Computer-assisted differential diagnosis and management

A number of divergent techniques have been utilized in attempts to produce computerized decision aids. Medical applications have been particularly prominent areas for these applications, due in part to the inability of other methodologies to adequately address problems in medicine. The techniques used include rule-based expert systems, pattern classification, database searching, statistical analysis, decision analysis, and most recently, neural network models. In this paper, a system is described which incorporates a number of these divergent techniques to produce a comprehensive system which can be utilized as an aid for differential diagnosis as well as an instructional tool for training of medical residents. The techniques used include rule-based expert systems, pattern classification, neural networks, and image analysis. The overall purpose of the system described is to illustrate the practicality of combining these diverse technologies in a single decision making system