In a Canadian trial that followed more than 2,000 homeless people living with mental illness for two years, the most effective intervention was not a new drug or therapy. It was a key.
The At Home/Chez Soi project, run across five cities from Vancouver to Moncton, tested Housing First: give people a permanent home quickly, without requiring them to first achieve sobriety or treatment milestones, and then offer support. By the final six months, reports of the national results put the share of Housing First participants in stable housing at around 60 percent, roughly double the 31 percent among those receiving usual services.
The phrase “housing is healthcare” has become a slogan. The argument of this piece is that it should become a budget line. The evidence is strong enough that health systems should be paying for housing interventions, not just endorsing them.
The biology of a bad home
Housing affects health through channels that are neither vague nor controversial.
Cold homes are one. The WHO Housing and health guidelines, published in 2018, make a strong recommendation that indoor temperatures be kept at a minimum of 18 degrees Celsius in cold seasons, with higher minimums possibly needed for vulnerable groups such as older people and young children. The guidelines cite an estimate of around 38,200 excess winter deaths a year attributable to cold housing across 11 European countries.
Damp, mould, crowding, injury hazards and insecure tenancy all have their own evidence. Instability matters too. A person who does not know where they will sleep next week cannot keep a medication schedule, attend a follow-up appointment or recover properly after surgery.
Homelessness is the extreme end of this spectrum. But it is the same spectrum.
What Housing First showed about cost
The usual objection to housing-led interventions is cost. Housing is expensive, and health budgets are already stretched.
The At Home/Chez Soi cost analysis, published in Psychiatric Services in 2020 for participants with high needs, found that about 69 percent of the intervention’s cost was offset by savings elsewhere, in areas such as emergency shelters. The net cost came to roughly 6,300 Canadian dollars per person per year, and the incremental cost per additional day of stable housing was around 42 Canadian dollars.
That is not free. But it means most of the money was already being spent, just on worse outcomes: shelter beds, emergency visits, police contact, hospital stays. The question is not whether to spend. It is whether to keep spending on the consequences of homelessness or on ending it.
Finland made it policy
Finland adopted Housing First as national policy in 2008. Its national homelessness survey, run by the Housing Finance and Development Centre (ARA), showed long-term decline from around 16,000 homeless people when the surveys began in 1986 to about 3,800 at the end of 2024.
The Finnish story has an honest footnote. The 2024 figure was an increase of 377 on the previous year, ending a run of more than a decade of annual declines. That matters, because it shows that progress depends on continued investment in housing supply and support. Housing First is a policy that must be funded every year, not a problem solved once.
Why health systems don’t pay
If the case is this strong, why do health budgets so rarely fund housing?
The answer is the classic wrong-pocket problem. Housing is paid for by housing ministries or municipalities. The savings from better housing land in health, justice and social services. The agency that would have to spend the money is not the one that gets the benefit.
Health systems also have a cultural reluctance. Clinicians and managers are trained to think of health spending as clinical services. Paying rent feels like mission creep. But health systems already pay for things that are not strictly clinical when the evidence supports it: transport to dialysis, meals for inpatients, social workers. Stable housing for people with high health needs belongs in the same category.
There is a risk on the other side, and it deserves to be named. If health systems start funding housing, governments may use that as an excuse to cut housing budgets, shifting costs onto health rather than adding resources. Any health investment in housing needs to be additional, not a substitute.
What health systems can do now
Even without major reform, there is a lot within reach.
- Screen for housing insecurity in primary care and hospital discharge planning, and act on the answer.
- Stop discharging to the street. Discharge protocols that allow patients to leave hospital into homelessness waste the care just delivered.
- Fund supportive housing places for high-need patients, jointly with housing agencies, and track health outcomes.
- Use health data in housing decisions, for example prioritising people whose conditions are worsened by damp or cold homes.
- Count the savings. Build shared budgets or joint accounting so the health benefits of housing investments show up where decisions are made.
The revolving door, costed in outline
To see why the wrong-pocket problem matters, follow an illustrative patient through a year without stable housing.
A man in his fifties with heart failure and depression is living in shelters and sometimes outdoors. He has no fridge for medicines, no reliable place to rest after a hospital stay, and no fixed address for appointment letters. Over a year, he might be admitted several times for fluid overload, each admission lasting days. Between admissions, he visits the emergency department when symptoms worsen, spends nights in shelters, and has occasional contact with police or ambulance services. His depression makes it harder to follow any plan.
Each of these services pays for its part. The hospital pays for admissions. The municipality pays for shelter beds. Emergency services pay for call-outs. None of them pays for the thing that would reduce all of those costs: a stable home with support.
