Women’s Health
Geographic Variation and Disparities in Hypertensive Disorders of Pregnancy Katherine Campbell* Katherine Campbell Lance Waller Penelope Howards Anne Dunlop Michael Kramer
Small area estimates of hypertensive disorders of pregnancy (HDP) (e.g. chronic vs gestational hypertension) are not readily available for county and local areas. Documenting local rates of disorders are of interest but generating statistically robust small area estimates can be difficult when local population and event counts are small. We use Bayesian spatial analysis to address challenges in describing county-level patterns of HDP subtypes and identifying characteristics of counties with high risk.
We abstracted birth certificate data for births to individuals ages 18 to 44 years in the US between 2009 and 2019. We model counts of each HDP subtype at the county-level via Poisson-Gamma Bayesian spatial model to estimate stable local rates of HDPs, account for spatial dependency, and identify counties exceeding the expectation of disease risk. We describe demographic and socioeconomic context based on county-specific measures to characterize counties identified as “high risk” (posterior probability of exceeding the national average) for each HDP.
Among 42 million live births in 2,869 counties, gestational hypertension rates were (7.5%) in 521 counties. Chronic hypertension rates were higher than expected (2.7%) in 513 counties and more consistently located in the southeastern US. The results revealed 204 counties exhibiting co-occurrence of high-risk for both chronic and gestational hypertension, but more than 300 counties showed subtype discordance. Differences in contextual factors were evident in counties with high-risk chronic hypertension compared to those without, whereas high-risk gestational hypertension was less strongly associated with racial density of counties and did not exhibit a clear correlation with rurality. This study underscores the question of combining HDPs for surveillance and research, given their distinct geographic distributions and the variation in their relation to place-based characteristics.