Socioeconomic and spatial inequities in U.S. county-level breast cancer mortality, 2018–2023: A Bayesian hierarchical spatial analysis.

Abstract

e23112 Background: Despite advances in screening and treatment, breast cancer mortality remains uneven across U.S. counties. Understanding how social, economic, and geographic factors shape these disparities is essential for designing equitable cancer-control strategies. Methods: County-level breast cancer mortality data (2018–2023) from CDC WONDER were linked with socioeconomic indicators from the PLACES and American Community Survey datasets. A Bayesian hierarchical Poisson model with a Besag–York–Mollié 2 (BYM2) spatial prior was estimated using Integrated Nested Laplace Approximation (INLA). Predictors included poverty, education, percentage of Black women ≥ 55 years, uninsured women ≥ 55 years, mammography screening, facility density, and metropolitan status, scaled for comparability. Model adequacy was evaluated using the Deviance Information Criterion (DIC), Watanabe–Akaike Information Criterion (WAIC), and Conditional Predictive Ordinates (CPO). Results: Higher poverty (Rate Ratio [RR] = 1.03, 95% CrI: 1.01–1.05), proportion of black women aged ≥ 55 years (RR = 1.05, 1.04–1.06), and higher uninsured rates (RR = 1.21, 1.19–1.22) were associated with elevated mortality. Higher educational attainment (RR = 0.94, 0.94–0.95), mammography screening (RR = 0.99, 0.99–1.00), and metropolitan residence (RR = 0.97, 0.95–0.99) were protective. Residual spatial dependence was modest (φ ≈ 0.06), with clusters mainly in the Southeast and Midwest. Conclusions: Socioeconomic disadvantage, insurance gaps, and racial composition remain key drivers of geographic variation in U.S. breast cancer mortality. Reducing these upstream determinants and improving screening access in persistently high-risk regions may meaningfully reduce county-level disparities.