Housing first for severely mentally ill homeless methadone patients
Effectiveness bank home page. Opens new windowResearch analysis

This entry is our analysis of a study considered particularly relevant to improving outcomes from drug or alcohol interventions in the UK. The original study was not published by Findings; click Title to order a copy. Free reprints may be available from the authors – click prepared e-mail. The summary conveys the findings and views expressed in the study. Below is a commentary from Drug and Alcohol Findings.

Links to other documents. Hover over for notes. Click to highlight passage referred to. Unfold extra text Unfold supplementary text
Copy title and link | Comment/query |

Housing first for severely mentally ill homeless methadone patients.

Appel P.W., Tsemberis S., Joseph H. et al.
Journal of Addictive Diseases: 2012, 31, p. 270–277.
Unable to obtain a copy by clicking title? Try asking the author for a reprint by adapting this prepared e-mail or by writing to Dr Appel at philipappel@oasas.ny.gov.

Homelessness is a significant obstacle to regular participation in methadone maintenance treatment, particularly among people leaving prison. This study in a major US city examines whether a ‘housing first’ programme could improve outcomes among this cohort.

Summary Homelessness is a significant obstacle to regular participation in methadone maintenance treatment, and this has been especially true for people released from short sentences in New York City jails. The Key Extended Entry Program (KEEP) in New York enabled people dependent on heroin to begin or continue with methadone treatment while in jail, and gave them medication on release from jail and an appointment to report to a community-based methadone maintenance treatment programme. Records revealed that a large majority (78%) kept their referral appointment; of the 22% who did not report, more than three-quarters were homeless and had co-occurring mental health problems. This equated to hundreds of people annually who were enrolled onto methadone in jail, but probably ceased this treatment when they left jail.

Key points icon

Key points
From summary and commentary

The Keeping Home project adopted a ‘housing first’ approach to addressing homelessness, mental health and substance use problems – providing access to housing and holistic community and psychosocial support services to people who were in receipt of methadone treatment, but without their housing and support being contingent on their engagement with treatment.

Compared to a similar cohort of people without access to housing first, Keeping Home participants had superior housing and methadone maintenance treatment outcomes two and three years after the first Keeping Home patient was housed.

The featured results are welcome for a set of people with complex problems who have found it challenging to exit homelessness and remain in treatment, and are generally consistent with evidence about the value of a housing first approach.

The Keeping Home project was developed to address the residential and support needs of people with both mental health and substance use problems, who were either:
• homeless and patients of KEEP, where homelessness was defined as living in a shelter or other indoor facility or on the streets/other public places;
• or methadone patients from outside KEEP with recent criminal justice involvement.

It adopted a ‘housing first’ approach to housing people with complex support needs and/or challenging behaviour – integrating housing and harm reduction support and delivering these regardless of current substance use.

The US Federal Department of Housing and Urban Development issued a grant in 2003 to Keeping Home, which covered the cost of 25 market-rate apartments plus assertive community treatment services including psychiatric, nursing, vocational, social, and peer support.

The featured study compared the outcomes of participants enrolled in Keeping Home with a cohort of people similarly struggling with mental health and substance problems and in receipt of methadone maintenance treatment between March 2005 and July 2006.

Housing first group: 31 people were recruited to the Keeping Home project from hospitals, drop-in centres, and jails near the time of their release; for this group only, participants had to have a diagnosis of a serious and persistent mental illness such as depression, schizophrenia, or bipolar disorder.

Comparison group: 30 people were selected from the Office of Alcoholism and Substance Abuse Services client database to form the comparison group; they were all enrolled in methadone treatment between 2005 and 2006, were homeless when they entered treatment, had a co-occurring mental health problem or had been treated for a mental illness, and had been involved with the criminal justice system (eg, had recently been incarcerated, or were on parole or probation); the mental health criteria were wider than for Keeping Home participants.

