Interobserver and intraobserver reliability of the measurement of shoulder internal rotation by vertebral level

Internal rotation is commonly measured as the vertebral level reached by the fully extended thumb. The purpose of this study was to evaluate interobserver and intraobserver reliability with the use of this method. Three male subjects were used for internal rotation measurement. Eleven orthopaedic surgeons and 2 physical therapists served as examiners. Each subject had a radiographic marker placed at a random vertebral level, and the subject's extended thumb was placed at this marker. All examiners then independently measured internal rotation based on vertebral level. To assess intraobserver reliability, this process was repeated twice. After all measurements were completed, an anterior-posterior radiograph of each subject was obtained to define the vertebral level of the marker. This process was repeated 2 additional times with the marker and subject's thumb positioned at different levels than in the previous examination. Intraclass correlation coefficients were calculated to determine reliability. Results demonstrated poor interobserver reliability and reasonable intraobserver reliability. The mean clinical measurement deviated from the mean actual measurement by 1 vertebral level. Despite being the standard method in which shoulder internal rotation is measured, measurement of internal rotation by vertebral level is not readily reproducible between observers.