CLINICAL DECISION MAKING AND ARTIFICIAL INTELLIGENCE

点击翻译按钮获取中文摘要

We propose to apply artificial intelligence methods to the study of clinical decision making and the creation of knowledgable expert medical consulting programs. This application proposes three related investigations: 1. To study the cognitive processes of physicians engaged in clinical decision making by collecting and analyzing transcripts of their volunteered and elicited responses while solving clinical problems; to describe the knowledge representations and problem-solving methods used by physicians in cases involving considerable uncertainty and considerable risk; to study the variation in skills and style of experts by degree of expertise and area of sub-specialization. 2. To develop new methods that allow expert programs to explain and justify their conclusions by arguing from fundamental medical facts and principles and reconstructing the path by which those bases have led to the program's behavior; to improve the application of automatic programming methods to allow the performance program to be generated automatically from these facts and principles. 3. To continue development of a decision analysis and sensitivity analysis system for clinical and policy analysis use on microcomputers; to distribute that system to a limited cohort of clinicians and investigators, and to study its use.