Measurement properties of an artificial intelligence video-based analysis IOS app for active range of motion and home-based use in population with different age groups

Background This study aimed to evaluate the validity and reliability of Artificial Intelligence (AI) video-based analysis (My Jump-Lab App) for self-reported active joint range of motion (AROM) of the upper and lower limbs and the influence of the age groups on the reliability of the app.Methods A cross-sectional study was conducted at the University of Valencia. A total of 65 healthy subjects with different age groups were included: younger group (≤35 years) and middle-aged group (≥55 years). For convergent validity, Pearson’s correlation coefficient (r) and limits of agreement (LoA) were used for the two devices. The intraclass correlation coefficient (ICC), standard error of measurement (SEM), and minimal detectable change (MDC) were analysed for reliability using home-based videos.Results For convergent validity, My Jump-Lab showed a good to excellent correlation (r = 0.638–0.958) and an acceptable LoA (<9.84%) for hip flexion and external rotation, knee flexion, and shoulder flexion, extension, and abduction in the younger group. In the older group, good/excellent correlations (ρ = 0.694 to 0.928) and acceptable LoA (<9.00%) were reported for all movements except hip external-rotation. For reliability, the older and younger groups showed moderate to good ICCs (0.575 to 0.723) and a moderate SEM for hip internal rotation and shoulder abduction and flexion, hip adduction, shoulder extension, elbow flexion, and ankle dorsiflexion.Conclusions IOS My Jump-Lab app is a valid and reliable tool for measuring AROM for 7 of the 13 movements analysed. AI video-based analysis may be suitable for home-based measurements across age groups, thereby enabling a low-cost digital tool for telerehabilitation.