Background/Purpose: Current guidelines encourage the measurement of rheumatoid arthritis (RA) disease activity to achieve a low disease state (treat-to-target). Many RA patients with documented moderate to high disease activity, however, remain on their current regimen. This study used clinical and administrative data of US Veterans, identified as having moderate/high disease activity (DAS28≥3.2), to identify patient and treatment characteristics associated with a major change in therapy.

Methods: Patients in the Veterans Affairs (VA) RA registry were included if they had: 1) moderate/high disease activity on the index date, 2) 18-months of VA activity prior to the index and 3) two or more prior DAS28 measures during the preceding 18-months (≥ three months apart). Patients were defined as having a major change in therapy if any of the following occurred within 7 days prior to 30 days after the index date: 1) initiation or escalation of DMARDs, 2) initiation or increase dose of prednisone and/or 3) ≥ 2 joint injections. Disease stability was determined during the observation period prior to the index date using an area under the curve calculation and compared to the index DAS28. Patients were categorized as improving (index DAS28 was ≥ 0.6 points lower), worsening (≥ 0.6 higher) or stable. Poisson multivariable regression was used to explore how demographic, clinical and treatment patterns influenced the initial finding that stability of disease activity was a strong predictor of a major treatment change.

Results: Of the 941 patients who met enrollment criteria only 41% had a major therapeutic change. Major therapeutic changes occurred in 50%, 37% and 30% with worsening, stable and improved DAS28, respectively. In multivariable models, patient demographics and individual components of the DAS28 score were not independently associated with a major therapeutic change. In the full model, the current DAS28, and oral steroids and non-biologic DMARDs in the past year increased the likelihood of a major change. Non-biologic DMARDs in the past 90-days decreased the likelihood of a major change. Table 1 displays the RR with “stable DAS28” as the comparison group after adjustment for demographic, demographic + DAS28 and clinical components and the full model, which added indicators of medication use, surgery, radiographs and comorbidity scores.

Conclusion: More than half the patients did not experience a major change in therapy despite moderate/high disease activity even among those patients with worsening disease. In crude analysis, the likelihood that RA patients with moderate/high RA disease received a change in therapy was highly associated with worsening disease activity, but this effect was diminished when accounting for current DAS28, indicators of non-biologic DMARDs and steroid use in the previous year. The decision to modify therapy is complex and we are conducting chart-review to better understand why providers are not modifying therapy.