Transcription

1 This article was originally published in a journal published by Elsevier, and the attached copy is provided by Elsevier for the author s benefit and for the benefit of the author s institution, for non-commercial research and educational use including without limitation use in instruction at your institution, sending it to specific colleagues that you know, and providing a copy to your institution s administrator. All other uses, reproduction and distribution, including without limitation commercial reprints, selling or licensing copies or access, or posting on open internet sites, your personal or institution s website or repository, are prohibited. For exceptions, permission may be sought for such use through Elsevier s permissions site at:

2 International Journal of Law and Psychiatry 29 (2006) The relationship between mandated community treatment and perceived barriers to care in persons with severe mental illness Richard A. Van Dorn a,, Eric B. Elbogen a, Allison D. Redlich b, Jeffrey W. Swanson a, Marvin S. Swartz a, Sarah Mustillo a Abstract a Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, United States b Policy Research Associates, Delmar, New York, United States Received 24 October 2005; received in revised form 11 August 2006; accepted 25 August 2006 Objective: In recent decades debate has intensified over both the ethics and effectiveness of mandated mental health treatment for persons residing in the community. Perceived barriers to care among persons subjected to mandated community treatment, and the possibility that fear of involuntary treatment may actually create or strengthen such barriers rather than dissolve them, are key issues relevant to this debate but have been little studied. This article explores the link between receipt of mandated (or leveraged ) community treatment and reasons for avoiding or delaying treatment reported by persons with severe mental illness. It also examines the potential moderating effect of social support on the association between mandated treatment experiences and barriers attributable to fear of involuntary commitment or forced treatment. Method: Data are presented from a survey of 1011 persons with psychiatric disorders being treated in public-sector mental health service systems in five U.S. cities. Logistic and negative binomial regression analyses were used to examine the association between mandated community treatment and perceptions of barriers to care, controlling for demographic and clinical characteristics. Results: Across sites, 32.4% to 46.3% of respondents reported barriers attributed to fear of forced treatment. Whereas 63.7% to 76.1% reported at least one non-mandate-related barrier to care; the mean number of non-mandated barriers to care ranged from 1.6 to 2.3 (range 0 7). Between 44.1% and 59.0% of participants had experienced at least one type of leveraged treatment. Persons experiencing multiple forms of mandated treatment were more likely to report barriers to care in comparison to those not reporting mandated treatment. Findings also indicated that social support moderates the relationship between multiple leverages (three or four forms) and mandate-related barriers to care. Conclusions: Perceived barriers to care associated with mandated treatment experience have the potential to adversely affect both treatment adherence and therapeutic alliance. Awareness of potential barriers to care and how they interact with patients' perceived social support may lead to improved outcomes associated with mandated treatment Elsevier Inc. All rights reserved. Corresponding author. Department of Psychiatry and Behavioral Sciences, Duke University School of Medicine, DUMC Box 3071, Brightleaf Square Suite 23-A, 905 West Main Street, Durham, NC 27710, United States. Tel.: ; fax: address: richard.vandorn@duke.edu (R.A. Van Dorn) /$ - see front matter 2006 Elsevier Inc. All rights reserved. doi: /j.ijlp

3 496 R.A. Van Dorn et al. / International Journal of Law and Psychiatry 29 (2006) Introduction A substantial proportion of persons with severe mental illness (SMI) face serious obstacles to obtaining treatment in the community. Individuals with these obstacles to treatment are at increased risk for problems related to substance abuse (Swartz et al., 2006), non-adherence with medication and other recommended treatment (Swartz et al., 1998), violence (Swanson et al., 1997; Swanson, Holzer, Ganju, & Jono, 1990; Swanson et al., in press), victimization (Hiday, Swanson, Swartz, Borum, & Wagner, 2001; Swanson et al., 2002; Van Dorn, Mustillo et al., 2005), and social stigma (Link, Phelan, Bresnahan, Stueve, & Pescosolido, 1999; Penn & Martin, 1998; Pescosolido, Monahan, Link, Stueve, & Kikuzawa, 1999; Van Dorn, Swanson, Elbogen, & Swartz, 2005). This clustering of problems may increase the likelihood that persons with SMI will discontinue treatment altogether and then suffer relapse of acute psychiatric disorder. Consequently, individuals with these difficulties are more often subjected to various kinds of legal and social pressures leverage, in effect to compel or influence them to participate in mental health treatment. Types of leverage commonly include: outpatient commitment, criminal justice diversion programs such as mental health courts, probation linked to treatment participation, representative payeeship for disability benefits, and subsidized housing (Monahan et al., 2001; Monahan et al., 2005). For example, a representative payee may make receipt of a person's money contingent upon treatment and/or medication compliance; a landlord may do the same with a subsidized-rent apartment. Although potentially effective at improving treatment adherence and clinical outcomes, mandated community treatment is ethically controversial (Bonnie & Monahan, 2005), insofar as it can be perceived as unnecessarily coercive (Elbogen, Soriano, Van Dorn, Swartz, & Swanson, 2005). Such laws and social policies may involve limiting personal autonomy in order to promote other social values such as public safety or beneficence (Schopp, 1996). Besides the ethical concerns raised by mandated community treatment, many consumers and advocacy groups claim that even the threat of leveraged treatment is injurious and serves as a barrier to help seeking (Campbell & Schraiber, 1989). Indeed, Swartz, Swanson, and Hannon (2003) found that more than a third of patients with schizophrenia and related disorders reported fear of coerced treatment (e.g., fearing involuntary commitment, contact with law enforcement officials or forced medication) as a barrier to seeking treatment. These mandate-related barriers to care have the potential to undermine not only future help seeking, but also therapeutic alliance and treatment adherence. Prior research has indicated that the possible benefits of mandated community treatment are significant (cf., Swartz et al., 2001); however, it is also conceivable that mandate-related barriers could derail any potentially beneficial outcomes. In addition to