Evaluation Report Housing First Pilot Evaluation

Executive Summary Background. Connecticut's Housing First program, initiated in 2009, offers permanent housing through rental assistance certificates and supportive services to individuals with serious mental illness who are being discharged from psychiatric hospitals, or who are homeless and at risk of hospitalization. It is named for the New York-based program which focuses on homeless populations and includes Assertive Community Treatment (ACT), the highest level of community care. However, the CT program is more similar to the standard of Permanent Supportive Housing, which does not include ACT level of care. Methods. The planned outcome evaluation compared program participants to people on the wait list for housing. However, many study participants services did not represent their study condition; program participants often did not receive housing in a timely way (especially because of delays in receipt of certificates), and wait-list people did receive housing. It is not surprising, therefore, that no significant differences emerged between the study conditions as originally assigned. In order to better measure the impact of housing, the outcomes at each wave were analyzed by evaluating the impact of being housed, whenever housing may have occurred, rather than by a simple group comparison. We applied Hierarchical Linear Modeling to the full sample (N=38) to consider changes over time with respect to each outcome over a 12-month period. Findings. This study found that being housed was associated with significant improvements in client satisfaction, as well as several life domains, including satisfaction with living situation, community integration, satisfaction with family involvement, isolation, and sense of choice and empowerment. Housing was also significantly associated with reductions in problems related to mental health functioning overall, and subdomains of impulsive/addictive behavior, psychosis, role functioning, and depression. Moreover, hospital days out of a 9-month period dropped from an average of 153.6 (s.d. 77.7) to an average of 3.5 (s.d. 8.5). The only domains in which no improvements were made were physical health, medication adherence, and substance abuse. However, it should be pointed out that lack of change in medication adherence and substance abuse represents a somewhat positive finding, since these problems might have been expected to increase following independent housing. Conclusions. Although this study followed only a small number of people, it appears that offering housing with supportive services can result in many clinical and life improvements, without increasing risk for these clients. The reduction in inpatient days through use of Housing First is especially notable.