A Novel Approach to Modelling the Prescribing Decision, Integrating Physician and Patient Influences

This paper describes a case study designed to demonstrate the feasibility of building a linked decision model based on the implications of distributed decision-making in healthcare, and thus to provide the ability to make quantified predictions of product offer performance. The approach taken was to adapt an existing conjoint-based forecasting tool (CAPMOD(tm)), (Brice et al. 2000). Our results show that there is a subset of product attributes on which physicians and patients perceive substantive differences in terms of their relative importance in their views of therapy alternatives. We also demonstrate that the observed differences in predicted share uptake between the separate, non-integrated physician and patient models and the integrated model do not necessarily follow from the observed differences in average relative importance between the two customer types, as would be the case for many existing simulation models. This additional insight into the decision-making process was possible through the use of a decision model which includes the key element of individual physician-patient linkage with an associated cut-off threshold. The paper describes the details of the approach and shows example outputs from the model. It will explore a number of interesting practical and theoretical issues that were encountered in the course of conducting this research.