Is segregation bad for your heart?: a natural experiment for residential racial segregation exposure during childhood and cardiovascular disease risk

Abstract

Background: Residential racial segregation is hypothesized to cause inequities in cardiovascular disease. Yet existing work is largely associational and has rarely tested longitudinal effects of childhood segregation on later CVD. Here, we leverage quasi-random urban planning decisions in Northern US cities prior to the Great Migration as an instrument for segregation as Black families moved North, providing a natural experiment. Methods: We followed Ananat’s work showing railroad track configurations serve as an instrument for segregation at the metropolitan area level. We linked Ananat’s data to the birthplaces of participants in the Panel Study of Income Dynamics (1968-1994, n=829 Black & 1,262 White), following them through 2023. Race-stratified accelerated failure time models used an instrumental variables design and cluster bootstrapping, progressively adjusting for other railroad features, socioeconomic status, and birth year. Results: Estimates suggest childhood residential segregation causally delays the onset of CVD or hypertension for White people but speeds it up for Black people, though confidence intervals and instrument strength varied across model specifications. Non-instrumented models directly adjusting for confounders yielded strongly attenuated results in the same direction. Conclusion: Segregation is likely a cause of Black-White CVD inequities, not just a correlate. Causal designs across the life course are needed.