Objective: The Disease Activity Score (DAS) is widely used in clinical trials. A DAS of 5.1 defines the level of severe rheumatoid arthritis (RA) and is the criterion for the initiation of anti-tumor necrosis factor therapy in the UK and The Netherlands. In North America, similar rules are sometimes imposed. However, it is not known how accurately the DAS characterizes RA activity. The present study was undertaken to determine the concordance between DAS scores and physicians' assessments of RA activity, to investigate factors relating to discrepancies, and to assess the suitability of using the DAS in individual patients.
Methods: Six hundred sixty-nine RA patients were assessed using the DAS and other clinical measures. A physician's global estimate of RA activity was performed using an 11-point predefined scale and a standard definition of disease activity.
Results: The DAS and physician global assessment had substantially different distributions of values. The level of agreement (Kendall's tau-a) between DAS scores and physician global assessments was 49% (95% confidence interval 45-53%), Lin's coefficient of concordance was 0.62, and the Bland-Altman 95% limits of agreement were -3.17 and 3.99. These results suggest poor-to-moderate concordance between the 2 measures of disease activity.
Conclusion: The DAS and the physician's assessment of RA activity do not approach, value, and weight RA variables to the same extent, suggesting that RA activity is not evaluated similarly by North American physicians and with the DAS. The scales do not have acceptable levels of concordance. There is too much inherent variability in the DAS and other RA scales (e.g., the Health Assessment Questionnaire) to recommend them as sole determinants of RA activity for clinical or regulatory purposes.