barriers associated with mandated treatment, it is likely that non-mandate-related barriers to care,in combination with mandated community treatment, also have the potential to affect clinical outcomes. Specifically, these non-mandate-related barriers including lack of transportation to attend appointments, living instability or homelessness, and lacking confidence that treatment is effective may compound other problems, such as substance abuse and poverty, which are commonly experienced by persons with SMI (Elbogen, Swanson, Swartz, & Van Dorn, 2005), both of which portend poor access to and utilization of mental health services. The perceived stigma of mental illness may also prevent some people from seeking treatment or accepting the reality of having a psychiatric disorder (Link, Struening, Neese-Todd, Asmussen, & Phelan, 2001; Perlick et al., 2001; Sirey et al., 2001). A history of treatment non-adherence generally that means there are some significant barriers whether internal or external in nature that are standing between the person with mental illness and the treatment they need (McEvoy, Applebaum, Apperson, Geller, & Freter, 1989). In short, people with mental illness who are under the most leverage or pressure to adhere with treatment would also be expected to perceive the greatest barriers including fear of coercion preventing them from seeking or obtaining mental health care. Despite scholarly speculation, there has been little empirical research examining what factors might be related to the perception of fewer barriers to mental health care by people who are most often subject to legal or other mandates. One hypothesis is that social support is associated with perceiving fewer barriers to care among people with difficulties in managing their mental health (Kessler & McLeod, 1985; Sherbourne, 1988). In particular, if a person with SMI perceives a supportive social network, family and friends may help to facilitate the patient's engagement in mental health treatment or to potentially assist in the implementation of legal mandates. Conversely, if a person has little or no social support, he or she may not alone be able to surmount the perceived barriers to mental health care. As a result, one would expect that social support would bolster chances of both fulfilling treatment mandates and overcoming perceived barriers to care.

4 R.A. Van Dorn et al. / International Journal of Law and Psychiatry 29 (2006) Finally, our use of the term barrier in this article refers to all of the various forces that, in effect, prevent or limit access to treatment for persons with mental illness. Some of these forces are external in nature, such as lack of transportation or health insurance. Others, however, are more internal, including a person's beliefs and attitudes about mental illness and the perceived effectiveness of treatment. Some barriers involve complex interactions between external and internal factors; for example, stigmatizing public attitudes and stereotypes of persons with mental illness may interact with an individual's own internalized experience of stigma, negative self-appraisal, feelings of shame, and fear of repercussions for disclosing one's mental illness and seeking help (cf., Van Dorn, Swanson et al., 2005). The current study thus has two aims: (1) examine the relationship between mandated community treatment and reports of mandate- and non-mandate-related barriers to care; and (2) determine if the presence of social support moderates the relationship between mandated community treatment and these two types of barriers to care. To accomplish these aims, we analyze data from a new survey of 1011 persons in treatment for psychiatric disorders in the public-sector mental health service systems of five U.S. cities (Monahan et al., 2005). 2. Study design The methodology of this study is described in detail in Monahan et al. (2005). In brief, approximately 200 outpatients from publicly funded mental health treatment programs were sampled from each of five sites: Chicago, IL, Durham, NC, San Francisco, CA, Tampa, FL, and Worcester, MA. Sample inclusion criteria were: age years, speaker of English or Spanish, had first mental health treatment episode at least 6 months ago, had at least one outpatient treatment encounter with a publicly-supported mental health service provider within the past 6 months. Persons treated only for substance abuse, and not for any another psychiatric disorder, were excluded. Otherwise, the inclusion criteria did not specify particular mental health diagnoses or level of acuity. At the Worcester, Tampa, and San Francisco sites, potential subjects were recruited sequentially in the waiting rooms of outpatient clinics of the community mental health centers. In Durham, a list of potentially eligible subjects was created from management information system data, and these patients were randomly selected to be approached for participation in the study. The Chicago site used both sampling methods, enrolling about half the sample using the waiting room approach and the other half using the eligibility list approach. Participants were enrolled after receiving a complete description of the study and providing written informed consent. All sites received approval from their respective Institutional Review Boards. Refusal rates varied from 2% to 13% across sites. A single structured interview, lasting about 90 min, was administered in person by a trained lay interviewer. Participants were paid twenty-five dollars for the interview. 3. Measures 3.1. Dependent variables Mandate-related treatment barriers Two yes no items were used to assess respondents' perceptions regarding barriers associated with mandated treatment (adapted from Duke Mental Health Study). These items included: (1) If you think about getting help for mental health, alcohol or drug problems, do you delay getting help because, 1 you think that if you went for treatment that you might be forced to take medicine or treatment that you don't want? and (2) Has fear of being involuntarily committed ever caused you to avoid treatment for mental health, alcohol or other drug problems? This outcome was measured as a dichotomous variable with respondents answering either of the above two questions affirmatively coded as one and compared to those respondents answering no to both questions Non-mandate-related barriers to care Seven yes no items were used to assess respondents' perceptions regarding barriers associated with non-mandated treatment (adapted from Duke Mental Health Study). These items included, (1) If you think about getting help for mental health, alcohol or drug problems, do you delay getting help because, 1 (1) you think the problem might get better 1 If participants claimed they never delayed in getting help, not applicable was entered but all questions were asked regardless. Thirty-two participants (3.2% of the entire sample) claimed that they never delayed seeking treatment. These participants are not included in these analyses.

