Using a naturalistic design involving consecutive referrals self-selected for family-oriented treatment, forty-two participants from a residential programme and twenty-five participants from a community-based programme were assessed on a range of alcohol-use and psychosocial measures before treatment. A proportion of these cases were assessed after treatment and at six months' follow-up. At six months' follow-up 79 per cent of both the residential and community groups were either abstinent or drinking moderately. However, more members of the residential group (75 per cent) were abstinent at follow-up compared with the community group (36 per cent).

In contrast, more members of the community group were moderate drinkers (43 per cent) at follow-up compared with the residential group (4 per cent). At six months' followup, compared with the community group, more members of the residential group showed a clinically significant reduction in recent negative consequences of drinking and psychological adjustment problems. Both groups made significant mean gains on indices of alcohol abuse and psychosocial adjustment but there were important intergroup differences. The residential group showed a greater mean reduction in recent negative consequences from drinking but the community group showed a greater mean reduction in the percentage of days' heavy drinking.