Movement-specific reliability of digital technologies for measuring active shoulder range of motion: a systematic review and meta‑analysis

Accurate measurement of shoulder range of motion (ROM) is essential in clinical diagnosis, rehabilitation, and sports science. The universal goniometer remains the clinical reference standard, but its accuracy depends on operator experience. Digital technologies—smartphone applications, inertial measurement units (IMUs), and markerless motion capture systems—offer alternatives, but reliability evidence for individual movements remains fragmented. We searched PubMed, Web of Science, Embase, the Cochrane Library, and EBSCOhost (CINAHL Complete, MEDLINE, SPORTDiscus) from January 2010 to October 2025. Eligible studies reported intra- and/or inter-rater ICCs of digital technologies for active shoulder ROM. Two reviewers independently screened studies, extracted data, and assessed quality using the COSMIN Risk of Bias checklist. ICCs were transformed using Fisher’s Z method and pooled using random-effects models (REML); standard errors were derived from reported 95% CIs. Pooling required at least three studies per technology–movement–rater combination. Sensitivity analyses excluded studies rated doubtful or inadequate. An adapted GRADE approach rated the certainty of evidence. Twenty-two studies (702 participants) were included, 18 in quantitative synthesis. Pooled estimates indicated good intra- and inter-rater reliability of smartphones across the four movements. IMUs showed favorable inter-rater reliability, but intra-rater estimates for abduction and flexion had extremely high heterogeneity (I² > 99%). Reliability of markerless motion capture varied by movement, with good intra-rater reliability for internal rotation but lower values for abduction and flexion. Shoulder extension did not meet the three-study threshold for any technology and was analyzed descriptively. Most combinations showed substantial heterogeneity, and the certainty of evidence was rated very low for all pooled analyses. Smartphone applications and IMUs generally achieve good intra- and inter-rater reliability for active shoulder ROM, whereas markerless motion capture shows movement-dependent reliability. Given substantial heterogeneity, inconsistent ICC model reporting, and very low certainty of evidence, these findings are hypothesis-generating rather than definitive guidance for device selection. Standardized reporting of ICC models, absolute measurement error, and testing protocols is needed. Not applicable.