barriers identified as primarily individual in nature are those focus ing on the person or family as the source. Once again, the source s and therefore responsibility for addressing the barrier, may include s ociety and programs; however, the language of the comments implied the individual as the greatest source. Among these is the area related to having children. One participant identified the barrier of h ow to take care of me and my daughter and to finally go out and find a job implying a deficit of information and personal resources related to childcare. Another participant recognized her fear of losing her child from the authorities. Many noted the lack of other resources including time, money, transportation, and work to obtain these resources. Those struggling with addictions cited their own denial, the shame and guilt of relapsing, and not knowing how to stop. Mental health barriers included emotional pain, continued symptoms of their diagnosis, a nd the link between their mental health and their addiction. Additionally, physical injuries were noted as another barrier to getting treatment services, a problem that was exacerbated by the lack of medical care caused by the lack of insurance. Luck was another issue named by someone who was able to get into a program. The lack of trust of others was also identified as barriers to care. Lastly, homelessness and the cycle of poverty were both named as barriers to getting care. Money and the cycle of po verty were noted in the comments of one who said: Actually money Not money for the cost of treatment but rather the need to work every day (sic ) through a labor pool, needed to just pay rent to stay just for that night. And then the circle begins again money to stay the night so labor pool that day to get the money to stay the night.
86 These barriers were not included in the BTI and remain essential to the continued dialogue about barriers to counseling and treatment services. Summary of O ther I denti fied B arriers Although many of the barriers noted in this section of the survey had been identified earlier, it offered a more detailed description of the difficulties that individuals face while living without a home. These responses also gave the opport unity to name barriers not listed on the BTI. Family and childcare issues were new and important issues that were not previously named, but were expected as the BTI was designed with the chronic population, who are persons, by definition, without children Other new barriers include not having a severe enough addiction, physical injuries, luck, as well as the general constructs of homelessness and poverty. Each of these will need to be included in studies that investigate perceived barriers in a comprehe nsive manner. Hypotheses Tests Results Five different hypotheses were tested using different methodologies. Hypotheses 1, 2, 4, and 5 were tested with logistic regression analyses and Hypothesis 3 was tested with a mixed analysis of variance of method. The results of each of these are discussed below. Hypothesis 1 Test R esults H0: There is no significant relationship between mental health service utilization and category of perceived barriers. This first hypothesis tested whether any of the three categor ies of perceived barriers (Soci etal Programmatic, or Individual) contributed to mental health counseling and treatment services Using the three subscale s of the BTI, the test was conducted using a logistic regression procedure with mental health counseli ng or treatment utilization as the binary dependent
87 variable. The omnibus test for the logistic regression model did not indicate statistical significance, 2( 3 N = 135) = 6.19, p = 103; however, one coefficient did show significance in the model. A s t his research explores a new area of the utilization research, the model wa s interpreted. Furthermore, as the overall effect size of the model was only .045 (Cox and Snell R2), a larger sample size would be necessary to generate the power necessary to conf irm the null hypothesis (Cohen, 1992). Any significance in the model should therefore be considered cautiously, and replicated with increased participants in future studies to determine its significance in the general homeless population. As seen in Table 4 13, the only coefficient to reach significance was the BTI P rogrammatic subscale (p = .031). The obtained odds ratio (Exp(B)) indicates that f or each point of increased P rogrammatic barriers score (a decrease in barriers ) the odds that the individual will utilize mental health services increases by a factor of 2.05 when controlling for individual and societal barriers The i ndividual bar riers coefficient approached but did not reach significance. Of related interest to the research hypothesis was t he relationship between the most salient category of barrier to mental health counseling and service utilization On the BTI is a single item which asks which is the biggest barrier of the three. A two -way contingency table analysis was conducted to dete rmine whether any of the three sources ( society, programs, or individuals ) were significantly related to ones current utilization of mental health services. The identified primary source and mental health service utilization approached, but was not found to be statistically significant, Pearson 2 (2 N = 1 30) = 5.14, p = 077, Cramrs V = .20. In total, the results of the logistic regression indicate d a reject ion of the null, thus showing partial support for Hypothesis 1.
88 Hypothesis 2 Test Results H0: There is no significant relationship between substance abuse service utilization and category of perceived barriers. To test the relationship between each of the barriers categories and substance abuse service utilization a logistic regression procedure wa s used. The omnibus test for the logistic regression model indicate d statistical significance, 2( 3 N = 135) = 10.93, p = 012. The overall model accounted for 7.8% of the variance in substance abuse service utilization. The individual barriers coeffici ent was the only variable to reach statistical significance ( p = .003). For each increased point on the Individual subscale (decrease in barriers) there was a 63.9% decrease in the odds of currently receiving substance abuse c ounseling or treatment servic es when controlling for programmatic and societal barriers. Interestingly, the Programmatic subscale coefficient approached but did not reach significance. Like the model in Hypothesis 1, with more participants, the increased power may have revealed stat istical significance. Table 4 14 shows all of the variables in the model. Using the single item of the BTI where participants identified the single -most largest source of barriers to treatment a two -way contingency table analysis was conducted to determine whether any of the three sources (society, programs, or individuals) were significantly related to ones current utilization of substance abuse services. The identified source and substance abuse service utilization were not found to be statistically significant, Pearson 2 (2, N = 130) = .14, p = .935, Cramrs V = .03. As before, the single test item did not show significance, but the significance found in the logistic regression analysis indicate d a reject ion of the null, thus showing partial support for Hypothesis 2.
89 Hypothesis 3 Test Results H0: There is no significant relationship between homeless status (chronic versus non-chronic) and category (individual, programmatic, or societal) of perceived barriers To determine if a significant relationship exists between chronic status and the three categories of barriers, a two -way (2x3) mixed analysis of variance was conducted. The independent variables included one between groups variable, homelessness status, with two levels (chronic, non-chronic) and one within subjects variable, barriers category with three levels (societal, programmatic, individual). Table 4 15 shows the mean and standard deviations for each of the combinations of variables. There was not a significant main effect between the chro nic and nonchronic groups ( F 1, 131 = 1.308, p = .255) ; h owever, there was a significant main effect of the barriers category ( F 2, 130 = 19.858, p < .001). P aired sample t tests (2 -tailed) indicated significant mean differences between the Social/Financial and P rogrammatic subscale s t (132) = 4.898, p < .001, as well as the Social/Financial and Individual subscale s t (132) = 4.943, p < .001. T he mean difference between the Programmatic and In dividual subscale s did not reach statistical significance t (1 32) = .798, p = .426. Furthermore, there was not a significant interaction effect between chronic status and the barriers category F 2, 262 = 1.262, p = .277. Similar to Hypotheses 1 and 2, a two -way contingency table analysis was conducted to determine w hether chronic status was significantly related to the single item question where participants indicated what they thought was the biggest source of barriers to treatment (society, programs, or the individual). The identified primary source and chronic st atus were not found to be statistically significant ly related Pearson 2(2, N = 129) = .34, p = .844, Cramrs V = .05.
90 In this case, neither the single item test nor the subscale s showed significance in the tests results, thereby failing to reject the n ull. There was no support for Hypothesis 3. Hypothesis 4 Test Results H0: There is no variance in mental health service utilization in the last 12 months explained by predisposing factors (gender, ethnicity, age, and duration of homelessness), enabling factors (societal barriers, programmatic barriers, individual barriers, natural supports, professional supports), and need factors (self -identified need and mental health rating). Eleven factors including four predisposing factors (gender, ethnicity, age, homeless duration), five enabling factors ( individual barriers, programmatic barriers, societal barriers, natural supports, professional supports), and two need factors (self identified need for mental health services, mental health rating) were analyzed wit h logistic regression procedures to determine their relationship with current mental health service utilization. As there were few participants who identified as Native American/Pacific Islander or Asian, they were combined with those identifying their et hnicity as other to create a single category. Each categorical variable included one dummy variable with which to make a base comparison. By having three categories of ethnicity instead of five, the overall total of coefficients in the model was reduce d from fourteen to t welve. The omnibus test of the logistic regression model showing the relationship between current mental health service utilization and the predisposing, enabling and need factors indicated statistical significance, 2 (12, N = 107) = 44.21, p < .001. The individual coefficients of the model are described in Table 4 16. T he only variable that significantly contributed to the model when controlling for all others was self identified need (p < .01) Those who self identif i ed the need for mental health services increased their odds of utilizing mental health services by 13.2
91 times over those who did not identify this need. In total, the model reflect ed 33.8% of the total variance in mental health treatment attempts (Cox and Snell R2), a large effect size (Cohen, 1992) Although self -identified need wa s the only variable to reach significance, it should be noted that this wa s based on controlling for the other variables in the model W hen they are analyzed separately, other variables ha d a significant independent relationship The Rao statistics as reported in SPSS indicate the variables that are likely to be significant when analyzed independently of the full equation. Table 417 shows how each of the variables found to be significant from the Rao score ( = .05) each contribute d to the likelihood of currently receiving mental health treatment services when analyzed with bivariate regression procedures. This analysis reveal ed that four variables (natural supports, professional support s, self identified need, and ones mental health rating) contribute d significantly to the mental health service utilization when analyzed independently of one another. None of the BTI subscales reached significance. So, although self identified need wa s the only significant factor in the full logistic regression model, it wa s clearly related to other significant contributors to utilization. In total, t he null was rejected, and the role of perceived barriers when controlling for the other factors appear ed to be minimal. Hypothesis 5 Test Results H0: There is no variance in substance abuse service utilization in the last 12 months explained by predisposing factors (gender, ethnicity, age, and duration of homelessness), enabling factors (societal barriers, p rogrammatic barriers, individual barriers, natural supports, professional supports), and need factors (recognizing the need to change and taking steps to change).
92 Eleven factors including four predisposing factors (gender, ethnicity, age, homeless duration ), five enabling factors ( individual barriers, programmatic barriers, societal barriers, natural supports, professional supports), and two need factors ( AMREC and Taking Steps factors of the SOCRATES Modified/Combined) were analyzed with logistic regress ion procedures to determine their relationship with current substance abuse service utilization. As in Hypothesis 4, ethnicity was combined into three categories, thus creating an overall total of t welve coefficients in the model. The omnibus test for t he logistic regression model indicated statistical significance, 2( 12, N = 94) = 41.051, p < .001. The model accounted for 35.4% of the variance in substance abuse service utilization (Cox & Snell R2), a relatively large effect size (Cohen, 1992). As se en in Table 4 1 8 only the need variable recognition contributed significantly to the model predicting substance abuse counseling and treatment utilization. Both professional supports and individual barriers approached, but did not reach significance. For each point increase on the AMREC factor the likelihood of utilizing substance abuse counseling or treatment increased by 9.9%. The Rao scores from SPSS on this model indicated that many factors contribute to substance abuse treatment attempts when lo oking at each independently. In addition to natural supports, professional supports, and the two need factors, individual barriers also reached significance according to its Rao score However, when the bivariate regression was conducted, individual barr iers slightly exceeded significance ( = .05), reaching .051. The other four factors are noted in Table 4 19. Once again, the model showed significance with only recognized need playing a significant direct role in this outcome. Therefore, the null was rejected but there was limited support as to the role of barriers in this model
93 Post -Hoc Analys e s Subsequent to these finding, some additional post hoc analyses were conducted to explore more fully the role of total perceived barriers among homeless ind ividuals. Building on the outcomes of the original hypotheses it became of interest to investigate the differences between those who were attempting to get services and those who were not. The research question was whether each type of barrier, when con trolled for predisposing, enabling, and need factors, impacted the behavior of seeking services instead of the outcome of obtaining the services. In effect, did these barriers deter individuals from trying to get services in the first place? To understan d this relationship, Hypothesis 4 and Hypothesis 5 were modified, respectively changing the dependent variables to mental health and substance abuse treatment attempts in the last 12 months. The new hypotheses were as follows: H 6 :There is no variance in m ental health treatment attempts in the last 12 months explained by predisposing factors (gender, ethnicity, age, and duration of homelessness), enabling factors (societal barriers, programmatic barriers, individual barriers natural supports, professional supports), and need factors ( self identified need and mental health rating ). H 7 :There is no variance in substance abuse treatment attempts in the last 12 months explained by predisposing factors (gender, ethnicity, age, and duration of homelessness), enab ling factors (societal barriers, programmatic barriers, individual barriers natural supports, professional supports), and need factors (recognizing the need to change and taking steps to change) All of the new hypotheses were analyzed utilizing logistic regression procedures as in the original hypotheses with each of the independent variables measured as before. The three independent variables, societal, programmatic, and individual barriers were measured with the
94 three subscales of the BTI. The depende nt variables measuring treatment attempts were also obtained from the BTI. The results are shown in the next s ections. Hypothesis 6 Test Results H0: There is no variance in mental health treatment attempts in the last 12 months explained by predisposing f actors (gender, ethnicity, age, and duration of homelessness), enabling factors (societal barriers, programmatic barriers, individual barriers, natural supports, professional supports), and need factors (self -identified need and mental health rating). The omnibus test of the logistic regression model showing the relationship between mental health treatment attempts in the last 12 months and the predisposing, enabling and need factors indicated statistical significance, 2(1 2 N = 105 ) = 64.701, p < .001. T he individual coefficients of the model are described in Table 4 2 0 When the three barriers subscales were used in the model, all three significantly contributed to attempts as did self identified need. For each point of increase o n the Social/Financial and P rogrammatic barriers score s (decreased barriers) there was a respective 49.4% and 76.8% decrease in the odds of attempting mental health utilization when controlling for all other variables Alternatively, for each increase in points on the Individual barriers score (decreased barriers), a homeless individuals was 3.74 times more likely to attempting these services, controlling for all other variables. In terms of need, those who self identify the need for mental health services increase d their odds of seeking this type of help by 66.4 times when controlling for all other variables. In total, the model reflect ed 46. 0 % of the total variance in mental health treatment attempts using the Cox and Snell R2. T he Rao score from SPSSs first iteration of the logistic regression model indicated that the variables likely to be significant when analyzed independently include d societal barriers,
95 professional supports self identified need and ones mental health rating. Table 4 2 1 shows how each of these identif ied variables contribute d to the likelihood of currently receiving mental health treatment services in a bivariate logistic regression analysis. In total, the role of barriers appeared more prominently than before and the null was rejected. Not all varia bles were significant, so there was only partial support for Hypothesis 6 Hypothesis 7 Test Results H0: There is no variance in substance abuse treatment attempts in the last 12 months explained by predisposing factors (gender, ethnicity, age, and duration of homelessness), enabling factors (societal barriers, programmatic barriers, individual barriers, natural supports, professional supports), and need factors (recognizing the need to change and taking steps to change). When conducting the omnibus test for the logistic regression model which shows the relationships between substance abuse treatment attempts and the predisposing, enabling, and need factors, there was statistical significance, 2 (1 2 N = 9 4 ) = 40.210, p < .001. As seen in Table 4 2 2 t wo o f the coefficients in this model contributed significantly to the dependent variable when controlling for all other variables; they were individual barriers and the recognized need for substance abuse treatment For each point of an increase on the Individual barriers score (indicating lower barriers), the odds of attempting to receive substance abuse treatment services decreased by 58. 3 % when controlling for all other variables Lower barriers indicated a reduced likelihood that services would be attemp ted. With recognized need, for each increasing point on the AMREC factor there was a 9 0 % increase in the odds that one would attempt to receive substance abuse treatment services when controlling for all other variables in the model. The Rao scores from SPSS on this model indicate d that many factors contribute d to substance abuse treatment attempts when looking at each independently. The Table 4 2 3 shows
96 the results of a bivariate analysis for each variable with a significant Rao score ( = .05) Inter estingly, gender, individual barriers both types of support, and both need variables prove d to be significant independent of one another in predicting those who are likely to attempt to receive substance abuse treatment services Those who were male wer e less likely to be attempting to receiv e substance abuse treatment services than were women Those with more supports and those with more need showed an increase in odds for seeking services. The significance of the model and at least two variables indi cated a rejection of the null and partial support for Hypothesis 7 In total, both perceived barriers and need play a significant role in predicting mental health and substance abuse service utilization. Hypothesis 6 showed a significant relationship betw een each of the three categories of barriers as well as self identified need and mental health treatment attempts when controlling for all other variables in the total model The outcomes of Hypothesis 7 revealed that individual barriers and recognized n eed significantly contributed to the total model of substance abuse treatment attempts To summarize the original and the post hoc analyses, there were many factors that significantly contributed to utilization and attempts to receive services. I t was det ermined that mental health service utilization is significantly determined by self -determined need in the total model; however, programmatic barriers, natural supports, professional supports, and ones mental health rating were found to have a contributing role in other analyses. Substance abuse service utilization was determined by recognition of need in the full model, with other analyses showing individual barriers, natural supports, professional supports, and need as determined by taking steps to chang e drinking and/or drug use all having a contributing role In terms of m ental health treatment attempts self identified need societal, programmatic, and individual barriers all
97 significantly predicted this outcome. Bivariate analyses revealed that natu ral supports, professional supports, and ones mental health rating all were contributing factors S ubstance abuse treatment attempts were determined significantly by recognition of need as well as individual barriers. The bivariate analyses showed that gender, natural supports, professional supports, and taking steps to change drinking and/or drug use were also significantly related to substance abuse treatment attempts The analyses did not find a significant relationship between the categories of barr iers and chronic status and there was not indication that duration of homelessness significantly contributed to either mental health or substance abuse service utilization, or mental health or substance abuse treatment attempts. Th ese findings and all of the results will be discussed in the next chapter.
