The reports of Valdez et al1 and Leibowitz et al2 mark the culmination of a landmark effort. Initiated over a decade ago, the Rand Health Insurance Experiment—the most ambitious and expensive randomized controlled trial ever conducted in health services research—was designed to produce estimates of the costs of various forms of national health insurance. Projected estimates of the costs of Medicaid and Medicare had been far too low and health policy experts hoped to obtain information that would prove closer to the mark in the (then) seemingly likely passage of some form of national health insurance.
Although national health insurance has not materialized, the impact of medical care costs on utilization of services and health status is still of great interest. The amount of "cost-sharing" (medical costs paid by patients) has increased dramatically in public insurance programs, and many health policymakers favor increases in private insurance programs. The central question now, however, is how much of the cost burden can be borne by patients without inducing reductions in utilization that are harmful to health.
The two papers1,2 in the May issue of Pediatrics reported on the impact of cost-sharing on children's utilization and health. The basic findings were similar to those for adults. In general, the higher the costs paid by families, the fewer the children receiving medical care and the fewer the services per user. The only exception was hospitalization of children aged 5 to 13, which was largely unaffected by cost-sharing; higher costs did appear to reduce hospitalization of younger children, as well as ambulatory care for children of all ages.