Accounting of Statistical and Expert Information in Medical Decision Support Systems

Abstract

The work is devoted to solving the actual scientific and technical problem of building a decision support system based on the implementation of the developed model of the diagnostic decision rule by means of modern information technologies. Based on the analysis of the methods used to build a diagnostic decision rule (DRU) in diagnostic decision support systems in medicine, the components of a combined DRU are proposed, which express two approaches to the formulation of a diagnostic conclusion: an objective component, which is based on the analysis of educational samples, and the subjective component, which is based on expert information about the structure of symptom complexes. The purpose of the study is the synthesis of the combined DRU based on the method of comparison with the prototype, which takes into account both the objective and subjective components of the diagnosis process. The paper developed a mathematical model of the combined DRU and substantiated the choice of its components. As an objective component, the method of comparison with a prototype was selected, in which the diagnosed conditions (a list of diagnoses in a given subject area of medicine) are represented by their prototypes in the space of signs. As a prototype of each class, the geometric center of the class grouping is calculated. Expert information on the structure of symptom complexes is formalized by presenting the symptom complex of each disease with numerical intervals of linguistic variables. Variants of taking into account expert evaluations about the structure of symptom complexes when calculating the coordinates of class prototypes (collective decision rules, weighting and summarization of evaluations) are considered. On the basis of the developed mathematical model of the combined DRU, the decision support system was designed and a comprehensive check of the developed system was performed on real medical data, which confirmed the effectiveness of the system.