A prospective, controlled trial to reduce polypharmacy in patients 65 years or older was carried out in the residents' clinic of a teaching hospital. Of 272 elderly patients surveyed, 89 (33%) were taking five or more prescription drugs. Recommendations to discontinue medications or to simplify regimens were formulated for 79 polypharmacy patients. Compared to 41 controls, the 38 patients whose physicians were informed of the recommended changes demonstrated a small but significant reduction in the mean number of drugs, the complexity, and the cost of their regimens. Physicians complied with eight (100%) of eight recommendations to simplify a dosage schedule, eight (62%) of 13 recommendations to substitute a new drug for the old one, and only eight (40%) of 20 recommendations to stop a medication (P = .04). Noncompliance usually resulted from patient refusal or from medications being prescribed by another provider. Whereas feedback to the primary physician is beneficial, more substantial reductions in outpatient polypharmacy may require overcoming patient barriers and limiting the number of prescribing physicians.