The Contribution of Contextual Factors to Racial Disparities in Severe Maternal Morbidity in California

Background: Non-Hispanic black (black) women are more than twice as likely to experience a severe maternal morbidity (SMM) event compared with non-Hispanic white (white) women. Individual characteristics do not the increase in SMM or the racial disparity in SMM. The aim of this study is to determine whether characteristics of women’s residential context are associated with the racial disparity in SMM. Methods: This is a secondary data analysis using hospital discharge and vital statistics data for delivery hospitalizations in California from 2008 through 2012 linked to data from the American Community Survey for the same time period. SMM was identified by algorithm developed by the Centers for Disease Control and Prevention to identify specific diagnostic and procedure codes. The associations between five dimensions of residential context and SMM in black and white women were estimated using stratified analysis and generalized estimating equations. The black-white disparity in SMM was characterized and assessed for potential modification by five contextual factors. Oaxaca-Blinder decomposition methods were used to estimate the contribution of group differences in contextual and individual characteristics to the black-white disparity in SMM. Results: High concentrated disadvantage, racial segregation, race and income segregation, and living in an underserved tract all were associated with increased rates of SMM, while high concentrated affluence appeared protective. After stratification by race, all associations were null. When stratified across levels of contextual variables, the black-white disparity in SMM remained the same. Age-adjusted overall and stratum-specific rate differences and rate ratios did not differ from unadjusted estimates. Combined, the five contextual factors explained only 8.0% of the black-white disparity in SMM. Maternal age group made a negative contribution of 18.4% to the disparity. Conclusion: Contextual factors did not appear to contribute meaningfully to the black-white health disparity in severe maternal morbidity. It is unclear whether there are genuinely no associations between these contextual factors and the disparity or whether the null findings were the result of the poor quality of census tract data in the delivery data or structural confounding. The complex interplay of biological, psychological, social, economic, political, and institutional factors that produce racial disparities between black and white women cannot be addressed solely through individual-level interventions. This research suggests that even addressing inequities in residential context would not be sufficient to eliminate the black-white disparity in SMM. We will likely need to address inequities at every level in order to achieve health equity for black women.