Background/Purpose:

Outcomes in ERA are largely thought to reflect individual prognostic factors and treatment strategy employed. We hypothesize that an additional factor may be the provider, who could influence outcomes through their educational and practice experience, approach to patient interactions, and practice resources. We have undertaken an analysis of remission outcomes across 11 ERA clinics, demonstrating significant variation in rates of remission. In this analysis, we incorporate patient characteristics, treatment strategy and investigator effect in understanding this variation.

Methods: Data were analyzed for ERA patients with >1 year of follow-up, enrolled at sites with >40 patients at baseline and >30 patients with 2 years of follow-up data. Multiple linear regression was applied to predict DAS28 remission at 12 months, with all continuous variables centered at the mean prior to entry in the models. Covariates in the model were patient characteristics (age, sex, symptom duration at initiation of treatment, education, comorbidities, smoking status, seropositive status and baseline DAS28 and HAQ scores); treatment strategy (oral methotrexate monotherapy, subcutaneous methotrexate monotherapy, methotrexate-based combinations, triple therapy, biologic therapy and steroid); and investigative site characteristics (volume of recruitment; academic or community practice; solo vs group; allied health resources available; teaching site; and the proportion of investigators at that site by sex, age cohort, location of training, training years and practice type (rheumatology alone or also providing internal medicine services).

Results:

1,633 participants with mean age 54 years, 73% female, and mean DAS28 4.9 (SD 1.4) were included. At 12 months, 49% of patients across all sites had achieved DAS28 remission (site range 22%-80%) and the frequency of sustained DAS28 remission over all follow-ups was 52% (site range 29-71%). There were significant differences between centers in the means of patient characteristics and frequency of use of treatment strategies (oral methotrexate monotherapy 8%, subcutaneous methotrexate monotherapy 6%, methotrexate-based combination therapy 36%, triple therapy 11%, and biologics 14%). 27% of sites were community-based practices, 45% systematically incorporated allied health professionals in the care plan, 82% were a teaching site, 64% were group recruitment sites with a mean of 5 investigators per site, and 62% of the investigators were female. In the model, investigator age, location of training, training year and practice type were excluded due to collinearity with the other characteristics. Patient characteristics of age, sex, comorbidities, smoking status, seropositive status, baseline DAS28 scores were significant factors in remission (all p<0.04). The specific treatment strategy had no significant effect, except for use of oral steroids. The presence of learners had a negative impact on remission (coefficient -0.50, 95%CI -0.76 to -0.24).

Conclusion: Patient prognostic factors contribute greatly to the variation in remission outcomes, with treatment strategy and investigator/investigative site factors having lesser contributory roles.

« Back to 2017 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/patient-characteristics-treatment-strategy-or-investigator-effect-an-analysis-of-factors-driving-variation-in-outcomes-in-early-rheumatoid-arthritis/