A US recovery model yet to be tried in Europe has proved its effectiveness in a rare randomised trial of a mutual aid intervention.
Democratically run by their residents, the USA has over 1200 Oxford Houses and others operate in Australia and Canada. Each houses six to 18 same-sex residents (some also house their children) who typically have achieved abstinence from alcohol or drugs through a short rehabilitation or detoxification programme and commit to maintaining it with the aid of fellow residents. Applicants are admitted by a vote of current residents. Those judged by housemates to have failed to comply with house rules including abstinence are evicted but may be readmitted after a month sober. The self-financing structure (residents pay all expenses and repay start-up loans from public bodies) permits unlimited stays (these average just over a year) and excess demand is typically met by opening another house.
Researchers recruited 150 adults from in-patient treatment units in Illinois who agreed to be randomly allocated to usual care (the control group) or to apply to Oxford Houses. Typically they were single, black, unemployed women in their late 30s, many with a history of imprisonment and mental health problems. Two reports document this first randomised controlled trial of the houses, the first1 focusing more on describing the intervention, the second2 a fuller account of the study and its outcomes.
All 75 Oxford House applicants became residents. Over half left within six months and by the final follow-up (two years after treatment discharge) just 5% remained. Importantly this means the later results reflect the persisting effects of having been prompted by the study to seek Oxford House residency. Regardless of their location, nearly 90% of the sample completed follow-up assessments.
Compared to the control group, over the follow-up period fewer Oxford House assignees were using alcohol or drugs or charged for a recent offence and more were employed. By the end fewer than half as many (31% v. 65%) were using alcohol or drugs, a third as many were in prison (3% v. 9%), and average earnings were $550 a month higher. All these differences were reported as statistically significant. Additionally, at two years 27% more Oxford House assignees had their own accommodation and nine more mothers had regained or retained custody of their children.
Longer stays were associated with better outcomes but residents who left within six months were still less likely than the control group to be using alcohol or drugs at the last follow-up (46% v. 65%). Though overall the houses were equally (in respect of criminal charges, more) effective for the younger half the sample (under 37), younger residents who left early did no better than the control group.
For a mutual aid intervention, this was an unusually rigorous test with convincing results, but there are some concerns. Unlike the national Oxford House population,3 nearly two thirds of the study sample were women (a function of the units where they were recruited4), raising concerns over representativeness. However, there was no significant indication that the houses benefited one gender more than another.
Presumably only people with no overriding accommodation or relationship commitments would have agreed to enter the study, confirmed by their generally single status and by the fact that just 16% of the control group exited treatment to their own homes and another 16% to a partner's. Where (as in Britain) welfare and housing safety nets are more robust, the control group may have been less disadvantaged. The study was unable to exclude the possibility that decent accommodation may in itself have raised the control group's outcomes nearer to those of the Oxford House group.
It was essential for the researchers to gain the trust of Oxford Houses over many years of collaborative working, raising the possibility of researcher allegiance influences, especially since it was difficult for interviewers to remain unaware of which group respondents had been assigned to.4
There is no indication of the severity of the sample's substance use or psychological problems at the time they entered the study, leaving an important question unanswered – whether severely dependent residents would have been able to comply with house rules. On the other hand, the fact that Oxford House assignees were aided by the researchers may mean they were less motivated than the typical resident who has to find and apply for a house under their own steam.
In the US context, for people without a home or able to move home for several months, Oxford Houses offer a way to preserve the gains achieved in short-term detoxification programmes at no cost to the wider society, which benefits from reduced substance use and criminality. In this role, they may in the UK provide an extended and accessible platform for sustaining recovery to supplement the limited supply and limited duration of expensive residential rehabilitation places.
More on Oxford Houses in the researchers' (at the time of writing) forthcoming book including residents' stories.
Thanks to Rowdy Yates of the University of Stirling for his comments on this entry in draft and to Leonard Jason for supplying further information. Neither bears any responsibility for the text including the interpretations and any remaining errors.
1 FEATURED STUDY Jason L.A. et al. Communal housing settings enhance substance abuse recovery. American Journal of Public Health: 2006, 96, p. 1727–1729.
2 FEATURED STUDY Jason L.A. et al. An examination of main and interactive effects of substance abuse recovery housing on multiple indicators of adjustment. Addiction: 2007, 102(7), p. 1114–1121.
3 Jason L.A. et al. The need for substance abuse after-care: longitudinal analysis of Oxford House. Addictive Behaviors: 2007, 32(4), p. 803–818.
Last revised 07 January 2008
STUDY 2004 Crack: making and sustaining the break
STUDY 2019 Efficacy and cost-effectiveness of an adjunctive personalised psychosocial intervention in treatment-resistant maintenance opioid agonist therapy: a pragmatic, open-label, randomised controlled trial