AHA Scientific Sessions 2026 Abstract · Nov 2, 2026
Shifting Structural Determinants of County-Level Cardiovascular Disease Hospitalization Burden Across Three Epidemiological Eras in Georgia: A Spatial Regression Analysis, 2017-2024
By Saint Julien B, Coles N, Obeng S et al.
BACKGROUND: County-level social determinants of health (SDOH) are established drivers of geographic cardiovascular disease (CVD) burden, yet whether COVID-19 fundamentally altered which structural factors predict CVD hospitalization remains unknown. We examined shifts in structural determinants of county-level CVD hospitalization burden across three epidemiological eras in Georgia, a state encompassing the nation's highest-burden CVD region. MethodS: We conducted a cross-sectional ecological analysis of all 159 Georgia counties in 2017 (pre-pandemic), 2020 (pandemic), and 2024 (post-pandemic). The outcome was the county-level age-adjusted deduplicated CVD hospitalization rate per 100,000 (Georgia OASIS). Twelve SDOH predictors spanning economic, healthcare access, behavioral, and demographic domains were derived from County Health Rankings. Moran's I confirmed significant spatial autocorrelation across all eras (z=8.36, 6.78, 8.91; all p<0.001). Lagrange Multiplier diagnostics confirmed spatial error dependence; spatial error regression was employed as the primary analytic approach. ResultS: Mean CVD rates were 1,113 (SD=309), 1,010 (SD=286), and 1,187 (SD=337) per 100,000 in 2017, 2020, and 2024; a 9.3% pandemic decline followed by a 17.6% post-pandemic rebound exceeding baseline by 6.6%. Black-White disparities widened (absolute gap: 406, 415, 421 per 100,000), with the relative gap peaking during the pandemic (46.2%) before partially narrowing post-pandemic (39.6%). Spatial error regression revealed significant era-specific structural predictors across all three years. Pre-pandemic, income inequality (β=71.3, 95%CI:22.9,119.7, p=0.004) and primary care physician access ratio (β=0.013, 95%CI:0.00,0.02, p=0.030) were dominant. During the pandemic, adult smoking (β=37.2, 95%CI:4.8,69.7, p=0.024) and uninsured rate (β=22.6, 95%CI:1.9,43.4, p=0.032) emerged significant while income inequality lost significance. Post-pandemic, physical inactivity (β=49.0, 95%CI:7.2,90.8, p=0.022) and rural percentage (β=-3.1, 95%CI:-4.8,-1.4, p=0.0003) became strongest predictors. Spatial parameters were large and consistent (λ=0.733, 0.707, 0.764; AIC improvement over OLS: 63.0, 51.2, 71.2). ConclusionS: Structural determinants of county-level CVD hospitalization shifted substantially across COVID-19. Income inequality and healthcare access lost significance while physical inactivity and rurality emerged as dominant post-pandemic drivers. While ecological design precludes individual-level causal inference, these findings challenge the assumption of temporal stability in SDOH-CVD associations and suggest pandemic-era disruptions may have reconfigured which community-level factors drive cardiovascular hospitalization, with critical implications for targeted prevention and resource allocation.