5 498 R.A. Van Dorn et al. / International Journal of Law and Psychiatry 29 (2006) by itself? (2) You are concerned about the cost? (3) You are unsure about where to go for help? (4) You think that going for help probably wouldn't do any good? (5) It is too difficult to get care because of distance or transportation problems? (6) You are concerned about what others might think if you went for help? And (7) you want to solve the problem on your own? This outcome was measured as an ordinal variable ranging from zero to seven Independent variables Leverage history Respondents' prior experiences with four types of leverage (via housing, representative payees, criminal justice or outpatient commitment; cf., Monahan et al., 2005) were modeled as a count variable of types experienced ranging from zero to three (the fourth category was collapsed into the third due to small cell size) Social support We used a single item to assess perceived social support. This item read, In times of trouble, can you count on at least some of your family or friends? Respondents indicating that this was true for them most of the time were compared to respondents indicating that they could count on family or friends hardly ever/never or only some of the time. Demographic variables included: age, gender (female = reference group), and racial status (nonwhite = reference group). Chart diagnoses were used to code participants' primary psychiatric disorder. These analyses compare psychotic disorder to all other disorders. Participants were asked about their alcohol and drug use during the past 30 days. Those who reported drinking any alcohol or taking any street drugs or non-prescribed psychoactive substances were asked follow-up questions from the CAGE Questionnaire (Allen, Eckardt, & Wallen, 1988). The CAGE Questionnaire is made up of four questions asking if people felt they needed to Cut down on their drinking, were Annoyed by people complaining about their drinking, felt Guilty about drinking, and if they need an Eye-opener in the morning. These same four questions were asked about drug use. For these analyses, we combined alcohol and drug abuse into a single dichotomous variable, coded 1 = one or more substance abuse symptoms and 0 = no substance abuse symptoms. The anchored version of the Brief Psychiatric Rating Scale (BPRS; Woerner, Mannuzza, & Kane, 1988) was used to assess current psychiatric symptoms. Possible scores range from 0 to 126, with higher scores indicating more symptomatology. Additionally, the Insight and Treatment Attitudes Questionnaire (ITAQ; McEvoy, Apperson et al., 1989; McEvoy, Applebaum et al., 1989) assessed participants' acceptance of their diagnosis and prescribed treatment (scores range from zero to twenty-two, with higher scores indicating higher levels of acceptance and agreement with diagnosis and treatment). The number of lifetime hospitalizations and the number of outpatient mental health treatment visits during the past month were included in the model as well; both of these factors were dichotomized above the median. Perceived coercion in mental health treatment was measured with a subscale of the MacArthur Admission Experiences Scale modified for outpatient use (Swartz, Wagner, Swanson, Hiday, & Burns, 2002). Scale items elicited participants' subjective experiences of force, threats, and other pressures to participate in outpatient mental health treatment. The scale includes fifteen items (1 = strongly agree, 5 = strongly disagree) and is calculated as the sum of responses indicating coercive experiences (α = 0.88 for current sample) divided by the total number of questions answered. Early onset of mental health or substance abuse problems was defined as respondents: (1) being told that they might need help for mental health or substance abuse problems; (2) seeking treatment for mental health or substance abuse problems; or (3) being admitted to a hospital for mental health or substance abuse problems, all prior to the age of sixteen. These independent variables were chosen for analysis based on prior clinical and epidemiological studies of risk factors related to perceived barriers to care and other outcomes for individuals with severe mental illness (Swanson, Swartz et al., 2006; Swanson, Van Dorn, Monahan, & Swartz, 2006; Swartz et al., 2003; Van Dorn, Swartz, Swanson, & Elbogen, 2005). We dichotomized covariates if warranted by their distribution or non-linear association with perceived barriers to care. All of the dichotomized variables were first tested in their original continuous distribution. If the categorical transformation revealed a stronger association with the dependent variable, we used the categorical version for the final analysis (cf., Farrington & Loeber, 2000). Finally, cross-sectional survey data do not provide clear evidence of cause and effect. However, by examining multivariate patterns of association in these data, we are able to refine the questions of interest to guide more definitive future studies of these issues.

6 R.A. Van Dorn et al. / International Journal of Law and Psychiatry 29 (2006) Statistical analysis Logistic regression was used to estimate bivariate and adjusted odds ratios for the dichotomous outcome (any reported barriers attributed to fear of involuntary treatment-mandate-related barriers.) A main effects multivariable model was estimated followed by a model that included the interaction between perceived social support and leveraged treatment and its impact on mandate-related barriers to care. For the ordinal outcome, non-mandate-related barriers to care, the Wilcoxon Mann Whitney Test was used to assess bivariate associations between the mean level of perceived barriers and the respective independent variables. The Wilcoxon Mann Whitney Test is appropriate for non-normally distributed outcomes (as is the case with this ordinal variable). Next, as perceptions of non-mandate-related barriers to care were modeled as a count variable, a negative binomial model was estimated due to the overdispersed nature of the outcome. The negative binomial model is similar to the Poisson model, but is used in cases of overdispersed count data (Hardin & Hilbe, 2001). We examined both deviance statistics from a Poisson model and alpha parameters from the negative binomial model to confirm that the negative binomial model was indeed the best fitting model. In the case of non-mandate-related