Increasing Access to Primary Care without Bumping Up Costs: Evidence from the Medicaid Fee Bump

Medicaid beneficiaries often have a more difficult time finding available health care providers who accept their insurance than patients with private coverage. This lack of access has been attributed to low reimbursement rates relative to other payers. For example, in 2012, the state-level ratio of Medicaid-to-Medicare physician fees was 0.59 for primary care providers. Limited access to primary care services and less use of preventive care adversely affect participants’ health and result in higher rates of emergency department and hospital use among Medicaid beneficiaries, which increase Medicaid costs.

We use data from the 2008-2015 Medical Expenditure Panel Survey (MEPS) to investigate the impact of changes in Medicaid payments to primary care physicians on the use of medical services. The variation in Medicaid physician payments comes from the Medicaid fee bump, a provision of the Affordable Care Act mandating that states raise Medicaid payments to match Medicare payment rates for primary care visits in 2013 and 2014. Because Medicaid payment generosity varied considerably across states before the mandated fee increase, the policy had a heterogeneous impact across states. While a few states had Medicaid payments that were unaffected by the fee bump, payment rates more than doubled in other states. We merge state-level Medicaid fee schedules for primary care services to the MEPS in order to calculate the magnitude of the fee change across states and time. We then regress different measures of healthcare utilization on the state-level fees, while using two different strategies to account for state policy endogeneity in the fee variable. First, we include state and time fixed effects in our models, so identification comes from changes in fee schedules within states over time. Second, we estimate a triple-difference specification that compares changes in Medicaid fees over time to changes in payments for privately insured patients over time. In addition, we include county-level controls for differences in the availability of medical resources such as the per capita supply of physicians, nurse practitioners, hospital beds, median household income, unemployment, and whether area is underserved for primary care services to reduce the potential for policy endogeneity.

In order to examine the impact of better access to primary care services, we estimate our models separately for all office-based visits, office-based visits for preventive care, outpatient department visits, emergency department (ED) visits, and inpatient admissions. In addition, we estimate the effect of generosity of physician payment levels on the site of ambulatory care received by Medicaid patients at the visit level and test whether higher physician payments are correlated with reductions in non-urgent ED care, as defined by the New York University ED Algorithm. Finally, we quantify the effect of Medicaid fee changes on healthcare spending to determine whether increasing Medicaid payments leads to lower overall medical care costs.

Preliminary results indicate that increasing Medicaid payments to primary care providers are associated with significant increases in office-based visits to physicians, but not midlevel providers, such as nurse practitioners or physician assistants. We also do not find any effects on outpatient department visits.