Determinants of infant and child mortality in Bangladesh

Abstract

Using the 1975-76 Bangladesh Fertility Survey d a ta and employing logit linear model analysis, this study examines first the effects of the three widely discussed demographic factors (maternal age at birth, birth order and preceding birth interval) on infant and child mortality in Bangladesh. It then examines how social factors, particularly mother’s education, could relate to child survival through the three so-called proximate determinants: mother’s age at birth, parity and birth interval in the framework for the analysis of child survival in developing countries suggested by Mosley and Chen (1984). The intra-family mortality relationship, th a t is, the influence of the survival sta tu s of the immediately preceding sibling(/s) on the mortality risk of the child in question is also investigated. The study also examines the effects of duration of breastfeeding together with the effects of birth intervals (both preceding and following) on the risk of dying at different stages of early life. Multivariate analyses reveal th a t both infant and child mortality risks are higher if the mother is under 20 years of age. The birth order of the child shows a positive relation with the risk of dying between ages one and Five years but higher mortality risks a t the extreme birth orders are observed during the period of infancy. Children born with a preceding birth interval duration of one-and-a-half years or less have considerably higher risks of dying in infancy as well as at ages between one and five years. A considerably higher risk of dying between the first and fifth years of life is also observed among children followed by another birth within one-and-a-half years. The analyses show th a t both infant and child mortality risks are significantly higher among mothers with no education than among mothers with some primary education. The b etter survival chance of children whose mothers had some education has resulted neither through maternal age a t birth nor the interval between successive births, but because the mothers had fewer births. This study also shows th a t the survival sta tu s of a child at age one year is one of the most important factors influencing the survival sta tu s of its next sibling. However, the preceding child’s survival has different roles during the period of infancy and at ages beyond infancy. The group of children with the immediately preceding sibling dead by age one year has a significantly higher risk of dying in infancy than the group with the preceding sibling alive. On the other hand, at ages between one and five years, children with the preceding sibling dead have a significantly lower risk than children with the preceding sibling alive. The same p a ttern has been found when the analysis has been extended to the survival sta tu s of the two immediately preceding siblings. While inadequate biological or reproductive conditions of the mother might explain the former result, the la tte r result is likely to relate to competition for family resources. The study further shows th a t the sex of the immediately preceding sibling is a significant factor influencing child mortality. The group of children preceded by a male sibling has a considerably higher risk of dying between the first and the fifth years of life th an the group preceded by a female sibling. This result is also likely to be related to competition*, the child in question faces harder competition from a preceding brother than from a preceding sister. Given the role of young girls in the society, it is also likely th a t a child is more likely to be cared for by its immediately preceding sister than by its immediately preceding brother. Based on the sample of last and next-to-last live births, the analysis of breastfeeding establishes it as one of the prime factors influencing mortality during the early months of life. However, no support for the hypothesis th a t breastfeeding is the intermediate factor through which preceding birth or subsequent pregnancy influences child survival can be obtained from the d ata. The analyses suggest th a t preceding birth interval, timing of the next conception and breastfeeding duration each influences mortality in its own way, but their relative importance as determinants of mortality varies with the age of the child.