Patients who were scheduled for an elective joint replacement or spinal procedure were randomly assigned prospectively to one of two groups for the management of postoperative pain: ninety-one patients (Group I) controlled the administration of a narcotic analgesic themselves and ninety-three patients (Group II) received intramuscular injections of a narcotic analgesic, as needed. The patients who controlled the analgesic used a smaller amount of the analgesic on the first postoperative day, but the over-all amount was not significantly different between the two groups. The group that received intramuscular injections reported less pain overall, according to one of three pain-assessment scales, and had more relief of pain over-all and on the first postoperative day, according to another scale. The patients who had had a total joint replacement and who controlled the analgesia walked farther on the first postoperative day than those who received intramuscular injections. There were no significant differences between the two groups with regard to the rate of complications, the arterial oxygen saturation levels during the first twenty-four hours after the operation, or the length of stay in the hospital. The nursing staff preferred the patient-administered method of analgesia, as it necessitated equal or less nursing time to assemble, initiate, and maintain than traditional intramuscular injections. The average cost per patient was $58.58 for the patient-administered analgesia and $22.45 for the intramuscular injections.(ABSTRACT TRUNCATED AT 250 WORDS)