"To What Extent Does Breastfeeding Explain Birth-Interval Effects on Early Childhood Mortality?"

Breast-feeding and child survival in Matlab, Bangladesh

“…The same holds true for the effect of previous birth interval suggesting that each of the covariates supplementation, previous birth interval and subsequent conception has its independent influence on childhood mortality. This contradicts the finding of Retherford et al (1989) which used WFS data from Nepal, possibly because the fate of the immediate preceding child is defined somewhat differently in the present study. Models 4 and 5 both show that children whose preceding sibling was dead at exact age 1 year experience higher mortality during childhood than children whose immediate preceding sibling was alive at that age.…”

“…Using World Fertility Survey (WFS) data, Palloni & Millman (1986) found that breast-feeding did not explain the effect of the preceding birth interval on early childhood mortality, but did partly explain the effect of the following birth interval. The WFS data from Nepal suggest that all the influence of the subsequent birth interval is through the cessation of breast-feeding (Retherford et al, 1989). Using Bangladesh Fertility Survey data Majumder (1989) reported that each of the covariates, breast-feeding, preceding and following birth interval, has an independent influence on early childhood mortality.…”

Covariates of infant mortality in China: An exploratory approach

“…In contrast, an earlier log-linear hazards model analysis of the covariates of infant and child mortality using the World Fertility Survey in Sri Lanka by Trussell and Hammerslough (1983) cut the maternal age into three groups (10-19, 20-34, and 35+) and found that death rates among children under 10 born to women less than 20 years of age were substantially (29 per cent) higher than for children born to women 20-34 years of age, as were death rates of children born to women aged 35 and over (21 per cent higher). When maternal age was treated as an interval variable, a U-shaped association of maternal age with child mortality was strongly supported in the hazards model analysis of child mortality in Guatemala by Pebley and Stupp (1987) and in Nepal by Retherford et al (1989).…”

“…Many previous studies have a Ushaped relationship between mother's age at the birth of child and infant mortality (Pebley and Stupp, 1987;Retherford et al, 1989;Koenig et al, 1990). This finding is strongly confirmed by the results from this study.…”

“…One of the methodological problems with application of these approaches has been the inadequate consideration of coding the explanatory variables in multivariate analysis. For example, mother's age at birth of child, an important determinant of infant mortality, is sometimes treated as a categorical variable (Trussell and Hammerslough, 1983;Hobcraft et al, 1985;Pebley and Stupp, 1987;Miller et al, 1992) and sometimes as a linear or quadratic interval variable (Pebley and Stupp, 1987;Retherford et al, 1989;Koenig et al, 1990), but justification was seldom given why a particular coding was chosen. Frequently, the choice of coding of such independent variables is done empirically or subjectively.…”