Andy Burnham’s first major policy announcement as Prime Minister was a commitment to ‘end rough sleeping’. No timeframe was given, and we are still awaiting policy details beyond the initial funding figure of £340m. Though the specifics are lacking, we can surmise from comments made during his recent leadership bid that this money could be dedicated to a ‘National Housing First’ strategy, and to expanding his ‘A Bed Every Night’ scheme, which offered housing to all rough sleepers in Manchester regardless of status or circumstance.
Ending rough sleeping is an admirable and ethical aspiration. It is also good politics, not least because the number of people it targets is less ambitious, in simple terms, than many might assume. Though the public may, in general, equate rough sleeping with homelessness, this is not how the homelessness sector conceives of the issue, nor is it the statute definition. Under the 1996 Housing Act, anyone in temporary accommodation, emergency placements, night shelters, hotels, hostels, sofa surfing or any other form of unstable, short-term accommodation is considered legally homeless. In December 2025, Shelter estimated that 382,000 people were homeless in England. By contrast, the Rough Sleeping Data Framework produced by the Ministry of Housing, Communities, and Local Government (MHCLG) estimates that 7541 people were sleeping rough in March 2026, with 3461 counted on a single night (the Ministry takes a snapshot approach to collecting these data because rough sleeping is variable: people will crash on friends’ sofas, use emergency accommodation, or occupy squats based on their specific circumstances or the weather).
However, the boundary between rough sleepers and those housed in dedicated accommodation is porous and defies public perceptions of homelessness. Watching The New Statesman’s coverage of Burnham’s inaugural speech, I was struck by the correspondent’s claim that rough sleeping is ‘the most visible form of homelessness’. This reflects a public misperception that people begging or engaged in ‘street culture’ are necessarily rough sleeping. As the manager of two complex needs hostels in an inner London borough, I can assure you that this is not always true. Rough sleepers are very visible, but the most vulnerable people in supported accommodation are often just as visible – and are often mistaken for the former.
If ‘ending rough sleeping’ means, in political terms, significantly reducing the visibility of vulnerable people on the street, a National Housing First Scheme, while crucially important, will not be enough. To end rough sleeping requires a broader strategy, with dedicated resources for substance misuse and mental health services – and significantly increased attention to the current landscape of homelessness and addiction support services.
Complex needs
Although long-term rough sleepers represent only a small fraction of the total homeless population, they often represent the most challenging cases – lack of available housing is only one of a suite of issues they face. Mental health problems, substance misuse, cognitive impairment, neurological diversity, and histories of complex trauma are all significantly overrepresented in the rough sleeping population. These individuals often present with multiple, co-occurring, and mutually reinforcing challenges – what is referred to within the homelessness sector as having ‘complex needs’. For example, a person may self-medicate the symptoms of their paranoid schizophrenia with heroin and crack cocaine. The physiological and psychological dependence these substances then create, as well as the life required to secure a consistent supply, exacerbate the symptoms of their psychiatric condition, which then requires further self-medication.
Not all the 33,000 supported- and single-homeless-specific bedspaces in England will go to the clients with the most complex needs, but many will: according to a 2024 report surveying 204 single homeless accommodation providers in England, 95% of providers support residents with a diagnosed mental health condition, and 100% support residents with drug and alcohol addiction. Substance misuse support is currently available in the community and often geared toward harm reduction. For example, opiate dependency is often treated by addiction services with a prescription for methadone, which alleviates the physiological symptoms of withdrawal without engendering a high. Often, the assumption is that patients will ‘use on top’ of their prescription. Proactive accommodations focus on getting residents scripted on methadone, as alleviating the physical pain of withdrawal is often a prerequisite for engaging in meaningful casework.
There is nothing inherently wrong with a harm reduction approach: people in supported accommodation are safer than those on the street; people on methadone suffer less than those who are not. After a period undergoing harm reduction treatment, many residents will be ready to move on to recovery options that aim to treat the issues themselves, rather than just their symptoms. The problem is that these options are, unfortunately, quite limited. Amenability to treatment is variable, and because new and traumatic crises continually occur in substance users’ lives, timely responses are essential. For many rough sleepers with substance abuse issues, rehab is the only effective route to addiction recovery: those who do not receive one of the few available rehab placements will often stay in the system for years – or return to the street. But even in well-resourced London boroughs, the wait time for a placement can be between six to nine months; supply simply cannot meet demand. This is not simply a function of a mercenary funding reality, but a national and local system design failure. For the taxpayer, residential rehab placements are initially more expensive than supported accommodation, though even then the cost differential narrows once residents’ Housing Benefit, Personal Independence Payment, and Universal Credit are taken into account. But in the longer view, and setting aside humanitarian considerations, successful recovery via rehab represents a significant cost reduction, and the scarcity of rehab placements for people who have experienced rough sleeping is an eye-wateringly expensive reflection of government short-termism.
Of course, not every client that receives a rehab placement will successfully realise and manage sobriety, and the co-occurring mental health or cognitive impairments that are common in the cohort makes success much less likely. If this is the case, why do homelessness professionals refer their clients into rehab services before seeking to address mental health concerns? Simply put, the current operating frameworks do not encourage a holistic view of clients’ conditions – mental health and substance abuse are frequently considered separate realms by their respective services. Many statutory mental health or learning disability services will either resist assessing clients who continue to misuse substances, or directly refuse engagement on that basis, despite guidance to the contrary – they feel that a client’s presentation can either be attributed to their substance misuse, or is not worth tackling until that misuse is brought under control. Yet, as should be obvious, substance misuse is often a direct consequence of rough sleepers’ mental health condition – they will likely never address their misuse without mental health support. We do have medium- to long-term supported accommodations for those whose primary care needs are mental health-related, but substance misuse is almost always an exclusion criterion.
