A perspective on the role of decision analysis in clinical practice

The upsurge in interest in clinical decision support techniques is not accidental. Clinical decision situations are more and more characterized by complexity and uncertainty, due to increasing diagnostic and therapeutic possibilities, the fast growth of medical knowledge, the increasing quality demands by the public, and the pressure towards cost-conscious decision making. On the other hand, information science is increasingly able to provide the tools for building decision-support systems. There are several approaches to the development of decision aids. The present paper is mainly concerned with one potentially very powerful method: decision analysis. Basically, decision analysis is a method for explicitating, structuring and analysing clinical decision situations for individual patients. Suggestions for therapy choice, and diagnostic and prognostic statements are obtained by integrating the probability and value judgments made during the analysis. A clinical decision analysis can conveniently be divided into four stages: defining and structuring the clinical problem, assessing the relevant probabilities and utilities, calculation of the preferred course of action, including sensitivity analysis, and formulation of the results of the analysis in a clinically useful way. An optimistic scenario of the future role of decision analysis in clinical medicine would describe--among others--decision-consultation departments, decision-oriented clinical data-bases and medical knowledge bases, research groups for development, testing, and updating of clinical strategies, decision-analysis courses in all levels of medical education, and decision-oriented clinical textbooks. Moreover, decision analysis will help in setting priorities in clinical research, in quality control and peer review, and in development of protocols for use by the paramedical personal. Some illustrations from applied decision analysis are given.