Reliability of classifications derived from Cyriax's resisted testing in subjects with painful shoulders and knees

Intrarater and interrater reliability.

Examine intrarater and interrater reliability of the resisted-testing component of Cyriax's selective tension testing for patients with painful shoulders and knees.

Clinicians make diagnostic and intervention decisions about lesions in contractile tissues based on resisted testing. Diagnostic and intervention decisions require reliable data gathering, especially when more than 1 physical therapist manages a patient. No studies have examined agreement of the results of the resisted tests used in selective tension testing, either within or between physical therapists, in subjects having pathology.

Subjects with pain in 1 knee (18 male, 22 female; mean age +/- SD = 31.8 +/- 9.5 years) or shoulder (21 male, 25 female; mean age +/- SD = 34.3 +/- 12.9 years) were examined twice. Referring diagnoses included ligament injuries, overuse syndromes, joint instability, and postsurgical symptoms, with some subjects seeking initial diagnosis. Two physical therapists used standardized positions to evaluate 2 knee motions or 6 shoulder and elbow motions. Evaluators applied maximal isometric manual resistance and rated the contraction as strong or weak while subjects identified the presence or absence of pain during the contraction. Evaluators did not interview the subjects and were masked to previous test results. Analyses included percentage of agreement, kappa coefficients, confidence intervals, and maximum kappa coefficients.

Intrarater kappa coefficients ranged from 0.44 to 0.82 and interrater coefficients ranged from 0.00 to 0.46. The small number of subjects who were classified as weak affected the kappa coefficients. In the intrarater condition, evaluators averaged 91% of maximum kappa for the knee and 66.5% for the shoulder. In the interrater condition, they averaged 60.4% of the maximum kappa for both the knee and the shoulder.

Based on 2 physical therapist evaluators with previous education in the selective tension system and an additional 6 hours of formal training on the methods, intrarater reliability of resisted tests was generally acceptable for the knee but not for the shoulder. Interrater reliability of these tests, however, was generally not acceptable. Results were limited by subjects who were younger and had mostly chronic conditions that were mildly to moderately severe and by the small subject samples in the analyses. Reliability might be improved by more intensive training of the evaluators and by standardizing the magnitude of the applied resistance and stabilization of the subjects.