Assessing the reliability and responsiveness of 5 shoulder questionnaires

The purpose of this study was to compare the test-retest reliability and re sponsiveness of 5 different shoulder questionnaires in a sample of patients with shoulder pain. Ninety-nine patients completed the Following shoulder questionnaires on 2 occasions: Subjective Shoulder Rating Scale, Simple Sho ulder Test, Modified-American Shoulder and Elbow Surgeons Form, Shoulder Se verity Index, and the Shoulder Pain and Disability Index. The Short Form-36 was also included. Test-retest reliability was assessed with intra-class c orrelation coefficients. Standardized response means were calculated to ass ess responsiveness. This procedure was done on 33 subjects who underwent ro tator cuff surgery or total shoulder arthroplasty and who believed that the y had improved between testings. All questionnaires had acceptable reliabil ity (coefficients >0.75) except the Subjective Shoulder Rating Scale (coeff icient = 0.71) and were more responsive (0.65 < Standardized response mean < 1.23) than the Short Form-36 (0.08 < Standardized response mean < 0.43) e xcept For pain (0.91). In this longitudinal study a direct comparison of 5 shoulder questionnaires was carried out. We found the Subjective Shoulder R ating Scale to have lower reliability and responsiveness. The other 4 quest ionnaires including 1 developed to measure whole extremity Function (Modifi ed American Shoulder and Elbow Surgeons Form) rather than the shoulder only were comparable with good reliability and responsiveness. The results indi cate that the shoulder questionnaires were more sensitive to change in pati ents with shoulder pain than the generic questionnaire (Short Form-36), and both types of questionnaires should be used in outcome evaluations.