barriers to care, the negative binomial model was a better fit than the Poisson model. Finally, for illustrative presentation of bivariate associations for both outcomes, continuous variables were dichotomized at the median (e.g., BPRS, perceived coercion, ITAQ). However, in the multivariable models, these same variables were analyzed in their original continuous state Pooling the data These data are drawn from five sites all of which contributed varying amounts of information for both the independent and dependent variables. Site-specific differences in both putative independent and dependent variables could result in outcomes that are overly influenced by site. For these analyses, in order to account for the effect of site, all regressions were run with site modeled as a fixed effect. Additionally, the Huber White sandwich estimator of variance to account for the clustering of observations within sites was used (Huber, 1967; White, 1982). Clustered Table 1 Cross-site profile of sample: means and percents Mean or percent SE Dependent variables Barriers Mandated barriers 40.55% Non-mandated barriers Independent variables Leverages Zero 48.66% One 28.98% Two 15.43% Three or four 6.92% Social support Count on family or friends most of the time 52.19% Demographic characteristics Age Male 50.35% Non-white 33.73% Clinical characteristics Psychotic diagnosis 45.00% Substance abuse 21.17% BPRS 32.21% 8.43 Insight (ITAQ) 18.24% 3.88 Perceived coercion in treatment Early onset of MH problems 51.33% Visits N 3/month 49.47% HospN 2 times 54.23%

7 500 R.A. Van Dorn et al. / International Journal of Law and Psychiatry 29 (2006) Table 2 Logistic regression analysis of effects on perceived mandate-related barriers to care Independent variables Bivariate associations Main effects model Interaction model OR (95% CI) OR (95% CI) OR (95% CI) Sum of leverages # 1 leverage 1.48 ( ) 1.14 ( ) 1.58 ( ) 2 leverages 2.60 ( ) 1.54 ( ) 1.76 ( ) 3 or 4 leverages 4.72 ( ) 2.37 ( ) 4.00 ( ) Social support 0.60 ( ) 0.70 ( ) 0.95 ( ) Soc Supp 1 leverage 0.52 ( ) Soc Supp 2 leverages 0.77 ( ) Soc Supp 3 or 4 leverages 0.30 ( ) Demographics Age ( 45 years) 0.82 ( ) 0.85 ( ) 0.84 ( ) Male 1.21 ( ) 0.92 ( ) 0.91 ( ) White 1.22 ( ) 1.38 ( ) 1.37 ( ) Clinical characteristics Psychotic diagnosis 1.28 ( ) 1.09 ( ) 1.09 ( ) Substance abuse 1.54 ( ) 1.21 ( ) 1.20 ( ) BPRS 1.06 ( ) 1.05 ( ) 1.05 ( ) Insight 1.01 ( ) 0.96 ( ) 0.96 ( ) Perceived coercion 2.27 ( ) 1.69 ( ) 1.68 ( ) in treatment Early onset of mental 1.15 ( ) 1.27 ( ) 1.30 ( ) health probs Visits N3/month 1.33 ( ) 1.09 ( ) 1.11 ( ) Prior hospitalizations (lifetime 4+) 2.34 ( ) 2.01 ( ) 2.03 ( ) # Compared to no leverages. Statistical significance: p b0.10; p b 0.05; p b 0.01; p b samples have larger variances than would occur with simple random sampling. Therefore, without adjusting for this, the precision of parameter estimates can be exaggerated and differences can appear statistically significant when they are not (Leaf, Myers, & McEvoy, 1991). Thus, all analyses, in addition to controlling for site as a fixed effect, use a robust variance estimator to account for design effects. All analyses were conducted using Stata 8.2 (StataCorp., 2003). 5. Results 5.1. Sample characteristics In Table 1, pooled means (and SEs) and percentages for sample characteristics are presented. Across the sites (pooled N=1011), 32.4% to 46.3% of respondents reported mandate-related barriers to care. Next, 63.7% to 76.1% of respondents reported at least one non-mandate-related barrier to care; the mean number of non-mandate-related barriers to care reported by respondents ranged from 1.6 to 2.3. Next, between 41.0% and 55.9% of respondents indicated not having experienced any of the four types of leverage, while between 25.3% and 36.0% reported one type of leverage, between 10.3% and 18.5% reported receiving two types of leverage, while between 5.9% and 8.0% reported experiencing three or four of the leverages at some time in their adult lives. Finally, between 39.0% and 61.2% of respondents indicated that they could count on family or friends most of the time. Regarding demographic characteristics across the five sites, the mean age of participants ranged from 41.3 to 46.7 years and the proportion of male subjects ranged from 32.4% to 64.5%. Finally, the proportion of respondents from minority (nonwhite) racial groups ranged from 28.5% to 64.0%. Regarding clinical characteristics, between 41.5% and 49.5% of respondents had a chart diagnosis of schizophrenia or another psychotic disorder, between 14.4% and 17.6% had bipolar disorder, and between 27.5% and 30.7% had major depression, while rates of substance abuse comorbidity ranged from 13.9% to 35.5%. Mean BPRS scores ranged

8 R.A. Van Dorn et al. / International Journal of Law and Psychiatry 29 (2006) Table 3 Cross-site bivariate associations with perceived non-mandate-related barriers to care Independent variables Mean number of barriers (0 7) Mean SD Z-value Significance Leverages Zero One Two Three or four Social support Count on family of friends most of the time Yes No Demographic characteristics Age ( 45 years) Above median Below median Gender Male Female Race White Non-white Clinical characteristics Psychotic diagnosis Yes No Substance abuse Yes No BPRS Above median Below median Insight (ITAQ) Above median Below median Perceived coercion in treatment Yes No Early onset of MH problems Yes No HospN 2 times Yes No Visits N 3/month Yes No Statistical significance: pb0.10; pb0.05; pb0.01; pb from 31 to 33 and mean ITAQ scores ranged from 18 to 19 across the sites; BPRS and ITAQ scores for this sample indicate moderate symptomatology and relatively high insight, both of which are consistent with prior outpatient samples of persons with severe mental illness (cf., Swanson et al., in press; Swanson et al., 2002). Mean scores assessing experiences with coerced mental health treatment ranged from one to five, indicating a broad range of experiences with and attitudes towards coercion; these experiences and attitudes range from coercion being viewed as no problem to a serious problem. Between 47.5% and 60.5% of respondents reported early onset of mental health or substance abuse problems. Finally, between 24.6% and 85.2% and 47.6% and 63.3% of respondents reported more than three outpatient visits per month and four or more lifetime hospitalizations, respectively.

