Design considerations of a medical expert system for differential diagnosis of low back pain (Eslbp )

Abstract

Low back pain is a communal musculoskeletal ailment that deprives many individuals worldwide of doing their daily and normal activities. With the absence of external biomarkers, most of the symptoms of low back pain diseases seem similar, making the diagnosis process quite difficult. Application of artificial intelligence is beneficial in this regard. The paper deals with the design of an efficient knowledge base and a reliable inference engine for a medical expert system for treatment of low back pain. As many low back pain diseases have common clinical signs, consideration of only the dissimilar patterns of the diseases in the design of knowledge base would surely overcome the problem of processing the same symptoms over and over. The acquired knowledge is represented with a discernibility matrix that captures only the disparities among low back pain diseases. An inference mechanism has also been proposed, which uses the discernibility matrix for offering the diagnostic conclusions in a timely manner. The designed system has been tested with patient records empirically selected from the repository of ESI Hospital Sealdah, Kolkata. The test results show that the diagnostic inference generated by the proposed inference engine conforms to the conclusions made by the expert physicians. © Springer Nature Singapore Pte Ltd. 2020.