Managed care advocates suggest that primary care gatekeepers may improve patient care and reduce costs. Training internal medicine residents in these gatekeeping functions has not been emphasized in most internal medicine programs. To determine if residents could perform gatekeeper functions and if patient costs and satisfaction would be favorable. Patients (n = 254) followed up by residents (n = 26) in continuity clinics at the Denver Veterans Affairs Medical Center in Denver, Colo, were divided into 2 groups. A control group of 128 was followed up by residents with no restrictions on appointments made for them. An intervention group of 126 patients were followed up by residents who had to approve all referrals made for their patients. A research nurse assisted with the approvals when the residents were rotating through other institutions. Utilization of resources, satisfaction with care, and health status were monitored over a 1-year period. A minor reduction of resource utilization was found in the intervention group, particularly in medication use. Significantly more visits were made to primary care providers in the intervention group (3.01 vs 2.59; P = .03). Patient satisfaction and health status were similar in both groups with a trend toward better satisfaction in the intervention group in some areas. Our study showed that residents can function as gatekeepers of highly complex patients and that satisfaction with care and utilization of resources does not suffer. Drug utilization and costs may be less when a gatekeeper exists.