Can OpenCap deliver valid and reliable kinematic data for motion analysis? A systematic review and three-level meta-analysis.

Abstract

Markerless motion capture systems have gained increasing interest as practical alternatives to gold-standard references systems in clinical and sports contexts. This systematic review and three-level metaanalysis aimed to evaluate the criterion validity of OpenCap and to systematically summarize the available evidence regarding its reliability. A literature search was conducted across Web of Science, PubMed, Scopus, and EBSCO. Among the 12 studies included in the systematic review, 11 provided sufficient data for the metaanalytic synthesis of criterion validity, encompassing 184 participants, from which 640 effect sizes (ES), 230 Fisher's Z values, and 1087 root mean square error (RMSE) values were obtained. OpenCap demonstrated a statistically significant, yet practically trivial effect compared to criterion devices (ES = -0.140; p = 0.021). Fisher's Z transformation indicated a good-to-excellent correlation with criterion devices (r = 0.845; p = 0.005). The pooled RMSE was 5.877°, which decreased to 5.197° after sensitivity analysis and further to 4.940° following trim-and-fill adjustment. In terms of reliability, test-retest consistency generally ranged from moderate to very good across many joint angles and tasks, although marked variability was observed in certain task-joint combinations, particularly in high-velocity movements and complex joint actions. OpenCap, as a smartphonebased markerless motion capture system, can provide valid and acceptable kinematic measurements when compared to criterion devices. However, its performance varies depending on task complexity and joint-specific demands, underscoring the need for evaluation across diverse populations, a wider range of task types, and within standardized methodological frameworks.