A controlled trial evaluated a program to decrease the unnecessary use of inpatient testing by medical residents in a university medical center. The program included education, concurrent feedback, and resident participation in program planning. Using specific criteria for 7,891 chart audits of patients who had repeat tests within seven days, the authors measured change in testing among 44 residents in the first year and 43 in the second year. There were no significant differences related to the program. They conclude that substantial overuse of diagnostic tests did occur, that it varied from hospital to hospital, and that the program could not overcome powerful counteracting influences. Future studies, using control groups and chart audits, should evaluate interventions other than education.