There were statistically significant differences between participants in the Keeping Home and comparison groups, which meant that the study could not offer a like-for-like comparison. Most notably, Keeping Home participants were older (average age of 46 vs. 40 years old) and more likely to enter the project from living on the streets or other public places (68% vs. 30%) as opposed to a homeless shelter.

The two main outcomes were retention in methadone maintenance treatment and retention in housing, for which the authors anticipated that Keeping Home patients would fare better. Changes in the treatment and housing status of Keeping Home participants were determined through interviews with assertive community treatment team staff in March 2007 (the first assessment date, two years after the first Keeping Home patient was housed) and June 2008 (the second assessment date, three years after the first Keeping Home patient was housed). The comparison group was monitored via the Office of Alcoholism and Substance Abuse Services client database.

Main findings

Methadone maintenance

March 2007. By the first follow-up point, two thirds (67%) of the comparison group had been discharged from methadone maintenance treatment, while the reverse happened with Keeping Home participants – 65% were still in treatment compared with 33% of their peers in the other group. The advantage for Keeping Home was statistically significant.

Of the 11 Keeping Home participants confirmed/judged to have discontinued methadone maintenance treatment, five reported that they had discontinued methadone maintenance treatment but stayed in their apartments. Six participants were discharged from the Keeping Home programme and presumed to have discontinued methadone maintenance treatment; two were transferred to inpatient settings where there were no methadone maintenance services, and two sadly died.

June 2008. A further two Keeping Home participants stopped methadone maintenance treatment between March 2007 and June 2008; one died and the other was transferred to a non-methadone, residential treatment programme. The remaining two transferred to other apartment/housing programmes, and as a conservative assumption the researchers recorded that they also stopped methadone maintenance treatment. As of June 2008, 16 of 31 Keeping Home participants were still in methadone maintenance treatment.

An additional four people in the comparison group left treatment, so only six of the original 30 were in treatment. The Keeping Home patient retention rate was more than double that of the comparison group (52% vs. 20%), and the difference statistically significant.


March 2007. Most Keeping Home participants (81%) were still in their apartments as of March 2007, compared with 37% of people in the comparison group who were housed and 17% who were living in private residences. The advantage for Keeping Home was statistically significant.

June 2008. Retention in housing among Keeping Home participants decreased between March 2007 and June 2008 to 68%, with four additional discharges. Only six people in the comparison group were retained in methadone maintenance treatment through June 2008, and of the three with a known residential status, just one was living in a private residence. Even if the three other participants with unknown housing status were housed, the difference of 68% Keeping Home versus 13% (four out of 30) comparison participants remained highly significant.

The authors’ conclusions

Compared to peers with similar problems in the domains of mental health, substance use, and criminal justice, Keeping Home participants had superior housing and methadone maintenance treatment outcomes two and three years after the first participant had been enrolled. At the final follow-up point, 52% of Keeping Home participants were still in methadone treatment compared with 20% of people in the comparison group, and 68% of those in the Keeping Home programme had retained their housing compared with 13% in the comparison group.

Differences in the groups other than their access or lack of access to the Keeping Home programme could have accounted for or contributed to these differences in outcomes. The policy of the housing first project was to recruit people experiencing chronic homelessness (ie, people who have complex and overlapping support needs, who find it difficult to exit homelessness) – by design selecting people who were particularly vulnerable. This was evident by the nature of the homelessness of Keeping Home participants (they were more likely to have been sleeping on the streets or public places) and the greater severity of their psychiatric diagnoses.

The assertive community treatment component of the intervention was probably essential in helping participants remain in their apartments and maintain their treatment regimens. Further research would be needed to determine the contributions of assertive community treatment services, accommodation within housing first programmes, and methadone maintenance treatment.

Overall it seems probable that the combination of interventions bound up in the featured housing first programme had a stabilising effect, supporting vulnerable people to remain in their housing and in receipt of methadone maintenance treatment. However, for some this may have arrived too late. Sadly, three people in the Keeping Home project died during the study period; the causes of death were not reported. Previous research indicates that homeless people are three to four times more likely to die prematurely than the general population, and those who live on the streets are at especially great risk for life-threatening diseases, trauma, and violence, and less likely to get the medical treatment they need.

