From Shelter to Home, Part One: The Housing First Evidence Base – Jennifer Bridge

This is the first of two linked papers on Housing First and its relevance to Jersey. This paper sets out what Housing First is, how it differs from traditional approaches to homelessness, and what the international evidence shows about its effectiveness. The companion paper, From Shelter to Home, Part Two: A Pathway for Jersey, considers what implementing Housing First would require in Jersey specifically: the housing, health, social care, legal and financial conditions that would need to be in place, and the risks and sequencing involved in getting there. 

Read together, the two papers make a single argument: Housing First deserves serious consideration in Jersey, but only if the Island is honest about the conditions required to make it work. 

Executive summary 

Housing First is an evidence-based approach to ending homelessness that inverts the traditional model. Conventional systems typically require people to demonstrate “housing readiness” through sobriety, engagement with services, or compliance with rules before they are offered permanent accommodation. Housing First removes that condition. It provides appropriate, stable, independent housing immediately, without preconditions, and then delivers flexible, person-centred support to the individual in their home or at an appropriate location for as long as they need it. 

Housing First programmes use integrated home-visiting teams: social workers, psychiatric nurses, housing officers, and substance-use specialists who visit tenants in their own homes to help manage daily challenges, connect them to mainstream services, and catch problems before they intensify. This is housing-centred social work, with support delivered where people live, respecting the dignity of their own space while making sure they are not isolated or forgotten. 

The model rests on a simple insight: a secure home is not the reward for resolving a crisis; it is the foundation on which recovery from a crisis becomes possible. 

First developed in New York in the early 1990s, Housing First has since been adopted internationally, particularly in Canada and across Europe. A review by Housing First Europe identifiedimplementation in 19 European countries, with Finland, Denmark, Norway and Sweden among the countries incorporating the model into homelessness responses. 

Housing First has now been tested in randomised controlled trials in multiple countries. The evidence is strong and consistent on some outcomes, and more mixed on others. People housed under Housing First stay housed at significantly higher rates than those managed through traditional staircase models, and their use of emergency services, hospitals, and the criminal justice system falls measurably. The evidence for improvements in health, substance use, and quality of life is more mixed, and the evidence for cost-effectiveness, while encouraging, is highly dependent on local context and implementation fidelity. Housing First is a set of principles, not a replacement for other services: it works alongside mental health provision, substance use treatment, and employment support, but it insists that none of those things can be made a condition of having somewhere safe to live. 

Introduction 

Considering Housing First for any new jurisdiction, including Jersey, means keeping three distinct propositions separate: 

1. Housing First is an evidence-based model. The evidence for this is strong, and it is the subject of this paper. 

2. Housing First could be adapted successfully to Jersey. This is plausible but needs careful analysis and is the subject of the companion paper. 

3. Jersey is currently ready to implement Housing First. This has not yet been demonstrated and is also addressed in the companion paper. 

This paper deals only with the first proposition. It sets out the model, the principles behind it, the trial evidence, and the limits of that evidence, including a cautionary lesson from a comparable field of practice about why evidence from one system does not automatically transfer to another. 

What Housing First is 

In practice, Housing First means that a person experiencing homelessness is offered a tenancy in ordinary, self-contained accommodation, not a hostel bed or a room in a shared facility, and, importantly, that offer is not contingent on sobriety or compliance with programmes. The tenancy is theirs from day one, with the same security of tenure as any other renter. A small team of dedicated workers then visits them at home, or, mindful of safeguarding considerations, at a suitable venue of their choosing, on their terms, for as long as they need support. That support is flexible: it might address mental health, substance use, debt, employment, or simply the practical challenges of sustaining a tenancy after years of instability. The home is not a reward for engaging with services. It is the foundation on which engaging with services becomes possible. 

Finland reduced long-term homelessness by an estimated 68% between 2008 and 2022 through a housing-first approach while addressing underlying issues. The model removes the requirement for many people to be “housing ready” before accessing permanent housing. People with complex needs often struggle to maintain regular attendance at meetings, abstinence, and self-management. Housing First treats stable housing as a human right and the foundation for recovery, treatment, employment, and social connection. 

National research conducted by Homeless Link found that Housing First services are established across every region of England, with a particularly strong presence in the South. The organisation’s 2020 survey reported that the South East contained 23% of all Housing First services in England, the highest proportion of any region, while London accounted for 20% of services. The research also identified substantial growth in the South West, where no services had responded to the 2017 survey but 13 services were operating by 2020. This evidence indicates that Housing First has become a significant component of homelessness provision across southern England, with delivery concentrated in London and the South East and expanding rapidly in the South West. [homeless.org.uk], [housingfirsteurope.eu] 

While published regional evaluations do not report reductions in homelessness comparable to Finland’s 68% reduction in long-term homelessness, evidence from England suggests that Housing First is effective at sustaining housing outcomes. National evaluations of major Housing First pilots found that 84% of participants sustained their tenancies after three years, while individual Housing First services commonly report housing retention rates of 80-90% after two years. Within London, the Single Homeless Project delivers Housing First services in Camden, Islington, Redbridge and East London, demonstrating how the model has been implemented locally to support people experiencing long-term homelessness and multiple disadvantage. 

