Knowledge Brief Health, Nutrition and Population Global Practice EARLY CHILDBIRTH AND UNDER-FIVE MORTALITY IN NEPAL Adenike Onagoruwa and Quentin Wodon June 2017 Child Marriage Series with Education Global Practice KEY MESSAGES:  In Nepal, about one in nineteen children die before the age of five; for children born of mothers younger than 18, the risk is higher.  Controlling for socio-economic and other characteristics, being born of a mother younger than 18 does not lead to a statistically significant increase in the risk of under-five mortality as compared to otherwise similar children born of older mothers.  Box 1: Brief and Series Primer Early childbirth is associated with health risks for children, including under-five mortality. This could be in part How is early childbirth defined? Early childbirth is defined in because young mothers often come from disadvantaged this brief as a child being born of a mother younger than 18. groups. But the age of the mothers may also play a role. Early childbearing is related to the practice of child marriage. Children of young mothers are at higher risk of Why a series on child marriage? Child marriage has significant negative impacts – not only for girls, but also for a morbidity and under-five mortality. range of development outcomes. Demonstrating these impacts will assist governments and others to make the case for Young mothers often come from disadvantaged socio- intervening to reduce the practice. economic backgrounds and early childbirths are more common in poor areas where prenatal care is limited and What are the topics discussed in the series? The series access to health facilities may be an issue. Girls who looks at the impacts of child marriage on health, population, have children early also tend to be less educated and education, employment, agency, and violence, among other may suffer from lack of agency within the household, outcomes. The welfare, budget, and non-monetary costs of child further reducing access to care. In addition, some girls marriage are estimated. Legal/institutional aspects and options to reduce the practice are also discussed. who give birth early may not yet be ready physiologically to give birth. The risk of obstetric fistula, for example, is What is the question asked in this brief? The question is: higher for young mothers. These various factors lead to How large is the impact of early childbirth on under-five mortality higher risks of under-five mortality for children born of and does it account for a large share of under-five deaths? young mothers. How large is the impact of early childbirth on under-five mortality? Does early childbirth accounts for How is the question answered? Econometric analysis of a large share of under-five deaths? These are the Demographic and Health Survey data is used to estimate the questions asked in this brief for Nepal. impact of early childbirth on under-five mortality. Page 1 HNPGP Knowledge Brief  Statistically, children from young mothers are two delivery at 18 to 34 years of age (in Nepal). The percentage points more likely to die before five. difference in risk of under-five mortality between mothers ages 18-34 and mothers above 35 is also not statistically The analysis is based on data from the 2011 significant. Demographic and Health Survey for Nepal. According to the survey, one in nineteen children die before the age of In Nepal, delivery at a young age does not increase the five (54 ‰). This represents a decline over time, since the likelihood of under-five mortality for the child in rate observed with the 2003 DHS was 61 ‰ and the rate comparison to a delivery at 18 to 34 years of age. with the 2001 DHS was 91 ‰. (Ministry of Health and Population (MOHP) [Nepal], New ERA, and ICF International Inc. 2012). Marginal effects do not change much when additional controls are used (extended model). There is thus some evidence that in the case of Nepal, after controlling for a In this brief we focus on whether all children identified in wide range of other variables, early childbirth may not the survey have died or not before reaching five years of contribute to under-five mortality, but prudence remains age, and the factors that affect that outcome. This implies relying on statistics computed in a different way from the needed when interpreting the results given the risk of official under-five mortality rate, but the idea and orders of omitted variable bias (box 2). magnitude are similar (to avoid a risk of confusion, we will Table 2: Impact of Early Childbirth on Under-five use below the term “under-five mortality” but not the term Mortality “under-five mortality rate”). As shown in table 1, 6.88 Age at first marriage Baseline Extended percent of children born of mothers younger than 18 die model model before reaching five years of age. The proportion, which Mother younger than 18 0.0201 0.0209 we refer to as under-five mortality, is 4.59 percent for Mother in 18-34 age bracket Reference Reference children born of mothers 18 to 34 years of age. For Mother older than 35 -0.00824 -0.00823 children of older mothers (35 and older), the under-five Source: Authors. mortality is at 3.30 percent. Levels of statistical significance: *** 1%, ** 5%, * 10%. Table 1: Under-five Mortality by Age of the Mother Box 2: Risk of Omitted Variable Bias Age of the mother Under-five Mortality (‰) In many countries, early childbirth appears to be positively Mother younger than 18 6.88 correlated with the risk of under-five mortality after controlling for Mother in 18-34 age bracket 4.59 other factors that may also contribute to mortality. This could Mother older than 35 3.30 indicate a causal effect. However, other variables correlated with Source: Authors. both early childbirth and under-five mortality not included in the analysis could be at the source of the correlation between early Controlling for other factors, early childbirth does not childbirth and under-five mortality. Because of the risk of omitted increase the likelihood of under-five mortality variable