Peer Review #2 of "A novel IMU-based clinical assessment protocol for Axial Spondyloarthritis: a protocol validation study (v0.1)"

Abstract

Clinical assessment of spinal impairment in Axial Spondyloarthritis is currently performed using the Bath Ankylosing Spondylitis Metrological Index (BASMI). Despite being appreciated for its simplicity, the BASMI index lacks sensitivity and specificity of spinal changes, demonstrating poor association with radiographical range of motion (ROM).Inertial measurement units (IMUs) have shown promising results as a cost-effective method to quantitatively examine movement of the human body, however errors due to sensor angular drift have limited their application to a clinical space. Therefore, this paper presents a wearable sensor (IMU) protocol that facilitates unrestrained orientation measurements in space while limiting sensor angular drift through a novel constraintbased approach. Eleven healthy male participants performed five BASMI-inspired functional movements where spinal ROM and continuous kinematics were calculated for five spine segments and four spinal joint levels (lumbar, lower thoracic, upper thoracic and cervical). A Bland-Altman analysis was used to assess the level of agreement on range of motion measurements, whilst intraclass correlation coefficient (ICC), standardised error measurement (SEM), and minimum detectable change (MDC) to assess relative and absolute reliability. Continuous kinematics error was investigated through root mean square error (RMSE), maximum absolute error (MAE) and Spearman correlation coefficient (ρ). The overall error in the measurement of continuous kinematic measures was low in both the sagittal (RMSE = 2.1°), and frontal plane (RMSE = 2.3°). ROM limits of agreement (LoA) and minimum detectable change were excellent for the sagittal plane (maximum value LoA 1.9° and MDC 2.4°) and fair for lateral flexion (overall value LoA 4.8° and MDC 5.7°). The reliability analysis showed excellent level of agreement (ICC > 0.9) for both segment and joint ROM across all movements. The results from this study demonstrated better or equivalent accuracy than previous studies and were considered acceptable for