Trends in the Supply, Utilization, and Payment for Primary Care Services in New Jersey FamilyCare before and after ACA Medicaid Expansion
Publication Date: 02/01/2019
Medicaid expansion under the Affordable Care Act (ACA) has raised questions about whether and how the program will be able to meet the rising demand for demand for services from new enrollees. The supply of primary care services is of particular concern. Numerous reports have warned about national shortages in the supply of primary care physicians overall and many practicing physicians do not participate in Medicaid.
This report examines trends in the supply of primary care services to NJ FamilyCare – which includes Medicaid and the Children’s Health Insurance Program (CHIP) – during a time of rising enrollment and multiple changes in primary care payment policy at the federal and state levels. It uses data from the NJ Medicaid Management Information system (MMIS) from 2012 through the first half of 2017 to document trends in key variables within regions that approximate primary care service areas. Variables examined include primary care visits, number of providers, use of midlevel providers (e.g., physician assistants, nurse practitioners), payments for services (measured by a primary care price index that preserves confidentiality of proprietary pricing information), and the market concentration of Medicaid HMOs and primary care providers within regions. Trends in treat-and-release emergency department (ED) visits, which are often used as a substitute mode of care by patients who face difficulty accessing primary care, are also examined.
The analyses center on primary care service areas, which are designated by 3-digit zip codes and weighted by NJ FamilyCare enrollment. For each service area in each period, a set of core primary care providers is defined based on the smallest number of unique primary care providers that accounts for at least 75% of all primary care visits. To minimize the influence of outlier zip codes, the report emphasizes median values of key variables. (Major findings are not sensitive to weighting, mean versus median values, or the use of 5-digit instead of 3-digit zip codes.)