Article Text

THU0269 Validity and Responsiveness of the Bristol Rheumatoid Arthritis Fatigue Multidimensional Questionnaire (BRAF-MDQ) in A Randomised Controlled Clinical Trial
  1. J. Kirwan1,
  2. G. Coteur2,
  3. E. Dures1,
  4. J. Nicklin1,
  5. J. Bryson3,
  6. S. Hewlett1
  1. 1Academic Rheumatology, Bristol Royal Infirmary, Bristol, United Kingdom
  2. 2UCB S.A., Brussels, Belgium
  3. 3UCB Pharma, Raleigh, United States


Background Measurement of fatigue in RA is important for patients and has been added to the RA core set of outcome measures.1 The BRAF-MDQ was developed specifically to measure RA related fatigue2, and includes a total score and Physical; Living; Cognition; and Emotion domains.

Objectives To evaluate the validity of the BRAF-MDQ in a new group of patients in the setting of a clinical trial, confirm its internal factor (domain) structure and document its sensitivity to change.

Methods Pooled data from a randomised controlled trial (RCT) in patients with active, moderate to severe RA of at least 6 months duration with inadequate response to anti-TNF therapy ( NCT01242488) were collected at baseline (BL), Week 10 (Wk10) and Wk12, including: patient and physician global assessments of disease (PtGADA & PhGADA); pain (VAS); disability (HAQ-DI), tender and swollen joint counts and the BRAF-MDQ. Spearman's correlation coefficients, Bland-Altman plots in 65 patients showing <10% change in PtGADA & PhGADA between Wk10 and Wk12 and confirmatory factor analysis tested construct validity, reproducibility and internal factor structure of the BRAF-MDQ respectively. Responsiveness was assessed by the effect sizes of changes from BL in BRAF-MDQ amongst clinical responders at Wk12 (based on 10mm improvement in PhGADA).

Results There were 219 (186 female) patients: mean age 55.5 years, disease duration12.6 years and BL DAS28 score 5.77. The proportion of missing item answers was very low (0-3%) and BRAF-MDQ total and domain scores could be calculated in 99% of questionnaires. At BL, BRAF-MDQ scores were normally distributed with <3% of total scores at the floor or ceiling. BRAF-MDQ total scores correlated with patient global, pain and HAQ at BL (r=0.49, 0.46 and 0.58) and at WK12 (r=0.63, 0.65 and 0.64) and did not correlate with tender or swollen joint counts or physician global (BL r=0.18, 0.02 and 0.16; Wk12 r=0.31, 0.21, and 0.38). Changes in the Physical and Living domains more closely related to changes in patient and physician global scores and DAS scores than did changes in Cognition and Emotional domains. Reproducibility (test-retest reliability) was high: r>0.87 for all total and domain scores and narrow Bland-Altman limits of agreement. Internal consistency was high: Cronbach's α was 0.97 for total scores and >0.82 for each domain at BL. Factor analysis confirmed a good fit for the four domains (factors) within the BRAF-MDQ: the Butler comparative fit index (CFI) was 0.92 indicating that the established structure within the BRAF-MDQ accounts well for the variation in the data. The effect sizes for BRAF-MDQ in clinical responders at Wk12 were very high (>0.7) for the Physical domain, were high (>0.5) for the Total score and the Living domain, and were moderate (>0.3) for the Cognition and Emotional domains.

Conclusions The BRAF-MDQ was completed well by study participants, related to appropriate measures of disease severity, retained its original factor structure, gave reproducible results in stable patients, and was responsive to clinical change, confirming its validity as a measure of RA fatigue.


  1. Kirwan et al. J Rheumatol 2007; 34: 1174-7.

  2. Nicklin et al. Arthritis Care and Research 2010; 62: 1559–1568.

Disclosure of Interest : None declared

DOI 10.1136/annrheumdis-2014-eular.1561

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.