An Intelligent Clinical Decision Support System for Analyzing Neuromusculoskeletal Disorders

This study presents a clinical decision support system for detecting and further analyzing neuromusculoskeletal disorders using both clinical and gait data. The system is composed of a database storing disease characteristics, symptoms and gait data of the subjects, a combined pattern classifier that processes the data and user friendly interfaces. Data is mainly obtained through Computerized Gait Analysis, which can be defined as numerical representation of the mechanical measurements of human walking patterns. The decision support system uses mainly a combined classifier to incorporate the different types of data for better accuracy. A decision tree is developed with Multilayer Perceptrons at the leaves. The system is planned to be used for various neuromusculoskeletal disorders such as Cerebral Palsy (CP), stroke, and Osteoarthritis (OA). First experiments are performed with OA. Subjects are classified into four OA-severity categories, formed in accordance with the Kellgren-Lawrence scale: "Normal", "Mild", "Moderate", and "Severe". A classification accuracy of 80% is achieved on the test set. To complete the system, a patient follow-up mechanism is also designed.