Background The interest in the use of patient-reported outcome measures in clinical practice and in clinical trials is increasing. The EULAR/ACR collaborative recommendations 2008 for rheumatoid arthritis (RA) included patient global assessment (PaGl), pain and fatigue (FTG) as key patient-reported outcome measures (ref.). For the clinician, it is essential to know whether an observed change in any disease measure reflects natural variation or “real change”. Natural variation may also be characterized as measurement error. Natural variation is assessed in individuals who are considered to be in “steady state”.
Objectives To examine natural variation of PaGl, pain and FTG in patients with stable RA during treatment with biological agents.
Methods 233 RA patients treated with a biological agent and with stable disease were identified in the Danish rheumatology registry (DANBIO). According to EULAR response criteria, stable disease was defined as a change in DAS28-CRP ≤0.6 between two consecutive visits. Paired data from a single set of such two consecutive visits were extracted for each patient. Data comprised PaGl, pain, FTG and physician global assessment scored on 0-100 VAS scales and DAS28-CRP, tender and swollen joint count, CRP and HAQ-DI. Variation of PaGl, pain and FTG, respectively, was assessed using the Bland-Altman method with calculations of lower and upper 95% limits of agreement (LLoA;ULoA) between two consecutive assessments and the corresponding bias (mean of individual differences). Associations between intra-individual inter-visit differences (Δ) in PaGl, pain and FTG and in other disease activity measures were evaluated by Pearson's linear correlation and stepwise multiple regression analyses.
Results Mean age was 60±15 years, mean inter-visit time duration 22±21 weeks, mean DAS28-CRP 3.1±1.2, mean PaGl 37±27 and mean ΔDAS28-CRP 0.0±0.3 (NS). LLoA;ULoA [bias] was -30;28 [-0.8] for PaGl, -35;36 [0.5] for pain and -36;38 [0.9] for FTG. No significant correlation was found between the absolute Δvalue of PaGl, pain or FTG and the inter-visit time duration (r =0.0 to -0.1, NS). ΔPaGl, ΔPain and ΔFTG were only weakly inter-correlated (r =0.33 to 0.39, p<0.01) and were not significantly correlated with change in any other single measure of disease activity. In a multiple regression analysis including Δ of all single disease activity measures and DAS28-CRP as independent variables, ΔPaGl was best predicted by Δpain (rpartial =0.45, p<0.0001), Δpain by ΔPaGl (rpartial =0.45, p<0.0001), and ΔFTG by ΔPaGl (rpartial =0.23, p<0.001). No statistically significant difference in ΔPaGl, Δpain and ΔFTG was found between males and females.
Conclusions Independently of time duration, PaGl, pain and FTG fluctuated substantially in RA patients who were in steady state according to the EULAR response criteria. In the daily clinic, a change in these patient-reported outcome measures of at least 30 on a 0-100 VAS scale may be considered as natural variation or “measurement error” with no relation to the disease activity.
Aletaha D et al. Arthritis Rheum 2008; 59: 1371-7.
Disclosure of Interest None declared