Housing First is best understood as a set of principles and an evidence base, not a single fixed programme design. It has been implemented differently in New York, Helsinki, Lisbon, Toronto, and Manchester, reflecting different housing markets, welfare systems, and legal frameworks. This adaptability is a strength, but it is also the source of an important caution addressed later in this paper: an intervention proven in one system cannot simply be transplanted into another and be expected to produce the same results. 

The seven principles 

Housing First services in the UK are generally organised around seven core principles. These are set out below in their general form. How they interact with Jersey’s specific legal, geographic and institutional circumstances is considered in the companion paper. 

1. People have a right to a home. People are offered housing as quickly as possible, based on suitability and stability rather than conditions or “readiness”, and they keep their tenancy even if they disengage from support. 

2. Flexible support is provided for as long as it is needed. Support must be flexible, open-ended, and available for as long as each individual needs it. It can fluctuate, intensify, or reduce over time, but it cannot disappear. Without this commitment, the model cannot function as designed. 

3. Housing is not dependent on accessing support. The home itself is not dependent on using support, and the provision of support follows the person if they move or a tenancy fails. 

4. Individuals have choice and control. People choose, within reason, the type and location of their housing, whether to use other services, and how, when, and from whom support is provided. 

5. An active engagement approach is used. Staff take responsibility for persistent, proactive contact, adapt the service around the person, keep caseloads small, and continue support as long as needed, including if someone temporarily loses their home. This needs an ongoing funding commitment. 

6. The service is based on people’s strengths, goals, and ambitions. Support works from the belief that positive change is possible, helps people identify and build on their strengths, develop skills and confidence, and connect with community and work opportunities. 

7. A harm reduction approach is used. People are supported across all dimensions of their lives to reduce the risks associated with substance use and self-harm, adopt safer practices, and work towards improved physical and mental health and wellbeing. 

Finnish Housing First programmes illustrate what this looks like operationally: integrated home-visiting teams of social workers, psychiatric nurses, housing officers, and substance-use specialists visit tenants in their own homes to help manage daily challenges, connect them to mainstream services, and catch problems before they intensify. This is housing-centred social work, with support delivered where people live, respecting the dignity of their own space while making sure people are not isolated or forgotten. 

It is worth being precise about what principle five requires, because this point is frequently misunderstood. The evidence does not simply show that a smaller caseload produces better results for an otherwise unchanged social work role. It shows that Housing First is a fundamentally different, multidisciplinary and relationship-based model of care, not a housing programme with a support worker added on. The justification for small caseloads is not that the job title changes; it is that the intervention itself is of a different kind. This distinction matters for how the model is presented to decision-makers and is developed further in the companion paper. 

The evidence 

Housing stability 

The evidence for housing retention under Housing First is the most consistent and robust finding in the literature. A 2019 systematic review and meta-analysis of randomised controlled trials, published in the Journal of Epidemiology and Community Health (Baxter et al., 2019)¹, found that people receiving Housing First support were more than twice as likely to achieve stable housing after 18 to 24 months compared with those receiving usual homelessness services. Specifically, participants in Housing First programmes were around 2.5 times more likely to remain stably housed (risk ratio = 2.46, 95% confidence interval 1.58-3.84), meaning that the positive effect was both substantial and statistically reliable. 

The same review found that people receiving Housing First used acute healthcare services less often than those receiving standard homelessness services. Participants in Housing First programmes had 37% fewer visits to emergency departments (incidence rate ratio = 0.63, 95% confidence interval 0.48-0.82) and 24% fewer hospital admissions (incidence rate ratio = 0.76, 95% confidence interval 0.70-0.83). These findings suggest that providing stable housing alongside ongoing support may reduce the need for crisis healthcare interventions. 

A 2022 evidence review² by the Australian Housing and Urban Research Institute (AHURI) reached a consistent conclusion from a wider body of international literature: housing retention rates under Housing First typically range from 66% to 90%, substantially higher than under “treatment as usual” approaches. Finland reduced long-term homelessness by 68% between 2008 and 2022 through national adoption of the model. 

