Can shoulder impairments be classified from three-dimensional kinematics using inertial sensors?

Abstract

Background: Magnetic Resonance Imaging and ultrasound imaging are often used for clinical decision making to confirm diagnosis and plan treatment in shoulder problems, but this can be costly and reporting the findings can take time. The aim was to determine whether range of shoulder motion during movement tasks measured using inertial sensors are capable of accurately discriminating between patients with different shoulder problems.
Methods: Inertial sensors were used to measure three-dimensional shoulder motion during six tasks of 37 patients on the waiting list for shoulder surgery. Discriminant analysis was used to identify whether the range of motion of different tasks could classify patients with different shoulder problems.
Results: The discriminant analysis could correctly classify 91.9% of patients into one of the three diagnostic groups based. The tasks that associated a patient with a particular diagnostic group were: subacromial decompression: abduction; rotator cuff repair with tears ≤5 cm: flexion and rotator cuff repair with tears > 5 cm: combing hair, abduction and horizontal abduction-adduction.
Conclusions: The discriminant analysis showed that range of motion measured by inertial sensors can correctly classify patients and could be used as a screening tool to support surgery planning.