98 Table 4 1. Demographics: gender and ethnicity Frequency (f) Percentage (%) Gender Male 97 66.9 Female 47 32.4 No response 1 .7 Ethnicity African American/Black 67 46.2 European American/White 60 41.4 Latina/Latino 3 2.1 Native American/Pacific Islander 9 6.2 Asian American 1 .7 Other 5 3.4 Table 4 2. Current place of residence Frequency (f) Percentage (%) Apartment/House 8 5.5 Car or v ehicle not intended for housing 3 2.1 Tent or self made shelter/Street/Woods 83 57.2 Emergency shelter 13 9.0 Transitional housing 13 9.0 Substance abuse treatment facility 10 6.9 Mental health treatment facility 4 2.8 Other 11 7.6 Table 4 3. Chro nic criteria variables Frequency (f) Percentage (%) Living without partner/spouse 113 77.9 Living without children 128 88.3 Four or more times homeless in last three years 41 28.3 One year or more of continuous homelessness 79 54.5 Physical disab ility 70 48.3 Mental health disability 51 35.2 Substance abuse disability 37 25.5 Total Chronic (meet criteria) 47 32.4
99 Table 4 4. Time in Alachua County Frequency Percentage Cumulative Cumulative (f) (%) (f) (%) 0 to six months 31 21.4 31 21.4 Seven to twelve months 11 7.6 42 29.0 Over one year 103 71.0 145 100.0 Table 4 5 Mental health need variables Frequency Percentage Cumulative Cumulative (f) (%) (f) (%) Need counseling or treatment services Yes 73 50.3 73 50 .3 No 72 49.7 145 100.0 Mental health rating Poor 26 17.9 26 17.9 Fair 43 29.7 69 47.6 Good 38 26.2 107 73.8 Excellent 37 25.5 144 99.3 No response 1 .7 145 100.0 Table 4 6. Number of persons willing to support Natural Supp orts Professional Supports Frequency Percentage Frequency Percentage (f) (%) (f) (%) Mental health counseling or treatment None 43 29.7 48 33.1 1 2 36 24.8 35 24.1 3 4 14 9.7 25 17.2 5 or more 48 33.1 34 23.4 No response 4 2.8 3 2.1 Substance abuse counseling or treatment None 51 35.2 50 34.5 1 2 29 20.0 35 24.1 3 4 13 9.0 25 17.2 5 or more 47 32.4 31 21.4 No response 5 3.4 4 2.8
100 Table 4 7. Current mental health and substance abuse service utilization Fre quency Percentage (f) (%) Are you currently receiving professional mental health counseling or treatment services? Yes 42 29.0 No 103 71.0 Are you currently receiving substance abuse counseling or treatment services? Yes 33 22.8 No 112 77.2 Individuals receiving either professional mental health and substance abuse counseling or treatment 51 35.2 Individuals receiving both professional mental health and substance abuse counseling or treatment 24 16.6 Table 4 8. A ttempts for substance abuse and mental health treatment services Prior 12 months Lifetime Frequency Percentage Frequency Percentage (f) (%) (f) (%) Mental health treatment services Yes 48 33.1 66 45.5 No 92 63.4 76 52.4 No response 5 3.4 3 2.1 Substance abuse treatment services Yes 43 29.7 63 43.4 No 99 68.3 79 54.5 No response 3 2.1 3 2.1 Attempting either MH or SA tx services 69 47.6 83 57.2
101 Table 4 9. Measurement reliabilities for BTI and SOCRATES (Modifie d/Combined) Scale/Factor N Minimum Maximum Mean SD Cronbachs Alpha BTI Social/Financial 138 1.00 5.00 3.08 1.39 .912 Programmatic 140 1.00 5.00 3.60 1.05 .918 Individual 139 1.00 5.00 3.70 1.09 .954 SOCRATES (Modified/Combined) AMREC 128 7 .00 35 .00 20.48 9.54 .911 Taking Steps 132 6 .00 30 .00 20.59 7.97 .895 Note: BTI = Barriers to Treatment Instrument; SOCRATES (Modified/Combined) = Stages of Change Readiness and Treatment Eagerness Scale, Version 8, Modified, drug and alcohol versions combined. Cronbachs alpha and SOCRATES (Modified/Combined) statistics includes only scales and factors with all items completed. BTI statistics (except alpha) include scale/subscales with missing items as follows: Social/Financial no more than 1 of 7 items missing; Programmatic no more than 2 of 17 items missing; Individual no more than 3 of 25 items missing Table 4 10. Ranking of barriers by mean for all cases Item (1 = Strongly Agree, 5 = Strongly Disagree) Mean SD Services and medicines were difficult to get because I couldnt afford the co payment. 2.67 1.77 It was hard to get services because I didnt have and couldnt get health care insurance. 2.81 1.75 I did not get medications because I did not have insurance coverage. 2.84 1.77 My o ther needs (like housing and food) were more important than treatment. 3.10 1.76 I didnt know about the services that were available to me. 3.11 1.66 I lost my health care insurance (e.g. Medicaid) so I didnt seek services. 3.17 1.75 Service providers do not care about the needs of homeless people. 3.23 1.66 I had to wait too long to get services. 3.24 1.65 I did not get mental health or addiction services because my insurance did not cover it. 3.24 1.80 I had no safe place to stay in order to beg in treatment. 3.27 1.80 Most providers would not accept Medicare/Medicaid. 3.27 1.69 I did not want to give up control of my own life. 3.36 1.78 I couldnt get transportation to get services. 3.38 1.71 Case workers change jobs so much that I was tired of having to start over with someone new. 3.40 1.63 I had no one to help me access treatment. 3.41 1.62 I kept relapsing. 3.50 1.73 I thought most program providers were too overworked to listen to my needs. 3.51 1.53
102 Table 4 10 Continued Item (1 = Strongly Agree, 5 = Strongly Disagree) Mean SD I enjoyed the lifestyle of using drugs and/or alcohol and didnt want to give it up. 3.53 1.72 Services (like detox, residential treatment) were always full. 3.53 1.60 Services were not available where I wa s living. 3.54 1.68 I didnt want to stop using drugs and/or alcohol. 3.54 1.69 The program would not let me work while I was in the program. 3.56 1.60 I was so discouraged and frustrated trying to find treatment that I gave up. 3.58 1.62 I thought ad diction and mental health programs were too strict and overly structured. 3.58 1.56 I did not want to take medications. 3.61 1.60 I had a previously bad experience with the service provider. 3.66 1.57 I thought that treatment programs required that I wo uld have to take mental health medications before entering. 3.69 1.57 I didnt try to get addiction treatment because I didnt think it would work. 3.71 1.59 I didnt want to admit to having a problem with alcohol and/or drugs. 3.71 1.61 I tried treatme nt before and it never worked on me. 3.72 1.56 I thought the program would make me feel bad about myself. 3.74 1.58 I didnt want anyone to find out that I was addicted. 3.75 1.64 I did not trust doctors, therapists and case managers. 3.75 1.49 My add iction to drugs and/or alcohol was too severe. 3.75 1.63 Programs shouldnt ask me to stop using alcohol and drugs. 3.79 1.70 I was not sure I had a mental health problem. 3.81 1.56 I didnt want to admit to having a problem with my mental health. 3.81 1.59 I didnt want anyone to find out that I had a mental health problem. 3.84 1.60 I didnt try to get mental health treatment because I didnt think it would help me. 3.85 1.47 I was too intoxicated to seek treatment. 3.91 1.60 I didnt want to atte nd a program where I had to quit using all drugs and alcohol. 3.91 1.59 I believed treatment programs were too religious. 3.92 1.46 I didnt think going to a program was safe. 3.92 1.54 I was not sure I had a problem with alcohol and/or drugs. 3.92 1.51 I didnt want to get help for my addictions and/or mental health problems. 3.93 1.49 I was afraid they would find out I had legal charges against me. 3.95 1.53 I didnt go because I thought I had to be able to read and write when I got there. 3.96 1.51 I was too mentally ill to seek treatment. 3.99 1.50 I kept losing my medication. 4.25 1.29
103 Table 4 11. Top barriers to treatment for persons seeking services Item (1 = Strongly Agree, 5 = Strongly Disagree) Mean SD Services and medicines were diffi cult to get because I couldnt afford the co -payment. 2.32 1.61 I did not get medications because I did not have insurance coverage. 2.49 1.64 It was hard to get services because I didnt have and couldnt get health care insurance. 2.51 1.67 I had no safe place to stay in order to begin treatment. 2.64 1.76 I lost my health care insurance (e.g. Medicaid) so I didnt seek services. 2.76 1.68 Most providers would not accept Medicare/Medicaid. 2.81 1.67 I did not get mental health or addiction service s because my insurance did not cover it. 2.81 1.77 My other needs (like housing and food) were more important than treatment. 2.84 1.75 I didnt know about the services that were available to me. 2.87 1.63 I had no one to help me access treatment. 2.94 1.56 I had to wait too long to get services. 2.94 1.67 Service providers do not care about the needs of homeless people. 2.96 1.68 Table 4 12 Bottom 5 barriers to treatment for persons seeking services Item (1 = Strongly Agree, 5 = Strongly Disagree ) Mean SD I didnt want to get help for my addictions and/or mental health problems. 3.91 1.46 I didnt want to attend a program where I had to quit using all drugs and alcohol. 3.91 1.63 I didnt go because I thought I had to be able to read and write when I got there. 3.99 1.45 I kept losing my medication. 4.13 1.34 I believed treatment programs were too religious. 4.14 1.26 Table 4 13. Logistic regression results for barriers categories on mental health utilization Variable B S.E. Wald Sig. Exp(B ) BTI Programmatic .72 .33 4.66 .031 2.052 BTI Individual .57 .29 3.71 .054 .567 BTI Social/Financial .20 .16 1.58 .210 .821 Note: R2 Nagelkerke = .063, R2 Cox & Snell = .045 Table 4 14. Logistic regression results for barriers categories on substance abuse utilization Variable B S.E. Wald Sig. Exp(B) BTI Programmatic .65 .37 3.14 .076 1.908 BTI Individual 1.02 .34 9.89 .003 .361 BTI Social/Financial .22 .18 1.47 .226 1.248 Note: R2 Nagelkerke = .118, R2 Cox & Snell = .078
104 Table 4 15. Mean and standard deviations of barriers for chronic and non -chronic groups Barriers Subscales Social/Financial Programmatic Individual Mean SD Mean SD Mean SD Chronic 2.92 1.30 3.62 .97 3.45 1.04 Non Chronic 3.17 1.41 3.63 1.06 3.79 1.09 No te: 1 = Strongly Agree, 5 = Strongly Disagree Table 4 16. Summary of logistic regression results for mental health service utilization Variable B S.E. Wald Sig. Exp(B) BTI Programmatic .04 .50 .01 .934 1.042 BTI Individual .09 .42 .05 .832 1.092 B TI Social/Financial .09 .24 .15 .696 .912 Gender (Male) .66 .69 .89 .345 1.926 Age .02 .03 .59 .442 .979 Total months since last residence .01 .01 .36 .549 .995 Ethnicity AfrAm/Black 1.05 1.14 .85 .357 2.870 Ethnicity EuroAm/White 1.80 1.07 2.84 .092 6.029 Natural supports MH .24 .31 .57 .449 1.265 Professional supports MH .50 .36 1.97 .160 1.646 Self identified need 2.58 .83 9.67 .002 13.176 MH rating .08 .34 .05 .821 .927 Note: R2 Nagelkerke = .474, R2 Cox & Snell = .338 Table 4 17. Bivariate logistic regression of significant variables on mental health service utilization R 2 Cox Variable ( n = 108) B S.E. Wald Sig. Exp(B) & Snell Natural supports MH .60 .19 9.97 .002 1.814 .097 Professional supports MH .93 .22 18.31 .000 2.535 .186 Self identified need 2.57 .59 19.32 .000 13.056 .226 MH rating .64 .22 8.34 .004 .529 .082
105 Table 4 1 8 Summary of logistic regression results for substance abuse service utilization Variable B S.E. Wald Sig. Exp(B) BTI Programmatic .5 2 .55 .89 .347 1.685 BTI Individual 1.00 .53 3.62 .057 .368 BTI Social/Financial .36 .29 1.56 .211 1.434 Gender Male .48 .81 .36 .550 1.620 Age .03 .03 1.04 .308 .969 Total months since last residence .01 .01 .68 .410 1.006 Ethnicity AfrAm/Bl ack .33 1.41 .06 .815 1.390 Ethnicity EuroAm/White .38 1.28 .09 .766 1.465 Natural supports SA .13 .31 .17 .679 1.136 Professional supports SA .67 .36 3.49 .062 1.962 SOC AMREC Factor .09 .04 4.93 .026 1.099 SOC Taking Steps Factor .08 .06 1.9 2 .165 1.084 Note: R2 Nagelkerke = .516, R2 Cox & Snell = .354 Table 4 19. Bivariate logistic regression of significant variables on substance abuse service utilization R 2 Cox Variable ( n = 95) B S.E. Wald Sig. Exp(B) & Snell Natural supports SA .55 .20 7.56 .006 1.731 .083 Professional supports SA .76 .23 11.11 .001 2.142 .123 SOC AMREC Factor .109 .03 13.34 .000 1.115 .159 SOC Taking Steps Factor .13 .04 9.82 .002 1.135 .128 Table 4 2 0 Summary of logistic regression results for mental health treatment attempts Variable B S.E. Wald Sig. Exp(B) BTI Programmatic 1.46 .59 6.07 .014 .232 BTI Individual 1.32 .51 6.63 .010 3.736 BTI Social .68 .29 5.70 .017 .506 Gender Male .82 .80 1.06 .304 .440 Age .02 .03 .32 .572 .9 83 Total months since last residence .02 .01 2.83 .092 .985 Ethnicity AfrAm/Black 1.44 1.16 1.55 .212 .237 Ethnicity EuroAm/White 1.33 1.20 1.23 .267 .265 Natural supports MH .28 .37 .57 .452 .755 Professional supports MH .62 .45 1.85 .174 1.8 50 Self identified need 4.20 1.02 17.06 .000 66.403 MH rating .24 .35 .49 .485 .784 Note: R2 Nagelkerke = .626, R2 Cox & Snell = .460
106 Table 4 2 1 Bivariate logistic regression of significant variables on mental health treatment attempts R 2 Cox Variable ( n = 105) B S.E. Wald Sig. Exp(B) & Snell BTI Social/Financial .48 .16 8.93 .003 .620 .089 Professional supports MH .44 .18 5.84 .016 1.557 .056 Self identified need 2.78 .55 25.71 .000 16.100 .286 MH rating .83 .22 13.62 .000 .437 .14 0 Table 4 2 2 Summary of logistic regression results for substance abuse treatment attempts Variable B S.E. Wald Sig. Exp(B) BTI Programmatic .07 .49 .02 .887 1.072 BTI Individual .87 .44 3.97 .046 .417 BTI Social/Financial .07 .26 .06 .801 .937 Gender Male 1.08 .71 2.30 .130 2.932 Age .01 .03 .08 .773 1.008 Total months since last residence .00 .01 .08 .784 .998 Ethnicity AfrAm/Black .34 1.17 .08 .772 1.405 Ethnicity EuroAm/White .72 1.12 .41 .522 2.056 Natural supports SA .42 .28 2.17 .141 1.514 Professional supports SA .51 .33 2.31 .128 1.656 SOC AMREC Factor .09 .04 4.90 .027 1.090 SOC Taking Steps Factor .05 .05 .90 .342 1.049 Note: R2 Nagelkerke = .487, R2 Cox & Snell = .348 Table 4 2 3 Bivariate logistic regression of significant variables on substance abuse treatment attempts R 2 Cox Variable ( n = 94) B S.E. Wald Sig. Exp(B) & Snell Gender Male .94 .45 4.33 .038 .389 .045 BTI Individual .67 .23 8.76 .003 .511 .098 Natural supports SA .67 .20 11.75 001 1.957 .129 Professional supports SA .63 .21 8.92 .003 1.872 .097 SOC AMREC Factor .09 .03 11.49 .001 1.094 .130 SOC Taking Steps Factor .09 .03 7.48 .006 1.096 .088
107 CHAPTER 5 DISCUSSION Overview of the Study The purpose of this study was to understand perceived barriers and their relationship to other factors that contribute to mental health and substance abuse counseling and treatment utilization. The study began with a focus on three categories of perceived barriers including societal, pr ogrammatic, and individual, in order to see their relationship to each type of service utilization. It then sought to determine the relationship between chronic status and the category of perceived barrier. Lastly, key predisposing, enabling, and need fa ctors were analyzed to see how well they predicted mental health and substance abuse service utilization among those who are homeless This chapter will discuss the results of the studys descriptive data, instrumentation, each hypothesis, and subsequent post hoc analyses. Furthermore, it will offer clinical, community, counselor training, and theoretical implications of the outcomes. F inally it will describe the limitations of the study and make recommendations for future research. Discussion of Descript ive Data The sampling methods of the study used convenience methods, gathering data from a single city located in the southeastern United States. Being diverse in gender, age, and ethnicity, the research team was able to collect information from a diverse group of homeless individuals from numerous collection sites. Although not from a national random sampling, the participants represented many of the same characteristics of the national homeless population. The gender mix wa s comparable as was the ethni c mix with the exception of Latina or Latino individuals, who were underrepresented. Generalizations to this population should be made with caution and with the recognition that barriers specific to this group (e.g. language) are not identified by the
