Background Patient-reported assessments have a considerable weight both in disease activity assessment in practice and response criteria in clinical trials of rheumatoid arthritis (RA). Discrepancies and disagreement on global disease activity between patients and physicians may result in disagreement in treatment decisions in routine care and impact assessment of treatment response in clinical trials as well.
Objectives We aimed to illustrate the amount of discrepancy between patient and doctor global assessments (PGA and DGA) in routine care and asses the determinants thereof.
Methods We used data from RA patients registered in the Turkish Rheumatoid Arthritis Registry (TRAV), a prospective registry established in 2010 in order to record clinical data at each visit of all RA patients seen at three academic rheumatology centers in a routine care setting. Currently 1923 RA patients are registered in TRAV. We used the initial visits of these patients to define a global assessment discrepancy (GAD) variable, GAD=PGA-DGA. We calculated the average GAD, overall correlation of PGA with DGA and constructed linear regression models to determine the predictors of PGA, DGA and GAD. Age, gender, physical function (MD-HAQ score), pain score (numeric rating scale; NRS), fatigue score (NRS), patient's joint assessment (PJA), physician reported tender (TJC) and swollen joint counts (SJC), erythrocyte sedimentation rate (ESR) and center of follow-up were included in these models as potential predictors.
Results The mean age (standard deviation) of the registered patients was 52.0 (12.8) and 1561 (81%) were females. The average PGA was 3.59 (2.99) and the average DGA was 2.85 (2.63) Overall correlation of PGA and DGA was low (0.18) and average GAD was 0.70 (95% CI: 0.53–0.87). Confidence intervals around GAD indicated that on average patients rated their current disease activity as significantly worse than that of physicians. In the regression analyses; MD-HAQ, pain scores and center were common predictors of both PGA and DGA. Additional predictors of PGA were age, fatigue scores and PJA whereas additional predictors of DGA were ESR, PJA, TJC and SJC. Predictors of GAD were pain, fatigue, TJC, SJC, ESR and treatment center (see table).
Conclusions We found a weak correlation between patient and doctor global assessments of overall disease activity in RA patients under routine care. Higher pain and fatigue scores and different treatment centers were associated with a PGA higher than DGA whereas increasing TJC, SJC and ESR were associated with a DGA higher than PGA suggesting that patients and physicians prioritized different aspects of RA in their global assessments. Variation of the discrepancy by centers warrants further research with respect to a possible impact in multi-center treatment trials.
Disclosure of Interest None declared