Background/Purpose: The Affordable Care Act proposes more widespread use of mid-level providers (MLPs), such as nurse practitioners and physician assistants, but little is known about the process and outcomes of the care they provide. This is of particular interest in rheumatology, where there is a predicted workforce shortage by the year 2025. We set out to compare the outcomes of care provided by MLPs for rheumatoid arthritis (RA) in rheumatology practices with that provided by rheumatologists alone.

Methods: This study was conducted in 7 rheumatology practices in the US – 4 with MLPs and 3 without. The primary outcome was RA disease activity, categorized as remission, low, moderate, or high, using standardized measures (e.g., DAS28, RAPID3, CDAI).  We abstracted the following information from medical records for patients with RA from the most recent 2 years: RA treatments, serologic status, disease duration, and disease activity measures. We compared patient characteristics and disease activity for visits across the MLP and rheumatologist only practices.  We performed a repeated measures analysis to compare disease activity for visits with MLPs versus those with rheumatologists. These contrasts were made with ordinal logistic regression and expressed as proportional odds ratios.  Sensitivity analyses also examined 1) the area under the curve (AUC) for disease activity over the 2-year study period (linear regression) and 2) the change in disease activity between visits (ordinal logistic regression).

Results: Records from 301 patients, including 1982 visits were reviewed: 1168 visits with MLPs and 814 with rheumatologists. Overall, patients had a mean age of 61 years and 77% were female. 69% of patients seen by MLPs and 70% of those seen by rheumatologists were seropositive. In the primary adjusted analysis, patients seen in MLP practices were more likely to have better disease activity (OR 0.32, p = 0.004, reduced probability for higher disease activity) than those seen in practices with only rheumatologists (see Table). Similar trends were observed in the AUC analysis.  However, there were no differences in the change in disease activity comparing patients seen in practices with MLPs versus rheumatologist only.

Conclusion: While non-randomized trials are subject to confounding by indication, patients seen in practices with MLPs for RA had reduced disease activity over a 2-year observation period compared with those seen in rheumatology only practices; although no differences were observed in the change in disease activity between visits. Two competing possibilities emerge that further research can help clarify: patients seeing MLPs are less sick to begin with or their care is associated with better disease control when MLPs are part of the team.

Table: Categorical disease activity measures compared across patients seen in practices with mid-level providers versus those without, based on adjusted regression models*