Oxford Houses Offer Both Recovery Benefits & Cost Savings

Oxford Houses are a type of abstinence-focused recovery residence that are democratically-run, where residents are entirely responsible for house decisions and maintenance. In this study, authors conducted the most rigorous evaluation of Oxford Houses – and recovery residences more generally – to date.


Oxford Houses are a type of self-sustaining recovery residence, first developed in 1975. They are non-professional, and require that residents are abstinent from alcohol and other drugs. While they are not affiliated with 12-step mutual-help organizations like Alcoholics Anonymous, members are traditionally encouraged – though not mandated – to attend meetings. Members pay rent, and can stay there as long as needed, provided they follow house rules. While there is a manual that lays the initial groundwork for a new Oxford House to aid in quality control, decisions around consequences for individuals who break rules are up to the other house members. House leadership positions are limited to 6 months so that members all have a chance to be decision makers.

There are approximately 2,000 Oxford Houses in the U.S. and other countries, supporting 24,000 individuals per year. This series of studies by Jason and colleagues – the most rigorous to date on this topic – examined the effects of Oxford House participation on several recovery-related outcomes among individuals who attended residential substance use disorder treatment. Study authors evaluated not only whether Oxford House participation is helpful, but also for whom it is most helpful (e.g., older vs. younger individuals), and, perhaps most importantly, whether Oxford Houses can help reduce the immense financial burden caused by the consequences of substance use disorder. Together, these studies provide the field with critical information on which to base policy-related, clinical, and personal decisions regarding recovery residences.


Authors assessed 150 individuals discharged from residential substance use disorder treatment in Illinois every 6 months for 2 years (i.e., assessments at 6 months, 12 months, 18 months, and 24 months). Half of the group (n = 75) was randomly assigned to the Oxford House condition, and the other half (n = 75) were randomly assigned to usual continuing care (i.e., the professional or mutual-help organization services received after treatment).


Apart from the initial random assignment to each of these conditions, participants were free to engage in other recovery support services as they wished. Thus, after individuals assigned to the Oxford House condition were brought to one of 20 residences across the state, current members voted on whether they could become a resident, as per Oxford House policy. Only one research participant was rejected by vote initially, though research staff subsequently brought this person to another house, who approved his/her residence.

Of note, members were able to stay or leave the residence voluntarily – 95% moved out of their respective Oxford Houses at some point over the 2-year study, for example. For those assigned to usual continuing care, case managers at the treatment center referred individuals to different combinations of outpatient treatment, mutual-help, and other community resources. The majority of usual care participants lived in their own home, or the home of a spouse/partner, relative, or a friend (67%). Nearly 20% lived in a non-Oxford, professionally staffed recovery residence.


The groups were compared on four primary outcomes across the entire 2-year study period:


  1. Substance use abstinence (yes or no in the past 6 months; self-reported alcohol and other drug use was validated by an identified significant other at the final, 2-year follow-up only)
  2. Criminal charge(s) (any: yes or no in the past 30 days)
  3. Employment (any full-time or part-time work in the past 30 days: yes or no)
  4. Self-regulation in daily life (36-item self-report scale where participants indicated the extent to which statements reflected who they are, such as “I am good at resisting temptation”, “I never allow myself to lose control”, and “I’m not easily discouraged”)


Given that individuals who are younger and/or have a co-occurring psychiatric disorder, often have poorer outcomes compared to older individuals and those with only a substance use disorder, respectively, authors also tested whether Oxford House participants did better or worse than usual care participants depending on their age (37+ years old versus 36 years old or younger) or whether they had a lifetime diagnosis of anxiety or mood disorder (according to the fourth edition of the diagnostic and statistical manual of mental disorders (DSM IV). Finally, just among Oxford House participants, they tested if individuals who stayed in the recovery residence for 6 or more months had better outcomes.

Oxford House and usual care participants were initially similar on the four primary outcomes (e.g., 6.7% in each group were abstinent before entering treatment) as well as on demographic characteristics, which is to be expected in a randomized trial of this size. Overall, 62% were women, and Black individuals were well represented, comprising 77% of the sample, compared to 11% White, and 8% Latino. The average participant had 12 years of education, corresponding with a high-school diploma, and 44% entered the study with a history of criminal justice system involvement. Six out of 10 participants had a co-occurring mood or anxiety disorder in their lifetime, while 28% reported a history of having taken psychiatric medication, 27% had attended inpatient substance use disorder treatment (before the residential treatment episode that preceded entry into the study), 28% attended outpatient treatment, and 8% had attempted suicide. Information regarding participants’ substance use history, including substance use disorder diagnosis, was not reported.


As illustrated in the figure below, Oxford House participants had better outcomes over time across the board, even when models adjusted for participant gender, age, and the presence of a co-occurring psychiatric disorder. In addition, Oxford House participants also had greater increases in self-regulation over time.


2-year outcomes for Oxford House vs. Usual Care groups

Only the effect of Oxford House vs. usual care on criminal charges depended on individual characteristics, such that those who were younger and had a co-occurring disorder did better in the Oxford House group. In other words, if they were older, and only had a substance use disorder (rather than both a substance use and mood/anxiety disorder), being in the Oxford House group or the usual care group made no difference with respect to the criminal justice outcome. For all three other outcomes (including abstinence), Oxford House outperformed usual care regardless of age or diagnostic status.

