Screening mammography is underutilized in many primary care practices. We designed a prospective, controlled trial to evaluate two strategies for improving the utilization of mammography in an academic general medicine clinic. We assigned teams of house officers to (1) physician audit with periodic feedback, (2) a visit-based strategy directed at both patient and physician, or (3) a no intervention concurrent control arm. After 6 months, the percentage of 50- to 74-year-old women meeting the standard of an annual mammogram was 36% for patients in the control arm, 62% for patients of feedback residents, and 54% for patients of the arm receiving the visit-based strategy. Patients of female providers were slightly more likely to meet the standard, but no effects were detected for patient race, new as opposed to follow-up patient, or higher frequency of clinic visits. We conclude that audit with feedback and a new visit-based strategy of a patient cue associated with a simplification of the ordering process each greatly improved the rate of utilization of screening mammography. Practitioners could reasonably choose the strategy most suited to their own situation.