The earliest studies of supportive housing found that between 75 and 85 percent of people who enter supportive housing remain housed one year later (Barrow et al., 2004; Lipton et al., 2000; Martinez & Burt, 2006; Wong et al., 2006). Over time, however, housing retention is diminished. Two years after entering housing, between 63 and 77 percent remained housed (Barrow et al., 2004; Lipton et al., 2000; Martinez & Burt, 2006). By three years, only 48 percent remained in housing in a California study with a high prevalence of individuals with substance use comorbidity (Martinez & Burt, 2006). At five years, 50 percent remained continuously housed in a study of nearly 3,000 New Yorkers in a range of high, moderate, and low intensity settings (Lipton et al., 2000).
The Pathways to Housing “Housing First” program, developed by New York City psychologist Sam Tsemberis in 1992, is one of the two program models that meet SAMHSA’s definition of permanent supportive housing. Through his work with a street outreach team, Tsemberis observed that the typical street-dwelling chronically homeless individual with severe mental illness and addictions was unable to meet the strict housing-readiness criteria of the prevailing stepwise continuum-of-care model that required treatment adherence and sobriety as a condition for obtaining housing. The Pathways program was developed to provide housing without preconditions. Individuals are offered housing in scatter-site community apartments, with services provided by assertive community treatment teams who implement a low-demand, individualized, harm- reduction treatment approach. A non-experimental study (Tsemberis & Eisenberg, 2000) and a controlled trial comparing the Pathways model to the continuum-of-care approach (Tsemberis et al., 2004) revealed housing retention rates of 80 percent or more. In the controlled trial, Pathways clients obtained housing earlier, remained stably housed longer over the two-year follow-up period, and reported greater perceived choice of housing.
More recently, findings on the efficacy of the housing-first model come from the "At Home/Chez Soi” two-year controlled trial of Housing First in five Canadian cities, funded by the Mental Health Commission of Canada. A “high need” sample of 950 homeless individuals with severe mental illness was randomly assigned to Housing First with assertive community treatment, or to treatment as usual. Housing First recipients were given a rent supplement, assistance in finding housing, and access to a community treatment team. Those assigned to the usual-care control condition had access to all other housing and treatment programs that existed in the community. At the two-year follow-up, 71 percent of Housing First participants had spent more time in stable housing, compared to 29 percent of participants in the usual-care control condition (Aubry et al., 2015a). In contrast to the usual- care group, Housing First participants who entered housing did so more quickly, had longer housing tenures at two years (281 days versus 115 days), and rated the quality of their housing more highly. In the final report of the At Home/Chez Soi study, 62 percent of high-need participants in Housing First were housed all of the time, 22 percent were housed some of the time, and 16 percent were not housed at all in the final six months of the two- year project. In the usual-care control condition, 31 percent were housed all of the time, 23 percent some of the time, and 46 percent were not housed at all (Mental Health Commission of Canada, 2014). Greater fidelity to the principles of Housing First (Tsemberis et al., 2004) was associated with improved housing outcomes (Davidson et al., 2014; Gilmer et al., 2014a; Goering et al., 2015).