The medical management of patients with acute respiratory illnesses was analyzed at two different clinics during a 14- to 21-month period. Patients received care from either physicians or physician-supervised physician's assistants (PA). The PAs used respiratory illness clinical algorithms to guide their choice of diagnostic tests and treatment. Illness outcome, patient satisfaction, and medical care cost data were obtained for all patients approximately two weeks after the index illness. Despite significant differences in patient population characteristics, illness outcomes were similar, regardless of the provider's educational background. Medical care costs, however, were highest for the physician's patients. For all patients, diagnostic tests contributed about one-third of the total direct costs, mainly because of chest x-ray and throat culture use. Sixty to eighty per cent of medication costs were due to nonprescription drugs used principally for symptom relief. The data demonstrate that the medical care delivered by these physician's assistants was as effective and less costly than the care provided by physicians. Reducing chest x-ray and throat culture use would have a significant economic impact, without adversely affecting medical care effectiveness.