Clinical Reliability and Diagnostic Accuracy of Visual Scapulohumeral Movement Evaluation in Detecting Patients With Shoulder Impairment

Document Type


Publication Date



BACKGROUND: Clinical investigation of shoulder injuries commonly utilizes visual evaluation of scapular movement to determine if abnormal or asymmetrical movements are related to the injury. To date, the intrarater reliability and diagnostic accuracy of visual evaluation of scapular movement among physical therapists are not known. PURPOSE: The aims of this study were to determine the clinical reliability and diagnostic accuracy of physical therapists visual evaluation of scapulohumeral movements when used to diagnose shoulder impairment. STUDY DESIGN: University based laboratory and an internet based survey. METHODS: Thirty-three physical therapists and 12 patient participants participated in this study. Reliability was measured as percent agreement and using the free marginal kappa statistic (κ) and Cronbach's alpha (α) for interrater and intrarater reliability respectively. Diagnostic accuracy variables such as sensitivity, specificity, likelihood ratios were calculated from contingency table analysis. RESULTS: Visual evaluation yielded the following (95% CI): diagnostic accuracy 49.5%, specificity 60% (56 - 64), and sensitivity 35% (29 - 41), positive and negative likelihood ratios were 0.87 (0.66 - 1.14) and 1.09 (0.92 - 1.27) respectively. Percent agreements of evaluation findings between sessions for static and dynamic symmetry were 69% and 68%, respectively. The alpha statistics for static and dynamic symmetry were both 0.51. Percentage agreement in determining the injured shoulder was 59%, with an alpha statistic of 0.35. CONCLUSION: Visual evaluation of scapular movements, without additional clinical information, demonstrated a poor to fair reliability and poor to fair diagnostic accuracy. CLINICAL RELEVANCE: The clinical utility of the use of isolated visual scapular evaluation is cautioned. More reliable and valid objective measures are needed for diagnosing shoulder impairment. LEVEL OF EVIDENCE: 2b, Exploratory cohort study.

This document is currently not available here.