Health, substance use and quality of life 

The evidence here is mixed. The same 2019 review found more limited and inconsistent evidence of improvement in mental health, substance use, and quality-of-life outcomes. This does not mean Housing First fails on these measures; it means the trial evidence has not yet established consistent effects in the way it has for housing stability. This is consistent with the model’s own logic: Housing First treats stable housing as the precondition for engaging with treatment, not a guarantee that treatment outcomes will improve on a fixed timetable. Programmes that pair housing with genuinely integrated clinical and social support tend to show stronger secondary outcomes than those that do not, which points to implementation fidelity, not the model itself, as the key variable. 

Cost-effectiveness 

The UK Government’s own cost-benefit analysis³ of the Housing First pilots in Greater Manchester, Liverpool City Region, and the West Midlands, published in 2024, found that for every £1.00 spent, just over £2.00 was saved in reduced use of emergency services, courts, and custody. The AHURI review reaches a consistent but more cautious conclusion: Housing First is most cost-effective for people experiencing chronic homelessness with complex and high needs, and cost-effectiveness is highly dependent on local implementation and context. Some of this return takes the form of cash-releasing savings; a significant proportion takes the form of cost avoidance and reduced demand on services that remain funded regardless. This distinction is addressed in full in the companion paper, because it matters to how the case could be made to unlock government funding. 

Implementation fidelity and housing supply 

Both the Baxter et al.¹ review and the AHURI² review identify the same two implementation risks. First, fidelity matters enormously: programmes that dilute the core principles, for example by reintroducing conditionality, raising caseloads, or under-resourcing the multidisciplinary team, produce weaker outcomes than high-fidelity implementations. Second, housing supply is often the binding constraint. Housing First cannot succeed if there is no housing to offer; a jurisdiction’s ability to deliver depends as much on its housing pipeline as on its clinical and social care capacity. 

A cautionary lesson from a related field 

There is a valuable parallel with Assertive Community Treatment (ACT), a service model developed for people with severe mental illness that, like Housing First, is intensive, team-based, and delivered assertively in the community rather than through office-based appointments. The ACT literature offers an important lesson: even well-evidenced interventions do not transplant automatically from one system to another. Outcomes depend on the existing service landscape, the fidelity of implementation, and local context. Trials that produced strong results in one health system have produced far weaker results when the same model was introduced elsewhere without equivalent resourcing or fidelity to the original design. 

This is an argument against assuming that results from Finland, Canada, or English pilot areas will automatically translate to other contexts, including Jersey, without deliberate, evaluated local adaptation. It is the strongest conceptual justification for testing any new implementation through a carefully evaluated, iterative approach rather than assuming success will simply follow from adopting the label. The companion paper sets out what that means in practice for Jersey. 

Responding to the critics 

Housing First has critics, and their arguments deserve consideration. 

Housing First stabilises housing but does not deliver recovery. This critique holds it to a standard rarely applied to other interventions. No single programme is expected to resolve addiction, mental illness, unemployment, and social exclusion simultaneously. The evidence consistently shows that stable housing is a precondition for engaging in treatment and support, not a substitute for it. If someone is housed, safe, and chooses to engage with services at their own pace, that is not failure; it is the foundation on which recovery can be built. 

Implementing Housing First will crowd out other approaches. This concern is largely unfounded. Treatment-led models, transitional housing, and prevention are not in competition with Housing First; a well-designed system incorporates all these approaches, targeted at people for whom they are most appropriate. 

The evidence base is selectively cited and overgeneralised. Many policy interventions face this challenge, and Housing First is one of the more comprehensively evaluated homelessness interventions, including randomised controlled trials such as At Home/Chez Soi⁴ in Canada and the Housing First Europe Hub studies⁵. The observation that evidence is thinner for families and young people is fair, and the appropriate response is to expand and adapt the evidence base rather than reject the approach where it is well-evidenced. The “cherry-picking” critique also cuts both ways: opponents routinely cite weak outcomes from low-fidelity implementations as evidence against the high-fidelity model. 

Implementation in tight housing markets with inadequate funding undermines fidelity. This is the most substantive criticism and deserves serious engagement considering the Jersey context. For Housing First to be effective, it must be accompanied by supply-side investment: building appropriate social housing, ensuring rents are affordable, and ensuring benefit levels cover actual housing costs. What happened in parts of California, where the label was adopted without the structural conditions and clinical partnerships that go with it, illustrates what happens when a jurisdiction adopts the name without the substance. On cost, the evidence from multiple jurisdictions shows that high-fidelity Housing First reduces use of emergency services, hospitals, police, and the criminal justice system: costs that are real but often invisible in conventional accounting. 

Some people prefer shelters. While some individuals may prefer shelters to individual housing, this critique overlooks evidence that most people, when genuinely offered an appropriately supported choice, prefer a secure tenancy. 

Fidelity versus failure. Much of the debate centres on underfunded versions of Housing First rather than the model as designed. Those who use weak results to advocate for a return to conditional, treatment-first models for all are promoting approaches that decades of evidence show produce worse housing outcomes for the most excluded individuals. 

