Residential Segregation and the Use of Primary Care Services

Research Objective: Does residential segregation explain disparities in the use of primary care services.

Study Design: Using data from the 2006 Medical Expenditure Panel Survey, the 2000 Census and the 2006 American Medical Association master file, we determined whether race/ethnic disparities in primary care service use were associated with residential segregation and the use of primary care services. We used 5 measures of primary care service use: 1) office based physician visits, outpatient department physician visits, office based non-physician visits, outpatient department non-physician visits, and having a usual source of care (USC). At the individual level we controlled for age, gender, marital status, insurance status, educational attainment, employment status, region, and health status. Segregation was measured by a set of indicator variables that denoted whether the zip code was predominately (50% or more) African American, predominately Hispanic, or predominately Asian. We estimated logistic and negative binomial regression analyses assuming primary care use has two parts: initiation and subsequent use.

Population Studied: We restricted are our sample to 15,689 adults living in MSAs for which we had zip code level data.

Preliminary Findings: Preliminary findings suggest that disparities in primary care use were associated with residential segregation and PCP availability especially for non-physician services. The impact varies by segregation measure and by race/ethnic minority group. In comparison to Whites, African Americans in low segregation areas were less likely to have an office based physician visit (OR=0.789, p=0.000). However, the race disparity was greater in high segregation areas (OR=0.674, p=0.000). The pattern was similar for office based and outpatient department visits to non physicians. For Hispanics, segregation improved use of primary care services. Hispanics in low segregation areas were less likely to have an office based physician visits (OR – 0.844, p = 0.038). However in high segregation areas the disparities disappears compared to whites in non predominately Hispanic areas. In fact, the Whites living in Hispanic communities were at a disadvantage compared to Whites living in non Hispanic communities (OR = 0.555, p=0.004). This pattern was similar for other primary care services measures.

Conclusions: Efforts should be made to improve access to primary care services in African Americans especially to those in predominately African American zip codes. However, Hispanic in non Hispanic communities are more vulnerable than those living in predominately Hispanic communities. Whites are at risk in Hispanic communities.