Smartphone Video-Based Cervical Range of Motion Assessment in Pediatric Torticollis: A Validity and Reliability Study

Abstract

Background: Accurate cervical range-of-motion assessment is important for diagnosing torticollis, planning rehabilitation, and monitoring clinical change. Smartphone video analysis may provide an accessible alternative to conventional measurement, but its measurement properties require evaluation. Objective: To determine the concurrent validity, intra-rater reliability, inter-rater reliability, and measurement error of smartphone video-based cervical range-of-motion assessment in children with torticollis. Methods: This cross-sectional measurement-property study included 30 children aged 6 months to 5 years with congenital or acquired torticollis. Cervical rotation, lateral flexion, flexion, and extension were assessed using smartphone-recorded video analysis and conventional goniometry. Pearson correlation coefficients evaluated associations between methods. Two assessors analysed the recordings for inter-rater reliability, and one assessor repeated the analysis after one week for intra-rater reliability. Intraclass correlation coefficients, standard errors of measurement, and 95% minimal detectable changes were calculated. Results: Smartphone measurements were 0.7°–0.8° lower than goniometric measurements, with no reported significant differences between methods. Correlations ranged from 0.88 to 0.95 (all p<0.001). Intra-rater ICCs ranged from 0.92 to 0.96 and inter-rater ICCs from 0.90 to 0.94. SEM values ranged from 2.1° to 2.5°, and MDC95 values ranged from 5.8° to 6.9°. Conclusion: Smartphone video analysis demonstrated strong associations with goniometry and high reliability for repeated analysis of stored recordings, although formal agreement analysis remains necessary before clinical interchangeability can be assumed