PURPOSE: To determine the effectiveness of a clinical-practice intervention in improving the control of pain in outpatients with cancer. METHOD: Between July 5 and September 30, 1995, a randomized, controlled trial of 510 cancer outpatients and 13 oncologists was conducted at 23 clinics in Indiana. All the patients completed assessments of their pain, their pain regimens, and the degrees of relief received; they were surveyed again by mail four weeks after their clinic visits. The intervention group's clinical charts contained a summary of the completed pain scales; the oncologists who treated these patients were instructed to review the summary sheet prior to an evaluation. This summary was not available for the oncologists treating the patients in the control group. Each patient's pain management index (PMI) was calculated: the patient's pain medication level was rated on a scale of 0 to 3; the patients's pain level was rated on a scale of 0 to 3 and then subtracted from the first rating. A negative PMI was interpreted as representing insufficient treatment. Data were analyzed with several statistical tests. RESULTS: In all, only 320 patients who reported cancer-related pain were used in the analysis: 160 to 260 in the control group and 160 of 250 in the intervention group. The groups were similar with respect to demographics, cancer sites, and performance status. A significant difference (p = .0162) in the physicians' prescription patterns was found. In the control group, prescriptions for 86% of the patients did not change, with no decrease in analgesic prescriptions; for 14% of the patients analgesic prescriptions increased. In the intervention group, analgesic prescriptions changed for 25% of the patients, decreasing for 5% and increasing for 20%. A decrease in the incidence of pain described as more than life's usual aches and pains was found for the intervention group (p = .05). No significant difference was found between the groups for the patients undertreated for pain, as measured by PMIs. CONCLUSION: Although analgesic regimens were altered significantly when the physicians understood more about the patient's pain, cancer pain management remains a complex problem. Future studies should focus on the long-term systematic incorporation of simple pain-assessment tools into daily outpatient oncology practices as well as on innovative ways to address other aspects of managing cancer pain.