This paper analyses the effects of timing and type of breastmilk supplementation on infant mortality using a discrete time analog of continuous time proportional hazards model. The data are taken from a prospective study which collected information on timing and type of breastmilk supplementation, mother’s nutrition at various stages of the study, and other associated factors of fertility and infant mortality.

The statistical analyses show that breastfeeding type at various stages of the child’s life is a significant predictor of infant mortality, even when variables such as mother’s sex, mother’s education, religion and SES are included in the model.

The study shows that infants breastfed at birth have better probabilities of survival relative to those who are never breastfed or are given liquid supplements very early in life. This effect remains significant even when mother’s nutrition at childbirth, which is used as proxy for birth weight, is controlled.