What the evidence does and does not show 

Bringing this together: the evidence for housing retention is consistently strong; the evidence for health, substance use, and quality-of-life outcomes is more mixed; the evidence for cost-effectiveness is encouraging but highly context-dependent; implementation fidelity matters enormously; and housing supply is often the limiting factor. This is the foundation on which this paper’s companion piece builds in assessing Jersey’s own circumstances. 

Conclusion 

The first of the three propositions set out at the start of this paper, that Housing First is an evidence-based model, is well supported by trial evidence, systematic reviews, and two decades of national experience in Finland and elsewhere. The evidence is strongest for housing stability, more mixed for health and social outcomes, and encouraging but context-dependent for cost-effectiveness. The ACT literature’s lesson that effective interventions cannot simply be transplanted between systems reinforces the case for careful, evaluated, locally adapted implementation rather than an assumption that what worked elsewhere will automatically work in a new setting. 

Whether Housing First could be adapted successfully to Jersey, and whether Jersey is currently ready to implement it, are separate questions. They are addressed in the companion paper, From Shelter to Home, Part Two: A Pathway for Jersey. 

Appendix: summary of key evidence 

Source	Design	Key finding
Baxter et al. (2019), J Epidemiol Community Health	Systematic review and meta-analysis of RCTs	Housed at 18–24 months: risk ratio 2.46 (95% CI 1.58–3.84). Emergency department visits: IRR 0.63 (95% CI 0.48–0.82). Hospitalisations: IRR 0.76 (95% CI 0.70–0.83). Mental health, substance use and quality-of-life outcomes: more mixed.
AHURI (2022) evidence review	Synthesis of Australian and international literature	Housing retention typically 66–90%, substantially above treatment-as-usual. Cost-effectiveness strongest for high and complex needs; highly context dependent. Fidelity and housing supply identified as principal implementation risks.
At Home/Chez Soi (Canada)	Multi-site RCT	Strong housing stability outcomes; informed Housing First Europe Hub and international replication.
Housing Finance and Development Centre of Finland (ARA, 2023)	National programme evaluation	Long-term homelessness reduced by 68% (2008–2022) through national Housing First adoption alongside sustained investment in social housing supply.
ICF / UK Ministry of Housing, Communities and Local Government (2024)	Cost-benefit analysis, three English pilot regions	Approximately £2.00 in reduced public service use for every £1.00 spent; benefits a mix of cash-releasing savings and cost avoidance.
Social Market Foundation (Salutin, 2025)	Policy analysis	Consistent with UK pilot cost-benefit findings; identifies fidelity and housing supply as principal constraints on scale-up.

Sources 

1. Baxter, A.J., Tweed, E.J., Katikireddi, S.V. and Thomson, H. (2019). Effects of Housing First approaches on health and well-being of adults who are homeless or at risk of homelessness: systematic review and meta-analysis of randomised controlled trials. Journal of Epidemiology and Community Health, 73(5), 379–387. 

2. Roggenbuck, C. (2022). Housing First: An Evidence Review of Implementation, Effectiveness and Outcomes. Australian Housing and Urban Research Institute (AHURI), Melbourne. 

3. ICF Consulting / Ministry of Housing, Communities and Local Government (2024). Housing First Pilots: Cost Benefit Analysis. UK Government. Published October 2024. 

4. Mental Health Commission of Canada. National Final Report: At Home/Chez Soi Project. 

5. Housing First Europe Hub (2025). Housing First Evidence and Country Data. 

6. Housing Finance and Development Centre of Finland (ARA) (2023). Homelessness in Finland 2022. ARA, Helsinki. 

7. Y-Foundation (2017). A Home of Your Own: Housing First and Ending Homelessness in Finland. Y-Foundation, Helsinki. 

8. Homeless Link, London. The Principles of Housing First. 

9. Salutin, G. (2025). Housing First: How to Radically Reduce Rough Sleeping This Parliament. Social Market Foundation, London. Published September 2025. 

10. Crisis (2022). For Everyone: The Final Report of the UK Collaborative Centre for Housing Evidence Review of Crisis’s New Plan to End Homelessness. Crisis, London. 

Biographical note 

Jennifer Bridge MBE is a member of the Policy Centre Jersey’s Management Committee and Lead Researcher. She is an accomplished Chair with extensive experience across the creative arts, third sector, and political spheres. Jennifer contributes regularly to public discourse through her columns and articles in local media, focusing on inclusion and open government. During the COVID-19 pandemic, she completed postgraduate studies in research methodologies, which has enhanced her evidence-based approach to public commentary. 

Jennifer is a co-leader of Jersey repair café – a community group where volunteers fix broken items like clothes, electronics, and bikes for free, promoting sustainability and reducing waste.