Concurrent charge feedback has gained widespread acceptance as a method of minimizing hospitals' losses under the Medicare prospective payment system despite the fact that its effect on patient outcomes, physician behavior, or charges has not been studied in depth. In a controlled trial on two medical wards in an academic medical center, the effect of daily charge feedback on charges was studied. Sixty-eight house staff and 16 teaching attending physicians participated during a 35-week period, taking care of 1057 eligible patients. No significant differences in charges were seen when all patients were included. Since 45% of patients had planned protocol admissions (diagnostic workups or protocol treatment) on which the house staff had little change to impact, a subgroup analysis was performed, excluding these patients. In the remaining patients, a highly significant reduction in mean total charges (17%), length of stay (18%), room charges (18%), and diagnostic testing (20%) was found. In-hospital mortality and preventable readmission within 30 days were similar on the two wards. It was concluded that charge feedback alone is effective in a teaching hospital for decreasing charges.