To assess the extent of inappropriate hospital use in pediatric inpatients, I modified the Appropriateness Evaluation Protocol (AEP)--developed to assess inappropriate hospital use in adults--to apply to children and used it to evaluate pediatric inpatients retrospectively for every 10th day, from July 1982 to July 1983, at the University of Wisconsin Hospital. Of 1098 patient-days evaluated, 21.4 percent were judged to represent inappropriate hospital use on the basis of the protocol's criteria. The rate of inappropriate use varied according to admitting specialty, ranging from 7 of 70 days (10 percent) for pulmonary medicine to 43 of 61 days (70 percent) for neurology (P less than 0.005). There was a tendency toward lower rates of inappropriate use in uninsured patients (6 of 44 days [14 percent] vs. 226 of 1038 days [22 percent] in patients with Medicaid or private insurance, P = 0.13), and rates were lower in younger children (74 of 432 days [17 percent] in children less than or equal to 5 years of age vs. 162 of 656 days [25 percent] in children greater than 5 years of age, P less than 0.005). There was no variation according to sex, day of the week, or month. Contrary to expectations, inappropriate use decreased with increased lengths of stay (for stays of 1 day, 8 of 13 days were inappropriate [61 percent]; for stays of 2 to 6 days, 118 of 410 days were inappropriate [29 percent]; for 7 to 13 days, 58 of 291 [20 percent]; and for greater than or equal to 14 days, 51 of 362 [14 percent], P less than 0.001). I conclude that there is a substantial rate of inappropriate hospital use in pediatrics and that such use is more likely during short admissions than during long ones. Cost-containment efforts directed at limiting the length of hospitalization may therefore not reduce inappropriate hospital use in this population.