Reliability and diagnostic accuracy of 5 physical examination tests and combination of tests for subacromial impingement.

OBJECTIVE To investigate the reliability and diagnostic accuracy of individual tests and combination of tests for subacromial impingement syndrome (SAIS). DESIGN A prospective, blinded study design. SETTING Orthopedic surgeon shoulder clinic. PARTICIPANTS Patients with shoulder pain (n=55, mean age=40.6y). INTERVENTIONS Patients were evaluated with 5 physical examination tests for SAIS: Neer, Hawkins-Kennedy, painful arc, empty can (Jobe), and external rotation resistance tests. Surgical diagnosis was the reference standard. MAIN OUTCOME MEASURES Diagnostic accuracy calculated with a receiver operating characteristic (ROC) curve and sensitivity, specificity, positive likelihood ratio (+LR), and negative likelihood ratio (-LR). A forward stepwise binary logistic regression analysis was used to determine the best test combination for SAIS. An ROC curve analysis was also used to determine the cut point of the number of tests discriminating between the presence and absence of SAIS. Kappa coefficients and percent agreement assessed interrater reliability. RESULTS The ROC analyses revealed a significant area under the curve (AUC) (AUC=.67-.72, P<.05) for all tests, except for the Hawkins-Kennedy. The tests with a +LR greater than or equal to 2.0 were the painful arc (+LR=2.25; 95% CI, 1.33-3.81), empty can (+LR=3.90; 95% CI, 1.5-10.12), and the external rotation resistance tests (+LR=4.39; 95% CI, 1.74-11.07). Tests with -LR less than or equal to 0.50 were the painful arc (-LR=.38; 95% CI, .16-.90), external rotation resistance (-LR=.50; 95% CI, .28-.89), and Neer tests (-LR=.35; 95% CI, .12-.97). The regression analysis had no specific test combinations for confirming or ruling out SAIS. The ROC analysis was significant (AUC=.79, P=.001), with a cut point of 3 positive tests out of 5 tests. Reliability was moderate to substantial agreement (kappa=.45-.67) for the painful arc, empty can, and external rotation resistance tests and fair strength of agreement (kappa=.39-.40) for the Neer and Hawkins-Kennedy tests. CONCLUSIONS The single tests of painful arc, external rotation resistance, and Neer are useful screening tests to rule out SAIS. The single tests of painful arc, external rotation resistance, and empty can are helpful to confirm SAIS. The reliability of all tests was acceptable for clinical use. Based on reliability and diagnostic accuracy, the single tests of the painful arc, external rotation resistance, and empty can have the best overall clinical utility. The cut point of 3 or more positive of 5 tests can confirm the diagnosis of SAIS, while less than 3 positive of 5 rules out SAIS.