Integrated treatment targeting co-occurring mental health and substance abuse problems were delivered through modified ACT services to improve housing stability, clinical outcomes, and quality of life for persons who were homeless or living in temporary and unstable housing (e.g., doubled up with friends).
- From March 2002 to December 2003, 560 persons with co-occurring mental health and substance use disorders, coupled with homeless or tenuous housing problems, were enrolled in the modified ACT study at Foundations Associates.
- More than 70% of persons enrolled in the program engaged in services lasting more than three months with almost 70% participating in IOP services specifically addressing co-occurring mental health and substance abuse treatment needs, although IOP treatment was not a requirement of the program.
- Program participants increased participation in community-based outpatient services while reducing utilization of costly emergency room services and inpatient hospital stays.
- While the program effectively engaged participants and appeared to increase treatment motivation by addressing basic needs and developing rapport, housing stability was drastically improved along with reduced mental health problems and successful reductions in substance use and rates of substance use relapse across 6- and 12-months following program enrollment.
- Quality of Life indicators consistently improved across all available QOL measures which included multiple subjective and objective ratings.
The intent of the project, Cooperative Agreements for the Development of Comprehensive Drug and Alcohol Systems for Homeless Persons, was to increase the effectiveness of services for homeless consumers with co-occurring disorders through a modified assertive community treatment (ACT) model utilizing key components of integrated treatment in combination with a stagewise approach. The general efficacy of ACT programs is well documented in the research literature for assisting individuals with severe mental illness (Bond et al., 2001; etc.) and improving housing stability and psychiatric symptoms for individuals with severe mental illness (Kenny et al 2005, etc.). Despite the extensive number of studies on the topic, relatively little is known about how ACT interventions impact treatment outcomes. Studies of integrated treatment interventions have also failed to provide clear evidence linking treatment processes with observed outcomes. As a result, there is good support for broad principles of integrated treatment, but we do not have an empirical basis to recommend specific treatment interventions, service dosages, staff training/composition, or other model characteristics. Based on the intended results of the program, key evaluation questions include:
1)Did program services reach the intended audience? The modified ACT program was intended for individuals with co-occurring disorders who were homeless or at-risk for homelessness. In addition to literally homeless, program services targeted individuals lacking stable housing (e.g., temporary doubling-up with friends or family members) as this instability coupled with co-occurring substance use and mental health problems required significant intervention to avoid long-term homelessness.
2)Did program services increase utilization of lower-cost community-based treatment and supports while reducing reliance on costlier institutionally based services?
3)Did program participants experience greater housing stability?
4)Were program participants engaged in treatment services? Many consumers were expected to show reluctance towards participation in outpatient treatment services that focus on substance abuse and mental health problems. Our program focused on meeting immediate consumer needs (e.g., food, shelter, etc.) and developing rapport using motivation approaches. Consumer trust in services and engagement in the program was expected to result in the majority of clients eventually participating in IOP services, although this was not a program requirement.
5)Did program participants reduce substance use, mental health problems, and quality of life?
The project modifies the assertive community team (ACT) model, demonstrated as efficacious for individuals with serious and persistent mental illness, adapting it to incorporate core practices empirically identified as effective in successful integrated treatment programs (Drake et al., 2001). Pairing derivate characteristics of ACT with key components of effective integrated treatment (i.e., assertive outreach, motivational enhancement, stagewise approaches, counseling/support, and long-term and comprehensive interventions), the model integrates core competencies by developing an integrated team of members with MH/SA experience, in addition to expertise in housing, vocational rehabilitation, outreach, and peer support.
The modified ACT program enrolled 560 participants from March 2002 to January 2004. Roughly half were male, 62% white and 37% black, and 68% were between 26 and 45 years of age. Almost half of participants did not complete high school (42%), 27% reported full or part-time work, and 56% reported less than $300 income in the last 30 days from all sources.
Illegal drug use was reported more frequently than alcohol use (57.8% compared to 49.4%), while cocaine use (42.1%) was more common than alcohol use to intoxication (39.2%). Compared to outpatient norms, participants reported more severe psychiatric symptoms overall and particularly with respect to phobia, paranoia, and psychosis symptom domains as measured using the BSI. General quality of life domains indicated lowest satisfaction with finances, while health related quality of life indicated poorer QOL for mental health compared to physical health related QOL.
