Improving our estimates: assessing misclassification of abortion accessibility in the United States

Quantify the misclassification of abortion accessibility resulting from calculations based on block groups, census tracts, ZCTAs, or counties versus census blocks.

We included 850 facilities from the 2018 ANSIRH Facility Database and Planned Parenthood website. Accessibility was the proportions of 18-44 year-old women living within a 30-, 60-, and 90-minute drive from an abortion-providing facility. We calculated accessibility using five different geographic units: census blocks, block groups, census tracts, ZCTAs, and counties. We estimated the potential magnitude of abortion accessibility misclassification resulting by calculating the difference between accessibility calculated using each geographic unit, respectively, for each drive time as compared to census blocks.

In this analysis, counties, the largest geographic unit considered, underestimated national abortion accessibility compared with census blocks by up to 24.21 percentage points; while block group-, census tract, or ZCTA-based national estimates of abortion accessibility in the US resulted in small underestimates relative to estimates constructed using census blocks.

Studies of population accessibility should use the smallest feasible geographic unit of analysis. While this study focused on abortion accessibility, our findings likely apply to other health services, particularly those with distributions like abortion care.