REVIEW OF METHODS FOR DESIGNING CLINICAL DECISION SUPPORT SYSTEMS

Abstract

Background. This study is devoted to the review of methods for designing medical decision support systems. Materials and methods. The main source of information was the Russian scientific electronic library elibrary and international databases of scientific articles. There were analyzed the works of Russian and foreign specialists who conduct research and development in the field of clinical decision support systems (СDSS). Results. As a result of the study, the method that is most effective in the development of a СDSS was identified on the base of the analyzed sources: fuzzy clustering. Moreover, the inefficiency of such methods as: k-means methods, Fuzzy C-Means and Gustafson-Kessel clustering is justified. The study revealed a number of drawbacks faced by developers when designing a СDSS. This study will allow us to avoid the problems associated with the development of the СDSS in the future and take into account the experience of using various AI methods in the design of the СDSS. Conclusions. Conclusions are drawn about the effectiveness of using the fuzzy clustering method and the rationality of using other methods (the k-means method, Fuzzy C-Means and Gustafson-Kessel clustering). On the basis of the identified problems in the development of the CDSS, an idea was formed about the possible problems of the existing CDSS.

Keywords

• clinical decision support system
• fuzzy logic
• fuzzy clustering
• expert systems
• artificial intelligence