Objective: To assess the responsiveness of the Rheumatoid Arthritis Disease Activity Index (RADAI) to increases in disease activity, using the occurrence of a flare of disease activity as an external standard.
Methods: A post hoc analysis was performed on data from a randomized, double-blind, controlled trial of methotrexate versus type II collagen in 92 patients with rheumatoid arthritis (RA). Responsiveness was analyzed by 1) correlating change in the RADAI score with change in the Disease Activity Score (DAS28), 2) determining the RADAI's ability to detect a disease flare by plotting a receiver operating characteristic (ROC) curve, and 3) using a responsiveness statistic, the standardized effect size (SES). The contribution of the single RADAI items to the change in total RADAI score was analyzed by the item score change in absolute value, the item responsiveness by the standardized response mean, and the correlation of item score change with total RADAI score change by Cronbach's alpha.
Results: Changes in the RADAI score correlated strongly with changes in the DAS28 (R2 = 0.70, P < 0.0001). The area under the ROC curve for the RADAI was 0.88 (95% confidence interval 0.78-0.95), which was similar to that for the DAS28. The SES for the RADAI was 1.56, which was also similar to that for the DAS28. The RADAI items of past global disease activity and morning stiffness contributed least to the total score change.
Conclusion: This study provides evidence that the RADAI is sensitive to relevant increases in disease activity in RA patients. The RADAI may complement clinical measures in clinical studies, or may be used as a proxy for disease activity in epidemiologic studies.