Comparison of Housing Alternatives for SMI in NYC
Updated June 2002

Co-Investigators: Carole Siegel, Ph.D., Judith Samuels, Ph.D., Kim Hopper, Ph.D., Sue Barrow, Ph.D., Kristine Jones, Ph.D., Frank Lipton, MD (New York City Human Resources Administration), Anthony Hannigan, MSW (Center for Urban Community Services), Sam Tsemberis, Ph.D. (Pathways to Housing).  

Field Researchers: Ilysa Berg, Alicia Diaz, Claire Haiman, Kinjia Hinterland, John Jost, Nicole Laborde, David Vine 

• To describe the characteristics and costs of two housing models for PSMD – supported housing (separation of housing and services, autonomous living) and community residences (structured housing with bundled services); 

• To compare the outcomes of propensity score “matched” PSMD in these models over an 18-month period.  

This project is a collaboration with the New York City Office of Health and Mental Health Services of the Human Resources Administration, the agency responsible for reviewing housing applications and tracking placement outcomes for individuals with SMI; the Center for Urban Community Services, an agency offering services through several sites and providing housing consultation services; and Pathways to Housing, a supported housing project serving homeless persons with SMI and other disorders. The project is an out-growth of the NY/NY Housing: An Examination of Tenure project conducted by this Center. The study compares different types of housing (supported housing vs. a housing continuum) for individuals with SMI by studying housing programs and consumer outcomes in two phases:

Phase I, a process evaluation, will examine the operations and ecology of five prototypical housing approaches: two that meet the SAMHSA definition for Supported Housing and three that collectively represent a housing continuum. Each will be studied in aggregate as a prototype, using available computerized data. In addition, an exemplary site of each prototype that represents a particularly "good example" of its class has been chosen for study in detail using on-site interviews and focus groups. Semi-structured interviews with program administrators will gather information on history, philosophy, organizational structure and operations. Focus groups with staff and tenants will yield information on how the program (or place) actually works. In addition, a profile of newly housed tenants in each housing type will be describe their background and characteristics, including perceived role in selecting the housing, hopes for the future, skills and deficits, substance abuse and residential histories.

In Phase 2, the exemplar sites (augmented by other sites) will be the settings for an outcome evaluation that compares the supported housing prototype to others on the continuum. A quasi-experimental study will be conducted to test the hypothesis that there is a difference in outcomes among those in supported housing vs. those in other housing situations. Outcomes will cover housing tenure, level of functioning, consumer satisfaction and costs. Mediating factors, such as service use, relations with property management, and perceived stake in choosing housing will be included in models that provide an explanation of outcomes.

RESEARCH ACTIVITIES AND RESULTS

Methods: Phase I of the study, a process evaluation in which the operations and ecology of five prototypical housing approaches falling into these categories were delineated, was completed in June 1999. Within two models of supported housing, distinctive differences in definition of the problem to be solved, target clientele, potential supporters, and neighborhood development strategy were observed and are expected to impact housing outcomes. A paper is under preparation on the "genealogies" of these sites. In Phase 2 we conducted fidelity assessments of the housing models according to the guidelines and procedures developed by the coordinating center. We completed data collection for our quasi-experimental study testing the hypothesis that there is a difference in outcomes among tenants in supported housing in comparison to those in community residences.  

Data collection for the study is now complete and extensive data cleaning and data set preparation were ongoing during 2001-2. 

Results: The study sample is comprised of 158 individuals; 76 (48%) who were placed in supported housing (SH), and 82 (52%) who were placed in community residences (CR). Our completed study database includes: 158 baselines interviews, 136 3-month, 134 6- month, 125 12-months and 88 18-months interviews.  Initial findings show that at 12 months, the percentage of people moving from housing to unfavorable conditions is significantly higher for those placed in a CR (46%) versus  those in SH (18%). In our initial analysis, three measures of quality of life have been examined: empowerment, housing satisfaction, and the Lehman Quality of Life measure. While housing satisfaction is significantly higher for those in SH at 6 months, there is no difference at 12 months, and there is no significant difference between housing groups on the other quality of life measures at 6 or 12 months. Those in SH had less use of crisis and/or emergency room services at 6 months. However there was no significant difference between the groups at 12 months. Community integration from self-reports is higher for those in SH than CR at 6 and 12 months. Of interest are preliminary findings that depression/anxiety as measured at baseline has a significant impact on measures of isolation, empowerment, housing satisfaction, and overall quality of life at both 6 and 12 months in all cases resulting in significantly poorer scores. Analysis of the 6, 12 and recently completed 18 month data are on-going. In addition, a cost analysis of each housing model has been completed that decomposes housing costs into shelter and service costs. While shelter costs are similar, not unexpectedly, service costs are higher in the community residences (CRs). NYC housing costs were higher han those of any other sites in terms of unadjusted dollars.  

In addition three site-specific studies were conducted and are all in the analysis stage. In “Opening the Black Box” – information on housing characteristics gathered from the cross-site study has been supplemented by information gathered through qualitative and ethnographic methods to better understand both the fidelity dimensions and other housing and locale characteristics that can impact upon tenant outcomes and the social good.  A paper has been submitted for publication (Hopper, Barrow 2002).

In “Longer Term Outcomes,” 90 tenants (35 in supported housing, 45 in community residences) who have resided in housing from 18-24 months have been randomly selected and interviewed.  Comparisons of outcomes of the Cross-Site short term follow-up cohort vs. those of the long-term cohort, on outcomes believed to take longer to manifest, such as empowerment and recovery, are in process. 

Data collection is also nearly complete for the third site-specific study “A cost-effectiveness study covering costs of housing, service in housing, service in the community, criminal justice system involvement and earned income” with analysis underway and reported separately.

            Initial findings were presented in June 2001 to the NY State OMH Best Practices Conference. In addition, a cost analysis of each housing model has been completed and a report has been submitted to the cross-site coordinating committee.

Knowing  the types of housing that work best for which types of people can lead to more informed decision making in the development and support of specific housing types. While our initial analysis shows that those in SH are more stably housed at 12 months, results in propensity score stratum suggest that certain persons do very well in SH while others may do equally well and sometimes better in CRs.   

INCLUSION  OF GENDER AND MINORITY SUBJECTS: SAMHSA Housing Study

*3 persons not interviewed at baseline

Data analysis is ongoing and will include intent to treat analyses and efficacy analyses. Missing data imputation methods, required for efficacy analysis, are under development. Transition state modeling will be used to understand pathways through housing, leading to a better understanding of the meaning of "appropriate" housing.

        Study results will be disseminated through published papers and conference presentations.

[Top]

HOME  

Cores: Negotiating Lives in Communities | Methods for MH Services Research | Systems Integration |  Promoting Recovery  |
Topics: Service Delivery Systems | Homelessness | Instrument Development | Managed Care | Mental Illness and Chemical Abuse | Multi-cultural Issues |  Improving Services Research | Recovery | Treatment Innovations |
Statistical Methods and Computer Programs