did housing first beat usual care in canada's at home/chez soi trial, especially for high needs?
“Housing first works much better than used well care especially for those that have the highest support needs those that have a primary psychologicalness.”
a guest, a psychiatrist and senior scientist at CAMH said on PsychEd: Educational Psychiatry Podcast, 2026-09-01, at 25:58. hear it in context → the speaker cited: the At Home/Chez Soi study in Canada.
the episode: PsychEd Episode 78: Mental Health and Homelessness with Dr. Vicky Stergiopoulos (at 25:58). we link the show's own page; we do not re-host audio.
first, the part that holds
the housing part holds up, and the gap is big. across the five cities, high-needs participants given housing first with assertive community treatment spent 71% of their time stably housed over two years, against 29% for usual care. the one long follow-up, which covers only the toronto site, found the advantage lasting six years and larger for high-needs participants than for moderate-needs participants.
what we found when we went and looked
we read the trial's published abstracts. housing first beat usual care on housing stability in both the high-needs arm (71% vs 29% of time stably housed, adjusted difference 42 points) and the moderate-needs arm (73% vs 32% across the whole trial). gains in quality of life and community functioning were small (standardized differences of 0.15 and 0.18 in the high-needs arm). there was no difference in psychiatric symptoms or substance use, and the moderate-needs arm showed no gain in generic quality of life. on "especially for high needs": the trial's own two-year review says both groups had "similarly superior housing outcomes". only the six-year toronto extension found a larger effect for high-needs participants (year-6 rate ratio 1.42 vs 1.13).
assertion by assertion
| assertion | verdict | what the record says |
|---|---|---|
| in at home/chez soi, housing first produced better housing outcomes than usual care | confirmed | high-needs arm (5 cities, n=950): 71% vs 29% of time stably housed over 24 months, adjusted absolute difference 42%, p<.01. housing came faster (73 vs 220 days) and tenancies lasted longer (281 vs 115 days). whole trial: 73% vs 32%. moderate-needs arm (4 cities, n=1,198): more days stably housed, though the size of the gap varied by site. limitation: all outcomes are measured over 24 months, except at the toronto site. [1][2][3] |
| housing first produced better outcomes than usual care beyond housing ("works much better") | overstated | quality of life and community functioning improved slightly in the high-needs arm (standardized mean differences 0.15 and 0.18). in that arm the quality-of-life gap showed at 12 months but was gone by 24. the trial's review reports no difference in psychiatric symptoms or substance use. the moderate-needs arm did not improve generic quality of life (EQ-5D difference 0.10, 95% CI -2.92 to 3.13, p=.95). the toronto high-needs report found no difference in health service use, community integration or substance use. [1][2][3][4] |
| the benefit was especially large for participants with the highest support needs | overstated | the two-year review of all sites: people with high needs receiving assertive community treatment and people with moderate needs receiving intensive case management "experienced similarly superior housing outcomes". the six-year toronto extension (n=575) found "a larger impact on individuals with high support needs than moderate support needs": rate ratios of 3.02 in year 1 and 1.42 in year 6 for high needs, against 2.40 and 1.13 for moderate needs. limitation: high and moderate needs were separate randomized arms with different support teams, so this is a comparison between two trials, not a tested interaction. [1][2] |
| the advantage persists over the long term | confirmed | toronto only: housing first stayed significantly ahead at every timepoint through year 6. the gap narrowed as usual-care participants found housing: high-needs usual care reached 60.33% of days stably housed in year 6, against 85.51% for housing first. |
both directions
for the speaker: the housing effect is large, replicated in five cities and both support-need arms, and lasted six years in toronto. high-needs participants given housing first with assertive community treatment had small but significant gains in quality of life and community functioning over two years. the only long-term data, from toronto, show a wider and longer-lasting housing gap for high-needs than for moderate-needs participants, which matches the speaker's "especially".
against: "works much better" fits housing, not clinical outcomes. no arm showed a difference in psychiatric symptoms or substance use. in the high-needs arm the quality-of-life gain was gone by 24 months. the moderate-needs arm showed no gain in generic quality of life at all. at two years, the trial's own review describes the housing benefit as similar in both need groups. the "especially for high needs" finding comes from a single-site extension, not the five-city main results. the housing gap also shrank over time as usual-care participants found housing.
what this cannot say
high-needs and moderate-needs participants were randomized in separate arms that received different support (assertive community treatment vs intensive case management). no direct, randomized comparison of effect size by need level exists, so "especially" can only be judged between two trials. long-term data come from toronto alone, a city with many resources. the publisher full texts (psychiatric services, jama, lancet psychiatry) blocked our fetches, so we worked from pubmed abstracts and open-access pmc articles. we also did not watch the episode: the quote reached us as a transcript with apparent errors ("used well care" is presumably "usual care"), and we judged the claim sentence we were given.
sources, each one fetched and read
- review of housing first for homeless people with severe mental illness, Canadian Journal of Psychiatry, 2015 (PMC4679127) What we read there: HF participants spent 73% of their time in stable housing vs 32% for TAU; "No differences emerged between the 2 groups regarding improvements in the severity of psychiatric symptoms or substance use"; high-needs (ACT) and moderate-needs (ICM) participants "experienced similarly superior housing outcomes"; QoL differences for ACT appeared at 12 months but not at 24 months. Fetched 2026-10-22 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC4679127/
- A Multiple-City RCT of Housing First With Assertive Community Treatment for Homeless Canadians With Serious Mental Illness, Psychiatric Services, 2016, PMID 26620289 (abstract via NCBI E-utilities) What we read there: 71% versus 29% of time in stable housing, adjusted absolute difference 42%, p<.01; time to housing 73 vs 220 days; quality of life ASMD=.15 and community functioning ASMD=.18, both p<.01. Fetched 2026-10-22 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/31601530/
- Effect of scattered-site housing using rent supplements and intensive case management on housing stability among homeless adults with mental illness: a randomized trial, JAMA, 2015, PMID 25734732 (abstract via NCBI E-utilities) What we read there: the intervention "resulted in increased housing stability over 24 months, but did not improve generic quality of life"; EQ-5D difference in mean changes 0.10 (95% CI -2.92 to 3.13), P=.95. Fetched 2026-10-22 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/25734732/
- How did a Housing First intervention improve health and social outcomes among homeless adults with mental illness in Toronto? Two-year outcomes from a randomised trial, BMJ Open, 2016 (PMC5030577) What we read there: high-needs HF+ACT vs TAU: mean difference in time stably housed 45.8% (95% CI 37.1% to 54.4%); no differences for health service use, community integration or substance use. Fetched 2026-10-22 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC5030577
speakers are described by role, not named. one sentence is quoted under fair use, with attribution and a link to the original at the timestamp; we do not re-host audio or video. nothing on this page is a reason to start, stop or change a medication. if you have the file behind anything marked unverifiable, the research board is where people can answer. adam