Review of 57 Studies Backs Housing First Over Treatment-First Models

Programs that provide permanent housing without requiring sobriety or treatment compliance as a precondition keep adults who are homeless and living with substance use and mental health needs housed far more reliably than treatment-first models, according to a rapid review of 57 studies published Wednesday in Addiction Science & Clinical Practice.

The review, led by Oluwaseyi Dolapo Somefun of the Centre for Addiction and Mental Health in Toronto with colleagues from the University Health Network, the Canadian Centre on Substance Use and Addiction and the University of Toronto, screened 13,393 records and synthesized two and a half decades of peer-reviewed evidence on what housing-based support actually delivers.

What the reviewers examined

The team searched five databases — CINAHL, Embase, Medline, APA PsycInfo and Web of Science — for studies published between January 2000 and April 2025. From the screened records, 57 studies met inclusion criteria: eight randomized controlled trials, 30 quasi-experimental studies, six qualitative studies, five mixed-methods studies and eight using other designs. Study quality was appraised with Joanna Briggs Institute tools, and the reviewers extracted data on program features, comparison conditions and outcomes covering housing stability, substance use, mental health and service use.

Four distinct program architectures emerged. Scattered-site models place participants in independent apartments with support delivered where they live. Single-site models concentrate housing and services in one building. Transitional housing offers time-limited accommodation as a bridge to permanence. Outreach support-to-housing programs begin with street-level engagement and move people gradually indoors. Each carries different assumptions about what a person needs first.

Where the evidence is strongest

On housing stability, the findings were unambiguous. Housing First models outperformed treatment-first and linear treatment-contingent programs, and treatment-as-usual comparison conditions, across randomized trials and observational studies alike. The reviewers describe the mechanism in plain terms: requiring someone to demonstrate sobriety or complete treatment before they can access housing creates a barrier many people with active substance use disorders cannot clear, leaving them cycling through shelters, hospitals and the street.

The picture was more complicated for substance use itself. Effects were heterogeneous across studies, with alcohol-related outcomes improving more consistently than illicit drug use. Some trials in scattered-site Housing First found meaningful reductions in alcohol consumption and alcohol-related harm; results for stimulants and opioids varied widely, and several studies found no significant difference from comparison conditions. The reviewers read that heterogeneity as a signal that housing alone, while transformative for stability, does not resolve the full spectrum of health needs among people with co-occurring mental health and substance use conditions.

Mental health outcomes showed a similarly mixed pattern — some studies found no difference in psychiatric symptoms, others documented reductions in psychological distress. The reviewers caution against treating those results as a failure, noting that trajectories in this population are shaped by trauma history, physical health, social connection and access to psychiatric care, and that the evidence at minimum shows Housing First does not worsen mental health outcomes.

Evidence on hospital use and cost pointed the same direction but more weakly. Several studies suggested Housing First participants spend fewer days in hospitals and emergency departments, with offsets that partially or fully cover the cost of housing and support — a case the reviewers describe as plausible but not yet ironclad.

What the authors recommend

The review lands in the middle of a policy fight. The authors note explicitly that rapid shifts in several jurisdictions have favored treatment-focused approaches that frame addiction care as a gateway to housing, a direction the United States has moved toward through the treatment-first homelessness guidance issued by federal health agencies in August.

Their recommendations point the other way. The evidence supports scaling up low-barrier Housing First while integrating expanded voluntary substance use treatment, mental health care, harm reduction services and wraparound supports into housing itself. The emphasis on voluntariness is deliberate: the reviewers found no support for coercive treatment mandates as a condition of housing, but strong support for making choice-based services readily available to tenants who want them.

They also mapped what is still unknown. Treatment-first models did report positive outcomes in some studies, but the evidence base for them was thinner and more heterogeneous, making head-to-head comparison difficult. Longer-term comparative research is needed, the authors write, to identify which specific components best support substance use health among a population with some of the highest and most expensive needs in the health system.