The inexperience of house staff has been offered as one explanation for the increased cost of care at teaching hospitals, but conclusive evidence for this has been lacking. We studied the relationship of house staff experience to the cost and quality of inpatient care in a large series of internal medicine patients at one teaching hospital. We defined house staff experience by the month of academic year during which the patient received care. Our measures of cost were length of hospital stay and total hospital charges, while our measures of quality were hospital deaths, hospital readmissions, and nursing home placement. Multiple linear regression analysis on 21,679 hospital discharges revealed increasing house staff experience to be associated with a significant decline in length of stay (95% confidence interval for b, -0.006 to -0.066 days per discharge per month of house staff experience) and total hospital charges (95% confidence interval for b, -0.002 to -0.017 log dollars per discharge per month of house staff experience). These findings constitute an estimated average decline of 0.43 days per discharge and +370 per discharge over the academic year. Logistic regression analysis found no relationship of house staff experience to hospital deaths, readmissions, or nursing home placement. These findings suggest that the process of training inexperienced physicians may represent an important source of inefficiency for teaching hospitals struggling in a competitive environment.