The efficacy of ketorolac, a non-steroidal anti-inflammatory drug, in the management of moderate to severe pain in adults, has led us to conduct a trial of this analgesic in children following tonsillectomy. Children were randomized to receive intramuscular (i.m.) ketorolac (1 mg/kg, EXP group, n = 45) or saline (CTL group, n = 42) at the completion of surgery. Intravenous (i.v.) fentanyl (0.5 micrograms/kg/dose) was administered in repeated doses postoperatively. Pain intensity was measured using both the Oucher and the Children's Hospital of Eastern Ontario Pain Scale (CHEOPS) to allow for comparison between self-report and behavioral measures of pain intensity. Severity of postoperative bleeding was measured using a 4-point rating scale. The EXP group had a significant reduction in total fentanyl dose (mean: 35.9 micrograms) compared to the CTL group (mean: 48.3 micrograms, t = -2.21, P < 0.03). There was a statistically significant decrease in pre-fentanyl CHEOPS scores in the Post-Anesthesia Care Unit (PACU) in the ketorolac group (F (2, 30) = 5.34, P < 0.01), but not in the saline group (F (2.24) = 2.46, P > 0.05). In the first hour postoperatively, the CHEOPS demonstrated significant decreases in pain intensity scores in response to opioids, in both groups. In the PACU, children were unable to provide a self-report of pain intensity potentially due to a variety of factors (e.g., emergence delirium, agitation, excitement, sedation, and/or pain). However, during the remainder of the postoperative stay, the photographic scale of the Oucher was a more valid measure of pain intensity than the CHEOPS.(ABSTRACT TRUNCATED AT 250 WORDS)