108 par ticipants. The incentives also appeared appropriate, respecting the time offered by the participants but not being too large as to encourage them to take the survey multiple time s The living conditions of most of the sample are characterized as in the woods or street, over an extended period of time, and in the same local county for at least a year. Over half (59.3%) of the sample were living outside of agencies, residing in the woods, streets, or in their cars. This figure is nearly identical to the number of unsheltered persons reported in the local county (Gainesville/Alachua County Office on Homelessness and the Alachua County Coalition for the Homeless and Hungry, 2008) By their self -reported information 32.4% of the population fits the HUD def inition of chronic homelessness. This is actually low compared to Kuhn and C ulhane s (1998) estimate of 50% at any given point and time. Nonetheless, the amount of time homeless is high, with a mean of nearly 3 years. Not only have many individuals in t he sample been homeless for an extended period of time, many have also lived in the local county for a while. With over two thirds of the sample being in the county for a year or more, the clear majority of experiences have been local. There are high needs for mental health and substance abuse services among the sample, as would be found in national samples. Over half indicated positively that they currently needed mental health services and nearly half rated their mental health as either fair or poor. The correlation was significant and moderate between the two measures r (144) = -.58, p < .001; however the correlation was not as high as may be expected. Some of those rating their mental health higher may have improv ed their mental health with the he lp of mental health counseling services and therefore recogniz e the need to continue them. This would cause them to rate current mental health as good or excellent and still maintain the need for services.
109 A large n umber of persons in the sample repor t ed not having family friends or professional s to support them if they needed mental health or substance abuse counseling services. It was not determined if this indicated general isolation from these persons or if these persons simply would not support receiving that type of care. It is noteworthy that two thirds of the population had a t least one supportive family member, friend, or professional if there was a need for either type of service. With nearly half indicating the need for mental health servi ces and two thirds indicating some level of support it is surprising to see that only 29% of the population was r eceiving mental health services and only 22.8% was receiving substance abuse services There was a large group of persons who self identify t he need for care and d id not get it. A number of persons in the sample (16.6%) were receiv ing both mental health and substance abuse services This may have been due to the fact that t he main mental health provider in the county offer ed both services an d even though their programs generally identif ied as primarily substance a buse or mental health oriented they often provide both. A comparison of those seeking either mental health or substance abuse treatment services (46.2%) with those actually receivin g services (35.2%) shows that there was a significant unmet need in the community. These figures are even more dramatic when considering that a number of persons currently receiving services did not indicate that they had attempted to get these services i n the last 12 months. This may be due to the fact that some were in a program that extend ed over one year. Actually, 39.6% of those who tried to get mental health treatment services were not currently getting them, and 44.2% who tried to get substance ab use treatment services were not currently getting them. Based on this information and the most current county
110 data (Gainesville/Alachua County Office on Homelessness and the Alachua County Coalition for the Homeless and Hungry, 2008), there are clearly ma ny who are not getting the help they need. Another important point is the fact that many persons are actually attempting to get services. With almost half (46.2%) attempting to get services and an additional 11% seeking services during their lifetime, man y have been motivated to seek care. These high rates would refute many stereotypes that homeless individuals fail to try to help themselves and confirm the fact that substantial barriers are keeping them from moving forward. Discussion of Barriers Many of the barriers identified by the total sample were the same as those identified by those attempting to get services. Among the total sample, the barriers with m ean scores falling below 3.00 were all related to co -payments and h ealth care insurance When l ooking at those who had attempted to receive services, th e list of barriers below 3.00 expanded. One barrier was the fact that p ersons seeking treatment services lacked safe housing for them to get help. This may reflect many concerns related to being vulnerable while living without a home including violence, having easy access to substances, or the general pressures of daily survival. The list also included other barriers such as the lack of program information, having other needs with higher prioritie s, access, waiting lists, and a general sense of providers not caring about those who are homeless Many of these barriers could be ameliorated at the program level Furthermore, i ncreased capacity may be achieved with additional grant writing and advoca cy at the local, state, and national levels. The barrier of other needs taking priority speaks to one of the reasons that Housing First programs have become more popular in recent years ( Tsemberis, Gulcur, Nakae, 2004). Without a place to stay and having basic needs met, one is less likely to invest time with a counselor or other similar provider.
111 Interestingly, the bottom barriers were revealing as well. One item I didnt want to get help for my addictions and/or mental health problem, had a mean scor e of 3.91 which indicated a higher level of disagreement by the participants. It appears that a lack of desire is one of the least of this populations problems. Furthermore, on average, the sample did not have a problem with programs asking them to quit using drugs and alcohol. Losing medications was not a problem, but this may be related to the fact that many were unable to get medications in the first place. Additionally, religiosity of the programs ranked low as a barrier, an indication that, on ave rage, the community may be providing enough secular based programs or that religious based programs are amenable to the homeless population. Being a sample taken in the South where religion tends to be strongly valued (Hunt & Hunt, 2001), faith-based prog rams may be less of a barrier than it may be in other areas of the country. It should be noted that just because an item is on the bottom of the list, does not mean that the item is not a barrier in the community. For instance, some of the participants were not able to read, a fact indicated by program administrators in agency settings where data was collected. Because there are fewer people who need reading assistance, this barrier may not appear as a high mean score in the larger sample. Indeed, in one case where the instrument was read aloud, the participant stated clearly with increased volume that he was able to read. It appears that social shame may be a factor and reading barriers may be underreported. Further investigation would be necessary t o determine if and how programs are accommodating for this need among the minority of persons who are unable to read. In total, the top barriers mark a starting point for the community to begin addressing barriers, with special considerations for the need s of smaller subpopulations.
112 In the qualitative section of the barriers instrument, a few new themes emerged that are important to broaden the discussions on barriers and help with creating new instruments that address the needs of more homeless individual s. In particular, it is important to recognize those barriers that are not faced by the chronically homeless population, namely childcare and family issues. M any programs in the sample we re designed for individuals, not couples or intact families with bot h male and female adults This can be particularly difficult if a family becomes homeless and housing is needed in addition to substance abuse or other mental health services. Considerations for physical disabilities also need to be addressed by programs ; and if provisions are already in place, it would be helpful to understand how homeless individuals continue to experience barriers in this regard. Furthermore, the idea that homelessness itself can be a perceived barrier is yet to be explored Money wa s noted by one participant as a barrier, and consistent with the literature (Sosin & Bruni, 2000) it could be both having money or the lack of money that may serve as an enabling factor. In summary, the sampling methods produced a diverse representation o f the local community, one that is similar to the national population of homeless individuals. There are substantial unmet needs among the population and a number of barriers which they have identified. In order to quantify the significance of these barr iers and substance abuse needs, two instruments were used, both of which are described in the next section. Discussion of Instrumentation To measure perceived barriers, this study used the Barriers to Treatment Instrument (BTI). Each subscale as well as t he total scale produced high reliability for this sample all exceeding .90 The Programmatic and Individual subscale s correlated h igh ly with one another r (138) = .78 p < .001, as many items are closely related For instance, the individual barrier, I had no
113 safe place to stay could be a program barrier if worded, Programs do not provide enough housing. Both the Programmatic and Individual subscale s correlated at lower level s with the Societal/Financial subscale r (137) = .48, p < .001, and r (136) = .38 p < .001, which suggests the more distinct nature of barriers that are attributed to be caused by society. The two factor s of the SOCRATES (Modified/Combined) also showed high reliability scores with this sample The two factor s were moderately correlated with one another r (126) = .5 5 p < .001. This is somewhat surprising given that Burrow -Sanchez and Lundberg (2007) found a low correlation rate (.11) between the two factor s. Their lower rate originated from an earlier version of the instrum ent that they subsequently recommended to shorten. The current study eliminated the two suggested items; it is uncertain how retaining those original questions would have affected the outcomes. Overall, both instruments performed well with high reliability rates. Discussion of Hypotheses The original five null hypotheses were designed to determine if there was a significant relationship between perceived barriers and mental health counseling and substance abuse counseling and treatment service utilization as well as chronic status. In the first two hypotheses, the three categories of barriers were tested in logistic regression analyses to determine their relationships with each of the two utilization variables. The third used a mixed analysis of variance method to determine the relationship between the three types of barriers and chronic status. Hypothesis 4 and 5 used logistic regression analyses to understand the relationship between predisposing, enabling (including perceived barriers), and need factors with each of the two utilization variables. The following sections discuss the results of each of these original hypotheses.
114 Discussion of Hypothesis 1 H0: There is no significant relationship between mental health service utilization and category of p erceived barriers. This hypothesis was established to determine if there was a direct relationship with the three types of barriers and mental health service utilization. The null for Hypothesis 1 was partially rejected as the coefficient for programmatic barriers significantly contributed to the model which predicted mental health utilization. Although the omnibus test did not show significance, the significant correlation coefficients were interpreted. This was done for two reasons. First, as the effe ct size of the model that included the three subscale s was only .045 (Cox & Snell R2) it is probable that the sample size was not large enough to capture the effects of these variables. Cohen (1992) recommends a sample size of 547 to have adequate power t o detect a small effect size with three independent variables when alpha is set at .05. The second reason for interpreting the results relates to the fact that the study of perceived barriers is relatively new especially as it relates to mental health se rvice utilization as understood with the Behavioral Model for Vulnerable Populations (Gelberg, Andersen, & Leake, 2000) Therefore, an y results that are marginally significant may prove helpful for future research. In this case, t he correlation coefficien t for programmatic barriers was a significant contributor to the model when controlling for individual and social barriers. Additionally, the relationship between individual barriers and mental health utilization approached significance, a factor that cou ld show significant results with a larger sample size. I ndividuals who perceived lower programmatic barriers were more likely to utilize mental health services. The reverse would also be true as persons perceiving greater barriers were less likely to be getting these services. These findings are consistent with the theor y of Emancipatory Communitarianism
115 (EC), which suggests that oppressive forces, especially those created by mental health professionals and their programs, prevent individuals from getting the help they need (Prilleltensky, 1997; Prilleltensky, 2000) Oddly enough, those programs which are in the business of trying to help others with mental health services are creating the barriers t hat inhibit the progress of those who are homeless EC recognizes the power imbalances between service providers and prospective clients as the root of these barriers. How counselors and programs can share this power will be discussed more fully in the Clinical and Community Implications sections of this cha pter. Interestingly, the BTI Individual subscale exceeded significance only fractionally (.054). The direction of the relationship between this factor an d mental health service utilization was as expected with persons experiencing lower barrier s within themselves were less likely use these services, or conversely, those blaming themselves were more likely to be receiving services. In effect, individual barriers did not prevent service usage, they promoted it. As many programs focus on the individual and how they have caused their problems, it is not surprising that individual barriers would be more prevalent among those getting help. As this analysis does not determine cau sality, it is possible that these clients began to focus on these barriers after receiving mental health services In either case, it is important that programs focus on their intervention and recruitment strategies. When programs focus on the individual as t he overall source of their problems, they may estrange homeless persons who appropriately recognize their problems as being caused by their environment, including the programs themselves Furthermore, interventions that promote beliefs that attribute problems to the individual may miss many of the environmental factors that contr ibute to ones circumstances.
