A new study out of the United States, led by UC Berkeley researcher Sydney Costantini, provides some of the strongest evidence to emerge in recent years that Housing First not only helps people stay housed but also plays a crucial role in saving lives. The findings align closely with what Canadian researchers have shown for over a decade through the At Home/Chez Soi project, while extending the evidence in an important new direction: long-term survival.
This research arrives at a critical moment. In the United States, the federal government has moved to dismantle Housing First funding in favour of treatment-based approaches. A similar shift is underway in parts of Canada. Alberta has been reshaping its approach to homelessness and addiction around a recovery-oriented, treatment-first model, including proposed legislation that would allow family members, healthcare professionals, or police to compel a person into treatment against their will.
In Ontario, the Safer Municipalities Act, passed in June 2025, gave municipalities and police greater authority to remove encampments from public spaces, a shift toward enforcement that has drawn widespread criticism from legal experts, housing advocates, and health organizations.
In this context, Costantini’s findings carry weight far beyond what they say about Housing First. They show, just as clearly, that Treatment First does not improve long-term outcomes. Costantini used data from the U.S. Department of Veterans Affairs to follow nearly 300,000 unhoused veterans with mental illness over three years, making it by far the largest study of its kind ever conducted. Participants who received permanent housing vouchers with no conditions attached had a 4.6 percentage point reduction in three-year mortality compared to those who received no intervention. Given that roughly 6.6% of the overall study population died within three years, this is a substantial effect, meaning Housing First prevented a meaningful share of all deaths that would otherwise have occurred in this group. The effect also grew stronger over time.
What makes this study significant
Previous research has consistently shown that Housing First keeps people housed. What has been harder to demonstrate is whether it improves long-term health outcomes and whether it outperforms other approaches. Costantini’s study gives us insights into both of these questions.
Rather than comparing Housing First to “standard of care,” a category that typically groups together people in treatment programs with those receiving no help at all, she separately tracked a third “treatment first” group: people in transitional housing with required mental health and addiction supports.
The table below summarizes the key differences between the two models, and what the study found:
| Housing First | Treatment First | |
| Housing type | Permanent, subsidized housing | Transitional communal housing |
| Duration | Ongoing, no time limit | Up to 2 years; short term and time-limited |
| Treatment required? | No | Yes (mental health & substance use) |
| Sobriety required? | No | Often |
| 3-year mortality effect | 4.6 percentage point reduction | No significant effect |
The results? Treatment First participants showed no lasting improvement in health outcomes. Mental health increased while people lived in the facilities, but those gains faded after they exited and did not translate into better long-term health outcomes.
How does Housing First save lives?
During the study, Housing First participants were significantly less likely to die from exposure to extreme heat and cold, from violence, and from infections associated with unsafe injection drug use, particularly Hepatitis C, which accounted for around 76% of the reduction in infection-related deaths. These are deaths directly caused by the absence of safe, stable housing.
That Hepatitis C finding is essentially an argument for harm reduction approaches to drug use. The evidence suggests that stable housing reduces the risky injection behaviours that spread blood-borne infections, such as needle sharing. Harm reduction is already central to Housing First practice, but this study provides some of the strongest quantitative evidence yet that it saves lives.
Interestingly, the survival benefit does not appear to come from increased access to mental health treatment. Housing First participants did not seek out more mental health care once they were housed, which challenges the common assumption that stable housing will naturally lead people to address underlying health issues. The evidence suggests that for many people, housing itself is the intervention.
We have all heard the argument that people need to address their mental health and addiction issues before they can be “ready” for housing. Costantini’s research found that requiring treatment readiness before providing housing does not improve outcomes and could instead cause harm by delaying the stability people need to survive.
Housing First is also cost-saving
The study found that reductions in inpatient psychiatric stays more than offset the cost of housing vouchers over the three-year study period, making Housing First not only the more effective intervention but also the more financially sustainable one.
This finding is consistent with Canadian evidence. The At Home/Chez Soi cost-effectiveness analysis, led by McGill researcher Eric Latimer, found that for people with severe mental illness, savings in emergency shelters, health visits, and incarcerations reduced the net cost of Housing First from around $20,000 to $6,300 per person per year. For the highest-need individuals, every $10 invested in Housing First generated $21.72 in cost savings.
A broader systematic review of 20 Housing First programs across Canada and the U.S. found societal cost savings of $1.80 for every dollar invested in Housing First, driven by reductions in healthcare, emergency housing, and justice system costs. Taken together, the Canadian and US evidence make a compelling financial case for Housing First.
Why this study matters for Housing First in Canada
Canada’s homelessness crisis continues to worsen. The number of people living unsheltered has more than doubled in recent years, increasing by 107% between 2020 and 2024. In Ontario alone, 85,000 people were experiencing homelessness in 2025. Indigenous Peoples, who are grossly overrepresented in these numbers, continue to bear the heaviest burden.
Meanwhile, Canada has been a global leader in research on Housing First. The At Home/Chez Soi project, still the world’s most extensive randomized control trial of Housing First, demonstrated that even people experiencing chronic homelessness, with the most complex mental health and addiction challenges, could be successfully housed and could thrive. That evidence helped shape housing policy across the country, including federal investment through the Reaching Home program.
Despite that investment, homelessness has continued to rise. This has led some to argue that Housing First has failed. The more accurate conclusion is that it has rarely been implemented with full fidelity to the model. This is not a reflection on the organizations and frontline workers delivering services, many of whom are doing exceptional work under significant pressure. The federal funding directives that flowed through the Reaching Home program were deliberately flexible, allowing community adaptation, but this also meant that strict adherence to Housing First principles was never required.
Fidelity to the model remains highly variable across the country. Even in Quebec, where the provincial government has invested beyond federal Reaching Home funding to expand access to Housing First, efforts are actively underway to improve adherence to the model. Persistent workforce challenges, including high staff turnover and insufficient training, compound the problem, making it harder for even well-resourced organizations to implement the model as intended. When the model has fallen short, it is often because it has not been fully resourced or implemented as intended.
Recently, the Mental Health Commission of Canada brought together researchers, clinicians, policymakers, and people with lived experience to chart a new path for Housing First in Canada. Their report called for renewed commitment to the Housing First approach, greater investment in the workforce and wraparound supports, and stronger leadership and public engagement. The MHCC report is also explicit that Housing First approaches must be adapted to be culturally grounded and aligned with Indigenous values, and that Indigenous-led efforts are critical to understanding and meeting the unique needs of Indigenous people and communities.
As encampments become more visible and politically contentious across the country, the temptation to respond with punitive measures rather than housing solutions is real and growing. Costantini’s research gives us the strongest evidence yet for why Housing First must remain a cornerstone of Canada’s response to homelessness, alongside the prevention strategies needed to keep people housed in the first place.