Internationally-embarrassing infant mortality rates, particularly among specific populations, prompt this investigation into factors across the heterogeneous United States population that correlate with infant health. Does the probability of an infant's surviving the first year of life differ across Medicaid policy levels? This question is assessed by a logistic regression model testing for unique effects of Medicaid policy differences on infant survivability, after controlling for individual maternal and infant factors. Newly available Vital Statistic data linking U.S. infant death records with their corresponding birth records provide individual information, far more accurate than previously available rates, for infant mortality by race, maternal education, use of early prenatal care, and birthweight. As there are no nation-wide data connecting specific infant deaths with Medicaid use, policy differences are tested by attaching aggregate data to each case according to the maternal State of residence. Measures of the poverty health policy environment in which the mother lived at the time of her infant's birth are: (a) whether or not the Medicaid program included the medical need option, (b) the percent of poverty population covered by Medicaid and (c) total Medicaid expenditures for the year per AFDC adult. In addition, an aggregate measure controls for the general poverty level environment. A second question is then addressed: For the infant mortality population, does Medicaid availability relate to age at death? A strong correlation with postneonatal mortality indicates the influence of poverty factors, while association with neonatal mortality stresses the importance of health policy, specifically availability of prenatal care and medical technology, to infant survival.