116 Also noteworthy is the fact that the model did not distinguish persons utilizing or not utilizing services based on societal barriers after taking into account the differences in programmatic and individual barriers The lack of statistical significance does not indicate a lack of societal barriers, it merely suggests that whether on the streets or in a program, the perception about the role of society in creating these barriers is largely the same. Discussion of Hypothesis 2 H0: There is no significant relationship between substance abuse service utilization and category of perceived barriers. The null for Hypothesis 2 was rejected as both the model predicting substance abuse service utilization and the individual barriers co efficient showed significance. Th e significance of the overall model was due in part to the fact that the overall effect size of the model was larger (Cox & Snell R2 = .078) than the one used in Hypothesis 1. Cohens (1992) effect size index reached .0 85, nearly twice that of the previous model. When looking at the individual correlation coefficients, the individual barriers were the more prominent in this case, with the significance level reaching .003. L ike Hypothesis 1, the direction of this relati onship was as expected; when there were more individual barriers there w as greater utilization of substance abuse counseling and treatment services. Compared to mental health service s i t appears that the emphasis on individual causality is even stronger in substance abuse programs. T hose persons living without a home who are trying to unhook the addictive power of mindaltering chemicals and heal from related emotional and relational wounds may not enter programs who fail to address the contextual oppres sive factors that are so evident in their daily lives The relationship between programmatic barriers and substance abuse utilization, when controlling for the other two types of barriers, approached but did not reach significance As
117 many of the particip ants were receiving both types of services, some of this effect may be related to the primary effects of those getting mental health services. Furthermore, like the previous hypothesis, there was not a statistically significant relationship between societ al barriers and substance abuse service utilization. Discussion of Hypothesis 3 H0: There is no significant relationship between homeless status (chronic versus non-chronic) and category (individual, programmatic, or societal) of perceived barriers. The r esults of the analyses for Hypothesis 3 failed to reject the null as there was no interaction effect; therefore, it did not support the expectations that the chronically homeless would experience each of the types of barriers differently than the non-chronically homeless population Moreover, there was no overall difference between the chronic and non -chronic groups across all barrier types. There was, however, a significant difference between the societal barriers when compared to either the programmatic or individual barriers. Both the chronic and non -chronic groups rated societal barriers higher (lower mean score) than the other two types of barriers. Even though the type of barrier and chronic status were not statistically significantly related, the analysis revealed that societal barriers, as measured by the BTI, are more strongly experienced than individual or programmatic barriers. One of the reasons the chronic population may not perceive barriers differently is that there are a number of non-chronic persons with extended periods or multiple episodes of homelessness who d id not fit the full criteria for chronic homelessness. In fact, over 20% of the sample had over one year of homelessness but did not have a disability or they lived with a partner or their children. So while not technically part of the chronic group, many had experienced similar barriers while living under similar circumstances.
118 D i scussion of Hypothesis 4 H0: There is no variance in mental health treatment attempts in the last 12 months explained by predisposing factors (gender, ethnicity, age, and duration of homelessness), enabling factors (societal barriers, programmatic barriers, individual barriers, natural supports, professional supports), and need factors (self -identified need and mental health rating). The null for Hypothesis 4 was rejected as the overall model was significant as was the correlation coefficient for self identified need. When other predisposing, enabling, and need factors were added to the model predictin g mental health service utilization, overall significance was clearly reached and the effect size improved greatly (Cox & Snell R2 = .338). This compared to effect sizes between .138 and .228 in previous studies of homeless individuals who are already get ting services ( Gamache, Rosenheck, & Tessler 2000; Lemming & Calsyn, 2004). A significant portion of this effect size appears to be from the need category, which is the only correlation coefficient to prove significant in the model when controlling for a ll other variables. When individuals without housing self -identify the need for mental health services, they are over 13 times more likely to obtain them, when controlling for gender, age, ethnicity, homeless duration, perceived barriers (individual, prog rammatic, and societal), support from family and friends, support from professionals, and rating ones mental health. Even when this variable alone is used to predict mental health service utilization, it still increases the odds of receiving services by 13 times. This finding is particularly important as previously the enabling factors had larger effect sizes when compared to need factors (Lemming & Calsyn, 2004). As the earlier studies only took samples from persons already receiving services, it follow s that need factors would be more important when homeless persons without need are included.
119 Other factors that appear to have a strong independent relationship with self identified need are both types of supports as well as the mental health rating. It i s of no surprise that when persons rate their mental health higher they would be less likely to utilize mental health services. The fact that natural and prof essional supports would all contribute to higher utilization confirm the results of previous stud ies ( Kertesz et al., 2006; Lemming & Calsyn, 2004; Sosin & Bruni, 2000); yet their lack of significance when controlling for need factors suggests that they are interrelated This connection may occur, in part, when individuals recognize their need for me ntal health services after talking to professionals, friends, and family supports. Statistically, all of the factors contribute d to the effect size, increasing it by .11 when all variables are included instead of just self -identified need Also notable is the fact that each category of perceived barriers did not contribute significantly to the overall model or to an independent bivariate regression model with mental health counseling and treatment service utilization as the dependent variable. Given the o utcome of Hypothesis 1, as more variables were introduced and there were fewer participants with usable instruments for these variables, the lack of statistical power increased the likelihood of a Type II error. The results of data from Hypothesis 1 revea led that that the effect size for barriers is relatively small; therefore, the use of 11 independent variables may be problematic. Yet, until a larger study is conducted, it appears that the barriers play a lesser role, if any, in preventing a person from getting mental health services. Furthermore any contribution s by these factors may be subsumed by the need factor. For example, p rogrammatic and societal barriers may be interrelated with need when a person does not successfully enter a program and then justifies their condition by stating that they do not need the services anyway. In that case, the barrier prevents entry, which then influences how the
120 individual perceives the need for services. Need may contribute to i ndividual barriers as a person exp eriences fewer barriers such as shame and general uncertainty when need is ignored or denied. If perceived barriers are contributing to mental health service utilization, their effects in relation to all of the variables in this hypothesis are yet to be f ully determined. Discussion of Hypothesis 5 H0: There is no variance in substance abuse treatment attempts in the last 12 months explained by predisposing factors (gender, ethnicity, age, and duration of homelessness), enabling factors (societal barriers, programmatic barriers, individual barriers, natural supports, professional supports), and need factors (recognizing the need to change and taking steps to change). The null for Hypothesis 5 was rejected as the model and the correlation coefficient recogniz ed need were both significant, results similar to Hypothesis 4. In this case, the effect size was even larger w ith the Cox & Snell R2 reaching .3 54. As before, need was a significant variable in the model, in this case recognition of ones substance ab use problems. Interestingly, it was the AMREC factor that showed significance in the model not the Taking Steps factor i n the SOCRATES (Modified/Combined) When the Taking Steps factor was analyzed in the bivariate analysis, the correlation coefficient w as significant yet the relationship for this variable was not as strong as the AMREC factor was to substance abuse service utilization O n e would have expect ed to see those who were taking steps to be the ones getting services and those who recognize d th e ir problems to be the ones who were preparing but not yet getting help. Having recognition as the more significant variable may indicate the short term nature of many treatment programs. Those who are currently utilizing services are those at the earli er stage of change, a number of whom will return to the streets and shelters if they are unable to secure income and
121 housing after treatment This may also indicate that there are a number of persons who are taking steps to quit using without the help of formal treatment programs One other reason for the close association between those who score high on the AMREC factor and those who are getting substance abuse services relates to recovery language. Many of the items on the AMREC factor are statements about being a problem drug-user/drinker or a drug addict/alcoholic. These statements relate closely to the language used by m any people speaking in a 12 Step meeting, My name is --, and I am an alcoholic/addict. Persons in a treatment setting that u ses a 12 -Step model and incorporates traditional meeting formats are likely to encounter this language and possibly embrace it. Therefore, the significant results of this factor may indicate conformity to program language as much as it is identifying re cognized need. The use of other need measures such as the Addiction Severity Index, which was used in Sosin and Bruni s ( 2000) study, or other symptom based instruments may help to resolve this question. As with Hypothesis 4, natural and professional supports had a direct predictive relationship with substance abuse service utilization that dissipated when controlling for all variables in the model. Indeed, one of the qualitative responses stated that having no contacts was a barrier to care, which woul d confirm the need to have personal connections to get into programs. These findings emphasize the need for counselors and other professionals to facilitate and be these connections, engaging in outreach services. They can provide not only a gateway for persons to enter but can also facilitate a smoother transition to getting care as a friendly face welcoming them into a strange environment. In fact, one of the case managers of a substance abuse program in this study had developed relationships with men and women living in the local shelter to recruit them into her program. This study included some these individuals.
122 The role of natural supports and professional supports is largely unstudied in the current literature with homeless adults; however, fami lies may play a significant role in helping prospective clients learn about services and facilitate their entry into programs. Counselors may also help this process if they are able encourage those who need care to make connections with existing supports. This information is also important for individuals without housing as they may also initiate some of these contacts prior to seeing a counselor Their family members, medical professionals, clergy, and friends may be resources that homeless individuals m ay approach in order to get needed care. To summarize, o f all five null hypotheses, one was not rejected, one was partially rejected, and three were not rejected. Of the four that showed some measure of significance, two revealed a significant relationshi p between at least one of the barriers subscale s. As indicated by the theory of Emancipatory Communitarianism (Prilleltensky, 1997) perceived barriers are internalized among those receiving services, a factor that will allow them entry into substance abus e programs. Furthermore, program barriers prevent mental health service utilization. Categories of p erceived barriers do not have a significant relationship with chronic status, an indication that duration and severity of homelessness have little impact on how these obstacles to care are experienced over time Self identified and recognized need are the most critical factors to contribute to utilization, both of which relate closely to nat ural and professional supports. Statistical power remains an issu e where effect sizes are low or the number of variables is high. All of these will be discussed in further detail in the following sections following the discussion of post hoc analyses.
123 Discussion of Post Hoc Analyses In order to determine the relationship of the three barriers variables to mental health and substance abuse service attempts, two addition hypotheses were tested. The se hypothese s modified Hypotheses 4 and 5 using mental health treatment attempts and substance abuse treatment attempts as th e dependent variables. The outcomes of these four hypotheses are discussed below. Discussion of Hypothesis 6 H0: There is no variance in mental health treatment attempts in the last 12 months explained by predisposing factors (gender, ethnicity, age, and duration of homelessness), enabling factors (societal barriers, programmatic barriers, individual barriers, natural supports, professional supports), and need factors (self -identified need and mental health rating). The nu ll was rejected for Hypothesis 6 and all three barriers coefficients significantly contributed to the model predicting mental health service utilization. When all three categories of barriers subscale s were used along the predisposing, other enabling, and need factors, the results were surprising. Not only were all three categories of barriers significant in the model, but their directions were reversed in comparison to the model presented in Hypothesis 1. Higher programmatic and societal barriers ( lower scores) predicted a greater lik elihood that these types of services were attempted. Alternatively, lower individual barriers predicted a higher likelihood that services would be attempted. Furthermore, self -identified need was an even stronger predictor of attempts when controlling f or the other variables When an individual sees the need for services, she or he is 66 times more likely to attempt to receive the services than when the need is not present.
124 It is difficult to determine why higher programmatic and societal barriers predi cted a greater likelihood that one would attempt mental health services. It is unlikely that these barriers actually motivated them or otherwise enabled them to seek care. More likely is the case that the barriers were experienced after seeking care. Wi thin this group of mental health treatment seekers are a number of persons who ha d attempted services but were not currently receiv ing them (n = 17). Among t hese are women and men who have attempted services and experienced the barriers so greatly that th ey could not get help. It is possible, that when controlling for need and the other factors in the model, the effect of these barriers were more pronounced. Essentially, after attempting services and not getting them, this group recognized society and pr ograms as creating the greatest obstacle s to their care rather than themselves. Having overcome personal barriers and tak en the initiative to get help, the individual was more likely to consider programmatic and societal barriers as preventing her or him fro m the next step in their care. Discussion of Hypothesis 7 H0: There is no variance in substance abuse treatment attempts in the last 12 months explained by predisposing factors (gender, ethnicity, age, and duration of homelessness), enabling factors (so cietal barriers, programmatic barriers, individual barriers, natural supports, professional supports), and need factors (recognizing the need to change and taking steps to change). The null for Hypothesis 7 was rejected as the overall model reached signifi cance as did two correlation coefficients, individual barriers and recognized need for substance abuse services. Similar to the results of the tests for Hypothesis 2, individual barriers contribute d significantly to the variance in the dependent variable, in this case, substance abuse treatment attempts. As in the earlier test, with lower individual barriers ( higher scores) there was a
125 decreased chance that substance abuse services would be sought when controlling for al l other variables in the model. Th ese results would suggest that persons who do not attribute their barriers to themselves are not even trying to get services in the first place The implication behind these findings is that substance abuse treatment programs have reputations among those who are homeless that they are in the business of helping persons overcome personal barriers. As programmatic and societal barriers are not predictors of substance abuse treatment attempts, one could conclude either that 1) these barriers do not exist, a premise that is not suggested by the descriptive data; or 2) that the programs remain largely silent on these issues. The message being delivered to the community in the latter c ase is that counselors and their programs do not recognize external forces as a primary cause of substance abuse problems. Those persons who express this perspective will be unsupported in their beliefs. Of further interest was the fact that males were less likely to seek substance abuse treatment services than were females, as re ported in the bivariate analyses. This factor was no longer significant when all of the factors were added to the model; however, it should not be overlooked. Homeless men are more likely to externalize their problems in comparison to homeless women (Jai nchill, Hawke, & Yagelka, 2000). This being t he case, programs may be dissuading men from entering treatment programs by assuming the individualistic philosophies which often neglect the role of ones environment in b oth their problems and their solutions In total, the post hoc analyses revealed that perceived barriers play a more significant role in mental health and substance abuse service attempts, when controlling for other predisposing, enabling, and need factors than mental health and substance abus e service utilization. It is at this early stage when a client has recognized the need for help that barriers are the most impacting.
126 This finding and each of those in the original hypotheses will be expanded upon in the next implication section s Clinic al Implications The results of this study are useful for clinicians as they try to promote a better quality of life for these marginalized citizens. The top barriers offer a starting point. As noted earlier (Table 4 11), the highest raking barriers noted by those seeking services included not knowing what services were available, waiting lists, and a general lack of care towards homeless persons. T o address these obstacles, there needs to be more share d information, expanded relationships, innovative cli nical approaches, and advocacy for more and improved resources. Improving the communication flow between programs and potential consumers is highly needed. Without knowing how to get help, those living without a home are not able to take advantage of the supports that currently exist. As personal connections are deemed important by potential consumers, it is important that counselors and other professionals meet those who are homeless and begin to build relationships that may continue through formal coun seling In order to do so, counselors who step outside of the traditional office and program settings will b e most effective (Kuhlman, 1994). In practice, c o unselors can begin by engaging persons in their living environments in coordination with existing outreach services who have already established posi t ive relationships with those without a home Information sessions could be offered at centralized sites that provide emergency services Handouts are helpful at these sessions, but they can not replace the human connections between service providers and potential clients (Wompold, 2001). By partnering with shelters and soup kitchens, counselors may gain access to those needing help and then begin to introduce themselves and their programs. Joining with these prospective clients over a meal at a soup
127 kitchen is one way to covey respect and care in an egalitarian relationship while sharing information about program services. Other ways of sharing information involve sharing power and building relationship s Program alumni can be powerful representatives to share the impact of mental health and substance abuse counseling services and offer encouragement and support throughout the healing process. Many of these alumni may have returned to the streets as th ey were unable to get employment due to substance abuse related criminal histories. Some may be hired to post or distribute flyers for information sessions ; but the most important aspect of their role, as suggested by the data, is that they become part of the network that supports positive change in the lives of those needing help. All of these activities should be done in concert with discussions with those who are currently homeless learning their insights on the best way to communicate about existing services A substantial outcome of this study is the recognition that programmatic and individual barriers impact service utilization differently. The fact that programmatic barriers play a role in mental health service utilization shows that counselors a nd the programs where they operate are part of the problem (Prilleltensky, 1997; Martn Bar, 1994). In addition to the lack of information, waiting lists and a general lack of care are program barriers more often exper ienced by those without a home. Wai ting lists are impacted by funding, which is generally recognized as a societal issue; however, service providers may improve the management of these lists as they improve communication with prospective clients and each other Those waiting for help can b e informed of their status on waiting lists as well as getting referrals for other programs (Rothbard, Min, Kuno, & Wong, 2004). More will be discussed on referrals in the Community Implications section.
