All live births that occur <20weeks contribute to Infant Mortality Rates (IMRs) as all result in death. We aim to describe racial variation in US IMRs attributed to infant deaths among previable (<20week) live born infants.METHODS:
Population-based retrospective cohort study of all US live births from 2007-2014, using CDC WONDER database’s Linked Birth/Infant Death Records. Data were stratified by maternal race and gestational age (GA) at birth. We compared the proportion of infant deaths within previable and periviable/viable (≥20weeksGA) categories. The US reported IMR (births at all GAs including <20weeks) was compared to an adjusted IMR (≥20weeks), and the difference (≥IMR) was calculated for each race/ethnicity.RESULTS:
Previable infant deaths represented 4.12%, 7.69%, and 4.99% of total deaths for non-Hispanic White, non-Hispanic Black, and Hispanic groups, respectively. The contribution of previable deaths to the total IMR of these racial/ethnic groups are .21, .89, and .26 per 1000 live births, respectively, P<.0001. Adjusted IMRs of the groups are 4.98, 10.85, and 4.69 deaths per 1000 live births.CONCLUSION:
There are significant racial/ethnic differences in the relative contribution of previable births to each group’s IMR in the US. Non-Hispanic Black mothers had the largest number of previable births and the largest IMR if a lower GA cutoff of 20 weeks is considered. One potential method of standardizing IMR on the regional, national, and international levels is to impose a minimum gestational age for infant mortality calculations. However, this approach may obscure the disproportionate contribution of previable births to IMRs among the non-Hispanic black population.