Validation of FLARE-RA, a Self-Administered Tool to Detect Recent or Current Rheumatoid Arthritis Flare

Objective. To validate the measurement properties and the detection performance of the FLARE-RA questionnaire in a longitudinal prospective study. Methods. To validate the FLARE-RA self-administered questionnaire, we conducted a prospective trial in rheumatoid arthritis (RA) patients to document: 1) content and construct validity by factor analysis, convergent validity by Pearson's correlation with routine assessment of patient index data (Routine Assessment of Patient Index Data 3 [RAPID-3] questionnaire), RA Impact of Disease (RAID) score, Disease Activity Score in 28 joints (DAS28), and Health Assessment Questionnaire (HAQ), 2) reliability (intraclass correlation coefficient [ICC] and Bland-Altman plot), and 3) feasibility of use. Patients were examined and questionnaires were collected at baseline and 3 months, and every week in between for RAPID-3. Results. We recruited 138 patients from 13 centers: 81.9% women, mean age 57.4 years, mean DAS28 2.9, mean C-reactive protein level 6.2 mg/liter, 84.4% rheumatoid factor positive, 78.0% anti-citrullinated protein antibody positive, and 78.8% with erosive disease. At baseline, the mean +/- SD FLARE-RA score was 2.3 +/- 2.3. The content and construct validity of FLARE-RA was good. A substantial floor effect, but no ceiling effect, was observed. Principal components analysis revealed 1 domain disentangled in 2 subdomains: physical and emotional. The FLARE-RA total score was correlated with the DAS28 (r=0.63, P<0.001), RAID (r=0.80, P<0.001), RAPID-3 (r=0.77, P<0.001), and HAQ (r=0.53, P<0.001). The ICC for reliability was 0.94 (95% confidence interval 0.92-0.96). Conclusion. The FLARE-RA self-administered questionnaire represents a valid and valuable instrument to detect RA flare between visits to the physician.