Objective: To determine the relative accuracy and sensitivity to change of 14 measures commonly used to assess arthritis activity in patients with rheumatoid arthritis (RA).
Methods: Twenty-four patients with RA were prospectively examined every 2 weeks for up to 60 weeks. At each examination, arthritis activity was assessed using 5 physician determined measures (global assessment, swollen joint count, weighted swollen joint count, tender count, weighted tender joint count), 3 patient determined measures (global assessment, pain score, duration of morning stiffness), 3 functional measures (Health Assessment Questionnaire Disability Index, grip strength, 50' walk time), and 3 laboratory measures [Westergren erythrocyte sedimentation rate (ESR), hemoglobin, platelet count]. Accuracy was determined by the degree to which changes in each measure over time were related to changes in other measures (i.e., correlational validity). Sensitivity to change was measured using standardized response means.
Results: Over the course of the study, each patient had at least 6 measures change more than 50% from their baseline values. The most highly intercorrelated measures were the physician global assessment (range of partial correlations r = 0.4-0.7), patient global assessment and pain scores (r = 0.2-0.7), the Disability Index (r = 0.3-0.7), and ESR (r = 0.2-0.4). Physician and patient global assessments, pain scores, and the Disability Index were more sensitive to change than other measures, while laboratory measures were generally less sensitive to change.
Conclusion: Based on the relative accuracy and sensitivity to change of these 14 clinical measures, the physician global assessment, a functional status questionnaire, and the patient global assessment or pain score should be the principal measures used to assess arthritis activity in patients with RA. Recognizing its limitations, the ESR could be included if a laboratory measure is needed.