128 The general impression that service providers do not care about those who are homeless may be improved by building these early relationships, but this may not be enough. What is named a lack of care may be a lack of understanding. Like the general population, counselors may also attribute the causes of homelessness back to the individuals experiencing this condition (Toro et al., 2007). The research focus on individual causes of homelessness rather than societal factors support such biases (Mallett, Rosenthal, & Keys, 2005; Tessler, Rosenheck, & Gamache, 2001). Clinicians who learn about the contextual factors that cause homelessness and explore their attitudes about those who are homeless will become more multicultural competent and better able to serve this population (Baggerly & Zalaquett, 2006). Th ese biases are also revealed in the positive correlation between individual barriers and substance abuse service utilization. Clinicians who fail to recognize ways to address contextual barriers instead focusing primarily on individual barriers may be le ss inclined to empathize with the experiences of persons living without a home who have experienced societal and programmatic barriers. By embracing a theoretical perspective that recognizes contextual forces and the power of the individual to disengage oppression, clinicians will be able to reach new clients and support change in the community. Emancipatory Communitarianism, for instance, is an approach that recognizes the causal role of these external forces as well as the fact that the individual needi ng help will need to take substantial responsibility to made the needed changes in self and community (Prilleltensky, 1997). In the pursuit of clinical approaches that consider contexts, it will be important that counselors use assessment instruments and c linical interviews that promote dialogue with new and prospective clients about the barriers to treatment services that they have experienced. Those identifying predominantly individual barriers may consider recognizing social systems
129 that impact their li ves through their families, employers, landlords, politicians, and the like. This perspective is not used to remove any power or responsibility in making choices to better their own lives. Actually, through activities that promote the development of crit ical consciousness, clients may be empowered to make independent decisions rather than being victims of a social will that often attempts to disempower those who have already been oppressed (Freire, 1970/2006; MartnBar, 1994; Montero, 2009) Finally, t he large impact of societal barriers among all homeless persons in this sample stresses the need for counselors to advocate in their communities, locally and nationally. They may advocate for more affordable housing, job training, and funding for programs (Baggerly & & Zalaquett, 2006). Politicians are not the only ones who need to be lobbied. Churches, businesses, and other institutions that can provide job and housing opportunities may be willing to support those in need when asked (Baggerly & Zalaquet t). Law enforcement officials can also be key societal representatives who can have a positive or negative impact on the lives of those living on the streets, in the woods, and in their cars. Counselors who develop relationships with these officials have the ability to inform and persuade them to support those who are homeless ra ther than harassing them (Notarangelo, 2005; Schroeder, 2002). By becoming advocates, counselors may develop greater empathy for homeless people and show that they indeed care ab out those who are living without a home. Community Implications The findings of this study have implications for community leaders, citizens with homes, and those without housing. Politicians and city officials have the ability to help those who are homel ess to positively engage services by supporting information networking and program coordination, such as the Access to Community Care and Effective Services and Supports
130 (ACCESS) programs offered in some cities (Rothbard, Min, Kuno, & Wong, 2004). The del ayed One -Stop Center a central support center for homeless individuals, is sorely needed in the city of Gainesville, the primary location of this study (The Same Old Song, 2008 ). Within such a program should be a centralized referral system that track s empty beds and facilitates referrals which would eliminate many of the information gaps and help those in need from having to get a phone in order to call multiple sites. Those offering referrals would need to be consumer friendly and trained in crisis management. Furthermore, these services would need to be advertised among community churches, non -profits, websites, and frequented sites. Funding for mental health and substance abuse services as well as basic living requirements need to be a priority for officials and the general public. With basic supports, many more individuals and families will feel safe enough to begin counseling and be able to transition into a therapeutic residential setting. Training for staff and clinicians are needed as is a general public awareness campaign. Without the support of the general public, even supportive officials will have difficulty allocating funding for services. The most significant perceived barriers noted on the list were those associated with society. Many identified health care insurance and co-payments as primary reasons for not being able to get needed care. Universal health care that includes support for mental health (including substance abuse) services is one way to offer continued support to ind ividuals who lose their jobs and homes and are experiencing great stress (National Health Care for the Homeless Council, 2008) For those with little to no income, co-payments should be waived, so that no obstacle exists to getting this type of care. If homeless citizens were able to maintain insurance, they would be able to get care with greater dignity and service providers could concentrate on their quality of care rather than trying to obtain short term grants. With more dignity and higher
131 quality of services, th ose living in crisis will have a better chance to exit homelessness ( Miller, 2001). As a part of the current study, the research team has begun to take steps to meet with participants, agencies, and community leaders. To date, the three inve stigators have met with the Mayor of Gainesville to determine the best places to share this information with city and county officials. She has provided her support in this regard. Furthermore, the investigators will supply brief reports to the agencies in the local community as well as to the participants. The latter conversation will encourage persons without housing to use this information to take action on their own behalf as well. There is much that all stakeholders can do to make a positive impact on the community. Counselor Training Implications In order to gain a better understanding of those who are homeless and to examine biases against that population, counselors would benefit from training and other forms of education. As noted previously, c ounselors are needed to build more supportive relationships and convey empathy for the struggles one e xperiences after losing a home. Training would not only benefit clients directly, but it would also impact the general public as counselors share informa tion about homelessness and the barriers to care. In order to raise counselor competencies when working with this population, training is needed that will help these mental health professionals examining their attitudes, learning the most current informati on, and advancing their skills ( Baggerly & Zalaquett 2006). This training not only occurs in the classrooms, but also in the shelters and other temporary residences of this population (Strawser, Markos, Yamaguchi & Higgin, 2000). By getting involved in the lives of this marginalized group of people, counselors will raise their first -hand knowledge of the
132 obstacles that low and noincome people face on a daily basis. Crisis management skills are needed, especially as many without a home have experienced significant traumas (Christensen et al., 2005; Jainchill, Hawke, & Yagelka, 2000). There is no research on the relationship between trauma and perceived barriers to mental health and substance abuse services ; however, professionals who are unprepared to handle re -emerging traumas will only create more barriers to care. Furthermore, counselors will benefit from training about strengths among individuals and in communities of homeless people. Many of these small communities can be mutually supportive in ways that families and society is not (Applewhite, 1998; Branch 2007). Communities also form in agencies where homeless adults show perseverance in overcoming personal and institutional difficulties (Fogel & Dunlap, 1998). The natural supports of many living without a home are those who share the same situation, some of whom will encourage each other in reduced substance use and gathering social support (Rew & Horner, 2003). Without training to look for these types of supports, counselors may impose their ow n ideals about having housed friends and professionals as the best type of supporters. There is a need for counselors in training to better understand their role among other professionals who serve people without housing. Educators are needed to inform these students on how to network with medical doctors, nurses, social workers, psychologists, and outreach workers who all play an important part in whether one uses counselors services or not. Furthermore, there needs to be more information taught on makin g referrals with these providers to offer comprehensive and integrated services. In addition, counselor educators are needed to teach students how to best advocate for their clients without housing. As they advocate in the community, they have a role in s haring
133 information about the causes of homelessness, negative biases against the population, and ways that society imposes barriers to care. Counselors may need to advocate with family members, teachers, clergy, and other people that have the potential to support or impair care. Sharing these broader perspectives with community agencies and faith-based organizations can all be ways that counselors can impact society in a positive way. They also have a role in advocating with government officials and need to understand the politics of a particular setting and how to strategize for community changes. Theoretical Implications This study also has important theoretical implications. Gelberg, Andersen, and Leake (2000) wrote very little about their enabling va riable perceived barriers to care when explaining the Behavioral Model for Vulnerable Populations (BMVP) This study make s strides to not only identify specific barriers, but also to use a measure which quantifies individual, programmatic, and societal ba rriers to explain substance abuse and mental health service utilization patterns among those without housing. The results show ed that perceived programmatic barriers appear to inhibit mental health service utilization. This finding supports the BMVP which theorizes that barriers serve to prevent utilization. In contrast, the findings revealed t hat perceived individual barriers positively impact substance abuse utilization This contradicts the premise of how barriers work in the model and would suggest t hat perceived barriers have a different role than other so called objective barriers Emancipatory Communitarianism (EC ; Prilleltensky, 1997 ) is a helpful theoretical perspective as it shows how biases and negative beliefs about ones self (i.e., indivi dual barriers) can be internalized. In these cases, barriers may not prevent ones entry into programs, it can serve as the passport. When agreeing with societal norms which blame
134 individuals for their circumstances related to addiction s and homelessness a homeless person may be closer in line with program philosophies and gain entry. In this case, the perceived individual barriers actually serve to enable utilization. It is important to note that individual barriers are linked closely to need, and that with more addiction severity, one will have both more need and more barriers (e.g. lacking a safe place to stay ). This study attempted to mitigate these forces by controlling for need. Indeed, when need was taken into account, much of this barriers af fect diminished The reduced statistical power may have also impacted the lack of significance, so it should not be assumed that the relationship between barriers and need is the only factor at work The difference in results between service utilization and service attempts highlights certain portions of the BMVP. The Model provides a feedback loop at the point of Outcomes, which returns to each of the prior factors, including enabling factors and subjective barriers. The results of this study show how the experience of not receiving services after a prior attempt can substantially change ones perception of barriers in comparison to those who cross the barriers and receiv e services. Negative experiences which prevent utilization appear to confirm the p erception that society and programs are the primary sources of these obstacles This study also contributes to the understanding of how large the impact is of barriers on support ing or prevent ing mental health and substance abuse service utilization. For mental health services, the effect size of these variables appear to be small and marginally significant ; it is a bit larger for substance abuse services. In contrast, need and support variables appear to have a greater impact on utilization than barriers This difference shows the power that supports have in overcoming potential barriers. More studies will be needed to understand this area of research and the role of supports in clarifying ones need for services as well as overcoming the
135 barriers which prevent utilization. Other suggested research as well as limitations are noted in the following section s Limitations As one of the first to test the impact of perceived barriers on mental health and substance abuse service utilization, this study has r evealed new insights, yet there remain limitation s related to study design and measure ments Future studies may address these limitations and may better understand the relationships between the co nstructs i n the current analyses Study D esign Limitations T hree limitations related to the study design were as follows: 1) convenience sampling methods, 2) the use of a small sample, and 3) the simple cross -sectional design. By using a convenience sample, the generalizability may not extend beyond the local are a of the survey. Many of the demographic factors were similar to national estimates of those who are homeless ; however, each community has differing political and social factors which may create different barriers. F unding and limited resources restricte d the study to its geographical area. In addition, the few ethnic groups besides African Americans and European Americans represented in the sample prevent generalizations to the s e populations as well. The small er sample size impacted the ability to ru le out Type II error s. As perceived barriers to mental health and substance abuse services had not previously been studies, the effect size and required sample size was unknown. Additionally, with as many as eleven variables, it was less likely that each o f the instruments would be fully completed for each participant so that all variables would be used in the logistic regression models. Most participants completed the surveys unassisted. It is expected that the item response rate may have increased if the surveys were read aloud, a more time consuming option that diminishes the privacy of the participants.
136 In addition, by using a cross sectional design, causality can not be fully determined. A l ongitudinal design would be necessary to fully understand how barriers change and are related to utilization over time. Measurement Limitations The measures used in this study presented some limitations including : 1) being self report measures, 2) using single item s to measure constructs 3) using the BTI, whic h is still in its developmental phase, and 4) using a modified version of the SOCRATES. As a ll were self report measures the information reported is not verified and could have been altered due to social desirability or hesitancy to provide sensitive inf ormation (Czaja & Blair, 2005). Some of the constructs were measured with single -items. This was done largely because the study collected many variables largely from non-clinical settings. A fuller understand ing of the relationships between some constru cts would need to come from other, more extensive measures. Additionally, s ome of these single -item measures could be modified to improve their construct validity and interpretability. For instance, it would have been appropriate to include clergy and ot her similar supports in the natural supports measure. Professional supports could have been more explicit about including medical doctors. As it stands, it is uncertain whether participants included these persons when filling out the survey. In addition it would be better to use a 7 point Likert scale with each point representing an additional person who offers support. By grouping numbers (e.g. 3 4 persons), it was more difficult to interpret the results. Even though the barriers instrument provided greater detail for this construct, the measure is still in the development phase. The reliability of the measure was high both in this study and the one used in its development. As noted earlier, it was developed for the chronically homeless
137 population diagnosed with both a substance abuse and other mental health disorder For this reason, there were a few barriers not included which could have impacted the results. Finally, b y using the SOCRATES (Modified/Combined), the study used a measure with high reliability that retains many of the same questions of the commonly used SOCRATES; however this modified version has been used less frequently. As noted in the results, if discrimination between the two constructs in this measure is important, it may be better to use the original forms developed by Miller and Tonigan (1996). In addition to using the SOCRATES (Modified/Combined), it would be helpful to include a face valid question asking if the participant deemed they had a need for substance abuse serv ices, similar to the mental health need question. While the SOCRATES (Modified/Combined) measures recognized need to change substance usage, it does not determine whether a person sees the need to seek counseling or other services to make these changes. More comprehensive measures of need and attempts to receive substance abuse and mental health counseling would be helpful to fully explore these constructs. Recommendations for Future Studies In order to ameliorate some of the sampling limitations, f utu re studies should be replicated in multiple locations with regional diversity. In addition, these studies may consider the impact of cultural on barriers particularly language barriers. They may decide to include veteran status as a variable a s this sub population is often able to get services that the general homeless population can not. Overall larger sample sizes would provide increased power to prevent Type II errors. Future studies should consider tracking individuals over longer period of time and use l ongitudinal designs. Furthermore, having researchers to read each survey would be desirable to
138 attempt to get more completed surveys. These strategies are more costly and time consuming and may increase social desirability; however, they will incre ase the item response rate and provide a greater ability to determine causality between the variables. A longitudinal study would also provide the opportunity to see how barrier attribution changes after a person engages counseling and treatment services. After initiating care, d o persons who attribute barriers to society and programs tend to disconnect with providers sooner than those who attribute barriers to themselves? Or do persons change their barrier attribution as they encounter positive relationships with service providers ? Future studies are needed to explore the need and attempts to receive counseling services. Although there are clear measures of diagnosed need, there are far fewer ways to quantify perceived need. The development of such ins truments would reveal a more comprehensive understanding of the utilization model and the role of other variables in changing perceived need. On a related note, f uture qualitative and quantitative studies should investigate ways that homeless individuals cl arify their need for services as well as establish positive supports. Often, these supports are considered from outside the homeless population, but it would also be important to understand how people without housing support one another. In this study, it was apparent that many were attempting to change their substance use patterns and were not currently receiving counseling to do so. The role of peers without housing in these attempts would be informative. In addition, it would be helpful to understa nd how supports improve ones chance in getting into a program. Do these persons encourage motivation as well as provide concrete resources such as money, transportation and the like? All of the questions could be answered in future studies to better u nd erstand how to help homeless persons get the assistance they need.
