Research That Matters (January 17 - 20, 2008)

Purpose: Estimates of the number of Americans without health care insurance continued to climb in the last decade, reaching over 45 million in 2006. African Americans are disproportionately represented in the uninsured population. Medicaid expansions are proposed as an attempt to expand health care coverage to the low-income and working poor. While expansions will reduce uninsurance rates, issues of access to primary health care for Medicaid-dependent populations remain. Provider preferences, in terms of both practice location and Medicaid participation, drive access to care. Because Medicaid eligibility is linked to income, recipients of public insurance tend to be concentration in areas with high poverty rates. These communities, plagued by crime and little economic investment, are not environmentally attractive to providers in terms of practice location. Further, reimbursement rates for Medicaid insurance providers continue to lag behind private insurance, creating a disincentive for providers to accept Medicaid. These factors combine to create communities with a dearth of primary care providers (PCPs). This paper examines the spatial relationship between communities of high Medicaid recipient concentration and PCP practice location.

Methods: Participating PCP lists were obtained for the 3 insurers contracted with the State of Pennsylvania's Office of Medical Assistance to provide coverage to Medicaid consumers in Philadelphia County. Office locations for Medicaid participating PCPs were geocoded and mapped using ArcView GIS. Community health centers and Emergency Departments were mapped as alternative care sites for Medicaid recipients. Concentration of poverty and concentration of African-Americans were chosen as proxies for high Medicaid recipient concentration. Census data was used to map the distribution of poverty and African Americans at the census block level. Public transportation routes were mapped, delineating potential transit routes to PCPs. Availability of PCPs within 2- and 5-mile radii of census block centroids was examined.

Results: No clear relationship between PCPs who accept Medicaid and concentration of poverty was discernable. Although many block groups with high poverty concentration had no providers within a 5-mile radius, the same was true for block groups of lesser poverty concentration. Additionally, for those census block areas with no PCPs that accepted Medicaid within a 2-mile radius, one was available within a 5-mile radius. A somewhat more distinct pattern emerged between African American concentration and Medicaid participating PCPs. Predominately African American communities had few PCPs scattered throughout. Also, providers located within a 5-mile radius of census block group areas with high African American concentration, were community health centers or hospitals, not PCPs.

Implications: The scarcity of PCPs located in predominately African American communities indicates that these Medicaid consumers are limited to care at health centers and hospitals. Given the connection between timely, consistent access to a regular care provider and positive health outcomes, this spatial relationship does not bode well for Medicaid recipients. If Medicaid expansions increase insurance rates but significant access barriers remain, such policy efforts may not impact health outcomes. Eligibility expansions must be coupled with efforts to encourage Medicaid participation and to locate practices in areas with a high concentration of Medicaid consumers.