Decision-theoretic approach to a knowledge-based clinical system

Abstract

Many researchers in the medical area are interested in developing methods for providing knowledge-based decision support to the practicing physician. Although they have demonstrated the potential to provide useful computer-based consultation, they have paid little attention to the modeling of patient specific preferences and tradeoffs about the quantitative values obtained from physicians. Furthermore, purely symbolic (nonquantitative) reasoning techniques have limited utility in the solution of many medical problems without the explicit consideration of the quantitative notions of uncertainty and tradeoffs. Thus, this research attempts to develop a knowledge-based clinical system for consulting disease treatment using decision trees, axioms of utility, and sensitivity analysis in a decision-theoretic approach. To implement the idea, DACS (Decision-Analytic Consulting System) was designed and a prototype was developed on the microcomputer. The operation of DACS is illustrated via a Cervical Cancer Treatment example. © 1990.