Correlates of Short Interbirth Intervals in Peninsular Malaysia: Their Pathways of Influence Through Breastfeeding and Contraceptive Use

Abstract

Recent research has shown that children born before and after short birth intervals run a considerably greater risk of dying in infancy or childhood than do others. This report investigates which women have short interbirth intervals, under what circumstances, and for what reasons. The analysis uses data from the Malaysian Family Life Survey to examine influences on the two main behaviors--breastfeeding and contraceptive use--that affect birth interval length, and assesses the the impact of these same variables on the probability of having a birth interval of less than 15 months. The analysis shows that many of the independent variables affect breastfeeding and contraceptive use in opposite directions, with no significant net effect on the likelihood of a short interval. For example, a woman's education is negatively related to the probability that she breastfeeds, positively related to the probability that she uses contraceptives, and has no significant effect on the likelihood that the interpregnancy interval is less than 15 months. Having a family planning clinic nearby is associated with less breastfeeding, offsetting whatever positive effects family planning clinics have on contraceptive use in terms of the percentage of birth intervals that are so short as to be detrimental to infant and child health. Hence, factors that increase contraceptive use do not necessarily reduce the incidence of short interbirth intervals, because they are also associated with reduced breastfeeding. We simulate the proportion of intervals that would be short for alternative combinations of breastfeeding and contraceptive use in the population and show that over the period covered by the data (1961-75), breastfeeding had a considerably greater effect on preventing short interbirth intervals than did contraceptive use.

Contraceptive Use, Birth Spacing, and Child Survival in Matlab, Bangladesh

“…This suggests that the prospective data from Matlab, Bangladesh, might provide a good alternative for randomized controlled trials (Philips et al 1982). Several studies using Matlab data investigated the determinants of infant mortality (DaVanzo and Starbird 1991;DaVanzo et al 2007 andHale et al 2009). These studies did not assess explicitly the magnitude of the effect of contraceptive use, however.…”

How Contraceptive Use Affects Birth Intervals: Results of a Literature Review

“…In some cases, a dramatic difference is observed-for example, birth-interval lengths increasing by 20 months (Janowitz and Nichols 1983). Several exceptions to this trend are found, however, notably in the studies by Ngianga-Bakwin and Stones (2005), who find that users had slightly higher odds of having short birth intervals; by Feyisetan (1990), who finds no significant difference in interval length among users and nonusers of contraceptives; and by DaVanzo and Starbird (1991), who observe that although contraceptive use reduces the odds of short birth intervals, the interaction between contraception and breastfeeding leads to dramatically increased odds of short birth intervals. Although these findings are noteworthy and deserve critical attention, the other seven multivariate analyses and the four univariate analyses show a positive association between the practice of contraception and longer birth intervals.…”

“…Bumpass and his colleagues (1986) find a clear relationship between contraceptive use and longer birth intervals in Korea. Analyzing data from Malaysia, DaVanzo and Starbird (1991) find contraceptive use to be protective against dangerously short birth intervals of 15 months or less; they note, however, that this effect is reversed by concurrent breastfeeding. They show that the interaction of breastfeeding and contraceptive use leads to dramatically increased odds of having short birth intervals.…”

“…Five articles presented results from multivariate regression analyses measuring contraceptive use compared with nonuse within a predefined "short" interval category. DaVanzo and Starbird (1991), Stevens-Simon et al (1999), andNgianga-Bakwin andStones (2005) apply logistic regression methods, Bumpass and his colleagues (1986) use multivariate life-table methods, and Sathar (1988) applies multiple classification analysis. 5 The cut-points of "short" birth intervals range between 15 and 24 months.…”

“…Although an observational design is, perhaps, inevitable, ideally it should be prospectively implemented. In this regard, prospective pregnancy-history data from Matlab, Bangladesh, have been analyzed in a number of studies concerning the determinants of infant mortality (DaVanzo and Starbird 1991;Hale et al 2006;DaVanzo et al 2007 and. These studies do not assess explicitly the magnitude of the effect of contraceptive use, however.…”