bias, the results cannot be considered as fully substantially. conclusive regarding a causal effect of early childbirth on under- five mortality. The difference in under-five mortality between children of young and older mothers does not necessarily imply a A number of results from the regression analysis not causal effect of the age at delivery, but it does suggest shown in table 2 are worth mentioning. The impact of the that early childbirth may contribute to under-five mortality. education of the mother on the likelihood of under-five To check whether controlling for other factors early mortality is not statistically significant. Similarly, the childbirth is indeed associated at the margin with higher impact of wealth on the likelihood of stunting is not under-five mortality, regression analysis is used (see the statistically significant, at least when measured through annex for details on the methodology). wealth quintiles. This is a bit surprising, but it could be due to the fact that households in the bottom quintiles are Table 2 provides key results with baseline and extended poor or near poor in a country like Nepal, whether a models. The interpretation of the coefficients is in terms of household is extremely poor or simply poor may not make marginal impacts in percentage terms. For example, a much of a difference on the likelihood that the child may statistically significant coefficient of 0.05 for a mother die. A few other effects are also statistically significant, as younger than 18 would indicate that children of very discussed in the more detailed study on which the brief is young mothers have a likelihood of under-five mortality based. five percentage points higher than children of older mothers, controlling for other characteristics. With the While the direct impact of early childbirth on under-five baseline specification, tables 2 indicates that deliveries at mortality is not statistically significant in Nepal, the a young age do not significantly increase the likelihood of inclusion of other independent variables as controls points under-five mortality for the children in comparison to a to the possibility of indirect effects of early childbirth on Page 2 HNPGP Knowledge Brief  under-five mortality. Because early childbirth may have an Because only a small share of deliveries are by mothers impact on other variables used as controls in the younger than 18, very few children deaths before five can regression, its overall effect on under-five mortality, be said to be directly due to early childbirth. including indirect effects through these other variables, may be larger than the (non-significant) direct effect Controlling for socio-economic and other characteristics, documented in table 2. For example, for some of the girls being born of a mother younger than 18 does not have a delivering a baby at a young age, early pregnancy may statistically significant effect on under-five mortality as have a negative effect on education enrollment and compared to otherwise similar children born of older attainment, which through the impact of education attainment on under-five mortality would lead to an mothers. This, together with non-significant indirect indirect effect of early childbirth on under-five mortality. In effects of early childbirth through education and wealth in Nepal, suggests that few children die directly or indirectly addition, early deliveries, by increasing the number of because of an early childbirth, even though some may. children that women have, may also contribute to lower labor force participation for women and thereby lower References household wealth, and in some countries higher under- five mortality (although for Nepal, wealth effects are again Bicego, G., & Ahmad, O. (1996). Infant and child mortality, not statistically significant). Demographic and health surveys comparative studies No. 20. Calverton (Maryland): Macro International Inc . Early childbirth may also affect under-five mortality indirectly, for example through lower education attainment Ministry of Health and Population (MOHP) [Nepal], New ERA, for mothers, but these effects are likely to be smaller. and ICF International Inc. 2012. Nepal Demographic and Health Survey 2011. Kathmandu, Nepal: Ministry of Health and Population, New ERA, and ICF International, Calverton, Simulations suggest that very few deaths among Maryland.. children under five are directly or indirectly due to early childbirth. Nasrullah, M., Zakar, R., Zakar, M. Z., & Krämer, A. (2014). Girl- child marriage and its association with morbidity and mortality of The last step in the analysis of the impact of early children under 5 years of age in a nationally-representative pregnancy and delivery on under-five mortality consists in sample of Pakistan. The Journal of pediatrics,164(3), 639-646. assessing the potential reduction in under-five mortality Prakash, R., Singh, A., Pathak, P. K., & Parasuraman, S. that could arise from eliminating early pregnancies and (2011). Early marriage, poor reproductive health status of deliveries. This can be done by predicting (i.e. simulating) mother and child well-being in India.Journal of family planning the likelihood that children who were born of mothers and reproductive health care, jfprhc80. younger than 18 would have remained alive if they had been born of older mothers. In other words, we are Raj, A., Saggurti, N., Winter, M., Labonte, A., Decker, M. R., considering the direct effects of the age of the mother on Balaiah, D., & Silverman, J. G. (2010). The effect of maternal under-five mortality, shifting in the data deliveries by child marriage on morbidity and mortality of children under 5 in young mothers to deliveries at a later age, and observing India: cross sectional study of a nationally representative sample. BMJ, 340. the difference that this makes for under-five mortality nationally. In the case of Nepal, because the effect of Raj, A., & Boehmer, U. (2013). Girl child marriage and its early childbirth on under-five mortality is not