139 When studying the role of predisposing, enabling, and need factors in predicting utilization and service attempts, it would be beneficial to analyze the data with methods that more closely follow the Behavior Model for Vulnerable Populations (Gelberg, Andersen, & Leake, 2000). Path analysis is one such method that would reveal both direct and indirect effects of the variables on the outcomes (Agresti & Finlay, 1997). The variables that proved to be significant in bivariate analyses in this study could then be more fully understood among all variables in the model. This study has revealed the need to research service attempts in relation to utilization and other variables. By including service attempts in a path analysis, one could see how feedback loops connect successful or unsuccessful utilization attempts back to perceived barriers, supports, and need. Additionally, qualitative studies exploring the differences between perceived actual b arriers would reveal how accurate barriers are and how misperceptions can occur. For example, if homeless individuals see the lack of transportation as a barrier to care, it would be helpful to know if there are transportation services available in the co mmunity. These services may be inaccessible for other reasons or simply unknown to the general community. In such a case, communication would be the barrier rather than transportation. As gender was determined to be a significant factor predicting substa nce abuse service attempts in a bivariate analysis wit hin this study, it would be important to more fully explore this variable in relation to others. As women, especially those with children, have access to more public funding and resources than men (Mor ris, 1997), this may affect service utilization. The effects of gender on housing, financial resources, and supports would inform counselors on how public policy and relational factors can impact service utilization.
140 Finally, portions of this study could be replicated involving community members more directly and at earlier stages. To increase participant involvement and critical consciousness, men and women without housing could be invited to participate in focus groups to discuss the role of barriers in their specific community. By investigating how politics and public policy contribute to barriers and service utilization, researchers can trigger a problematization (Montero, 2009) whereby individuals would be invited to think critically about the existi ng structures that contribute to their difficulties in ways that they may not have done so previously. As part of this conversation, participants would also be invited to critically explore the history of homelessness in their community and how attitudes and policies have changed over time. Quantitative analyses would then be built upon the findings of these qualitative studies, and the findings returned to the community. Such procedures would even more closely follow the philosophies of Liberation Psych ology and Emancipatory Communitarianism ( Freire, 1970/2006; Martn Bar 1994; Montero 2009; Prilleltensky, 1997) Conclusion In this study, 145 adult men and women without housing provided information to understand how mental health and substance abuse c ounseling and treatment services were impacted by perceived barriers as well as other predisposing, enabling, and need variables. Three categories of perceived barriers were studied including societal, programmatic, and individual, each of which produced significant results in some, but not all of the hypotheses. Furthermore, the participants indicated, on average, the most and least significant barriers they faced in the prior 12 months. This chapter detailed the implications of each hypothesis test resu lt along with clinical community, training, and theoretical implications, as well as limitations and recommendations
141 for future research. The impact of counselors and other supports with those who are homeless are undisputable. This study has shown that barriers play a significant role in th ese relationship s as these helpers can hinder or help those without a home to get the services they need. Furthermore, it is apparent that all in society may play a role as we support or discourage public policies th at enable the betterment of individuals and families who live on the streets, shelters, and other places instead of a home. Although there are limitations to the current study, future research can mitigate many of these and further the knowledge of ways t hat this population may overcome the obstacles that inhibit their wellbeing.
142 APPENDIX A BARRIER S TO TREATMENT INSTRUM ENT In the last 12 months, 1 ) D id you try to get addiction treatment services? ____ Yes ____ No If Yes, how may times? ______ 2 ) D id you tr y to get mental health treatment services? ____ Yes ____ No If Yes, how may times? ______ In your lifetime, 3 ) Ha ve you ever tried to get addiction treatment services? ____ Yes ____ No If Yes, how may times? ______ 4 ) H ave you ever tried to get mental health treatment services? ____ Yes ____ No If Yes, how may times? ______ Each of the following statements relates to barriers you faced in the last 12 months when trying to get addiction and/or mental health treatment services Circle the number which best i ndicates how much you agree with each statement on a scale from 1 (strongly agree) to 5 (strongly disagree). Social /Financial Strongly Strongly Agree Disagree 5 ) It was hard to get services because I didnt have and couldnt get health care ins urance. 6 ) I lost my health care insurance (e.g., Medicaid) so I didnt seek services. 7 ) I did not get medications because I did not have insurance coverage. 8 ) Services and medicines were difficult to get because I couldnt afford the co payments. 9 ) I did not get mental health or addiction services because my insurance did not cover it. 10) Most providers would not accept Medicare/Medicaid. 11) Services were not available where I was living. Programmatic 12) I didnt want to attend a program where I had to quit usi ng all drugs and alcohol. 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
143 Strongly Strongly Agree Disagree 13) Programs shouldnt ask me to stop using alcohol and drugs. 14) I thought the program woul d make me feel bad about myself. 15) I thought that treatment programs required that I would have to take mental health medications before entering. 16) Services (like detox, residential treatment) were always full 17) I thought most program providers were too ove rworked to listen to my needs 18) Service providers do not care about the needs of homeless people. 19) I thought addiction and mental health programs were too strict and overly -structured 20) Case workers change jobs so much that I was tired of having to start ov er with someone new. 21) I believed treatment programs were too religious 22) I didnt try to get addiction treatment because I didnt think it would work. 23) I didnt try to get mental health treatment because I didnt think it would help me. 24) I didnt go becau se I thought I had to be able to read and write when I got there. 25) I didnt know about the services that were available to me. 26) The program would not let me work while I was in the program. 27) I had to wait too long to get services. 28) I had a previously bad experience with the service provider. Personal 29) I was too intoxicated to seek treatment 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
144 Strongly Strongly Agree Disagree 30) My addiction to drugs and/or alcohol was too severe. 31) I didnt want to stop using drugs and/or alcohol. 32) I didnt want to get help for my addictions and/or mental health problem s. 33) I was too mentally ill to seek treatment 34) I did not trust doctors, therapists and case managers 35) I did not want to take medications 36) My other needs (like housing and food) were more important than treatment 37) I didnt want anyone to find out that I wa s addicted. 38) I didnt want anyone to find out that I had a mental health problem. 39) I didnt want to admit to having a problem with alcohol and/or drugs. 40) I didnt want to admit to having a problem with my mental health 41) I enjoyed the lifestyle of using drugs and/or alcohol and didnt want to give it up. 42) I was not sure I had a problem with alcohol and/or drugs. 43) I was not sure I had a mental health problem. 44) I was so discouraged and frustrated trying to find treatment that I gave up 45) I tried treatment before and it never worked on me. 46) I kept losing my medication 47) I kept relapsing 48) I had no safe place to stay in order to begin treatment 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
145 Strongly Strongly Agree Disagre e 49) I didnt think going to a program was safe. 1 2 3 4 5 50) I had no one to help me access treatment. 1 2 3 4 5 51) I couldnt get transportation to get services. 1 2 3 4 5 52) I was afraid they would find out I had legal charges against me. 1 2 3 4 5 53) I did not want to give up control of my own life. 1 2 3 4 5 54) What do you think is the biggest source of barriers to treatment? (pick one) ___ The programs that provide treatment ___ The individual seeking treatment ___ Society who funds treatment In the last 12 months what types of medical and/or mental health services did you try to receive? Tried, but Did not did not try to Received receive receive Medical (physical health) ____ ____ ____ Prescription Medications ____ ____ ____ Addiction Counseling ____ ____ ____ Mental Health ____ ____ ____ H IV services ____ ____ ____ Job training/placement services ____ ____ ____ Social Security Disability Insurance payments ____ ____ ____ Other (please describe): ____ ____ ____ __________________________________________________________________ __________________________________________________________________ __________________________________________________________________
146 APPENDIX B GENERAL QUESTIONNAIR E Please tell us about yourself in the following questions. Residential Information 1. Where do you currently reside? (please check one) ___ apartment/house ___ c ar or vehicle not intended for housing ___ tent or self -made shelter ___ emergency shelter ___ transitional housing ___ substance a buse treatment facility ___ mental health treatment facility ___ street/woods ___ other location not intended for housing ___ other please describe_____________________________ 2. Do you have a spouse (wife/husband) or partner who currently resides wit h you? ____ yes ____ no 3. Do you have any children who currently reside with you? ____ yes ____ no 4. How many times have you been homeless in the last three years? ____ 0 3 times _____ 4 or more times 5. How long has it been since you last had a residence (excluding treatment, hospitals, shelters, or transitional housing) ____ months and ____ years
147 Supports and Services 6. Among your family and friends how many would support your receiving substance abuse counseling or treatment services? (check one) ___ None ___ 1 -2 ___ 3 -4 ___ 5 -6 ___ 6 or more 7. Among professional service providers you know (case managers, counselors, psychologists, psychiatrists, nurses), how many would support your receiving substance abuse counseling or treatment se rvices? (check one) ___ None ___ 1 -2 ___ 3 -4 ___ 5 -6 ___ 6 or more 8. Among your family and friends how many would support your receiving mental health counseling or treatment services? (check one) ___ None ___ 1 -2 ___ 3 -4 ___ 5 -6 ___ 6 or more 9. Among p rofessional service providers you know (case managers, counselors, psychologists, psychiatrists, nurses), how many would support your receiving mental health counseling or treatment services? (check one) ___ None ___ 1 -2 ___ 3 -4 ___ 5 -6 ___ 6 or more 10. Are you currently receiving substance abuse counseling or treatment services? ____ yes ____ no 11. Are you currently receiving professional mental health counseling or treatment services? ____ yes ____ no 1 2 Do you believe that you currently need professional mental health counseling or treatment services? ____yes ____ no 13. How would you rate your current mental health? (check one) ___ Excellent ___ Good ___ Fair ___ Poor
148 Demographic Information 1 4 What is your gender? ____ Male ____ Female 1 5. What is your age? _____ 16. What is your ethnicity? (please check one) ___ African American/Black ___ European American/White ___ Latina/Latino ___ Native American/Pacific Islander ___ Asian American ___ Other: ___________________ 17. Do you have any physical health problems that are disabling? ____ yes ____ no 18. Do you have any mental health problems that are disabling? ____ yes ____ no 19. Do you have any substance abuse problems that are disabling? ____ yes ____ no Thank you for you r participation!
149 APPENDIX C PERSONAL DRUG AND AL COHOL USE QUESTIONNAIRE (SOCRATES: Version 8 Modified) INSTRUCTIONS: Please read the following statements carefully. Each one describes a way that you might (or might not) feel about your drug and/or alcohol use For each statement, circle one number from 1 to 5, to indicate how much you agree or disagree with it right now Please circle one and only one number for every statement. NO! Strongly Disagree No Disagree ? Undecided or Unsure Yes Agree YES! Str ongly Agree 1. If I dont change my drinking/drug use soon, my problems are going to get worse 1 2 3 4 5 2. I have already started making some changes in my drug use/drinking 1 2 3 4 5 3. I was using drugs/drinking too much at one time, but Ive managed to change my drug use/drinking 1 2 3 4 5 4. Sometimes I wonder if my drinking/drug use is hurting other people 1 2 3 4 5 5. I am a problem drug user/drinker 1 2 3 4 5 6. Im not just thinking about changing my drinking/drug use, Im already doing somet hing about it. 1 2 3 4 5 7. I have already changed my drug use/drinking, and I am looking for ways to keep from slipping back into my old pattern. 1 2 3 4 5 8. I have serious problems with drinking/drug use 1 2 3 4 5 9. My drinking/drug use is causing a lot of harm 1 2 3 4 5 10. I am actively doing things now to cut down or stop using drugs/drinking 1 2 3 4 5 11. I know that I have a drug use/drinking problem 1 2 3 4 5 12. I am a drug addict/alcoholic 1 2 3 4 5 13. I am working hard to change my drinking/drug use 1 2 3 4 5
150 APPENDIX D OPEN ENDED RESPONSES TO OTHER BARRIERS EXPERIENCED Packet Response s (sic. ) 002 I have had no funds to keep getting there. No transportation. No Home or Tent. 004 Having to keep in touch with Meridian Mental Health Se rvice. Emily Blaine is one worker and Leah Vail is another. 006 First Meridian wouldnt help me because I was not in system. I had to go through Shands. Services I needed from Meridian I was told I couldnt recieve because not enough history of abuse ( alcohol & drugs) problems. 008 Time money 010 N/A 011 I dont have a place to live and no money to support myself 015 N/A 017 0 018 N A 019 Lack of Resources 020 Home school. Tunnel location. Popes Creek. 021 No openings available! 023 No insuri nce. No help. Bottom line. 024 Havent had too many barriers considering Ive never had drug/alcohol mental health issue. 025 Tring to get disablity & tring to get insurance Medicade. 028 No work 029 Programs should relate to the person where they are coming from, not where you came from. Today is different. Personal and Individual basis. 030 NA 031 Myself deniel 032 I do not drink or take drugs 034 None 035 Lost my job so help! Me get a new job 036 My unemployment compensation 039 N/A 040 Pr ide, shame and guilt from last relapse 041 None V.A. Hosp. 043 None 044 You have to be in trouble or jail. They dont care that you know already. 045 Get a job 049 Trying to get SSI for the last 9 years 051 They need to help more than what they are d oing. Told us for years they were trying to help make easier and havent done it yet. 052 None V.A. Hosp 054 Transportation & notification 056 Waiting to get in 057 Im so afraid I will relapse again and jail the program. Its almost easier not to eve n try.
151 Most inpatient programs arent long enough. Really, though, I dont know how not to use. Part of that is my mental illness. 060 No 064 NA 065 More income would have helped me get services 068 Transportation problems. Lack of financial 069 Stil l going through the panic attacks and shit. 070 My insurance is a problem. Went to mental health on 3/11/09. Crying depressed confuessed upset all the time anymore 071 Because of old injuries and no insurance I do not receive a proper health check out on my back/neck injury. I also feel it should be a little easier or not so long of a waiting list for housing. Help for small income familys. 072 N/A 073 N/A 074 I was very lucky I was at my end living in a garage then I found Meridean and the hope progr am things are slowly getting better 075 Knowing the resources are available 079 How to take care of me and my daughter and to finally go out and find a job 083 You got to be willing to help your own self. 084 Transportation 086 No 087 None 088 Havin g my heart broken is the one thing that has hurt the worst. I loved these women and they all went off with someone else. 089 Pass the buck to other agencies! 090 None but homeless 092 I have yet to seek help for any addiction or mental health because I do not have any kind of drug addiction and my mental state of mind is good. 093 None dont have a place to stay Help 094 My self will wanted to smoke crack and not stop. So it would not allow me to stop and seek help. 097 Trying to find work. 098 No contacts 099 I dont feel I have any mental problems & do not use drugs or alcohol. I just like living outdoors & traveling around the US. 101 I wish it was California and I had the 514 102 Everyone is OK 103 No funding 104 No transportation 106 Ill egal Mexicans have all the real labor jobs all over America. Get rid of all illegals !! 112 First, admitting I had a problem. Thinking about what other would say about my mental condition. Lack of trust in others and providers. 117 Mental illnesses are not currently understood or cured by doctors. In my case, the treatment made, in time, my illness much, much worse after for a short time making me feel somewhat better. Resperdal Consta is toxic and leads, I am sure, to alzhermers. Tell somebody plea se. Now
152 118 No barriers were encountered. 119 Since Im a veteran, blessed that I dont have these problems of getting help. I do know that the state of Fla help peoples that want help. 120 I feel good about my self 121 Actually money Not money for the cost of treatment but rather the need to work every day through a labor pool, needed to just pay rent to stay just for that night. And then the circle begins again money to stay the night so ---labor pool that day to get the money to stay the night. 125 NA 127 Where Im currently at they seem to not know what there doing and some dont seem to care. Its to confinded and we never get out. 130 The amount of people waiting to get in. 131 There were no barriers because I wasnt trying to get addic tion services. 132 There have not been none. 133 None 136 The funding for treatment has run low, you cant stay as long as you want. 138 Paying for my treatment 139 My last date to be here in Bridgehouse is April 30, 2009 I have yet to find employment and find somewhere to live But as of now Im jobless and homeless Ive applied for Social Security disability for my HIV and AIDS status Please can someone help me Im so afraid that if Im not prepared for living in society Im gonna be on the streets or back on drugs crack cocaine and drinking. 140 I looked for a re hab to go to. No one would accept me because I had no insurance. There is not enough funding or housing for people trying to get help for substance abuse. 141 I try to get some help but all they said tomorrow I willy need help. 144 N/A 145 Barriers would be a fear of D.C.F. loosing my child. I also stayed on drugs/alcohol because I had no place to lay my head.