9 502 R.A. Van Dorn et al. / International Journal of Law and Psychiatry 29 (2006) Table 4 Logistic regression analysis of effects on perceived non-mandated-related barriers to care Independent variables IRR Model 1 (95% CI) 5.2. Mandate-related barriers to care: bivariate and multivariable models IRR Model 2 (95% CI) Sum of leverages # 1 leverage 1.12 ( ) 1.19 ( ) 2 leverages 1.26 ( ) 1.17 ( ) 3 or 4 leverages 1.25 ( ) 1.47 ( ) Social Support 0.79 ( ) 0.83 ( ) Soc Supp 1 leverage 0.88 ( ) Soc Supp 2 leverages 1.14 ( ) Soc Supp 3 or 4 leverages 0.61 ( ) Demographics Age ( 45 years) 0.92 ( ) 0.91 ( ) Male 0.96 ( ) 0.95 ( ) White 1.08 ( ) 1.08 ( ) Clinical Characteristics Psychotic diagnosis 0.87 ( ) 0.87 ( ) Substance abuse 1.07 ( ) 1.07 ( ) BPRS 1.02 ( ) 1.02 ( ) Insight 0.99 ( ) 0.99 ( ) Perceived coercion in treatment 1.19 ( ) 1.18 ( ) Early onset of mental health probs 1.12 ( ) 1.13 ( ) Visits N3/ month 1.22 ( ) 1.23 ( ) Prior hospitalizations (lifetime 4+) 0.92 ( ) 0.92 ( ) /lnalpha Alpha # Compared to no leverages. Statistical significance: p b0.10; p b 0.05; p b 0.01; p b Table 2 presents bivariate and adjusted odds ratios for both the main effects and interaction models for respondents reporting mandate-related barriers compared to those reporting no mandate-related treatment barriers. First, unadjusted odds ratios indicate a significant and positive relationship between leverage and perceived mandate-related barriers to care. Specifically, compared to those with no history of leveraged treatment, those with one, two or three/four episodes of prior leveraged treatment were more likely to report barriers by a factor of 1.5, 2.6 and 4.7, respectively. The multivariable model (second column of Table 2) shows that reporting mandate-related barriers to care was significantly more likely among participants with three or more leverages (and a trend p=0.07 for those reporting two leverages), while there was a significant and negative relationship between social support and mandate-related barriers to care. In the demographic domain, mandate-related barriers to care were significantly associated with younger age and white race. In the clinical domain, mandate-related barriers to care were significantly associated with psychotic symptomatology (higher scores on the BPRS), perceived coercion in treatment, early onset of mental health problems, four or more lifetime hospitalizations, and less insight. We included an interaction term to explore the potential of social support to moderate the relationship between mandated community treatment and mandate-related barriers to care. This analysis (column 3 in Table 2) found that participants reporting social support and one or three/four leverages were less likely to also report mandate-related barriers to care. These findings indicate that social support moderates the original positive relationship between leverages and mandate-related barriers to care, particularly for those respondents with one or three/four leverages. Findings for the remaining covariates in the interaction model were consistent with those found in the main effects model Non-mandate-related barriers to care: bivariate and multivariable models Table 3 first presents ANOVA results for non-mandate-related barriers to care and the different levels of leverage experienced. The bivariate relationship between these two variables is significant; specifically, the more leverages

10 R.A. Van Dorn et al. / International Journal of Law and Psychiatry 29 (2006) endorsed, the more non-mandate-related barriers endorsed. Table 3 then presents the bivariate Wilcoxon Mann Whitney Tests between perceptions of non-mandate-related barriers to care and the remaining independent variables. Multivariable associations for the ordinal variable (non-mandate-related barriers to care) were tested using negative binomial regression procedures. The first model assesses main effects, while the second model includes the hypothesized moderating impact of social support and experiences with leverages. Model one (Table 4) shows that those with two types of leverages compared to those with no such leverages, are more likely to report increased barriers to care. The same model also shows a significant and negative relationship between social support and non-mandate barriers to care. Findings from the demographic and clinical domains are largely the same between the bivariate and multivariable models. Specifically, there was a negative relationship between age and non-mandate-related barriers to care. In the clinical domain, those with a psychotic diagnosis were less likely to report non-mandate-related barriers to care. However, there was a positive relationship between experiences with coerced mental health treatment, high scores on the BPRS, and frequent outpatient visits each month and perceptions of non-mandate-related barriers to care. Model two (Table 4) adds the interaction terms between perceptions of social support and number of leverages. In this model the interaction between reporting social support and having experienced three or four leverages tends toward significance ( p =0.07). All other covariates that were significant in model one of Table 4 remained so in model two. 6. Discussion This study examined sociodemographic and clinical factors that may impact perceived barriers to care barriers attributed to fear of involuntary treatment as well as other kinds of barriers among persons with severe mental illness. In the present study, there are at least two findings of note. First, persons who experienced multiple forms of leverage were more likely to report encountering barriers to care (especially mandate-related barriers) in comparison to persons who did not experience leverages. Results indicated that as experiences with leverages increased, so to did the likelihood of reporting mandate-related barriers to care. Specifically, the likelihood of reporting mandate-related barriers to care increased from 1.5 to 2.6 to 4.7 for those reporting one, two or three/four instances of leveraged treatment, compared to those with no prior leveraged treatment. On average, for the non-mandate-related barriers, persons who did not experience leverages via the social welfare or legal systems reported less than two barriers, whereas those who experienced three or four leverages reported more than three barriers. Second, we found that social support moderates the relationship between multiple (three or four forms) leverages and perceived barriers to care. Multivariate analysis confirmed this pattern for different types of barriers to care, especially mandate-related barriers to care. These findings can be interpreted several ways. On the one hand, persons with more barriers to care are more likely to be non-compliant with treatment, and will thus be subject to treatment mandates and imposed leverages from professionals and non-professionals (e.g., parents, spouses). On the other hand, persons who are non-compliant (regardless of barriers) and subject to formal and informal mandates (or pressures) may have a higher number of perceived barriers to care because of the treatment mandate. For example, a person who has been mandated to treatment via outpatient commitment has an obligation to attend treatment and take prescribed medications; if noncompliant, the consequences can be significant, such as having to return to an inpatient setting (Monahan et al., 2001). Thus, people who are subject to these specific, pre-set requirements may perceive it more difficult to adhere and perceive more obstacles to doing so; there is some evidence for this. In the larger study utilizing these data (Monahan et al., 2005), persons with more outpatient visits were more likely to have experienced at least one form of leverage. In the present study, we also found that persons with more outpatient visits reported more barriers to care (non-mandaterelated). Thus, it may be that persons who are pressured and/or required to attend treatment more often are the ones who are most likely to encounter barriers. In other words, for people who are mandated to attend treatment on a regular ongoing basis, it may be harder for them to pay for repeated appointments, to find transportation, etc. Regardless of whether leverage precedes perceived barriers, there are important implications from our finding that social support moderates the relationship between multiple leverages and perceived barriers to care. In previous studies, social support among persons with mental illness has been studied, but the relationship between barriers and social support has only been examined in a few studies. Lam and Rosenheck (1999) surveyed homeless persons with serious mental illness and found that (1) younger people and persons with psychotic disorders had less social support, and (2) social support and access to services were positively related. Likewise, in the present study, younger age (for both outcomes) was associated with a higher number of perceived barriers, and certainly barriers to care influence access to care.

