Breast-feeding pattern and its impact on post-partum non-susceptible period of currently married women(15-49) in an Indian state using multivariate hazard model and life table techniques

Shankar Dihidar, Indian Statistical Institute

Determination of breast-feeding patterns along with its duration and its variation among different subgroups of population is a vital activity. A number of studies have established the role of breast-feeding in reducing the child mortality, fertility to mothers and its impact on post partum amenorrhoea(PPA). In this context our mission is two-fold. Using National Family Health Survey, 1998-99 data for West Bengal, an Indian State, we intend to study differentials in early initiation of breast-feeding and in providing the first milk from breast and breast-feeding pattern among women belonging to various socio-cultural groups. Secondly, we determine the differentials in PPA among those women and find out the role of biological and socio-cultural variables on resumption of post-partum menstruation using Cox’s proportional hazards model and life table techniques. In this study it has been observed that only one third of the children began breastfeeding within one hour of birth and no more than 49 percent of women in any group and about half of the children began during their first 24 hours of life. In case of giving colestrum, as many as 78 percent of the women who breastfed had squeezed the first milk from the breast before they began nursing their babies. Results from multivariate hazards analysis reflect the relative risk of return of menses at time t of women who are breastfeeding at time t compared to women who are not breastfeeding at that time is 0.43 if 0 < t < 8 months and 0.99 if 8 < t < 12 months. Thus, breastfeeding during first eight months significantly influences the duration of PPA as at time point during the interval the rate of return to menses is reduced by 57 percent if the woman is breastfeeding at that time compared to the rate of women not breastfeeding at that time