Journal of Health Economics 5 (1986) 253-276. North-Holland zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPO MEDICAL STAFF SPECIALTY MIX AND HOSPITAL PRODUCTION Gail A. JENSEN* Universi’ty of Illinois at Chicago, Chicago, IL 60680, USA Michael A. MORRISEY” University of Alabama at Birmingham, Birmingham,AL 35294, USA Received November 1985, final version received May 1986 This paper analyzes the role of medical staff characteristics in determining different dimensions of hospital output. Using a set of flexible functional form production functions, and adjusting for hospital case mix, we examine the output contribution of physicians and other inputs, and the influence that physicians in different specialties have on the productivity of other physicians as well as on other labor and capital- We also examine the input substitution possibilities available to hospitals, and where possible, we compare our estimates to those obtained by other researchers. We find that physicians have numerous significant effects on production and conclude that physicians are an important input that should not be ignored in empirical cost and production function studies for hospitals. 1. Introduction Physicians play a unique role in the production of hospital care, By contributing their own time to activities which improve health outcome they serve as an essential factor of production. Also by managing their patients’ use of hospital services they play an important role in determining the productivity and allocation of other resources. This dual role as both an input and manager has long been recognized in theories of hospital behavior. The Pauly and Redisch (1973) physicians’ cooperative model explicitly incorporates physicians in the hospital production function, and posits that they choose inputs to maximize average revenue per medical staff member. Feldman (1976) and Harris (1977) also acknowledge their multiple role. In their views, the hospital is actually two separate firms or divisions of medical *We wish to thank Roger Feldman, David Dranove, and particularly an anonymous referee for helpful comments on a previous draft. An earlier version of this paper was written while both authors were with the Hospital Research and Educational Trust (HRET), Chicago: we thank them for their continued interest. The opinions and conclusions of this paper do not necessarily reflect those of our a*;ganizations or of the HRET. 0167-6296/86/$3.50 0 1986, Elsevier Science Publishers B.V. (North-Holland)