11 504 R.A. Van Dorn et al. / International Journal of Law and Psychiatry 29 (2006) The current findings are also consistent with results of a study of the interaction between social support (size of the social network) and well-being when examined by level of housing-related stress among persons with SMI (Earls & Nelson, 1988). For clients with high levels of housing stress, having a small social network was associated with higher levels of negative well-being. A similar relation between social support and negative well-being was not found for those with low levels of housing stress. Earls and Nelson's finding is similar to the interaction found in the present study concerning the effects of leverage and social support on perceptions of barriers to care. Thus, there is converging evidence that, among persons with severe and persistent mental illness, negative experiences such as housing stress and impediments to services can be moderated by the number, and perhaps the quality, of social support systems. These analyses suggest that distinct, but overlapping profiles are linked to perceptions of both mandate- and nonmandate-related barriers to care in persons with severe mental illness. Overlapping characteristics include a negative association between both social support and age and positive relationships between scores on the BPRS and perceived coercion. Early onset of mental health problems, multiple prior hospitalizations and white race were uniquely and positively associated with mandate-related barriers to care, while insight was negatively associated with the same. Finally, having a psychotic diagnosis and reporting more than three mental health visits per month were uniquely associated with non-mandate related barriers to care. There are limitations to the study that should be addressed. Because of the cross-sectional nature of the data, our data do not allow for definitive answers regarding whether treatment mandates and leverages lead to more perceived barriers to care, or vice versa. However, this is an important question for future research and a prospective, longitudinal study would provide more insight into the relations between perceived barriers to care, social support, and treatment leverage experiences. Additionally, although we controlled for site-level differences, the five data collection sites were not chosen at random and two slightly different recruitment methods were used to obtain the samples. The samples also differed from broader, nationally representative, survey samples of individuals receiving specialty mental health services in the community (cf., Monahan et al., 2005). For example, our samples included higher proportions of males, individuals with psychotic disorders, and high-frequency outpatient service users compared to respondents from the NIMH National Comorbidity Study (Kessler, 2003). Further, since two of these variables (psychotic disorders and high-frequency outpatient service users) were significantly associated with non-mandate-related barriers to care, it is possible that our samples' estimates of these barriers are somewhat different than would have been obtained in a random national survey of persons in treatment for mental disorder. Finally, our measure of social support was assessed with one item. While this single-item allowed for the assessment of both availability of and satisfaction with social support, future research could utilize a more finely detailed measure that assesses other aspects of social support not captured in the current research. In sum, the present study is one of a few to examine the complexities surrounding treatment mandates and impediments to accessing such treatment. We find a significant association between experiencing leveraged community treatment and reporting more barriers to care both mandate-related and non-mandate-related barriers. Yet the causal direction of this relationship is unclear. This finding may partly reflect clinicians' practice of targeting leverage to persons who otherwise would not participate in treatment on their own, due to some combination of internal resistance, impaired functioning, and external obstacles such as inadequate transportation and lack of accessible services. Leverage is thus offered as a sort of remedy to overcome these various barriers to care. Ironically, however, efforts to use leverage to improve treatment adherence may, in some cases, have the opposite effect by further alienating some individuals and strengthening the internal barriers that keep them from participating voluntarily in treatment. Notably, we find that social support can serve to moderate these deleterious effects of leverage. One can only speculate regarding the mechanisms involved, but perhaps family and friends provide greater access to, and ongoing support for treatment, and in so doing buffer the potential negative consequences of coercion. It is also possible that social support was provided by clinicians or other formal service providers (Faccincani, Mignolli, & Platt, 1990; Jacoby & Kozie-Peak, 1997; Shaw et al., 2000), given that support networks can include anyone who provides such support in a time of need (Cunninghan & Barbee, 2000). If so, supportive clinicians might be in a good position to help mitigate the potential negative consequences of perceived barriers associated with leveraged community treatment. Further, it is possible that the leveraged community treatment orders could serve as a means to mobilize the consumer's existing social support network. To the extent that clinicians become aware of the consumer's available social supports, and are able to integrate those resources into the implementation of the mandated treatment, this would seem to reduce perceived barriers and increase the likelihood of successful outcomes. However, the most intuitively plausible explanation may simply be that supportive social networks whether they include family, friends, clinicians, or emanate more broadly from other community or religious affiliations contribute to a greater quality of life for the

