Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

The recent uptick in infant mortality has sparked much concern among policy makers and public health officials. However, there is relatively little research examining the cause of the increase in infant mortality rates after decades of falling mortality rates. There are a number of clinical pathways through which maternal obesity can be associated with an increase in infant mortality, most notably through the greater difficulty of detecting fetal congenital anomalies. However, obese women are more likely to be poor and have other conditions that are associated with poor birth outcomes. Hence a strategy to deal with this difficult confounding problem is essential. We employ an instrumental variables approach using male obesity rates in the same area to control for local factors associated with female obesity rates that are plausibly uncorrelated with unobserved factors affecting infant mortality. We use data on obesity rates calculated from the Behavioral Risk Factor Surveillance System (BRFSS); data on infant mortality come from the National Vital Statistics System (NVSS); data on births come from the Natality Data. We also use data from the Linked Birth and Infant Death Data. Our preliminary results suggest that state-level obesity rates for women of child-bearing age is predictive of infant mortality rates for congenital anomalies though not for all-cause infant mortality. Encouragingly for our instrumental variables strategy, male obesity rates are also predictive of increases in infant mortality rates for congenital anomalies.