EXPERT SYSTEMS FOR CLINICAL DIAGNOSIS.

The process of medical diagnosis may be decomposed into the following problem solving tasks: Initial findings accumulation and integration, hypothesis generation and evaluation, goal setting for attention management, information sources evaluation and selection, sorting of new findings by goals, hypotheses and other relevant disorders, evidence integration, and summary attempts. Our approach to structuring the knowledge for medical diagnosis is presented, followed by a description of a clinician-oriented control strategy for diagnostic expert systems. This approach represents a culmination of ten years experience gained in the research and development of large scale medical expert systems, i. e. , MEDAS, RDAS and TOXDAS.