12 R.A. Van Dorn et al. / International Journal of Law and Psychiatry 29 (2006) consumer. Insofar as that quality of life is attributed also to participation in mental health treatment, then the perceived value of treatment may be enhanced, so that despite receiving formal or informal pressures to stay in mental health treatment, the consumer may not experience such pressures, or the treatment itself, as something to be feared or avoided in the future. Perceived barriers to care can be problematic from the point of view of consumers as well as providers of mental health services. For consumers, these barriers can delay or prevent receiving beneficial treatment in the community. For providers, the barriers may sometimes thwart their best attempts to deliver high-quality services to a high-priority population with great needs. In this regard, results of this study suggest that leveraged community treatment has both an upside and a downside. The disadvantages lie in the potential of leverage to actually increase barriers related to consumers' fear of coercion in future encounters with the treatment system; but this must be weighed against the advantages of engaging some consumers in treatment who otherwise will forego its benefit. Finally, efforts to foster and strengthen consumers' natural social support networks, in addition to support from clinicians, may go a long way toward mitigating the potentially negative effects of leverage. Acknowledgments This work was supported by the John D. and Catherine T. MacArthur Foundation Research Network on Mandated Community Treatment and the National Research Service Award Postdoctoral Traineeship from the National Institute of Mental Health (T32MH19117) sponsored by Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill for Dr. Van Dorn. The authors thank John Monahan, PhD and Paul Appelbaum, M.D. for their comments on an earlier draft of this manuscript. References Allen, J. P., Eckardt, M. J., & Wallen, J. (1988). Screening for alcoholism: Techniques and issues. Public Health Reports, 103, Bonnie, R., & Monahan, J. (2005). From coercion to contract: Reframing the debate on mandated community treatment for people with mental disorders. Law and Human Behavior, 29, Campbell, J., & Schraiber, R. (1989). The Well-Being Project: Mental health clients speak for themselves. Sacramento: California Network of Mental Health Clients. Cunningham, M. R., & Barbee, A. P. (2000). Social support. In C. Hendrick & S.S. Hendrick (Eds.), Close relationships: A sourcebook (pp ). Thousand Oaks, CA: Sage. Earls, M., & Nelson, G. (1988). The relationship between long-term psychiatric clients' psychological well-being and their perceptions of housing and social support. American Journal of Community Psychology, 16, Elbogen, E. B., Soriano, C., Van Dorn, R. A., Swartz, M. S., & Swanson, J. W. (2005). Consumer perspectives on the use of disability funds to leverage adherence to mental health services. Psychiatric Services, 56(1), Elbogen, E. B., Swanson, J. W., Swartz, M. S., & Van Dorn, R. A. (2005). The impact of depressive symptoms and social instability on medication nonadherence in psychotic disorders. Journal of Nervous and Mental Disease, 193(10), Faccincani, C., Mignolli, G., & Platt, S. (1990). Service utilization, social support, and psychiatric status in a cohort of patients with schizophrenic psychoses: A seven-year follow-up study. Schizophrenia Research, 3, Farrington, D. P., & Loeber, R. (2000). Some benefits of dichotomization in psychiatric and criminological research. Criminal Behaviour and Mental Health, 10, Hardin, J., & Hilbe, J. (2001). Generalized linear models and extensions. College Station, TX: Stata Press. Hiday, V. A., Swanson, J. W., Swartz, M. S., Borum, R., & Wagner, H. R. (2001). Victimization: A link between mental illness and violence? International Journal of Law and Psychiatry, 24(6), Huber, P.J. (1967). The behavior of maximum likelihood estimates under nonstandard conditions. Paper presented at the Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability. Jacoby, J. E., & Kozie-Peak, B. (1997). The benefits of social support for mentally ill offenders: Prison-to-community transitions. Behavioral Sciences and the Law, 15, Kessler, R. (2003). The national comorbidity survey. In M. Tsuang, M. Tohen, & G. Zahner (Eds.), Textbook in psychiatric epidemiology, (2nd ed.). New York: John Wiley and Sons. Kessler, R. C., & McLeod, J. D. (1985). Social support and mental health in community samples. In S. Cohen & S.L. Syme (Eds.), Social support and health (pp ). Orlando, FL: Academic Press. Lam, J. A., & Rosenheck, R. (1999). Social support and service use among homeless persons with serious mental illnesses. International Journal of Social Psychiatry, 45, Leaf, P., Myers, J., & McEvoy, L. (1991). Procedures used in the Epidemiologic Catchment Area Study. In L. N. Robins & D.A. Reiger (Eds.), Psychiatric disorders in the America: The Epidemiologic Catchment Area Study (pp ). New York: Free Press.