153 APPENDIX E INFORMED CONSENT Protocol Title: Barriers to Substance Abuse and Mental Health Services among Persons without Housing (UFIRB # 2009 U 118) Please read this consent document carefully before you decide to participate in this study. Purpose of the research study : The purpose of this study is to examine the barriers to substance abuse and mental health services among adults without housing. You do not have to have a mental health or substance abuse problem in order to participate. What you will be asked to do in the study: You will be asked to complete three questionnaires. One is about any experiences you may have had with substance abuse and mental health services. The second is about your thoughts about alcohol and drug use. The third one asks general questions about you and people who may support you. Time re quired: 20 minutes Risks and Benefits: There are no anticipated risks with this experience. The expected benefits are limited, but this information may help you become more aware of barriers that prevent you or others from getting substance abuse and mental health services. The information from all of the surveys may also help agencies and public policy makers understand the need to reduce these barriers. Compensation: You will be given a small gift for your participation, which may include food, cloth ing, or other items not to exceed $5 in value. Confidentiality: You will not put your name on the questionnaires, and so your responses will be anonymous, and not even we will be able to link your name to your responses. This signed informed consent will be kept separately from the questionnaires that you complete. All of this information will be kept in a locked file in the Department of Counselor Education at the University of Florida. Your name will not be used in any report. Voluntary participation: Your participation in this study is completely voluntary. There is no penalty for not participating. Right to withdraw from the study: You have the right to withdraw from the study at anytime without consequence. You have the right not to answer any que stion asked of you on these surveys. Whom to contact if you have questions about the study: Principal Investigator : Faculty Supervisor : Michael Brubaker, PhD Candidate Michael T. Garrett, PhD Department of Counselor Education Department of Counselor Educa tion College of Education College of Education University of Florida University of Florida 1202 Norman Hall 1202 Norman Hall PO Box 117046 PO Box 117046 Gainesville, FL 32611 7046 Gainesville, FL 32611 7046 352 562 6099 352 392 0731 x 356 Whom to contac t about your rights as a research participant in the study: UFIRB Office, Box 112250, University of Florida, Gainesville, FL 32611 2250; ph 352 392 0433. Agreement: I have read the procedure described above. I voluntarily agree to participate in the procedure and I have received a copy of this description. Participant: ______________________________________ Date: _________________ Principal Investigator: ______________________________ Date: _________________
154 APPENDIX F RECRUITMENT SCRIPT Barriers to S ubstance Abuse and Mental Health Services among Persons without Housing Script: Hi, my name is ______________________ and I am a researcher with the University of Floridas Department of Counselor Education. We are looking for individuals without hou sing to participate in our study about service barriers. It will take about 20 minutes and we are offering a small gift for their time. Do you know anyone who would be willing to do the survey?
155 APPENDIX G SOLICITATION FOR IN -KIND DONATIONS Hello Couns elor Ed! Are you committed to engaged research? Do you want to help people who dont have a home? If so, we are inviting you to join us in an outreach and research project to benefit the homeless. A group of us (TaJuana Chisholm, Niyama Ramlall, and Mi chael Brubaker) are conducting research over the month of March, investigating what are the barriers that prevent persons without housing from getting substance abuse and mental health treatment services. We would like you to join us as community partners We are embracing a theoretical perspective called Emancipatory Communitarianism which calls us into community through mutual responsibility, shared resources, and shared information. Each of us, as research investigators, is dedicating our time and res ources to persons living on the streets and shelters, a highly marginalized population. We are asking you to join us by contributing items that will be distributed to persons who take the survey, a way to respect their time. Consistent with our theoretic al perspective, the information will be summarized and presented to the homeless community as well as key stakeholders such as the Mayor and various service providers. If you are willing to be a part, here are some of the items we are collecting: Tube socks Chocolates (small bars a rarity on the streets) Pens Candles Lighters Vienna sausages Also, we need items to give to women in shelters including: Lotion (small bottles) Chapstick/lip gloss Emory boards Tweezers Hair ties Face powder Durable combs Shampoo/conditioner Sanitizer Other general items that would be appropriate in clude: Small first aid kits Bus cards Deodorant Floss Tooth brushes Phone cards (not to exceed $5 in value) Nail clippers There will be a box in the Counselor Education Of fice to receive these items. Also, as you make a donation, please feel free to leave your name and email so we can thank you. Your contributions will make a real difference. We hope you will join us! Michael Brubaker, Principle Investigator Niyama Raml all, Co Investigator TaJuana Chisolm, Co Investigator
156 LIST OF REFERENCES ACA (American Counseling Association). (2005). Code of ethics Alexandria, VA. Author. Retrieved November 12, 2006, from the World Wide Web: http://www.counseling.org/Resources Agresti, A. & Finlay, B. (1997). Statistical methods for the social sciences (3rd ed.). Upper Saddle River, NJ: Prentice Hall. American Society of Addiction Medicine. (2001). ASAM patient placement criteria for the treatment of substance -related disorders (2nd ed.). Chevy Chase, MD: Author. Andersen (1968). Behavioral Model of Families Use of Health Services. Research Series No. 25. Chicago, IL: Center for health Administration Studies, University of Chicago. Andersen, R.A. (1995). Revisiting the behavioral model and access to medical care: Does it matter? Journal of Health and Social Behavior, 36, 1 10. Andersen, R. & Newman, J.F. (1973). Societal and individual determinants of medical care utilization in the United States. The Milbank Quarterly, 83(4), 1 28. Applewhite, S. L. (1998). Coping and survival skills of homeless veterans. Journal of Applied Social Sciences, 23, 1321. Baggerly, J. & Zalaquett, C. P. (2006). A descriptive study of single adults in a homeless shelters: Increasing counselors knowledge and social action. Journal of Multicultural Counseling and Development, 34, 155167. Barris, M. (2008, January 29). Foreclosure filings surged 75% in 07 as subprime mess grew. Wall Street Journal (Easter n Edition), p. D.3. Retrieved July 25, 2008, from Wall Street Journal ProQuest database (Document ID: 1420031311). Barrow, S. M., Herman, D. B., Cordova, P., & Struening, E. L. (1999). Mortality among homeless shelter residents in New York City. American Journal of Public Health, 89, 529534. Benda, B.B. (2006). Survival Analyses of Social Support and Trauma Among Homeless Male and Female Veterans Who Abuse Substances. American Journal of Orthopsychiatry 76, 7079. Benda, B.B., DiBlasio, F.A. & Pope, S.K (2006). Spiritual well being, relationships, and work satisfaction in the treatment of homeless veterans with alcohol/other drug problems. Alcoholism Treatment Quarterly, 24 109124 Bhui, K., Shanahan, L., & Harding, G. (2006). Homelessness and mental i llness: A literature review and a qualitative study of perceptions of the adequacy of care. International Journal of Social Psychiatry, 52, 152165.
1 57 Bluestein, D.L., McWhirter, E.H., & Perry, J.C. (2005). An emancipatory communitarian approach to vocationa l development theory, research, and practice. The Counseling Psychologist, 33, 141179. Bonin, J.P. (2007). Predictors of mental health service utilization by people using resources for homeless people in Canada. Psychiatric Services, 58, 936941. Bradford D.W., Bradley, N.G., Kim, M.M., Kaufman, J.S., & Weinberger, M. (2005). Can shelter -based interventions improve treatment engagement in homeless individuals with psychiatric and/or substance misure disorders? A randomized controlled trial. Medical Care, 43, 763768. Branch, A. (2007, January 21). The hidden homeless: Dont have a roof. Dont have walls. But Ive got my freedom. Fort Worth Star -Telegram Retrieved July 1, 2009, from AccessMyLibrary database Brodie, M., Weltzien, E., Altman, D., Blendon, R.J., & Benson, J.M. (2006). Experiences of Hurricane Katrina evacuees in Houston shelters: Implications for future planning. American Journal of Public Health 96, 14021408. Burnette, M.L., Ilgen, M., Frayne, S.M., Lucas, E., Mayo, J., & Weitlauf, J.C. (2008). Violence perpetration and childhood abuse among men and women in substance abuse treatment. Journal of Substance Abuse Treatment, 35, 217222. Burrow Sanchez, J.J. & Lundberg, K.J. (2007). Readiness to change in adults waiting for publicly funded s ubstance abuse treatment. Addictive Behaviors, 32, 199204. Burt, M.R. (2004). Homelessness, Definitions and Estimates of. Encyclopedia of Homelessness Thousand Oaks, CA: SAGE Publications. R etrieved November 23, 2008, from http://www.sage -ereference.com/ homelessness/Article_n67.html Burt, M.R., Aron, L.Y., Douglas, T., Valente, J., Lee, E. & Iwen, B. (1999). Homelessness: Programs and the people they serve Washington, DC: Urban Institute. Calloway, M.O. & Morrissey, J.P. (1998). Overcoming service barrie rs for homeless persons with serious psychiatric disorders. Psychiatric Services, 49, 15681572. Christensen, R.C., C.C. Hodgkins, L.K. Garces, K.L. Estlund, M.D. Miller & R. Touchton. (2005). Homeless, mentally ill and addicted: The need for abuse and tra uma services. Journal of Health Care for the Poor and Underserved 16, 615 621. Chronister, K.M. (2006). Social class, race, and ethnicity: Career interventions for women domestic violence survivors. American Journal of Community Psychologists, 37, 175182. Chronister, K.M., Wettersten, K.B., & Brown, C. (2004). Vocational research for the liberation of battered women. The Counseling Psychologist, 32, 900922.
158 Connolly, C. (2005, October 3). 9th Ward: History, Yes, but a Future?; Race and Class Frame Debate on Rebuilding New Orleans District :[FINAL Edition]. The Washington Post, p. A.01. Retrieved July 13, 2008, from Washington Post database. (Document ID: 905610281). Cosgrove, L. (2006). The unwarranted pathologizing of homeless mothers: Implications for re search and social policy. In R.L. Toporek, L.H. Gerstein, N.A. Fauad, G. Roysircar & T. Israel (Eds.) Handbook for social justice in counseling psychology: Leadership, vision, and action (pp. 200 214). Thousand Oaks, CA: Sage Publications. Crane, M., Byrne K. Fu, R., Lipmann, B., Mirabelli, F., Rota -Bartelink, A., Ryan, M., Shea, R., Watt, H. & Warnes, A.M. (2005). The causes of homelessness in later life: Findings form a 3 nation study. Journal of Gerontology: Social Sciences, 60B S152S159. Crane, M. & Warnes, A.M. (2007). The outcomes of rehousing older homeless people: A longitudinal study. Ageing & Society, 27, 891918. Czaja, R. & Blair, J. (2005). Designing surveys: A guide to decisions and procedures (2nd ed.). Thousand Oaks, CA: Pine Forge Press. Dame Sheila McKechnie. (2004 January 5 ). Times Online Retrieved from http://www.timesonline.co.uk/tol/comment/obituaries/article842101.ece DeNavas-Walt, C., Proctor, B.D. & Smith, J. (2007). Income, poverty, and health insurance coverage in the United States: 2006 (U.S. Census Bureau Publication No. P60233) Washington, DC: U. S. Government Printing Offi ce. Retrieved May 6, 2008, from http://www.census.gov/prod/2007pubs/p60 233.pdf Dooley, D. (2001). Social research methods (4th ed.) Upper Saddle River, NJ: Prentice Hall. Downey, R. G. & King, C. V. (1998). Missing data in Likert ratings: A comparison of replacement methods. The Journal of General Psychology, 125, 175191. Erlenbusch, B., OConnor, K., Downing, S., & Phillips, S.W. (2008). Foreclosure to homelessness: The forgotten victims of the subprime crisis. Washington, DC: National Coalition for the Homeless. Retrieved July 25, 2008, from http://www.nationalhomeless.org/housing/foreclosure_report.pdf Evans, G. (2004). The environment of childhood poverty. American Psychologist, 59, 7792. Felix, A.D. & Wine, P.R. (2001). From the couch to the street: Appli cations of psychoanalysis to work with individuals who are homeless and mentally ill. Journal of Applied Psychoanalytic Studies, 3, 1732. Fogel, S J. & Dulap, K. M. (1998). Communal living and family life in transitional shelters. Journal of Applied Social Sciences, 23, 3 11.
159 Folsom, D.P., Hawthorne, W., Lindamer, L., Gilmer, T., Bailey, A., Golshan, S., Garcia, P., Unutzer, J, Hough, R. & Jeste, D.V. (2005). Prevalence and risk factors for homelessness and utilization of mental health services among 10,340 patients with serious mental illness in a large public mental health system. American Journal of Psychiatry, 162, 370376. Foster, R.P. (2007). Treating depression in vulnerable urban women: A feasibility clinical outcomes study in community service sett ings. American Journal of Orthopsychiatry, 77, 443453. Freire, P. (2006). Pedagogy of the oppressed. (M.B Ramos, Trans.) New York: Continuum International. (Original work published 1970) Gainesville/Alachua County Office on Homelessness & Alachua County C oalition for the Homeless and Hungry. (200 8 ). Homeless conditions in Gainesville: Results from the January 200 8 Point -in Time Census & Survey of Alachua County residents without housing. Gainesville, Florida: author. Gamache, G., Rosenheck, R.A. & Tessle r, R. (2000). Factors predicting choice of provider among homeless veterans with mental illness. Psychiatric Services, 51, 10241028. Gelberg, L., Andersen, R.M., & Leake, B.D. (2000). The behavioral model for vulnerable populations: Application to medical care use and outcomes for homeless people. Health Services Research, 34, 12731302. Harter, L.M., Berguist, C., Titsworth, S., Novak, D. & Brokaw, T. (2005). The structuring of invisibility among the hidden homeless: The politics of space, stigma, and ide ntity construction. Journal of Applied Communication Research, 33, 305327. Hartman, L., Little, A., Ungar, M. (2008). Narrative inspired youth care work within a community agency. Journal of Systemic Therapies, 27, 4458. Hunt, L. L. & Hunt, M. O. (2001). Race, religion, and religious involvement: A comparative study of Whites and African Americans. Social Forces, 80, 605631. Jainchill, N., J. Hawke & J. Yagelka. (2000). Gender, psychopathology, and patterns of homelessness among clients in shelter -based TCs. American Journal of Drug and Alcohol Abuse 26, 553567. Kanabanow, J. (2008). Getting off the street: Exploring the process of young peoples street exits. American Behavioral Scientist, 51, 772788. Kertesz, S. G., Larson, M. J., Horton, N. J., Win ter, M., Saitz, R., & Samet, J. H. (2005). Homeless chronicity and health related quality of life trajectories among adults with addictions. Medical Care, 43, 574585. Kertesz, S.G., Larson, M.J., Cheng, D.M., Tucker, J.A., Winter, M., Mullins, A., et al. (2006). Need and non-need factors associated with addiction treatment utilization in a cohort of homeless and housed urban poor. Medical care, 44 225233.