Importantly, when looking only at Oxford House participants, individuals who stayed there for 6 or more months had much better abstinence rates (84 vs. 54%). This added benefit of a 6-month or longer stay was especially true for younger individuals – not only on the abstinence outcome, but on the three other recovery outcomes as well. Given that employment may be a particularly important outcome for young adults (given that substance use may interrupt the achievement of typical adult milestones), it is important to note, for example, that 94% of younger patients with 6+ months were employed at 2-year follow-up vs. 56% who stayed for less than 6 months.

Are Oxford Houses also Cost-effective?

The public health significance of these findings are further enhanced by data from a related study by the same research team, who evaluated cost-effectiveness of Oxford Houses in this sample of individuals. They examined 129 of the 150 individuals that had sufficient data to carry out the analyses.

Authors measured costs based on the following:


They measured financial benefits from both personal and societal perspectives based on the following:


Despite greater average costs per participant across 2 years ($3200 more), financial gains for Oxford House participants far outweighed costs ($32,200 more), primarily driven by reduced illegal activity. All told, the net benefit of being assigned to the Oxford House condition versus usual care was $29,000 per person during the 2-year study.


This series of studies on Oxford Houses by Jason and colleagues is the most rigorous evaluation of recovery residences to date. Overall, for individuals completing residential substance use disorder treatment, Oxford Houses provided substantially greater benefit over time, not only in terms of abstinence rates but also employment and criminal justice outcomes as well.

Furthermore, stays of at least 6 months seem to provide added benefit, particularly for younger individuals. Emerging adults (e.g., ages 18-29) are often at greater risk for relapse, in part due to their riskier social networks, which have a greater density of alcohol and other drug users, compared to older individuals. Oxford Houses, and potentially other recovery residences, may offer a substance-free community that helps promote engagement in recovery-related activities.

In another study specifically on 18-24 year old individuals who attended residential treatment, for example, recovery residence participation was uniquely related to better abstinence rates during the first post-treatment year, beyond participation in residential treatment, outpatient treatment, and mutual-help organizations. It is worth highlighting also that longer Oxford House stays in this study were associated with extremely high rates of employment for younger individuals, who may otherwise struggle to meet important adult milestones like financial independence.

  1. Individuals for this study were recruited after being discharged from residential treatment. More research is needed to evaluate the benefits of Oxford Houses for other types of individuals.
  2. Oxford Houses are a specific type of recovery residence, with fairly rigorous levels of quality control, and a specific democratically-run system of house governance. While other studies have examined different types of recovery residences (e.g., Sober Living Homes), less is known about whether staying in these other types of residences produces similar recovery benefit.


Next steps might include examining how Oxford Houses help their residents. In a related but separate study of Oxford House residents, Jason and colleagues found increased confidence to abstain from substance use (e.g., self-efficacy) and lower social support for substance use in one’s social circle were associated with higher abstinence rates over time. As such, these variables might be good initial candidates to examine. Randomized trials like this one, but focused on other types of recovery residences, are also warranted.


  • For individuals & families seeking recovery: For someone attending, or being discharged from, residential substance use disorder treatment, Oxford House participation after treatment is not only likely to improve one’s chances of recovery, employment, and reduced criminal activity, but, on balance, it is also likely to help financially as well. While more research is needed on how long one needs to stay at the Oxford House residence, some preliminary research suggests added benefit with at least a 6-month stay.
  • For scientists: This series of studies is the most rigorous evaluation of recovery residences to date. They suggest not only that Oxford Houses are beneficial, but also cost-effective, recovery support services. Jason and colleagues provide a helpful model by which other recovery support services may be evaluated, including a cost-benefit analysis to understand the potential societal impact of linking more individuals to these widely available services available in the communities in which individuals live.
  • For policy makers: Oxford Houses not only help promote recovery-related benefit, but also help address the financial burden of substance use disorder. Based on this series of studies by Jason and colleagues, policies that support greater access to, and linkage with, recovery residences are warranted.
  • For treatment professionals and treatment systems: This series of studies is the most rigorous evaluation of recovery residences to date. They suggest not only that Oxford Houses are beneficial, but also cost-effective, recovery support services. Particularly for providers in residential treatment settings, Oxford Houses should be considered a first line continuing care resource to which patients can be referred after discharge. More research is needed on the recovery-related benefit offered by other types of recovery residences. Based on preliminary data where studies show recovery residence participation could be, but is not necessarily, responsible, for better outcomes other types of recovery residences also appear to be promising post-treatment recovery support services.


Jason, L. A., Davis, M. I., & Ferrari, J. R. (2007). The need for substance abuse after-care: Longitudinal analysis of Oxford House. Addictive Behaviors32(4), 803-818.

Lo Sasso, A. T., Byro, E., Jason, L. A., Ferrari, J. R., & Olson, B. (2012). Benefits and costs associated with mutual-help community-based recovery homes: The Oxford House model. Evaluation and Program Planning35(1), 47-53.


Sign Up for the Bulletin

  • This field is for validation purposes and should be left unchanged.