Where a caseworker can secure a psychiatric diagnosis – and perhaps a Care Act assessment that defines the client’s primary support needs as mental health-related – the options for long-term housing remain limited. In many cases, even when the care team accepts that a client has co-occurring psychiatric and substance misuse needs – and that they will not recover in homelessness supported accommodation – there is still nowhere to place them. There exist residential care settings that cater to ‘dual diagnosis’ clients, but placements are extremely limited. Such situations frequently result in a return to rough sleeping.
Even where a client’s mental health reaches such a crisis point that psychiatrists and mental health professionals agree they require detention under the Mental Health Act, such interventions are often brief and ineffective. Anecdotally, it appears as though stigma about substance abuse and external pressure to free up beds are resulting in many missed opportunities for support. Clients will be discharged to hostels or the street after only brief intervention, despite their stay under section representing their longest period of sobriety in years. I have seen people in such circumstances become almost immediately non-compliant regarding prescribed medication, and rapidly return to chronic substance misuse.
What comes first?
None of which is to say that a Housing First strategy, scaled to the national level, could not make a difference. At its simplest, Housing First (HF) is the idea that the reality of a person’s homelessness should be addressed before its causes: that an effective rough-sleeping strategy will focus on housing even the most complex-needs individuals prior to addressing any mental health or substance misuse issues. Security of tenure (usually in the form of an assured tenancy) is key to the definition. Housing First is also built on a philosophy of community integration, and while a few programmes are situated in a single building, most HF options are self-contained flats separated from each other and dispersed into the community. Mental health, substance misuse, and other support is then co-ordinated by the HF programme’s caseworkers – with housing security a key priority.
Housing First’s proponents argue that it is the best-evidenced means of ending rough sleeping. Many studies stress its superior record on tenancy sustainment (the length of time residents remain in housing and off the streets), with an international average of about 80% retaining secure accommodation over 1-2 years. The evidence for improvement in other areas, such as psychiatric stability and wellbeing, substance misuse reductions, greater community engagement, and physical health is more mixed: some studies suggest no statistically significant gains in support areas beyond tenancy sustainment; others suggest modest improvements in some but not others. MHCLG’s analysis of England’s three HF Pilot Schemes (which included Burnham’s in Greater Manchester) suggested modest improvements in respondents’ senses of safety, isolation mitigation and self-reported mental wellbeing – but no statistically significant decrease in substance misuse.
One cause of this inconsistency is a lack of clarity, regulation and oversight regarding Housing First’s operating principles. What counts as ‘Housing First’? There is no international body that regulates which projects can claim the term. Given both the sector’s chronic underfunding and HF’s status as innovative policy – indeed, now raised to national strategy – providers and policy makers may be motivated to apply the HF label liberally, wherever they see its status as a useful means of securing funding and resources. This could lead to huge local discrepancies in provisions across the country, all under the same HF policy branding. Already, some HF provisions are timebound at two years (with expectations that residents will then move on), while others support residents as long as necessary. Some have access to dedicated psychiatric provision, others do not. The success of Housing First’s ability to address support needs beyond housing will be dependent on its proponents’ capacity to offer specialist services where needed, as well as meaningful integration and coordination with mental health, substance misuse and housing specialists.
Another approach to housing our most vulnerable
Housing First may be an effective means of reducing rough sleeping, but it is not a panacea and cannot work in isolation. Mental health and addiction services in this country require reform to better deal with dual diagnosis. There have been improvements in multi-agency co-operation in recent years, mostly led by homeless accommodation providers, but there needs to be more dedicated support for those with combined mental health and substance use needs. The success of the GMHF pilot was likely due to the integration of mental health and dual diagnosis specialists into its staffing model, as well as the long-term nature of its provision. And assuming the 79% tenancy sustainment rate of the GMHF is nationally replicable, where does that leave the remainder? These will likely be the most vulnerable clients, whose needs exceed the capacity of a HF service to safely manage in independent accommodation – even with very intensive support. We need more specialist dual diagnosis services to support these people, both as medium-term rehabilitative placements, and long-term residential settings for the most vulnerable.
Estimates for those currently placed in dedicated homeless accommodation vary. As of 2025, Shelter estimates the number at 16,000, though this is likely an underestimate. Many of these institutions do good work, and they sweat blood to keep highly vulnerable people safe with limited resources. The current offer of an additional 1200 HF homes likely does not come close to what will be needed to supply the population of individuals whose needs prevent them from being stably housed, and most likely we will be reliant on supported homelessness accommodations for many years to come.
Even if HF becomes the mainstream option for homeless accommodation, existing providers will still be vital as medium-term respites and triage centres. Indeed, many of the residents in these accommodations will not require HF in order to leave homelessness permanently – many will have needs manageable enough to be referred into the private rented sector, or light-touch floating support schemes like the GLA’s Clearing House. A massive programme of social housebuilding would also, obviously, help with both prevention and relief, especially if we are to avoid the next generation of homelessness.
Nevertheless, many who interact with the current provision for homeless accommodation do eventually return to the street. This fact alone should be argument enough for a more innovative approach to housing our most vulnerable. The ability to provide a roof over everyone’s head is good, it is right, and it is achievable – but the scale of the crisis will require a far more comprehensive approach, and a willingness to engage the problem on a systemic scale.
Michael Jones is a service manager of supported accommodations for homeless people with complex needs in London, and a PhD candidate in sociology at University College London, researching homelessness and adult literacy.