Objective: Fatigue is an important symptom in patients with RA. Measurement of fatigue in clinical trials and in clinical practice requires scales that are reproducible, sensitive to change and practical. This study examined the reliability and sensitivity to change of fatigue and its relative independence as an outcome measure in RA.
Methods: Successive patients referred to the rheumatology clinic at St Vincent's University Hospital and Our Lady's Hospice were evaluated. Clinical assessments were undertaken at baseline and 3 months after commencing TNF-alpha blockade. Fatigue was measured using an 11-point numeric rating scale (NRS). Sensitivity to change when compared with current core set outcome measures was determined by calculation of the standardized response mean (SRM). Multiple regression analysis was employed to determine the independent variance of fatigue scores relative to the core set.
Results: Forty-nine patients were evaluated. At baseline, mean (s.d.) fatigue scores were 6.7 +/- 2.1. At 3 months, fatigue scores had fallen to 4.3 +/- 2.6 (P < 0.001). Test-retest intraclass correlation coefficient for the NRS was 0.79 (P < 0.008). Fatigue was ranked third for relative sensitivity to change as shown by SRM: pain, 1.37; tender joint count (TJC), 1.09; fatigue, 0.92; swollen joint count (SJC), 0.86; HAQ, 0.82; CRP, 0.69; and patient global health (GH), 0.25. The relative independent variance in fatigue of 22% was higher than that of the core set: TJC, 20%; pain, 19%; SJC, 16%; GH, 8%; HAQ, 7%; and CRP, 8%.
Conclusions: This study demonstrates that measures of fatigue are reliable and sensitive to change, and should be considered for inclusion as a core outcome measure in RA.