13 506 R.A. Van Dorn et al. / International Journal of Law and Psychiatry 29 (2006) Link, B. G., Phelan, J. C., Bresnahan, M., Stueve, A., & Pescosolido, B. A. (1999). Public conceptions of mental illness: Labels, causes, dangerousness, and social distance. American Journal of Public Health, 89(9), Link, B. G., Struening, E. L., Neese-Todd, S., Asmussen, S., & Phelan, J. C. (2001). Stigma as a barrier to recovery: The consequences of stigma for the self-esteem of people with mental illnesses. Psychiatric Services, 52(12), McEvoy, J. P., Apperson, L. J., Appelbaum, P. S., Ortlip, P., Brecosky, J., Hammill, K., et al. (1989). Insight in schizophrenia: Its relationship to acute psychopathology. Journal of Nervous and Mental Disease, 177(1), McEvoy, J. P., Applebaum, P. S., Apperson, L. J., Geller, J. L., & Freter, S. (1989). Why must some schizophrenic patients be involuntarily committed? The role of insight. Comprehensive Psychiatry, 30(1), Monahan, J., Bonnie, R. J., Appelbaum, P. S., Hyde, P. S., Steadman, H. J., & Swartz, M. S. (2001). Mandated community treatment: Beyond outpatient commitment. Psychiatric Services, 52(9), Monahan, J., Redlich, A., Swanson, J. W., Robbins, P. C., Appelbaum, P. S., Petrila, J., et al. (2005). Use of leverage to improve adherence to psychiatric treatment in the community. Psychiatric Services, 56, Penn, D. L., & Martin, J. (1998). The stigma of severe mental illness: Some potential solutions for a recalcitrant problem. Psychiatric Quarterly, 69 (3), Perlick, D. A., Rosenheck, R. A., Clarkin, J. F., Sirey, J. A., Salahi, J., Struening, E. L., et al. (2001). Stigma as a barrier to recovery: Adverse effects of perceived stigma on social adaptation of persons diagnosed with bipolar affective disorder. Psychiatric Services, 52(12), Pescosolido, B. A., Monahan, J., Link, B. G., Stueve, A., & Kikuzawa, S. (1999). The public's view of the competence, dangerousness, and need for legal coercion of persons with mental health problems. American Journal of Public Health, 89(9), Schopp, R. F. (1996). Communicating risk assessments: Accuracy, efficacy, and responsibility. American Psychologist, 51, Shaw, W. S., Patterson, T. L., Semple, S. J., Halpain, M. C., Koch, W. L., Harris, M. J., et al. (2000). Use of community support services by middleaged and older patients with psychotic disorders. Psychiatric Services, 51, Sherbourne, C. D. (1988). The role of social support and life stress events in use of mental health services. Social Science & Medicine, 27(12), Sirey, J. A., Bruce, M. L., Alexopoulos, G. S., Perlick, D. A., Friedman, S. J., & Meyers, B. S. (2001). Stigma as a barrier to recovery: Perceived stigma and patient-rated severity of illness as predictors of antidepressant drug adherence. Psychiatric Services, 52(12), StataCorp. (2003). Stata Statistical Software: Release 8.0. College Station, TX: StataCorp LP. Swanson, J., Estroff, S., Swartz, M., Borum, R., Lachicotte, W., Zimmer, C., et al. (1997). Violence and severe mental disorder in clinical and community populations: The effects of psychotic symptoms, comorbidity, and lack of treatment. Psychiatry, 60(1), Swanson, J. W., Holzer, C. E., Ganju, V. K., & Jono, R. T. (1990). Violence and psychiatric disorder in the community: Evidence from the Epideiologic Catchment Area Surveys. Hospital & Community Psychiatry, 41(7), Swanson, J. W., Swartz, M. S., Elbogen, E. B., Van Dorn, R. A., Ferron, J., Wagner, H. R., et al. (in press). Facilitated psychiatric advance directives: A randomized trial of an intervention to foster advance treatment planning among persons with severe mental illness. American Journal of Psychiatry. Swanson, J. W., Swartz, M. S., Essock, S. M., Osher, F. C., Wagner, H. R., Goodman, L. A., et al. (2002). The social-environmental context of violent behavior in persons treated for severe mental illness. American Journal of Public Health, 92(9), Swanson, J. W., Swartz, M. S., Van Dorn, R. A., Elbogen, E. B., Wagner, H. R., Rosenheck, R., et al. (2006). A national study of violent behavior in persons with schizophrenia. Archives of General Psychiatry, 63, Swanson, J. W., Van Dorn, R. A., Monahan, J., & Swartz, M. S. (2006). Violence and leveraged community treatment for persons with mental disorder. American Journal of Psychiatry, 163, Swartz, M. S., Swanson, J. W., & Hannon, M. J. (2003). Does fear of coercion keep people away from mental health treatment? Evidence from a survey of persons with schizophrenia and mental health professionals. Behavioral Sciences and the Law, 21(4), Swartz, M. S., Swanson, J. W., Hiday, V. A., Borum, R., Wagner, H. R., & Burns, B. J. (1998). Violence and severe mental illness: The effects of substance abuse and nonadherence to medication. American Journal of Psychiatry, 155(2), Swartz, M. S., Swanson, J. W., Hiday, V. A., Wagner, H. R., Burns, B. J., & Borum, R. (2001). A randomized controlled trial of outpatient commitment in North Carolina. Psychiatric Services, 52, Swartz, M. S., Wagner, H. R., Swanson, J. W., Hiday, V. A., & Burns, B. J. (2002). The perceived coerciveness of involuntary outpatient commitment: Findings from an experimental study. Journal of the American Academy of Psychiatry and the Law, 30(2), Swartz, M. S., Wagner, H. R., Swanson, J. W., Stroup, T. S., McEvoy, J. P., Canive, J. M., et al. (2006). Substance use in persons with schizophrenia: Baseline prevalence and correlates from the NIMH CATIE study. The Journal of Nervous and Mental Disease, 194, Van Dorn, R. A., Mustillo, S., Elbogen, E. B., Dorsey, S., Swanson, J. W., & Swartz, M. S. (2005). The effects of early sexual abuse on adult risky sexual behaviors among persons with severe mental illness. Child Abuse & Neglect, 29, Van Dorn, R. A., Swanson, J. W., Elbogen, E. B., & Swartz, M. S. (2005). A comparison of stigmatizing attitudes toward persons with schizophrenia in four stakeholder groups: Perceived likelihood of violence and desire for social distance. Psychiatry, 68, Van Dorn, R. A., Swartz, M. S., Swanson, J. W., & Elbogen, E. B. (2005). Perceived fairness and effectiveness of leveraged community treatment among public mental health consumers in five U.S. cities. International Journal of Forensic Mental Health, 4, White, H. (1982). Maximum likelihood estimation of misspecified models. Econometrica, 50, Woerner, M. G., Mannuzza, S., & Kane, J. M. (1988). Anchoring the BPRS: An aid to improved reliability. Psychopharmacology Bulletin, 24,