Decision support system for diagnosing Rheumatic-Musculoskeletal Disease using fuzzy cognitive map technique

Abstract Rheumatic-Musculoskeletal Disease (RMD) is a leading cause of disability worldwide. It causes inflammation of the connecting body structures, affects joints, tendons, ligaments, bones, and muscles, and is responsible for thousands of deaths per year worldwide. It is prevalent in Africa. In Nigeria, RMD constitutes 10–15% of rheumatology cases in most clinics, with a ratio of 2.4:1 (Female:Male), killing over 1652 people per year. The similarity of its symptoms with other diseases often cause misdiagnosis at an early stage among infected patients, and existing computational techniques used for diagnosis cannot address the confusability of the symptoms. Hence, there is an inability to determine the causality of symptoms. Moreover, a dearth of Rheumatologists prevents many RMD infected persons from obtaining an early and accurate diagnosis. Therefore, the need arises to develop a decision support system (DSS) for diagnosing RMD, such as by using a fuzzy cognitive map (FCM) technique. Our study focuses on the development and implementation of FCM-based DSS for RMD diagnosis (RMD-FCMDSS). RMD-FCMDSS serves as a software tool complementing physician decision-making. Our evaluation of RMD-FCMDSS suggests that it has an improved diagnostic value as compared to earlier traditional and conventional medical methods. Performance results indicate an 87% accuracy, 90% sensitivity, and 80% specificity. RMD-FCMDSS was tested on limited data due to a dearth of Rheumatologists in Nigeria and a limited population of rheumatic patients interacted with; thus in future work we shall need to increase the test sample. Hybridization of Fuzzy-Logic and Cognitive Mapping techniques helps to guarantee accuracy, specificity, sensitivity, and also speed diagnosis.