Clinical reliability of shoulder function assessment in patients with rheumatoid arthritis.

A model for functional assessment and a dynamic test of the shoulder joint were designed and tested for normal variation and clinical inter- and intra-rater reliability. The functional assessments, which covered four common shoulder functions, were compared with assessments of pain, recordings of active motion range and the results of a Health Assessment Questionnaire, in eight patients with rheumatoid arthritis according to the ARA criteria. Intra-rater reliability was satisfactory for all four functions and inter-rater reliability was satisfactory for the hand-raising and hand-to-opposite-shoulder functions but less so for hand-behind-back and hand-to-neck. A second test-retest study in 15 patients, with a slight modification of one of the functional tests, confirmed the results and improved the reliability of the modified test. The reliability of the dynamic test and of the active motion range measurement was less satisfactory or not satisfactory. No significant correlation was found between shoulder functional assessment and the Fries index, but there were positive significant correlations between active motion range and shoulder functions. It is concluded that the method presented for evaluating shoulder functions has satisfactory reliability and in the first test-retest study was more reliable than conventional motion range measurement of the shoulder joint.