Substance Abuse as the “Missing Piece” in Outcomes Related to Housing First: Findings from the New York Services Study

Objective. The ‘housing first' approach for homeless mentally ill adults, which began in the early 1990s, garnered support from a quantitative four-year experimental trial showing improvements in housing stability, consumer choice and sense of mastery (Greenwood et al., 2005). These findings led to the endorsement of ‘housing first' as an ‘evidence-based practice' by Federal agencies (e.g. SAMHSA, HUD) in 2007. One noteworthy group difference--reduced use of drugs or alcohol--has remained elusive. Yet even non-significant group differences indirectly support ‘housing first' since its practice of harm reduction places few limits on substance use (compared to the strict sobriety requirements of the mainstream, or ‘treatment first' approach). This symposium paper revisits the issue by drawing upon qualitative data from the New York Services Study to address the following research questions: 1) are there substance use differences in ‘housing first' vs. ‘treatment first' consumers? 2) How (if at all) do group members differ in their attitudes about substance use and substance abuse treatment?

Methods. This report analyzed in-depth interviews with 27 ‘housing first' and 48 ‘treatment first' clients who were interviewed at 0-, 6- and 12 months after enrolling in their programs; data collection took place from 2005 to 2007. Of all eligible program enrollees who met study criteria (Axis I serious mental disorder and a history of homelessness and substance abuse), only one refused to give formal consent. Interviews inquired about service use, housing situation, social support, and substance use. Case study analyses were used to classify participants dichotomously as abstinent or substance-using for the 12 month period in order to carry out statistical analyses. Qualitative analyses used Boyatzis' method (1998) of coding and thematic development related to substance use.

Results. Chi square tests revealed much higher rates of abstinence in the ‘housing first' group (65% vs. 30%; p=.004) as well as a lower likelihood of using substance abuse treatment (7% vs. 46%; p=.001) and of going AWOL from their programs (11% vs. 54%; p=.000). Logistic regression analyses controlling for race and sex revealed strong group differences favoring housing first clients (OR=3.4; 95%CI=1.15-10.00). Qualitative themes included: stable housing as a platform for recovery, challenges in finding ‘positive' relationships and maintaining sobriety as an individual journey.

Conclusions/Implications. Findings are discussed in terms of ‘housing first' clients' ability to avoid using substances amidst harm reduction policies and despite living in drug-available neighborhoods. Concurrently, the dominant approach has difficulties in retaining clients and avoiding the higher treatment costs associated with relapse. Implications for practice with this vulnerable population are discussed in the context of recent and ongoing policy changes at the Federal and state levels.