Cost offsets of housing first programs in 5 Canadian cities

Background Few large and rigorous economic evaluations of Housing First programs have been conducted. Objectives To evaluate the effects on costs at one year of a Housing First intervention in five cities: Vancouver, Winnipeg, Toronto, Montreal and Moncton. Method As part of the At Home study, over 2,200 homeless people with mental illness entered a randomized trial of Housing First (HF) compared to treatment as usual (TAU). Participants with high needs (HN) received a more intensive clinical intervention than participants with moderate needs (MN). Service use and residential data were obtained through questionnaires administered at 6- and 3- month intervals. Actual costs of the intervention were used to replace self-reports, rendering our estimates of cost offsets more conservative. Total costs of health, social and justice service use, excluding medications, were estimated for three consecutive six-month periods starting six months prior to baseline. Results Follow-up rates exceeded 80% overall and were modestly higher in experimental than TAU groups. On average, by the 6-12 month period, HN participants receiving the intervention cost $47,754 compared to $42,992 for TAU participants ($4762, or 11%, more). For MN participants the corresponding figures were $36,585 compared to $31,998 ($4,587, or 13%, more). Discussion Savings from reductions in use of shelters and other resources by HF participants largely offset intervention costs. Considering also effectiveness results reported separately, It would be highly cost-effective to generalize Housing First programs to a much greater number of homeless people with mental illness in Canada.