To investigate the effect of performance-based financial incentives on the influenza immunization rate in primary care physicians' offices. Randomized controlled trial during the 1991 influenza immunization season. Rochester, New York, and surrounding Monroe County during the Medicare Influenza Vaccine Demonstration Project. A total of 54 solo or group practices that had participated in the 1990 Medicare Demonstration Project. All physicians in participating practices agreed to enumerate their ambulatory patients aged 65 or older who had been seen during the 1990 or 1991 calendar years, and to track the immunization rate on a weekly basis using a specially designed poster from September 1991 to January 1, 1992. Additionally, physicians agreed to be randomized, by practice group, to the control group or to the incentive group, which could receive an additional $.80 per shot or $1.60 per shot if an immunization rate of 70% or 85%, respectively, was attained. The main outcome measures are the 1991 immunization rate and the improvement in immunization rate from the 1990 to 1991 influenza seasons for each group practice. For practices in the incentive group, the mean immunization rate was 68.6% (SD 16.6%) compared with 62.7% (SD 18.0%) in the control group practices (P = .22). The median practice-specific improvement in immunization rate was +10.3% in the incentive group compared with +3.5% in the control group (P = .03). Despite high background immunization rates, this modest financial incentive was responsible for approximately 7% increase in immunization rate among the ambulatory elderly.