This study compares the effectiveness of two methods of opiate reduction in 108 chronic pain patients during a 4 week inpatient pain management programme, and at 1-month and 6-month follow-up. Patients chose either the patient-controlled reduction (PCR) or cocktail reduction method, aiming to complete withdrawal by discharge. Use of opiates and other drugs was recorded, and psychological measures taken, at admission, at discharge, and at follow-ups. Patients who opted for the cocktail reduction method started at higher morphine equivalents (P < 0.001), were less confident in their ability to cope without medication (P < 0.05), and rated their everyday activities a more disrupted by pain (P < 0.05). At discharge, 89% of the cocktail group were abstinent from opiates compared with 68% of the PCR group (P < 0.05). By 1-month follow-up, the advantage of the cocktail method had disappeared, with no significant differences between the two groups in mean opiate dose, nor in the proportion of abstinent patients. This was the result of a greater return to opiate use in the cocktail group, with abstinence rates remaining unchanged in the PCR group. By 6-month follow-up, abstinence rates for the groups were equivalent, with 55% of patients remaining off opiates. By this stage, however, non-abstinent cocktail group patients were taking significantly larger doses of opiates than PCR patients (P < 0.05), although in both groups, the dose was well below admission level. Admission opiate dose level was the best predictor both of abstinence at discharge and of subsequent opiate dose level in non-abstinent patients. This study demonstrates that both reduction methods can produce substantial reduction in opiate use by severely impaired chronic pain patients with long medication histories.