A patient acceptability study was conducted using patient-controlled intranasal diamorphine. Patients undergoing nonemergency orthopaedic or gynaecological surgery self-administered intranasal diamorphine for 24 h postoperatively. Pain, pain relief, sedation, respiratory rate, nausea and vomiting were assessed regularly. After 24 h, patients and their attending nurses completed a questionnaire assessing satisfaction and practical aspects of the technique. Satisfaction was reported as good or complete by 69% of patients and 69% of nurses. Pain relief was assessed as better than expected by 45% of patients and better than normal by 50% of nurses. Seventy-nine per cent of patients would be pleased to use patient-controlled intranasal diamorphine again and 89% of nurses would be happy for their patients to use it again. Sedation was uncommon and mild and there were no episodes of significant respiratory depression. Fifty-three per cent of patients reported no nausea and 74% did not vomit at any stage. There were seven withdrawals, four due to problems with the device and three due to therapeutic problems. The nasal spray may need modification to improve reliability. However, we found patient-controlled intranasal analgesia an effective technique, which was well tolerated by patients and nurses and was without unpleasant side-effects. Further work to determine how it performs compared with intramuscular or intravenous analgesia is now needed.