OP0239 PSYCHOMETRIC EVALUATION OF FUNCTIONAL ASSESSMENT OF CHRONIC ILLNESS THERAPY-FATIGUE IN PATIENTS WITH ACTIVE ANKYLOSING SPONDYLITIS

Abstract

Background:Due to its impact on their health-related quality of life (QoL), fatigue is considered a core domain of disease assessment in patients (pts) with ankylosing spondylitis (AS). Psychometric data analyses in pts with rheumatoid and psoriatic arthritis have previously demonstrated that the Functional Assessment of Chronic Illness Therapy-Fatigue (FACIT-F) is a reliable and valid measure in these pt populations,1,2 but there are limited data supporting the psychometric validity and reliability of the FACIT-F scale in adult pts with active AS.Objectives:To further evaluate the psychometric properties of the FACIT-F scale in adult pts with active AS.Methods:This post hoc analysis used data from the Phase (P)2 (NCT01786668, Study 1) and P3 (NCT03502616, Study 2) studies of tofacitinib in pts with active AS. Second-order confirmatory analysis (CFA) evaluated the measurement model of FACIT-F with two domains (Experience and Impact) and the total score at baseline (BL) and the studies’ respective primary analysis time points, Week (W)12 (Study 1) or W16 (Study 2). At the same time points, internal consistency reliability was examined by Cronbach’s Coefficient alpha (α), and convergent validity was assessed through correlations of FACIT-F domain and total scores with a set of pt-reported outcomes (PROs): Pt Global Assessment of Disease Activity (PtGA), total back pain/nocturnal spinal pain due to AS, Short Form-36 Health Survey (SF-36v2), Bath AS Functional Index, Bath AS Disease Activity Index and ASQoL. Test-retest reliability was assessed by calculating Intraclass Correlation Coefficients (ICCs). The known-groups validity assessment was derived from an anchor-based repeated measures longitudinal model, with PtGA scores as the anchor (PtGA represented pt state from ‘no disease activity’ to ‘very active disease’ [PtGA=0 and 10, respectively]), and FACIT-F domain/total scores as the outcome. Estimations of ability to detect change and meaningful within-pt change (MWPC) were derived from anchor-based repeated measures longitudinal models, with change from BL in PtGA scores as the anchor and change from BL in FACIT-F domain/total scores as the outcome.Results:The CFA model fit the data well (Bentler’s Comparative Index ≥0.92 at BL, W12 [Study 1] and W16 [Study 2]), supporting the measurement model of the FACIT-F scale in pts with AS. Across time points in each study, the FACIT-F domain and total scores demonstrated excellent internal consistency (Cronbach’s Coefficient α ≥0.88), and correlations between FACIT-F domain/total scores, and all PROs assessed generally exceeded 0.40. Generally, in both studies, the largest correlations (0.62–0.85) were between FACIT-F domain/total scores and the SF-36v2 vitality domain score and ASQoL. Test-retest reliability was acceptable for all FACIT-F domain/total scores (ICC ranged from 0.75–0.89 in both studies). An approximately linear relationship was observed between FACIT-F total scores and PtGA scores in both studies (Figure), as well as for Experience and Impact domain scores. The differences in FACIT-F domain/total scores were large and statistically significant between ‘no disease activity’ and ‘very active disease’ pt groups (standardised effect size ≥1.17), supporting known-groups validity. Ability to detect change was evidenced by an approximately linear relationship between changes in PtGA and FACIT-F domain/total scores in both studies; when pts experienced a change in PtGA, values for FACIT-F domain/total scores changed accordingly. Conservatively, MWPC was estimated as 6.3 for FACIT-F total score, and 2.8 and 3.6 for FACIT-F Experience and Impact domain scores, respectively.Conclusion:This quantitative analysis of data from two clinical studies of tofacitinib demonstrates the validity and reliability of the FACIT-F scale in adult pts with active AS. Therefore, these findings support the use of FACIT-F in AS studies.