Unacceptably high childhood mortality is evident in many developing countries including Bangladesh. Parents poor socioeconomic background in our country like their counterparts in other underdeveloped or developing ones do ascribe such painful death events more to their ill-luck than to their real causes because of ignorance, illiteracy, carelessness, absence of essential healthcare services or their non-affordability to seek such care or service. Just as painful for the parents, such death are matters of great concern for a nation as well. Childhood deaths whittle away a nation’s prospective manpower resources. But the fact remains that such new-born mortality is in most cases preventable.

This study attempts to identify the factors responsible for high childhood mortality rates and at the same time points out where and why such rates are higher or lower. This study owes much of its data to the 2007 BDHS survey under the National Institute of Population Research & Training (NIPORT). Besides estimating the mortality rates, attempts have been made to uncover the risk factors associated with the neonatal, infant and under-five deaths. The study has at the very outset, resorted to the chi-square test in the bivariate section for identification of risk factors. Then in the multivariate section, the well-known Cox’s hazard model has been used to determine the effect of the selected variables on the reference deaths.

According to the analysis, death of male children outnumber female by 4 (42-38) in the case of the neonatal, by 4 (62-58) in the case of infant and by 2 (77-75) in the case of the under-five, per thousand live births. Evidently therefore child mortality rate is higher for males. This study finds that children born to comparatively older mother outlive those born to younger ones. In the neonatal case, children born to mother aged less than 30 years are 1.88 times likely to die during first month of their birth as compared to children born to mothers aged 20-29 years. Also childhood mortality is negatively associated with birth intervals. Mortality rate is higher at short birth interval than at long birth intervals. The study further reveals that child mortality is higher at first birth as well as the seventh order or above. In the multivariate analysis, division presents itself as a significant factor for only under-five mortality. For other two mortality, age groups, i.e. neonatal and infant, division is not a significant factor. Mother’s education level significantly affects childhood mortality. Children born to educated mother are less vulnerable to deaths at ages related to three groups under study. The study further shows that, children born to illiterate mothers are more than 1.5 times more likely to die before completion of first month after birth compared to that of mothers from higher education group. Women and households in the BDHS data have been categorized into different socioeconomic levels using an index of household assets. The study depicts that mortality rate for neonatal, infant and under-five is the highest (41, 60 and 78 deaths per 1000 live births respectively) among the poorest asset groups and the lowest (25, 37 and 44 deaths per 1000 live births respectively) among the richest asset groups. Multivariate analysis shows that variables like asset quintile, current age of mother, previous birth interval, birth order, mother’s education level etc are also significant in this regard. Most of the significant factors found in the multivariate analysis are demographic ones.