Slavery, homeownership, and contemporary perinatal outcomes in the southeast: a test of mediation and moderation.

The objective of this study was to estimate the effect of geographic variation in historic slavery on perinatal outcomes (chronic hypertension, hypertensive disorders of pregnancy [HDP], very preterm birth [VPTB], or very low birth weight birth [VLBW]) among Black people living in states where slavery was legal in 1860; and to test mediation by Black homeownership. We linked data on the proportion of enslaved residents from the 1860 Census to natality data on outcomes (2013-2021) using resident county. The percentage of Black residents in a county who owned their home was a potential mediator. We fit log binomial models to estimate risk ratios (RRs) representing total and controlled direct effects (accounting for Black homeownership) of the proportion enslaved on outcomes, accounting for potential confounding, using marginal structural models. Among 2 443 198 included births, 8.8% (213 829) experienced HDP, 4.1% (100 549) chronic hypertension, 3.3% (81 072) VPTB, and 2.6% (62 538) VLBW. There was an increase in chronic hypertension and VPTB risk, but not HDP or VLBW, in counties with a 10% greater proportion of enslaved residents in 1860 (adjusted RR [95% CI], chronic hypertension: 1.06 [1.02-1.1]; VPTB: 1.02 [1.00-1.05]; HDP: 1.00 [0.98-1.02]; and VLBW: 1.01 [1.00-1.03]). There was not evidence of mediation by Black homeownership. We conclude that historic slavery remains relevant for perinatal health.