Estimating the Burden of Prematurity on Infant Mortality: A Comparison of Death Certificates and Child Fatality Review in Ohio, 2009-2013.

Introduction Infant mortality is a key population health indicator, and accurate cause of death reporting is necessary to design infant mortality prevention strategies. Death certificates and child fatality review (CFR) both track leading infant causes of death in Ohio but produce different results. Our aim was to determine the frequency and characteristics of differences between the two systems to understand both cause of death ranking systems for Ohio. Methods We linked and analyzed data from death certificates and CFR records for all infant deaths (aged < 1 year) in Ohio during 2009-2013. Death certificate and CFR cause of death assignments were compared. Kappa statistic was used to measure concordance. Death certificate-CFR cause of death pairs were plotted to identify common concordant and discordant pairs. Results A total of 5030 infant deaths with death certificate and CFR records were analyzed. The most common discordant cause of death pair was other perinatal condition on the death certificate and prematurity by CFR (1119). Specific injury categories had higher concordance (kappa 0.71-1.00) than medical categories (kappa 0.00-0.78). Among 456 deaths categorized as sudden infant death syndrome on death certificates, approximately 50% (230) were categorized as missing, unknown, or undetermined by CFR. Discussion Linking death certificate and CFR causes of death provided a more robust understanding of infant causes of death in Ohio. Separately, each system serves distinct and valuable purposes that should be reviewed before selecting one system for ranking leading causes of infant mortality.