This study reports a 9-18 month follow-up of a randomised controlled trial of pain management programmes for chronic, non-malignant pain. Twenty-two inpatients, 18 outpatients and 12 control subjects completed the follow-up assessments. Significant treatment effects were demonstrated by the inpatient group on pain ratings, the Pain Behaviour Checklist, and General Health Questionnaire, with similar effects demonstrated by the outpatient group on the former 2 measures. The findings were confounded by higher inpatient scores at pretreatment, in comparison with the 2 other conditions. There was a high drop-out rate of subjects, particularly from the control condition which illustrates the limitations of controlled group designs in this area. Analgesic use, activity levels and pain ratings were also evaluated using the criteria for 'success' described by Malec et al. (1981). Results indicated that 68% of inpatients, 61% of outpatients and 21% of control subjects met all 3 criteria. Both treatment programmes were effective in returning patients to paid employment, whilst 3 control group patients gave up work. The cost-benefit implications of these changes are discussed. We conclude that pain management programmes contribute substantially to the rehabilitation of chronic pain sufferers.