Time of birth, breast milk feeding and health outcomes in the neonatal intensive care unit

A foundational assumption of many infant health policies is the health advantage of breastfed infants over formula fed infants. However, because of ethical and practical concerns, evaluating the efficacy of breast milk feedings on infant health outcomes is difficult. It is impossible to randomize breast milk feeding among infants; further, full-term infants face environmental and familial confounders that cloud the true impact of breast milk. Premature infants are a population whose environment is strictly determined by hospital policy, and whose health inputs and outputs are rigorously monitored. For these reasons, this study uses a three-year sample of 90% of admits to California Neonatal Intensive Care Units (NICU) that includes obstetrical diagnoses, birth characteristics, incidence of adverse health outcomes and long-form vital statistics data. This is the first paper in the literature to apply econometric techniques to this particular, exhaustive data set. Past research suggests that NICU admits are particularly susceptible to the health benefits of breast milk feedings, particularly with gastrointestinal outcomes. Past research by this author and others also suggests that breast milk feeding is more likely among White, affluent and educated mothers. Therefore, this study employs an instrumental variables (IV) approach that exploits the random variation in hour of birth as an instrument for breast milk feeding. The author’s past qualitative research implies that infants born in the middle of the night are exposed to less-educated—in terms of supporting breast milk feeding and ensuring immediate expression of breast milk—clinicians and a lack of certified lactation support. The preliminary first-stage results confirm the qualitative findings—infants born between 10pm and 4 am are less likely to be receiving breast milk at discharge home, adjusting for delivery type and a full range of controls. Preliminary second-stage results indicate that infants fed breast milk are less likely to experience serious gastrointestinal illnesses compared to formula-fed infants. Breast-fed infants also had significantly longer rates of stay. No significant differences have been found in respiratory illnesses or growth rate. The policy implications of the first-stage results are discussed, as are the cost implications of longer lengths of stay in the NICU.