FACTORS ASSOCIATED WITH DECISIONS TO ADJUST THERAPY FOR RHEUMATOID ARTHRITIS PATIENTS IN MODERATE TO HIGH DISEASE ACTIVITY

Y. Shaw 1,*, C.-C. H. Chang 2, H. Eng 1, I. Metes 2, S. R. Wisniewski 1, M. S. Roberts 1, M. C. Levesque 2
1Graduate School of Public Health, 2School of Medicine, University of Pittsburgh, Pittsburgh, United States

Background: The American College of Rheumatology (ACR) recommends treating rheumatoid arthritis (RA) to the target of low disease activity or remission with traditional and biologic disease-modifying anti-rheumatic drugs (DMARDs). However, significant numbers of RA patients do not receive care consistent with these recommendations. Previously, we found that age, race, physician age, current use of a biologic, Disease Activity Score-28 joint (DAS28-CRP), and RA duration were significantly associated with decisions to adjust DMARD therapy for RA patients with moderate to high disease activity. Here, we present an update to our findings with new covariates added to the analysis, including income and comorbidities, which were adjusted for to prevent confounding bias in the effects of age and race.

Methods: Data was drawn from the University of Pittsburgh Rheumatoid Arthritis Comparative Effectiveness Research (RACER) observational registry (2010-13) on visits where patients had moderate to high disease activity for at least 3 months. Therapy adjustment was defined as adding, switching, or increasing the dose of oral or biologic DMARD therapy. To assess this binary outcome, a generalized linear model with logit link was used. Generalized estimating equations (GEE) with clustering by patient were used to account for correlation between different visits for each patient. Variables significant in univariable GEE analyses (Wald test on coefficient has p0.05. The model controlled for patient demographic and disease characteristics: age, race, gender, RA duration, DAS28-CRP, Short Form 12 (SF12) physical and mental components, Health Assessment Questionnaire (mdHAQ), and two new variables (income and Charlson comorbidity index).

Results: There were 562 visits for 255 patients with moderate to high disease activity for at least 3 months. Therapy was adjusted at 23.1% of visits. The odds ratios (OR; [95% CI]) of adjusting therapy were decreased by longer disease duration (0.965; [0.943,0.987]), physician age (0.955; [0.932,0.978]), current use of biologic therapy (0.394; [0.240,0.648]), and lower DAS28-CRP (0.640; [0.455,0.900]). Age and race were no longer significant.

Conclusion: In the RACER observational cohort, 76.9% of RA patients with moderate/high disease activity for at least 3 months did not have DMARD therapy adjusted. Physician age, duration of RA, lower DAS28, and current use of biologic therapy decreased the likelihood of therapy adjustment. The fact that age and race were no longer significant upon the addition of Charlson comorbidity index and income to the model suggests that these new variables may partially explain the previously observed effects of age and race. These results can be used to improve treat-to-target strategies in clinical practice. Our future research will explore how other factors such as prior use of DMARDs/corticosteroids, insurance coverage, and risk aversion may influence treatment decisions for RA patients with moderate/high disease activity.