The impact of primary care gatekeeping on selected patterns of physician use was examined among Medicaid beneficiaries in two demonstration programs. The evidence indicates that beneficiaries enrolled with gatekeepers were significantly less likely to see specialists when compared with unenrolled beneficiaries in comparison groups. Primary care visits increased to offset these reductions only when gatekeepers were paid on a fee-for-service basis. Increased overall reliance on primary care physicians as opposed to specialists was also observed in the gatekeeper programs. Findings also indicate that enrolled beneficiaries received care from fewer sources than they had prior to enrollment. Although these changes in patterns of use have the potential to assure access to a more stable and structured system of care, the clinical and long-term economic consequences of such changes remain unknown.