Now imagine the same man in supported housing. He has a fridge, a bed, a mailbox and a support worker who helps him keep appointments. His heart failure is still serious, and he may still be admitted. But many admissions can be avoided when medicines are taken reliably and early warning signs are noticed. The cost of his housing and support falls on one budget. The savings are spread across several others.
That is why housing investments often look expensive to the agency paying and cheap to the system as a whole.
The strongest counterargument
There is a serious case against health systems taking on housing, and it deserves engagement.
Health budgets are already under strain, and every dollar spent on housing is a dollar not spent on clinical care. Health systems have no expertise in managing property or tenancies. Housing shortages are structural, driven by land, planning, construction and interest rates, and health spending cannot fix them. Some argue that Housing First works best for a specific group, people with high needs and serious mental illness, and that evidence for broader populations is weaker. Others point out that housing outcomes improve more reliably under Housing First than some health outcomes do.
Parts of this hold. Health systems should not become landlords, and they cannot solve housing supply alone. The strongest evidence for Housing First is indeed for people with high needs, and claims should not be stretched beyond that. Housing stability is the most consistent outcome, and health improvements can be slower and harder to measure.
But the argument does not support the conclusion that health systems should stay out of housing. It supports a narrower role: health systems partnering with housing agencies, funding support services and contributing to supported housing places for patients whose health is most affected by homelessness. That is targeted, evidence-based and within health’s competence. It does not require running housing markets.
Second-order effects
Hospital flow improves. Patients who cannot be discharged because they have nowhere safe to go occupy beds that others need. Housing options for these patients improve flow for everyone.
Clinical effort is not wasted. Treatment delivered to someone who returns to the street is often undone within weeks. Stable housing protects the value of the care already given.
Mental health and substance use care become possible. Many treatments require stability, routine and a safe place to store medicines. Housing makes those treatments more feasible.
Children’s health benefits. Families in insecure or poor-quality housing face higher risks of respiratory illness, injury and disrupted schooling. Housing policy is child health policy.
Climate adaptation matters. As heatwaves become more frequent, poorly insulated and overcrowded housing becomes a health risk in summer as well as winter. The WHO guidelines address high indoor temperatures too.
What this means for different decision-makers
For health ministries, the priority is joint planning and joint budgets with housing and social services, with shared targets for groups such as people with serious mental illness who are homeless.
For hospital leaders, the practical steps include systematic screening for housing insecurity at admission, early discharge planning with housing partners, and tracking how many patients are discharged to no fixed address.
For primary care leaders, housing conditions such as damp, cold and overcrowding can be recorded and linked to referral routes for repairs, energy support or rehousing.
For funders and investors, supported housing for high-need groups is a candidate for blended funding that recognises savings across several agencies.
For clinicians, the habit worth building is asking where a patient will sleep tonight, and treating the answer as clinically relevant information.
What to watch next
Watch whether Finland’s 2024 rise in homelessness proves to be a blip or a turning point, and what it reveals about the conditions needed to sustain Housing First. Watch for health systems that experiment with funding supported housing directly and publish their results. And watch whether cold and heat in homes become a routine part of public health planning, as the WHO guidelines recommend.
Why “housing readiness” was the wrong idea
It is worth remembering what Housing First replaced. For decades, many homelessness services followed a staircase model: people had to prove they were ready for housing by completing treatment, staying sober or moving through temporary accommodation first. Each step was a test, and failing any step sent people back down.
The logic seemed prudent. In practice, it asked people to achieve stability without the one thing that makes stability possible. Someone with a serious mental illness and no home is being asked to manage their health under the worst possible conditions, and is then judged not ready because they could not.
Housing First inverted that sequence. The Canadian trial did not show that housing solves everything. It showed that housing does not have to be earned first, and that people with high needs can keep a tenancy when given one with support. That is a lesson health systems should recognise, because it mirrors a familiar clinical principle: stabilise the patient before expecting them to engage with treatment.
The same logic applies to discharge. A hospital that sends a patient with heart failure back to a shelter is, in effect, asking them to manage a complex condition in conditions where almost nobody could. Treating housing as part of the care plan is not generosity. It is clinical realism.
The takeaway
The evidence from Canada’s randomised trial and Finland’s national policy points the same way. Stable housing improves outcomes for people with high needs, and much of its cost is recovered from services that would otherwise absorb the consequences.
Saying “housing is healthcare” is easy. Putting housing in the health budget is the test of whether we mean it.
Does your health system fund housing directly, or does it stop at the hospital door? We’d like to hear how it works where you are.
