Birth Spacing, Breastfeeding Pattern and Child Survival in Bangladesh

Abstract

Birth spacing is a major determinant for evaluating family building process due to its direct relationship with fertility. The mean duration of successive birth intervals is obviously related to the fertility rate, the longer the interval consequently the lower the fertility. Breast feeding and post-partum amenorrhea (PPA) have also been a matter of rapidly growing interest, as they are important not only for maternal and child health but also for its fertility reducing effect through PPA. Hence for a complete understanding of the process of family building, in Bangladesh, it is imperative to analyze the birth spacing, breastfeeding pattern and PPA as well as their differentials and determinants in context of female population of Bangladesh. This study also examines the influence of birth spacing on child survival. The data used for the completion of this work, is extracted from the Bangladesh Demographic and Health Survey (BDHS) conducted in 2003-2004. The BDHS recorded enormous data on complete birth history of 11440 ever married women of the age group 10-49, which are very useful for studying birth interval. The study considers first to fifth birth intervals as it covers most of the range of fertility experience of Bangladeshi women. Since data on breastfeeding were available only for the last child, the study of breastfeeding and post partum amenorrhea is based on information for last birth. The study reveals that the distribution of first to fifth birth interval is largely positively skewed lying somewhere between 12 to 23 months for first births and for subsequent birth in the duration 24 to 35 months. Both Chi-square statistic and Cox proportional hazard model demonstrate that age, women age at marriage, couples education, respondent work status, residence, socio-economic status, contraceptive use and watch TV have significant influence on first birth interval but subsequent birth showed miscellaneous results. Moreover, survival status of previous child showed highly significant effect on second to fifth birth intervals. But the variable religion has no significant orientation at all. To study the quantum and tempo of fertility life table technique is employed and demonstrates that the mean birth interval for marriage to first birth is 24 months but for subsequent birth the interval lies between ranges 30 to 32 months. The differential analysis of quantum and tempo of fertility reveals that educated women have shorter first but longer subsequent birth interval. Urban mother have shorter first birth interval is lower but longer subsequent birth interval is higher than rural mother. The birth interval is higher (25 months) when age at marriage is less than or equal to 15 years, but lower (21 months) when age at marriage is 15 years and above. But for higher order births age at marriage has no differential effect. The summarized results speculate that child survival status affects timing of birth not only for first two or three orders but also for higher orders, but at a smaller pace, for women whose previous child is death birth interval is much lower than those whose previous child is alive. Determinants and differential of breastfeeding and post partum amenorrhea are also performed using life table analysis and Cox Proportional Hazard model. The results indicates that breastfeeding is virtually universal (98.3 percent) and homogeneously prolonged in Bangladesh. The mean duration of breastfeeding is about 32 months. To see the effects of socio-demographic variables on breastfeeding we fitted three model- the first one to see the effect of demographic variables, second one to see the effects of socio­economic variables and lastly to see the combined effect of these two variables. The study results divulge that demographic variables have more influence than other variables. Finally the proportional hazard analysis has identified that administrative division, Religion, maternal education, working status, current age, age at marriage, parity; Contraceptive use and place of delivery have significant effect on the duration of breast feeding. The overall mean duration of PPA was found to be 8.51 months and the length increases with increase in the parity. The mean length of PPA significantly varies by place of residence, region, mother's education, work status, sex of previous child, breastfeeding status, parity, delivery status and age of mother. The Cox proportional hazard model suggested that duration of breastfeeding has strong positive influence on duration of PPA among all included explanatory variables. The study also analyzed the relationship between the length of preceding birth interval and child survival and their effect on age specific probability of death of index child. The preceding birth interval and child survival are significantly correlated and probability of survival is much lower for less preceding birth interval (<12 months) and also a lesser extent at higher birth interval (84+ months). To see the effect of preceding birth interval and socio-demographic factors on child survival status four separate logistic regression models (neonatal, post neonatal, mortality between 12-35 months and child mortality) are fitted. These models reveal that among the socio-demographic variables, preceding birth interval, breast feeding status and mother's education has strong significant effect on mortality but the variables birth order and mother's age at birth has little or no significant effect. Therefore, the study results emphasize that it is need to encourage women to have longer birth interval, not only to limit family size, but also to guarantee good health of mother and the child. Education is a consistently a dominant factor and such formal and informal education should encourage women to differ marriage and prolong breast feeding.