Key Outcome Measures
The primary goal of the project is to reduce homelessness, substance use, and psychiatric symptomatology, along with utilization of institutionally based services (i.e., inpatient psychiatric care, inpatient substance dependency treatment, and emergent medical and psychiatric services), and to improve daily functioning, quality of life, and systems of social support.
- Housing: Individuals at risk for homelessness were those who did not have enough money for housing in the past month (QOL item 10.C), indicating difficulty with housing situation and avoiding homelessness in the last 30 days (COFD Activities item 8), and multiple days spent in inpatient, residential, treatment setting, jail, other institution or someone else’s home in the last 30 days. Literally homeless includes individuals currently housed in a shelter or living on the street.
- Substance Use: Substance use frequency is based on days of self-reported use of the last 30 days for alcohol use, alcohol use to intoxication (greater than four drinks), and other illegal drug use. Substance use abstinence is a dichotomous variable indicating any substance use in the past 3 months.
- Psychiatric Severity: The Brief Symptom Inventory (BSI; Derogatis, 1993) was used as an indicator of psychiatric severity. The BSI is a subset of the SCL-90-R with similar validity and reliability and a shorter administration time.
- Service Utilization: The Treatment Services Review (TSR) provides information on service utilization and service profiles.
Associated Outcome Measures
- Substance Abuse Symptom Severity: The Triage Assessment for Addictive Disorders (TAAD) was used to screen for symptoms of a possible current DSM-IV diagnosis for alcohol or other drugs. The TAAD has 16 items that address drug dependence and 19 that address alcohol dependence. The instrument assesses both dependence and abuse by establishing a pattern of behaviors and consequences rather than relying on a pattern of use.
- Quality of Life: The subjective measure of overall quality of life was measured with the short form of the Lehman Quality of Life Interview. Quality of Life was measured by 22 items rating satisfaction across seven domains: living arrangements, family, social relations, leisure activities, finances, legal and safety issues, and health. Two items rated overall life satisfaction. A mean of the items was obtained. Higher scores indicate greater life satisfaction.
- Functioning and Well-Being: The objective measures of the quality of life from the short form of the Lehman Quality of Life Interview were used to assess functioning and well-being in four domains: perceived overall functioning, social interaction, financial well-being, and perceived health status. Social interaction is the frequency of contact with family members and friends rated on a scale of 1 not at all’ to 5at least once a day.’ Financial well-being is the average of four items addressing whether the participant had enough money to cover daily living expenses such as food and clothing. Perceived overall functioning is a single item rating overall functioning in home, work, school, and social settings. Health status is a single item rating overall health. Both items are rated on a scale of 1excellent’ to 5 poor.’
- Overall Health Related Quality of Life: The Medical Outcomes Study Short-Form 12-Item Health Survey (SF-12) was used to measure overall health related quality of life based on subjective measures of physical and mental health functioning. Mental health and physical health indexes are represented by the Mental Component Summary (MCS) and Physical Component Summary (PCS) from the SF-12.
- Treatment Motivation: The University of Rhode Island Change Assessment (URICA; Prochaska & DiClemente, 1992) was used to measure the motivational state of the consumer during various phases of the treatment and recovery process.
Following informed consent, each participant completed the baseline and follow-up interviews conducted six and twelve months after the baseline date.
Program participants were invited to participate in the evaluation study and enrolled following signed informed consent with 560 participants enrolled from March 2002 to January 2004.
Of the 560 participants who enrolled in modified ACT program services, 287 also received IOP integrated treatment services at Foundations Associates for an average of 23 sessions. An additional 103 participants received residential integrated treatment services at Foundations Associates. Only 171 participants (30.5%) did not participate in additional IOP or residential integrated treatment programs at Foundations Associates. It should be noted that these integrated services were offered by Foundations Associates in addition to services offered within the multidisciplinary modified ACT team (e.g., specialized treatment groups, individual and group counseling) during the course of the three-year grant funded program. The average duration of treatment services was almost 8 months, including 71% who participated in treatment services for more than three months. Participants averaged more than 4 case management contacts per month.
During follow-up data collection activities, reported services received (in the last 30 days prior to each interview baseline, 6- and 12-months) is consistent with utilization patterns indicated by administrative data. At the 6-month follow-up, 54% reported outpatient service utilization in the last 30 days. This percentage declined to 34% reporting access to outpatient services in the 30 days prior to the 12-month follow-up.
Did program services reach the intended audience?
