IV. Clinical Consultation Systems Medical Decision Support Systems and Clinical Research Data Bases: A. Medical Decision Support and Artificial Intelligence Methodologies: INFERNET—A Computer-Based System for Modeling Medical Knowledge and Clinical Inference

摘要

A new system for computer-based modeling of clinical inference is demonstrated. Within this system, medical knowledge may be represented as a network of “cause and effect” relationships among states (symptoms, signs, physiological states, and diseases). The strength of the relationships are quantified with conditional probabilities. Given a set of patient findings, the system will compute probabilities of diseases that may “cause” those findings. The algorithm used is based on Bayes Rule, but it is unique in the way that it accounts for the presence of conditional non-independence of observations, and for the presence of multiple diseases in the same patient. The purpose of this paper is to demonstrate the use of the system, and in particular, to demonstrate how these two features are clinically important in diagnostic inference, and how without them, a Bayesian approach will produce inferences that conflict with a common-sense understanding of diagnosis.