Findings logo commentary The featured paper is one of very few studies that have focused on people with substance use problems (and their substance use outcomes) in housing first programmes. Indeed, the bulk of research on housing first to date has been concerned with people who have mental health problems, and its application in a North American context.

The findings were promising, showing superior methadone and housing outcomes among people in housing first compared to peers without access to housing first. These differences between the two groups were very substantial and statistically significant at both follow-up points, indicating that the differences were unlikely to be due to chance. However, as the researchers acknowledged, the design of the study meant that they could not directly attribute these differences to the intervention.

The ‘gold standard’ for determining whether an intervention actually causes the desired changes is to use a randomised controlled trial format, whereby researchers randomly allocate participants to two or more groups – an intervention versus an alternative intervention, a ‘dummy’ intervention, or no intervention at all. This type of design is not common in the housing first evidence base. What studies tend to do instead is compare the outcomes of people entering housing first under real-world circumstances versus people participating in treatment as usual in the local area, meaning that the characteristics and motivations of the two different groups may be different to begin with, and changes/improvements in their housing, health, and social outcomes may not come (solely) from the support they received. The present study recruited a sub-population of people eligible for housing first – people who met specific criteria in the domains of substance use, mental health, and criminal justice – and compared their outcomes to people also burdened by mental health and substance use problems, but not necessarily to the same degree, and due to this perhaps not eligible for such an intervention. The main consequence of not using a randomised controlled trial format was that the results could not speak to the absolute effectiveness of housing first. However, even if randomisation was not possible or desirable in this case, confidence in the study findings could have been improved if the researchers had compared a group of people eligible for the intervention who were exposed to the intervention, versus other people also eligible for the intervention but who were not exposed to it.

Another limitation of the study was the use of fixed follow-up points for all participants in May 2007 and June 2008 (two and three years after the first Keeping Home patient was housed), despite enrolment in the intervention being staggered. Due to this the study does not provide a picture of longer-term housing and treatment outcomes two and three years after being in the programme, only two and three years after the first Keeping Home patient entered the programme. To illustrate this point, there were only 25 market-rate apartments in the housing first programme but 31 participants, meaning that the last participants would not have had the opportunity to be two or three years into the project when they were followed up.

During the study period, three participants in the Keeping Home project group sadly died, whereas there were no reported deaths among people in the comparison group. All who died had stopped methadone maintenance treatment, but it was not clear under what circumstances. There was also no information about the causes of death; they may have been drug-related deaths, they may not. Internationally the evidence is strong that being in treatment – and especially for opioid users, being in a substitute prescribing programme such as methadone maintenance therapy – helps prevent overdose deaths. Leaving methadone treatment may have played a role in the deaths in the featured study, but if so, without further information it is unclear why there were three deaths in the housing first group when a greater proportion of people in the comparison group had left methadone treatment (six versus 16 still in treatment at the final follow-up point), leaving a greater number of the comparison group at risk. As noted by the authors, the housing first group may by design have been more vulnerable to begin with.

The seven principles of housing first

  1. People have a right to a home.
  2. Flexible support is provided for as long as it is needed.
  3. Housing and support are separated.
  4. Individuals have choice and control.
  5. Staff are responsible for proactively engaging their clients; making the service fit the individual instead of trying to make the individual fit the service.
  6. The service is based on people’s strengths, goals and aspirations.
  7. Staff support people who drink and take drugs to reduce immediate and ongoing harm to their health.