Baseline characteristics for consumers who received less than three months of services were compared to those who engaged in services for at least three months. Initial descriptive analyses identified potential predictor variables based on chi square statistic for categorical variables and t-tests for continuous variables. All variables that approached statistical significance (p value at or below 0.10) were included in further analysis using binary logistic regression (see Table 1). Given the exploratory intent of this analysis, all potential predictor variables identified in the previous step were entered as a block in the regression analysis to explore the extent to which at least marginally significant variables predicted program retention.
Of the categorical predictors included in the initial tests of potential predictors, only Housing Status = street and Ethnicity = black were significant baseline predictors. None of the continuous variables were significant predictors. According to subsequent binary logistic regression analyses, black participants were 1.7 times more likely to engage in at least three months of program services. Living on the street was not a significant predictor of engagement when controlling for the effect of ethnicity.
Did program services increase utilization of lower-cost community-based treatment and supports while reducing reliance on costlier institutionally based services?
Utilization of emergency room and inpatient hospital services declined steadily after enrollment in the modified ACT program (see Figure 1). During the 30 days prior to program enrollment, emergency room and inpatient services were higher than outpatient services utilization. As expected, outpatient services increased at follow-up with a sharp increase at 6 months and slight decrease at 12-months. At the same time, high-cost service utilization declined steadily at each follow-up, despite the slight decrease in outpatient services from 6-months to 12-months. This pattern suggests that the program successfully engaged clients in outpatient services while reducing reliance on high cost services which were utilized prior to enrollment in the modified ACT program.
Did program participants experience greater housing stability?
Evidence of greater access to permanent housing arrangements following enrollment in the program were strongly supported by several housing outcome indicators (see Figure 2). Almost 70% of participants were housed in their own or someone else’s apartment room or house 12 months after enrollment compared to approximately 55% at baseline. The percentage living in a shelter or on the street decreased from 11% at baseline to 7% 12 months following enrollment in the program. Those who indicated not having enough money for housing decreased from 60.2% at baseline to 32.8% at 12 months. As expected, given these improvements in accessible housing situations, participants reported far greater satisfaction with housing.
Were program participants engaged in treatment services?
More than 70% of program enrollees engaged in more than three months of services. According to URICA measures of treatment readiness, participants who engaged in services for more than three months also increased treatment readiness over time compared to those who were engaged in services for no more than three months (see Figure 3). Readiness for change (RFC) scores are calculated by adding contemplation, action, and maintenance, while omitting precontemplation scores. As a result, RFC scores indicate that engaged participants were more likely to progress beyond precontemplation over time. Interestingly, precontemplation scores for engaged participants were higher at baseline compared to those who did not engage in services, while these scores at follow-up were higher for the participants who did not engage in services.
Did program participants reduce substance use, mental health problems, and quality of life?
The frequency of substance use (see Figure 4) and relapse rates (see Figure 5) were reduced at both 6- and 12-months following enrollment in the program. For participants who reported illegal drug use at baseline, relapse rates were 35% 6-months and 28% 12-months later. The majority (56%) of participants who were using illegal drugs at baseline were able to avoid relapse at both 6- and 12-months. Similar positive outcome trends were similar across all major substance use categories.
Similarly, the percentage of participants reporting mental health problems declined at each follow-up compared to the initial baseline interview (Figure 6). Severity of psychiatric symptoms was also reduced at both follow-up interviews. Relative to outpatient treatment norms, BSI score percentiles dropped across all measured psychiatric domains (Figure 7).
Quality of life (QOL) was also associated with positive outcomes at both follow-up interviews compared to baseline indicators. Relative to baseline subjective ratings of satisfaction across multiple domains, all corresponding follow-up measures indicated greater satisfaction (Table 2). Objective QOL ratings provide further evidence that improved satisfaction with finances was due to greater access to income that covers basic needs such as food, clothing, and housing (Figure 8). Relatively consistent frequency of social and family interactions, coupled with increased satisfaction with family and social relationships, appears to indicate more appropriate and positive relationships rather than simply increasing interactions with others. Many participants were involved in dysfunctional and sometimes abusive relationships which clearly did not support recovery. Improved relationships and interactions is a critical step towards sustainable treatment improvements.
Acknowledgement and Disclaimer
This project described was supported by Grant Number TI12964 from the Substance Abuse and Mental Health Services, Center for Substance Abuse Treatment. The contents are solely the responsibility of the authors and do not necessarily represent the official views of SAMHSA/CSAT.