Therapy selection in an expert medical consultation system for ocular herpes simplex

This paper describes a general scheme for selecting therapies which can be used in expert medical consultation systems. The scheme consists of a topological sorting procedure within a general production rule representation. The procedure is used to choose among competing therapies on the basis of precedence rules. This approach provides a natural way in which therapy choices can be automatically explained. The precedence rule has the capability to summarize a great many facts elegantly and to insure that the conclusions reached will be mutually exclusive. Precedence rules are computationally more efficient in both storage and time than an equivalent set of production rules. As the number of competing therapies increases this computational advantage increases. An expert consultation system for diagnosis and therapy planning of ocular herpes simplex has been implemented using this approach and examples of the system's output on actual cases of this disease are given.