Abstract
Research Objective
Currently, services for Medicaid patients in ambulatory care are predominantly delivered by primary care physicians (PCMDs). This study compared the quality of care and practice patterns among primary care physicians, nurse practitioners (NPs), and physician assistants (PAs) in ambulatory care settings to indicate gaps in services that could be filled if primary care services were delivered by NPs and PAs.
Study Design
Using Medicaid claims data of the New York State Medicaid Data Warehouse from 2016 to 2018, multivariate regression analyses were conducted to estimate the impact of receiving primary care by NPs/PAs versus receiving primary care by PCMDs.
Literature reviews were conducted to select 7 patient‐level outcome variables, including 3 quality of care indicators (smoking cessation, depression, and hyperlipidemia) and 4 practice pattern indicators (general examination, ultrasonography, medication, and referral). Both design‐based and model‐based inferences were utilized to test on 7 key outcome variables (software SPSS 25 and Hierarchical Linear and Nonlinear Modeling (HLM) 7, respectively).
Effects of practitioner type on quality of care and practice patterns were examined using weighted regressions and presented as adjusted odds ratios (AORs) for each outcome.
In addition to the AORs, predictive margin modeling was also used to estimate the probability of each outcome occurring among PCMDs, NPs, and PAs while controlling for other outcomes.
Population Studied
Analyses included over 23 million Medicaid claims from 21,392 PCMDs, 3,492 NPs, and 1,951 PAs on more than 6 million unique Medicaid patients in New York State from 2016 to 2018.
Principal Findings
The distribution of practitioners’ characteristics was largely similar across practitioner type, although a greater proportion of NPs were female compared with PCMDs and PAs.
On 4 of the 7 outcomes—depression treatment, hyperlipidemia treatment, general examination, and medication—no statistically significant differences were detected in NP/PA‐provided primary care compared with PCMD‐provided.
On the remaining outcomes, Medicaid patients seen by NPs were more likely to receive smoking cessation treatments than those seen by PCMDs (P ≤ .05). Medicaid patients seen by PAs were more likely to receive primary care ultrasounds and physician referrals than those seen by PCMDs (P ≤ .01).
Conclusions
Among the outcomes studied, primary care services provided by PCMDs, NPs, and PAs are comparable in ambulatory care settings in New York for Medicaid patients.
Implications for Policy or Practice
A greater use of NPs and PAs is likely to produce equivalent quality primary care services and may improve primary care in underserved areas where have difficulties hiring primary care physicians.