Physicians and medical innovation

Previous attempts to model some aspects of physician behaviour include those of Evans, Sloan and Feldman and Wolfson. It is suggested that the introduction of knowledge as a distinct element in a microeconomic model of physician behaviour is preferable to the inclusion of a variable called 'discretionary influence' or 'quality of care' in the physician's utility function. This is because the properties of functions containing either of these variables appear to be indeterminate. By comparison the properties of the knowledge constraints can be specified with some confidence. The factors affecting a physician's demand for treatment on behalf of patients are identified as (1) the physician's objective function, (2) his knowledge and (3) the availability of medical resources. Furthermore, the knowledge element can be sub-divided into two parts: the set of prior probabilities and the set of likelihood functions. The former may be identified with the physician's local knowledge, whereas the latter may be associated with the physician's medical training. A significant fraction of the growing demand for hospital care has been attributed to changes in medical technology. During the late fifties and afterwards 'more cases became treatable' and physicians, it is argued, cannot resist the 'technological imperative'. The paper shows that the model may be used to generate testable hypothesis regarding the adoption by physicians of both process and product innovations. The discussion of the physician's medical knowledge is fundamental to the inducement mechanism. The policy instruments available to achieve an optimal diffusion of innovations are reviewed.