160 Klein, J. (2005, October 3 ). Lets have an antipoverty caucus. Time, 166 (14), 29. Retrieved Janua ry 16, 2009, from Wilson Web database Kuhn, R. & Culhane, D.P. (1998). Applying cluster analysis to test a typology of homelessness by pattern of shelter utilization: Results from the analysis of administrative data. American Journal of Community Psychology, 26, 207232. Kuhlman, T.L. (1994). Psychology on the streets: Mental health practice with homeless persons New York: J. Wiley & Sons. Lam, J.A. & Rosenheck, R. (1999). Social support and service use among homeless persons with serious mental illness. International Journal of Social Psychiatry, 45, 1328. Lee, B.A. & C.J. Schreck. (2005). Danger on the streets: Marginality and victimization a mong homeless p eople. American Behavioral Scientist 48 10551081. Lemming, M.R. & Calsyn, R.J. (2004). Utility of the behavioral model in predicting service utilization by individuals suffering from severe mental illness and homelessness. Community Mental Health Journal, 40, 347364. Levy, J. (2004). Pathway to a common language: A homeless outreach perspective. F amilies in Society, 85, 371378. Lydersen, K. (2008, June 14). Iowa flooding rivals 1993 deluge; Midwestern storms swamp Cedar Rapids and Des Moines. The Washington Post p. A.1. Retrieved August 12, 2008, from Washington Post, ProQuest database. (Document ID: 1494676911). MacKnee, C.M. & Mervyn, J. (2002). Critical incidents that facilitate homeless peoples transition off the streets. Journal of Social distress and the homeless, 11, 293 306. Maisto, S.A., Conigliaro, J., McNeil, M., Kraemer, K., OConnor, M., & Kelley, M.E. (1999). Factor structure of the SOCRATES in a sample of primary care patients. Addictive Behaviors, 24, 879892. Mallett, S., Rosenthal, D., & Keys, D. (2005). Young people, drug use and family conflict: Pathways into homelessness. Jour nal of Adolescence, 28, 185199. Marshall, S. & Souccar, M.K. (2005, August 15). Troubled homecoming. Crains New York Business, 21 (33), 1, 13, 15. Martijn, C. & Sharpe, L. (2006). Pathways to youth homelessness. Social Science & Medicine, 62, 1 12. Mart n Bar, I. (1994). The role of the psychologist. In A. Aron (Ed. & Trans.) & S. Corne (Ed.), Writings for a liberation psychology (pp. 33 46). Cambridge, MA: Harvard University Press.
161 McCarthy, J.F., Blow, F.C., Valenstein, M., Fischer, E.P., Owen, R.R., B arry, K.L., Hudson, T.J., & Ignacio, R.V. (2007). Veterans affairs health system and mental health treatment retention among patients with serous mental illness: Evaluating accessibility and availability barriers. Health Services Research, 42, 10421060. M cKinney -Vento Homeless Assistance Act of 1987, Pub. L. No. 100177 11302, 101 Stat. 482 (1987). Mel i to, R. (2003). Values in the role of the family therapist: self -determination and justice. Journal of Marital and Family Therapy, 29, 3 11. Melnyk, K.A. M. (1988). Barriers: A critical review of recent literature. Nursing Research, 37, 196201. Melnyk, K.A.M. (1990). Barriers to Care: Operationalizing the variable. Nursing Research, 39, 108112. Milby, J.B., Schumacher, J.E., Vuchinich, R.E., Freedman, M.J ., Kertesz, S. & Wallace, D. (2008). Toward cost -effective initial care for substanceabusing homeless. Journal of Substance Abuse Treatment, 34, 180191. Miller, A.B. & Keys, C.B. (2001). Understanding dignity in the lives of homeless persons. American Jo urnal of Community Psychology, 29, 331 354. Miller, M.D., Hodgkins, C., Estlund, K.L., & Brubaker, M.D. (2008). Perceived barriers to treatment for the chronically homeless. Poster presentation at the Treatment for Homeless Program, Technical Assistance Wo rkshop, Substance Abuse and Mental Health Administration (SAMHSA), Arlington, Virginia. Miller, W.R. & Tonigan, J.S. (1996). Assessing drinkers motivation for change: The stages of change readiness and treatment eagerness scale (SOCRATES). Psychology of A ddictive Behaviors, 10, 8189. Mitka, M. (2006). Chronic homeless in intervention spotlight. Journal of the American Medical Association, 295, 23442345. Morrell Bellai, T., Goering, P.N., & Boydell, K.M. (2000). Becoming and remaining homeless: A qualitat ive investigation. Issues in Mental Health Nursing, 21, 581604. Morris, J.M. (1997). Chronic homelessness among workingage adults: personal problems and public assistance. Journal of Social Distress and the Homeless, 6, 5769. Montero, M. (2009). Methods for liberation: Critical consciousness in Action. In M. Montero & C. C. Sonn (Eds.), Psychology of liberation: Theory and applications (pp. 73 91). New York: Springer. Murali, V. & Oyebode, F. (2004). Poverty, social inequality, and mental health. Advance s in Psychiatric Treatment, 10, 216224.
162 National Coalition for the Homeless. (200 8 ). Questions and Answers about the Chronic Homeless Initiative. Washington, DC: author. Retrieved December 27 2008, from http://www.nationalhomeless.org/publications/chronic/chronicqanda.html National Coalition for the Homeless. (2007a, June). Homeless Veterans (NCH Fact Sheet #14). Washington, DC: author. Retrieved May 6, 2008, from http://www.nationalhomeless.org/publications/facts/veterans.html National Coalition for the Homeless. ( 2007b June). Who is homeless? (NCH Fact Sheet #3). Washington, DC: autho r. Retrieved May 6, 2008, from http://www.nationalhomeless.org/publications/facts/Whois.pdf National Coalition for the Homeless. ( 2007c June). Why are people homeless? (NCH Fact Sheet #1). Washington, DC: author. Retrieved July 15, 2008, from http://nationalhomeless.org/publications/facts/Why.pdf National Health Care for the Homeless Council (2008). Universal health care and homelessness. Nashville, TN: author: Retrieved May 14, 2009, from http://www.nhchc.org North, C.S., Pollio, D.E., Perron, B., Eyrich, K.M., & Spitznagel, E.L. (2005). The role of organizational characteristics in determining patterns of utilization of services for substanc e abuse, mental health, and shelter by homeless people. Journal of Drug Issues, 35, 575591. Notarangelo, A. (2005, October 25). Concord plans to clear out homeless camps. CBS 5. Retrieved July 19, 2007, from cbs5.com Overton, S. L. & Medina, S. L. (2008). The stigma of mental illness. The Journal of Counseling & Development, 86 143151. Pack Brown, S.P., Thomas, T.L., & Seymour J.M. (2008). Infusing professional ethics into counselor education programs: A multicultural/social justice perspective. Journal of Counseling & Development, 86, 296302. Prilleltensky, I. (1997). Values, assumptions, and practices: Assessing the moral implications of psychological discourse and action. American Psychologist, 52, 517535. Prilleltensky, I. (2000). Bridging agency, theory and action: Critical links in critical psychology. In T. Sloan (Ed.), Critical psychology: Voices for change (pp. 67 81). New York: St. Martins Press. Prilleltensky, I. & Nelson, G. (2002). Doing psychology critically, Making a difference in divers e settings. New York: Palgrave MacMillan. Reid, S., Berman, H. & Forchuk, C. (2005). Living on the streets in Canada: A feminist narrative study of girls and young women. Issues in Comprehensive Pediatric Nursing, 28, 237256.
163 Rew, L. & Horner, S.D. (2003) Personal strengths of homeless adolescents living in a high risk environment. Advances in Nursing Science, 26, 90101. Rivlin, L.G. & Moore, J. (2001). Home -making: Supports and barriers to the process of home. Journal of Social Distress and the Homeles s, 10, 323336. Rollinson, P. (2007). Homelessness beyond the big cities; Overlooked America. Planning, 73, 2023. Rosen, M.I., McMahon, T.J., Lin, H., & Rosenheck, R.A. (2006). Effect of social security payments on substance abuse in a homeless mentally i ll cohort. Health Services Research, 41, 173191. Rosenstock, I.M. (1966). Why people use health services. The Milbank Memorial Fund Quarterly, 44, 94127. Rothbard, A.B., Min, S.Y., Kuno, E., & Wong, Y.L.I. (2004). Long term effectiveness of the ACCESS pr ogram in linking community mental health services to homeless persons with serious mental illness. Journal of Behavioral Health Services & Research, 31, 441449. Schroeder, C. (2002, June 8). Homeless camp cleanup defended: Health report supports city. The Cincinnati Enquirer. Retrieved July 19, 2007, from http://www.enquirer.com Sklarew, B., Twemlow, S.W., & Wilkinson, W.M. (2004). Analysts in the Trenches: Streets, Schools, War Zones. Hillsdale, NJ: Analytic Press. Sosin, M. R. & Bruni, M. (2000). Persona l and situational perspectives on rejection of a homelessness and substance abuse program: An exploratory study. Social Work Research, 24, 1627. Stainbrook, K.A. & J. Hornik. (2006). Similarities in the characteristics and needs of women with children in homeless family and domestic violence shelters. Families in Society 87 5362. Stein, J.A., Andersen, R., & Gelberg, L. (2007). Applying the Gelberg-Andersen Behavioral Model for Vulnerable Populations to health services utilization in homeless women. Jou rnal of Health Psychology, 12, 791804. Stein, J.A., Leslie M.B. & Nyamathi A. (2002). Relative contributions of parent substance use and childhood maltreatment to chronic homelessness, depression, and substance abuse problems among homeless women: Media ting roles of self -esteem and abuse in adulthood. Child Abuse & Neglect 26, 10111027. Steinhaus, D.A., Harley, D.A. & Rogers, J. (2004). Homelessness and people with affective disorders and other mental illnesses. Journal of Applied Rehabilitation Counse ling, 35, 3640.
164 Steinhauer, J. (2002). Bloomberg plans more housing aid for the homeless. New York Times (Late New York Edition) A1, B7. Strawser, S., Markos, P. A., Yamaguchi, B. J. & Higgins, K. (2000). A new challenge for school counselors: Children who are homeless. Professional School Counseling, 3, 162171. Sumerlin, J.R. (1999). Cognitive affective preparation for homelessness: Quantitative and qualitative analyses of childhood out of -home placement and child abuse in a sample of homeless men. Psy chological Reports, 85, 553573. Swick, K.J. (2006). Helping homeless families overcome barriers to successful functioning. Early Childhood Education Journal, 33, 195200. Tessler, R., Rosenheck, R., & Gamache, G. (2001). Gender differences in self -reporte d reasons for homelessness. Journal of Social Distress and the Homeless, 10, 243254. The same old song. (2008 October 15). Gainesville Sun. Retrieved April 16, 2009, from www.gainesville.com Thompson, S.J., McManus, H., Lantry, J., Windsor, L., & Flynn, P (2006). Insights from the street: Perceptions of services and providers by homeless young adults. Evaluation and Program Planning, 29, 3443. Thompson, S.J., McManus, H. & Voss, T. (2006). Posttraumatic stress disorder and substance abuse among youth who are homeless: Treatment issues and implications. Brief Treatment and Crisis Intervention, 6, 206217. Thompson, S.J., Pollio, D.E., Eyrich, K., Bradbury, E., & North, C.S. (2004). Successfully exiting homelessness: Experiences of formerly homeless mentall y ill individuals. Evaluation and program planning, 27, 423431. Toro, P.A., Tompsett, C.J., Lombardo, S., Philippot, P., Nachtergael, H., Galand, B., Schlienz, N., Stammel, N., Yabar, Y., Blume, M., MacKay, L., & Harvey, K. (2007). Homelessness in Europe and the United States: A comparison of prevalence and public opinion. Journal of Social Issues, 63, 505524. Tsemberis, S., Gulcur, L., & Nakae, M. (2004). Housing first, consumer choice, and harm reduction for homeless individuals with a dual diagnosis. A merican Journal of Public Health, 94, 651656. Tsemberis, S., Moran, L., Shinn, M., Asmussen, S.M., & Shern, D.L. (2003). Consumer preference programs for individuals who are homeless and have psychiatric disabilities: A drop in center and a supported hous ing program. American Journal of Community Psychology, 32, 305317. Urban Institute. (2001, October). What will it take to end homelessness? Retrieved April 12, 2004, from http://www.urban.org/urlpr int.cfm?ID=7281
165 The United States Conference of Mayors (2007). Hunger and homeless survey: A status report on hunger and homeless in Americas cities. Retrieved May 6, 2008, from http://w ww.usmayors.org/hhsurvey2007/hhsurvey07.pdf United States Department of Health and Human Services Substance Abuse and Mental Health Services Administration. (2008). Homelessness Provision of Mental Health and Substance Abuse Services Retrieved on Dec ember 2, 2008, from http://mentalhealth.samhsa.gov/pu blications/allpubs/homelessness U. S. Department of Housing and Urban Development Office of Policy Development and Research (2004, January). Strategies for reducing chronic street homelessness. Retrieved April 16, 2007, from http://www.huduser.org/publications/homeless/chronic_homeless.html U. S. Department of Housing and Urban Development Office of Community Planning and Development (2007, February). The annual homeless assessment report to congress. Ret rieved April 14, 2007, from http://www.huduser.org/publications/povsoc/ Vera, E.M. & Speight, S.L. (2003). Multicultural competence, social justice, and counseling psychology: Expanding our roles. The Counseling Psychologist, 31, 253272. Wilde, M.H., Alba nese, E.P., Rennells, R., & Bullock, Q. (2004). Development of a student nurses clinic for homeless men. Public health nursing, 21, 354-360. Wise, J.B. (2005). Empowerment practice with families in distress. New York: Columbia University Press. Wompold, B .E. (2001). The great psychotherapy debate: Models, methods, and findings. Mahway, N.J.: Lawrence Erlbaum Associates. Wong, Y.L I. & Piliavin, I. (2001). Stressors, resources, and distress among homeless persons: A longitudinal analysis. Social Science & M edicine, 52, 10291042. Zlotnick, C., Tam, T. & Robertson, M.J. (2003). Disaffiliation, substance use, and exiting homelessness. Substance Use & Misuse, 38, 577599.
166 BIOGRAPHICAL SKETCH Michael Brubaker received his M.S. in p rofessional c ounseling from Georgia State University, his M.Div. from Emory University, and his B.S. in Business Administration from the University of North Carolina at Chapel Hill. He was hired as a program director for a faith based non -profit serving the homeless and subsequently be came a Certified A ddictions Counselor as well as a National Certified Counselor Michaels clinical experiences have extended into serving impaired professionals, in -home counse ling, and clinical supervision. T h e socially marginalized including those liv ing in poverty, sexual minorities, and Nativ e Americans, are among his research priorities. Taking a social justice and strengths approach Michael has address ed both clinical and social biases which compromise counseling relationship s To counter these forces, he has stressed the importance of communal supports, spirituality and the significance of the helping relationship in improving mental health and overall wellness. As a practitioner and researcher, Michael has presented in multiple professional co nferences including the American Counseling Association (ACA), the National Association for Counselor Education and Supervision (ACES), the Association for Specialists in Group Work (ASGW), and the Southern Association for Counselor Education and Supervisi on (SACES). In addition, he has written for the Journal of LGBT Issues in Counseling, the Journal for Specialists in Group Work Sexual Addiction and Compulsivity and Addiction Professional