In the study reported here, the author compared the hospital charges of three groups of patients in one combined university/community hospital: private patients with a community attending physician; teaching patients with a community attending physician; and teaching patients with a faculty attending physician. Control of case mix differences was accomplished by selecting four surgical diagnosis related groups, modifying these groups through use of the International Classification of Diseases procedure codes, and using patient demographic, health status, and severity of illness control variables. The results of the study indicated that teaching patients with a faculty attending physician had significantly higher hospital charges, while teaching patients with a community attending physician had similar charges as compared with the private, nonteaching patients. The higher charges were attributable to increase ordering of diagnostic tests and differential use of supplies and equipment. The level of resident supervision by attending physicians appeared to be influential in determining total charges incurred by the teaching patients.