Intelligent Diagnosis and Predictive Rehabilitation Assessment of Chronic Ankle Instability Using Shoe-integrated Sensor System

Ankle sprains, the leading injuries in the emergency department that affect people worldwide, often leading to chronic ankle instability (CAI) characterized by recurring pain and weakness. However, challenges are presented in accurately identifying CAI-related abnormal gait patterns and assessing rehabilitation effects. Traditional plantar pressure systems lack portability and can only be used in limited specific actions, while a few early proposed portable systems have demonstrated insufficient accuracy. Besides, no previous studies have yet focused on assessing rehabilitation effects, which is crucial to providing the treatment selection and rehabilitation evaluation of CAI. Considering this, we propose a novel approach to improve the diagnostic process for CAI. A Shoe-Integrated Sensor System (SISS) which can accurately capture gait data during various activities was implemented. We collected and processed level walking data from 80 CAI patients diagnosed by professional experts and 42 healthy individuals using the system, including feature extraction and filtering algorithms. An artificial intelligence diagnosis was applied to the data, achieving a classification accuracy of 93.39% and an area under the curve (AUC) of 0.959, satisfying the clinical requirements for accuracy. Furthermore, a novel methodology was proposed to assess the level of patient rehabilitation. The validation results of rehabilitation status prediction demonstrated highly consistent results with doctors' diagnoses. Due to the significant impact of gait data in assisting the diagnosis of various neurological and musculoskeletal diseases that result in gait abnormalities, the proposed system can also be extended and utilized in other similar medical fields for diagnosing and real-time monitoring, promoting the development of smart healthcare.