Exploring Priority Parameters in Physiotherapist Decision Models for Tailoring Robot-Aided Rehabilitation

Abstract

Robot-aided rehabilitation is a promising method for treating both musculoskeletal and neuromuscular disorders, as the robotic systems can be equipped with sensors capable of deciphering the user needs, thus enabling therapists to tailor the rehabilitation program accordingly. These robotic platforms often rely on specific rules or Artificial Intelligence algorithms to reduce the therapists’ workload by automating parameter tuning. However, it is crucial to understand the reasoning behind a therapist’s decision to modify tunable parameters. This understanding enables the algorithm to better adapt to the patient’s individual needs. The selection of parameter settings for robot adaptation depends on the therapists’ experience, practice, goals, and strategies. To assess the consistency in decision-making among physiotherapists and determine if the choices align across different therapists with respect to the parameters, this study aims to propose a methodology for exploring priority parameters. Moreover, it was validated in a real clinical setting. Eight orthopedic patients, along with their physiotherapists, were enrolled in the experimental study aimed at gathering the clinicians’ decisions regarding modifications to the Assistance Level provided to each patient based on motor performance, physiological activity, and subjective scores. Two statistical analyses revealed that the therapists’ choices are related to both the patients’ subjective perceptions of exertion and pain, as well as to measurable objective parameters derived from physical and cognitive workload, as measured by wearable sensors.