A Decision Support System for the Diagnosis of Breast Cancer Using Fuzzy Logic and Case Based Reasoning

Abstract

Background The use of Clinical decision support systems helps improve medical diagnosis and also minimize diagnostic errors. Older diagnosis systems have proved cumbersome to use and avail limited success in identifying the correct diagnosis in complicated cases like breast cancer at early stages. Objectives The objectives are to design, develop, and assess a clinical decision support system that offers a suite of services for early detection of breast cancer. Methods The CDSS prototype was developed based on cased based reasoning and fuzzy logic artificial intelligence technologies. The functionalities of the CDSS were developed iteratively through requirement- development cycles using enterprise-grade software-engineering methodology. Within each cycle, the acquisition of clinical knowledge was done by a health informatics engineer and a team of oncologists. The research involved 50 case records at St. Francis Mission hospital, Kasarani whose final diagnosis had already been ascertained as breast cancer. The patient symptoms from the records were manually entered in to the system so as to determine how often the CDSS would suggest the correct diagnosis. In addition to this, the speed at which data entry could be done and results recovered were evaluated. Results The clinical decision support system suggested the correct diagnosis in 48 of the 50 cases (96%). Manual data entry took less than a minute while results were provided within 2–3 seconds. Conclusions The CDSS prototype suggested the correct diagnosis in almost all of these complex cases during testing and evaluation. The prototype therefore merits evaluation in more natural settings and clinical practice.