Geographical variation in infant mortality due to congenital heart disease in the USA: a population-based cohort study. (July 2021)

• Record Type: Journal Article
• Authors: Udine, Michelle L; Evans, Frank; Burns, Kristin M; Pearson, Gail D; Kaltman, Jonathan R
• Is Part Of: Lancet. Volume 5:Number 7(2021)
• Page Start: 483
• Page End: 490
• Publication Date: 2021-07
• Subjects: Pediatrics -- Periodicals; Children -- Health and hygiene -- Periodicals; Adolescent medicine -- Periodicals; Teenagers -- Health and hygiene -- Periodicals; 618.920005
• DOI: 10.1016/S2352-4642(21)00105-X
• Languages: English
• ISSNs: 2352-4642

Abstract

Summary: Background: Little is known about geographical variation in infant mortality due to congenital heart disease (CHD) and the social determinants of health that might mediate such variation. We aimed to examine US county-level estimates of infant mortality due to CHD to understand geographical patterns and factors that might influence variation in mortality. Methods: This US population-based cohort study used linked livebirth–infant death cohort files from the US National Center for Health Statistics from Jan 1, 2006, to Dec 31, 2015. All deaths attributable to congenital heart disease in infants in a given year were included. We used hierarchical Bayesian models to estimate rates of infant mortality due to congenital heart disease for all US counties. We mapped model-based estimates to explore geographical patterns. Covariates included infant sex, gestational age, maternal race and ethnicity, percentage of the county population below the poverty level, and proximity of the county to a US News & World Report 2015 top-50 ranked paediatric cardiac centre. Findings: From 2006 to 2015, 40 847 089 livebirths occurred, of which there were 13 988 infant deaths attributed to congenital heart disease, with an unadjusted infant mortality rate due to CHD of 0·34 per 1000 livebirths (95% CI 0·34–0·35). Kentucky and Mississippi had the greatest proportions of counties with a predicted rate of infant mortality due to CHD above the 95th percentile. All counties in Connecticut, Massachusetts, and Rhode Island had a predicted rate below the fifth percentile. In the model, lower mortality risk correlated with closer proximity to a top-50 ranked paediatric cardiac centre (odds ratio [OR] 0·890, 95% credible interval [CrI] 0·840–0·942), whereas higher mortality risk correlated with higher levels of poverty (OR 1·181, 95% CrI 1·125–1·239). Interpretation: Substantial geographical variation exists in infant mortality due to CHD in the USA, highlighting the potential importance of bolstering care delivery for infants from economically deprived communities and areas remote from top-performing paediatric cardiac centres. Funding: None.