Validity and Reliability of a 2-Dimensional Artificial Intelligence Camera Software Application to Assess Hand, Wrist, and Forearm Range of Motion.

PURPOSE: The purpose of this study was to investigate the psychometric properties of a 2-dimensional artificial intelligence (AI) camera software application (app) for range of motion measurements in the upper extremity. We hypothesized that the 2-dimensional AI-camera app would have excellent inter-rater reliability and high concurrent validity with manual goniometry. METHODS: Fifty patients were prospectively recruited. Patients aged 18-75 years with upper-extremity pathology and no immobilization were eligible for inclusion. For each patient, 27 range of motion measurements were performed, beginning with manual goniometer measurement by a fellowship-trained hand surgeon (MD) and then the 2-dimensional AI-camera app. The primary outcomes were the inter-rater reliability of the measurements (intraclass correlation coefficient [ICC]) and concurrent validity (proportion of measurements falling within the clinical agreement parameters) between raters. The measurement time was recorded. RESULTS: The mean patient age was 49.2 years. The overall inter-rater reliability was excellent (ICC 0.99; 95% CI 0.98 to 0.99). Forearm pronation, forearm supination, and wrist extension were the only range of motion measurements without excellent inter-rater reliability (ICC < 0.9). Bland-Altman analysis showed that 81% (22/27) of motions had high concurrent validity, with low concurrent validity seen with forearm supination, wrist extension, long finger total active motion (TAM), ring finger TAM, and small finger TAM. The mean time to completion of all measurements using the 2-dimensional AI-camera app was 60 seconds, compared to 273 seconds by the MD. CONCLUSIONS: The 2-dimensional AI-camera app demonstrated excellent inter-rater reliability and high concurrent validity for the majority of motions; however, the reliability and validity were both lower for forearm supination and wrist extension. CLINICAL RELEVANCE: In upper extremities with pathologic conditions, the overall reliability and validity of the 2-dimensional AI-camera app render it potentially useful to providers in the clinic or as a remote monitoring tool, with limitations in the consistency and accuracy of certain motions.