Health disparities in severe maternal morbidity: selection bias, intersectionality, and hypertensive disorders of pregnancy
Embargo Date
2029-01-31
OA Version
Citation
Abstract
Severe maternal morbidity (SMM), defined as unexpected outcomes of labor and delivery, affects more than 50,000 birthing individuals annually in the United States, and is characterized by persistent health disparities across various social factors, such as race/ethnicity and socioeconomic status. These patterns have continued over time despite evidence that many SMM events are preventable. Therefore, a deeper understanding of both social and clinical drivers, along with careful consideration of potential sources of systematic error, is critical to inform effective interventions. Using data from the Pregnancy to Early Life Longitudinal Data System, which links live birth and fetal death certificates to hospital delivery discharge records in Massachusetts, we analyzed deliveries from 1998 to 2021 to address key methodological, social, and clinical gaps in research on health disparities in SMM. In the first study, we estimated the effect of SMM, compared with no SMM at the first hospital delivery, on the risk of SMM at the second hospital delivery and applied inverse probability of continuation weighting to quantify the magnitude and direction of bias introduced by selective attrition. A history of SMM was associated with an excess of more than 500 SMM events per 10,000 in-hospital deliveries at the second delivery, underscoring the importance of tailored care in future pregnancies. We also found that Black birthing people were disproportionately affected, with rate differences (RD) exceeding 800 per 10,000 in-hospital deliveries. Selection bias was minimal and generally resulted in a slight downward bias of <1%. In the second study, guided by intersectionality, we used interaction analyses to examine the association between occupying one or more marginalized identities compared with only privileged social identities and the risk of SMM. Multiply marginalized groups (i.e., racially and ethnically minoritized individuals with socioeconomic disadvantage), as well as Black individuals with socioeconomic advantage, experienced excess SMM events (RD 27–150 per 10,000 in-hospital deliveries) compared with White birthing people with socioeconomic advantage. These findings highlight the importance of addressing multiple intersecting social positions in efforts to reduce disparities in SMM. In the third study, we used mediation analysis to estimate the proportion of the Black–White racial disparity in SMM that would be eliminated through the prevention of hypertensive disorders of pregnancy (HDP). HDP accounted for 38.5% of the Black–White racial disparity in SMM, indicating that prevention of HDP may represent a potential pathway for reducing part of the racial disparities in SMM. Together, these three studies advance our understanding of various methodological, social, and clinical factors underlying health disparities in SMM.
Description
2026
License
Attribution-NonCommercial-NoDerivatives 4.0 International