Breastfeeding Behavior Differs Significantly by Insurance Status at an Urban University Hospital [78

Abstract

INTRODUCTION: Although breastfeeding rates have been increasing nationally since 1990, lower rates of breastfeeding are associated with certain demographic factors. To determine whether insurance status is an independent predictor of breastfeeding behavior, we performed a retrospective cohort study in a mixed-race urban population. METHODS: We identified women delivering singleton term neonates at Thomas Jefferson University Hospital between July and December 2013 from labor and delivery records. Self-reported breastfeeding status was abstracted from medical records at postpartum follow-up and was analyzed as exclusive breastfeeding, any breastfeeding, or no breastfeeding. Data were analyzed with logistic regression to compare rates of breastfeeding between women with Medicaid and those with private insurance. RESULTS: Data for 656 women were reviewed; 405 women who completed postpartum follow-up within 8 weeks and had feeding methods recorded were included in the study. Rates of breastfeeding were significantly lower in the Medicaid population (any breastfeeding, odds ratio [OR] 0.42, 95% confidence interval [CI] 0.29–0.65). A significant difference remained after accounting for interaction and confounding by race, age, parity, and marital status (any breastfeeding, OR 0.53, CI 0.29–0.77). When stratified by race, white women on Medicaid had the lowest probability of breastfeeding (P<.01 of interaction by race). CONCLUSION: Among patients delivering at an urban academic hospital, women on Medicaid were significantly less likely to breastfeed than were those with private insurance. Although African-American women had the lowest breastfeeding rates overall, the greatest differential by insurance was among white women. This suggests that efforts to improve breastfeeding should focus on low-income women of all races.