statistically association with national rates of HIV, maternal health, and significant, direct effects are not (statistically speaking) infant mortality across 97 countries.Violence against present. women, 19(4), 536-551. Conclusion Raj, A., McDougal, L., & Rusch, M. L. (2014). Effects of young maternal age and short interpregnancy interval on infant Early pregnancy and delivery may contribute to the risk of mortality in South Asia. International journal of gynaecology and obstetrics: the official organ of the International Federation of under-five mortality for children, directly, or indirectly. This Gynaecology and Obstetrics, 124(1), 86. brief has provided estimates of the direct impact of early childbirth on under-five mortality in Nepal using the latest DHS survey. About one in nineteen children under the age of five die; but for children born of mothers younger than 18, the risk of under-five mortality is higher by 2.3 percentage points than the risk for children of older mothers. Page 3 HNPGP Knowledge Brief  Annex: Methodological Note It is worth noting that the specifications used for modeling the correlates of under-five mortality are more parsimonious in terms There is an existing literature on the relationship between early of the independent variables included than the specifications pregnancy (as well as child marriage) and the risks of infant, used in a separate analysis by the authors for malnutrition using child, and under-five mortality. Much of the literature focuses on stunting as the main measure of interest. In principle, given that South Asia (e.g., Bicego, 1996; Adhikari, 2003; Raj, 2010; Raj et malnutrition is a key factor leading to premature death for al., 2013; Raj and Boehmer, 2013; Prakash et al., 2011; children, one could argue that all correlates of under-five Nasrullah et al., 2014). The results suggest that children born of malnutrition should also be used as correlates of under-five young mothers are indeed at higher risk of under-five mortality. mortality. The issue however is that the rate of under-five mortality in countries is much smaller than the rate of stunting. Statistics comparing under-five mortality according to the age of The models used tend to perform less well when very few of the the mother may hint at the relationship between the two. But for observations take on a positive value for the dependent variable assessing marginal impacts, regression analysis is needed. Part (i.e., dying before the age of five). In addition, when many more of the literature relies on survival models, which provide variables are used, there is also a much higher risk of perfect information on how long children survive given their correlation (prediction) between some of the variables and the characteristics. The models are censored, in that if a child has dependent variable, in which case the independent variable will survived beyond five years of age, s/he is considered as having be dropped from the model as well as the observations for which avoided under-five mortality. One may also rely on simpler logit the outcome was perfectly predicted. This calls for being or probit models to analyze under-five mortality. In this case, the somewhat parsimonious in the specifications. focus is on whether the child has survived, or not, as opposed to how long the child has survived. Given the focus in this brief on In addition to what is often done in the literature, this brief also the contribution of early pregnancies and deliveries to under-five assesses the potential reduction in under-five mortality that mortality, results from the probit regressions will be presented, could arise from eliminating early pregnancies and deliveries, or, as they provide a simple way to provide those estimates. said differently, the share of deaths for children under five that can be attributed to early childbirth according to the results. Different specifications are estimated to assess the robustness Finally, in term of interpretation, it is important to mention the of the results to changes in the econometric models. Overall, the risk of omitted variables bias, as noted in Box 3 of the brief. main results are robust to different specifications. For the baseline model, the independent variables are: (1) the age of the mother at the time of delivery by categories; (2) the child’s This brief was produced by a World Bank team as part of the Economic gender; (3) whether the child had siblings born at the same time Impacts of Child Marriage study. The synthesis report under the study (multiple birth); (4) the birth order of the child and the child’s birth was produced jointly with the International Center for Research on weight by categories; (5) the length of time between the child’s Women. The study benefitted from support from the Bill & Melinda Gates birth and a previous birth for the mother; (6) whether the delivery Foundation, the Children’s Investment Fund Foundation, and the Global took place in a health facility; (7) the location of the child by Partnership for Education. Comments from colleagues and peer reviewers are gratefully acknowledged. The opinions expressed in this region and by urban-rural category; (8) the education of the brief are those of the authors only and need not reflect the views of the mother; (9) whether the household has access to an improved World Bank, its Executive Directors, of the countries they represent. water source and improved sanitation; (10) the wealth quintile of the household; (11) whether the household practices polygyny; and finally (12) indicators of decision-making power for the mother in the household. The Health, Nutrition and Population Knowledge Briefs of the World Bank are a quick reference on the essentials of specific HNP-related topics summarizing new findings and information. These may highlight an issue and key interventions proven to be effective in improving health, or disseminate new findings and lessons learned from the regions. For more information on this topic, go to: www.worldbank.org/health. Page 4