An Effectiveness Bank hot topic on the overlapping experiences of homelessness and substance use problems includes a section on housing first. As this details, the housing first model originated in the United States, based on the view that a home of your own (and, within normal constraints, of your own choice) is a right regardless of whether you have accepted help with (still less succeeded in) resolving substance use and other problems. Through the Keeping Home housing first project, participants were offered assertive community treatment, a highly intensive, holistic, and integrated form of mental health support in the community. This may be desirable in, but not necessarily typical of, housing first projects. Another factor that arguably set the featured study’s iteration of housing first apart was that it recruited only people prescribed methadone maintenance therapy, and had the implicit aim to retain people in this treatment. As such, it is possible that these participants had different levels of motivation, support, and stability than the general run of people who would be suitable for (or the target group for) housing first.

Altogether, there were at least three different components to the programme under study that individually and/or collectively could have had an impact on the trajectories of participants: (1) assertive community treatment; (2) housing first accommodation; and (3) methadone maintenance. While this may make it all the more difficult to discern the effectiveness of a housing first programme, the positive outcomes achieved may serve as an example of the lengths that housing first interventions need to go to give people with overlapping, chronic and complex needs a ‘fighting chance’ of staying engaged with treatment and remaining in housing.

Overall, housing first has been shown to improve stability of housing, even when substance use stays relatively unaffected, and is more effective at maintaining engagement with homeless people than abstinence-based services. In the UK specifically, there is evidence that it works well in reducing long-term homelessness, and can successfully engage with people who have very long-term histories of contact with other forms of homelessness service (without their homelessness having ever been resolved).

That housing first should replace other homelessness services has not yet been borne out by the evidence. However, governments across the UK have accepted it as part of the solution to homelessness. For the UK Government, housing first is part of the plan to end rough sleeping by 2024, and is being supported initially with £28m of funding for housing first pilots in Greater Manchester, the West Midlands and Liverpool; the Welsh Government is funding 10 housing first projects and endorses housing first as a solution to long-term chronic rough sleeping; and the Scottish Government is piloting housing first in five of its biggest cities (Aberdeen, Dundee, Edinburgh, Glasgow, and Stirling) through the Housing First Pathfinders programme, which is “working to make Housing First a reality on a much bigger scale than what we’ve so far known”.

The strong evidence that housing first can end homelessness among people with multiple and complex needs applies to (but is not limited to) the population of people who are homeless with substance use problems. However, there has been a lack of clarity in the discourse about what the ultimate goal of housing first is for this client group; specifically, what yardstick we should use to judge the success of housing first.

Is the aim first and foremost to ensure that vulnerable people become and stay housed?
Or should the goal be for people to show movement on their substance use outcomes – moderating their use, showing less dependence, becoming abstinent, using in a less risky way etc?

Given these questions, findings from the featured paper and other studies – limited though they may be – will be pertinent to conversations in the UK about the value of extending the availability of accommodation and support under the banner of housing first for people with substance use problems.

Last revised 22 May 2020. First uploaded 24 April 2020

Open Effectiveness Bank home page

Top 10 most closely related documents on this site. For more try a subject or free text search

STUDY 2005 Addressing medical and welfare needs improves treatment retention and outcomes

STUDY 2009 The Drug Treatment Outcomes Research Study (DTORS): final outcomes report

REVIEW 2017 An evidence review of the outcomes that can be expected of drug misuse treatment in England

STUDY 2011 A good quality of life under the influence of methadone: a qualitative study among opiate-dependent individuals

REVIEW 2015 Heroin on trial: systematic review and meta-analysis of randomised trials of diamorphine-prescribing as treatment for refractory heroin addiction

STUDY 2010 The Andalusian trial on heroin-assisted treatment: a 2 year follow-up

REVIEW 2012 New heroin-assisted treatment: Recent evidence and current practices of supervised injectable heroin treatment in Europe and beyond

DOCUMENT 2017 Drug misuse and dependence: UK guidelines on clinical management

DOCUMENT 2012 Medications in recovery: re-orientating drug dependence treatment

STUDY 2009 A practical clinical trial of coordinated care management to treat substance use disorders among public assistance beneficiaries