Background/Purpose: The goal of ERA treatment is remission but many patients do not achieve this state due to patient factors and perhaps differences in treating physicians. Studying treatment variation can lead to adopting best practices. Our objective was to investigate whether site differences occur and have an effect on outcome in ERA, and if so, to determine whether site size and/or differences in treatment explain the variability.

Methods: Sites from the CATCH database that had >40 patients at 6 months after enrolment were studied.  Included sites were randomly renumbered and assigned a letter with investigators blinded.  Patient data was used to calculate remission by several definitions (DAS28 < 2.6, SDAI ≤ 3.3, CDAI ≤ 2.8) and to determine treatment and treatment changes.  Regression models included site as a variable and confounding baseline characteristics (HAQ, DAS28, serology, presence of erosions, cigarette smoking, age, gender, symptom duration, and SES) that had a p-value < 0.10 in univariate analyses. 

Results: Of the 1138 baseline patients, 798 and 640 patients had data at 6 and 12 months respectively.  Baseline descriptive statistics revealed that (mean (SD) or %): age 52(17) years; 72% female; 23% had erosions; 54% either were current or ever smokers; 37% anti-CCP positive; 51% RF positive; disease duration 187(203) days; HAQ 0.9(0.7); DAS28 4.5(1.4).  Regression analyses showed that site is an important predictor for mean changes in DAS28 (p ≤ 0.000), increase in DAS28 (p ≤ 0.004), DAS28 remission (p ≤ 0.000), CDAI remission (p ≤ 0.000), and SDAI remission (p ≤ 0.022).  Regression analyses including treatment showed that increases in medication was the strongest predictor of bad outcome (p < 0.05) and caused site to lessen some of its predictive ability.

Conclusion: The two largest sites had the biggest mean changes in DAS at 6 and 12 months. Site is a predictor for ERA outcome. The fastest and best indicator for low DAS / remission was initial treatment with combination DMARDs and for 12 months the latter approach or initial treatment with parenteral methotrexate at 20-25mg per week. We cannot say if an unmeasured factor accounted for better changes in DAS at the largest sites.