County-Level Outliers in Risk-Adjusted Preterm Birth Among Young Women in the United States

Objective: To assess county-level variation in risk-adjusted preterm birth rates among young women and to detect counties with outlying rates. Design: Population-based cross-sectional study. Setting: All counties in the United States with at least 50 births. Participants: U.S. natality data for women younger than 25 years with singleton live births residing in the United States (N = 1,701,436). Methods: We estimated and mapped the predicted probability of risk-adjusted preterm birth using U.S. Restricted-Use Detail Natality (2018–2019) and American Community Survey data. We excluded births that occurred more than 320 km (199 miles) from the residential county. We used multivariable logit models to control for individual sociodemographic variables, health factors, and residence in high-poverty counties. We summarized county, state, and regional variation using dispersion measures. We identified geographic hotspots or counties with preterm birth rates at or above the 90th percentile and geographic safety zones or counties with preterm birth rates at or below the 10th percentile. Results: The overall crude preterm birth rate was 8.4%. In the South, where 46% of young women lived, the mean risk-adjusted preterm birth rate was 9.1%. Risk-adjusted preterm birth rates at the 10th percentile ranged from 3.0% to 9.1%, and at the 90th percentile they ranged from 6.6% to 11.6%. We created regional maps to illustrate the variation. Conclusion: Risk-adjusted preterm birth rates in many hotspot counties are about two times higher than the national average, whereas rates in safety zones are nearly half the national average. More research is needed to understand key socioecological factors, including prenatal care, that might explain variation in county-level risk-adjusted preterm birth rates.