Intra-observer reliability of commonly used outcome measures in rheumatoid arthritis

Objective To determine the reliability of some commonly used outcome measures in patients with rheumatoid arthritis. Methods We studied 22 consecutive patients with rheumatoid arthritis enrolled in a clinical trial in a tertiary care center. The study design consisted of a test-retest, in which the same rheumatologist evaluated all of the patients twice, with an interval between evaluations of 90 to 120 minutes. Statistical analysis of the data consisted of calculation of the weighted Kappa (kappa w) and the intraclass correlation coefficient (ICC). Results For the Ritchie articular index, kappa w = 0.83, ICC = 0.49 p < 0.0001; tender joint count, kappa w = 0.82, ICC = 0.49, p < 0.0001; physician 's global assessment, kappa w = 0.79, ICC = 0.48, p < 0.0001; disease activity score, kappa w = 0.79, ICC = 0.49 p < 0.0001; utilities, kappa W = 0.71, ICC = 0.48, p < 0.0001; swollen joint count, kappa w = 0.7, ICC = 0.47 p < 0.0001; patient's global assessment, kappa w = 0.58, ICC = 0.44 p < 0.0001; pain kappa w = 0.45, ICC = 0.41, p < 0.0001. Conclusions The reliability of most of the outcome measures was good. It was higher for those measurements evaluated by a rheumatologist and for the composite indexes. Those requiring patient participation need to be improved.