A four-way single-blind crossover study was used to compare the efficacy and tolerance of four non-steroidal anti-inflammatory drugs. In addition, pain intensity was compared during the day, at night, at rest, on walking, in the most painful joint, and with the patients most painful activity. Ninety-six patients with rheumatoid arthritis took single daily doses of controlled release naproxen (N), diclofenac S.R. (D), indomethacin S.R. (I) and standard piroxicam (P). The greatest changes from baseline after treatment were seen in those patients with the highest initial pain measurement scores. Assessments of pain in the morning, in the most painful joint and the most painful activity were more discriminating than those at noon or at rest. Of the treatments, 'N' and 'P' were the most effective in reducing pain, with statistically significant differences from baseline. 'I' was the most effective in reducing morning stiffness. Adverse experiences were generally mild, occurring more frequently on 'I' than on other treatments.