A total of 693 children between the ages of 0 and 13 years were randomly assigned to either a staff model health maintenance organization (HMO) or to one of several fee-for-service insurance plans in Seattle to evaluate differences in medical expenditures and health outcomes. Although the fee-for-service plans varied the amount of cost sharing (0 percent to 95 percent), all children were covered for the same medical services, for either three or five years. No differences in imputed total expenditures were observed for children assigned to the HMO or any of the fee-for-service plans. Children with cost-sharing fee-for-service plans were perceived (by their mothers) to be in better health overall than those assigned to the HMO. No significant differences regarding physiological outcomes (e.g., visual acuity, hemoglobin level) were observed between the two groups. The results of this experiment neither strongly support nor indict fee-for-service or prepaid care for children.