POS0266-HPR PATIENT-PERCEIVED ASPECTS OF RA FLARE EVOLVE OVER TIME, AS REFLECTED BY THE FLARE-RA QUESTIONNAIRE: POST-HOC ANALYSIS OF TAPERA

Background: Flares are common in rheumatoid arthritis (RA). While flares negatively impact clinical and patient-reported outcomes, different aspects of disease activity may constitute a flare to patients. Flare Assessment in RA (FLARE-RA) is a patient-reported questionnaire aiming to detect active or recent RA flares (1). During its validation, arthritis and general health subscales were identified and the instrument was adapted from 13 questions (1-6 Likert scale) to 11 questions (0-10). Objectives: To investigate which patient-perceived aspects of flare are assessed by FLARE-RA in the context of a TNFi-tapering trial, using exploratory factor analysis (EFA). Methods: Patients with RA in DAS28CRP/ESR-remission (≥6 months) and treated with etanercept 50 mg weekly (≥1 year) were included in the 12-month TapERA (Tapering Etanercept in Rheumatoid Arthritis) trial between 2012 and 2014. Participants completed 3-monthly FLARE-RA questionnaires. The first and final follow-up visits (M3 & M12) were analyzed. Missing data were imputed with multiple imputation (n = 10). Sampling adequacy was assessed by Kaiser-Meyer-Olkin (KMO) and correlations between variables were evaluated with Bartlett’s sphericity test. Spearman correlation matrices were constructed in each of the 10 imputed datasets. The pooled matrix was then analyzed by EFA with principal component extraction and promax-rotation. EFA aims to identify clusters of questions that elicit similar responses because of association with the same underlying latent (not observable) constructs/factors. Results: Sixty-six patients (68% female) with a mean age of 55 years (SD 13) and a mean disease duration of 14.8 years (SD 9) completed a total of 330 FLARE-RA questionnaires. Sampling adequacy was acceptable (KMO = 0.94) and correlation between items was sufficient for factor analysis (p Table 1 shows the results of EFA in TapERA compared to the validation study (1). Factor loadings indicate how strongly each item correlates with its underlying factor. EFA of the full 13-item FLARE-RA at M3 revealed 3 factors: Arthritis, General health and a Medication factor relating to management of flare. The Arthritis factor explained the largest proportion of variance (31%). EFA at M12 showed the same underlying factors, but a less robust factor structure (cross-loadings >0.3) and a larger proportion of variance explained by the General health factor (33%). Conclusion: FLARE-RA assessed similar patient-perceived aspects of RA flare within the context of a TNFi-tapering trial when compared to the validation study, including a Medication factor reflecting use of both glucocorticoids and analgesics. This underlines the usefulness of FLARE-RA in providing a multi-faceted view of patients’ conceptions of RA flare. However, these aspects and their relative importance do seem to evolve over time. Further research is needed to assess if this is due to the influence of time or specific to the studied population/tapering setting. References: [1]Fautrel B, et al. Validation of FLARE-RA, a Self-Administered Tool to Detect Recent or Current Rheumatoid Arthritis Flare. Arthritis Rheumatol. 2017;69(2):309–19 Disclosure of Interests: None declared