Lattice-Based Fuzzy Medical Expert System for Management of Low Back Pain: A Preliminary Design

Low back pain (LBP) appears to be the foremost contributor to the years lived with disability globally, depriving many individuals across the nations of leading daily activities. Diagnosing LBP is quite challenging as it requires dealing with several clinical variables having no precisely quantified values. With the goal to design a reliable medical expert system for assessment and management of LBP, the research offers a lattice-based scheme for efficient representation of relevant medical knowledge, proposes a suitable methodology for design of a fuzzy knowledge base that can handle imprecision in knowledge, and derives a fuzzy inference system. A modular approach is taken to construct the fuzzy knowledge base, where each module is able to capture interrelated clinical knowledge about medical history, findings of physical examinations, and pathological investigation reports. The fuzzy inference system is designed based on the Mamdani method. For fuzzification, the design adopts triangular membership function; for defuzzification, the centroid of area technique is used. With the relevant medical knowledge being acquired from the expert physicians, a working prototype of the system has been built. The prototype has been successfully tested with some LBP patient records available at the ESI Hospital, Sealdah. The designed prototype is found to be clinically acceptable among the expert and non-expert physicians.