Specificity and responsiveness of patient-reported and clinician-rated outcome measures in the context of elbow surgery, comparing patients with and without rheumatoid arthritis.
Dawson J., Doll H., Boller I., Fitzpatrick R., Little C., Rees J., Carr A.
OBJECTIVES: To examine and contrast the ability of patient-reported and clinician-rated measures, reflecting different levels of specificity, to detect differences in outcomes between patients with and without rheumatoid arthritis (RA), at 6 months following elbow surgery. METHODS: One hundred and four consecutive patients/elbows self-completed the Oxford Elbow Score (OES), Disabilities of the Arm, Shoulder and Hand (DASH) and SF-36 general health questionnaires prior to elbow surgery. A surgeon assessed the standard Mayo Elbow Performance Score (MEPS) and recorded patients' diagnoses. Assessments were repeated 6 months following surgery in an outpatient clinic. Patients also completed pain "transition" and global satisfaction items. Patients who did not attend completed their assessment by post. RESULTS: Twenty-three (22%) patients had RA. These patients were more likely than other patients to have bilateral elbow problems and to have total elbow replacement (85.7% versus 10.5% P<0.001). Patients with RA had more severe preoperative elbow-specific scores, and experienced a large and significantly greater amount of change in elbow function, as measured by the OES function (P=0.002) and pain scales (P=0.013). The surgeon-assessed elbow-specific MEPS score also detected a large and significant difference between the two groups (P<0.001). However, these differences were not detected by the upper limb specific DASH, by any SF-36 general health dimensions, or by transition or satisfaction items. CONCLUSIONS: The OES performed well in assessing surgical outcomes in patients with RA. Neither the upper limb specific DASH nor the SF-36 is specific or responsive enough to warrant its exclusive use as an outcome measure for elbow surgery. LEVEL OF EVIDENCE: Level II (prospective non-randomised study).