Knowledge for Development Policy Brief: Malawi May 2016 Education Global Practice Early Childhood Development: Health, Nutrition, and Population Global Practice Situation Analysis for Malawi Anne Bakilana, Corrina Moucheraud, Christin McConnell, and Rifat Hasan Teenage pregnancies have potential nega- child mortality rates have been declining tive consequences for the next generation. steadily since 1992 (figure 1a), children born Children born to adolescent mothers are par- to adolescent mothers still experience higher ticularly at a disadvantage with regard to their mortality than children born to mothers in health, nutrition, cognitive, and socioemo- their 20s and 30s (figure 1b). tional development. The early years, especially The decline in neonatal mortality the first 1,000 days, are crucial for lifetime (deaths within 28 days of birth) has not health, learning, and productivity. Particularly been as fast, as it declined from 41 deaths for the most vulnerable children and families, per 1,000 live births in 1992 to 31 in 2010. early childhood development (ECD1) is a high- Similar to infant and childhood mortality, return investment. This policy brief presents neonatal mortality is higher for babies born evidence on the health, nutrition, and overall to mothers under 20 years of age than for development of children in Malawi with a babies born to mothers in their 20s. Even focus on those born to adolescent mothers. after controlling for key demographic and so- cioeconomic variables, the odds of mortality Trends in ECD are approximately 70 percent higher for in- Health and Nutrition fants born to adolescent mothers than for in- Mortality rates among children are declin- fants born to older mothers; similar elevated ing in Malawi. Infant mortality (deaths of odds exist for deaths of children under 5 children under the age of 1 year) and child- years of age by mother’s age.2 hood mortality (deaths of children under Children born to women with higher edu- the age of 5 years) have fallen substan- cational attainment have lower mortality tially and are expected to continue to de- before age 5 than children born to women cline. The infant mortality rate in 2010 was with less education (figure 2)—but as mor- an ­ estimated 66 deaths per 1,000 live births, tality rates declined over time, the gap in a decline from 135 in 1992, while under-5 neonatal and infant mortality between mortality rate is an estimated 112 deaths per women with different education levels, has 1,000 live births—down from 234 in 1992. considerably narrowed. In 2010, the infant Data show that while neonatal, infant and mortality rate was 71 per 1,000 babies born to 1   ECD is defined as an orderly, predictable process along 2   Results of a multivariate regression model using a continuous path, in which a child learns to handle more 2013–14 MICS data on whether infant (under-1) and complicated levels of moving, thinking, speaking, feeling, under-5 mortality are associated with mother’s and relating to others. Physical growth, literacy and adolescence, ethnicity, religion, region of residence, numeracy skills, socioemotional development and urban or rural residence, household wealth quintile, level readiness to learn are vital domains of a child’s overall of mother’s educational attainment, sex of the child, development, which is a basis for overall human twinship, and birth order. The full findings are available development (MICS 2014). from the authors upon request. 1 Figure 1. Under-5 Mortality Rates, Overall and by Mother’s Age at Birth, 1992–2010 a. Neonatal, infant and under-5 mortality b. Neonatal, infant, and under-5 mortality rates, rates, overall 1992–2010 by mother’s age, 2010 250 160 140 Number of deaths per 1,000 live births Number of deaths per 1,000 live births 200 120 100 150 80 100 60 40 50 20 0 0 lit t S S S S lit l ta r-5 ta n ta ta DH DH DH DH or Infa y or a y y or e m eon lit m Und 00 04 10 92 20 20 N 20 19 m Mother's age at birth Mother's age at birth Neonatal mortality <20 20–29 30–39 40–49 Infant mortality rate Under-5 mortality rate Source: Demographic and Health Surveys. Note: Neonatal mortality is death in the first 28 days of life; infant mortality is death before one year of age. Figure 2. Infant and Under-5 Mortality in Malawi, Risks for poor nutrition are higher among by Mother’s Level of Education, 1992–2010 children born to adolescent mothers 138 Children born to adolescent mothers are 2010 mortality rate 94 129 at higher risk for stunting, especially se- Under-5 2000 vere stunting, which can affect cognitive 1992 development and educational outcomes in later childhood and adolescence.1 These higher stunting rates may be explained by the Infant mortality 71 2010 65 76 strong intergenerational relationship between rate 2000 stunting and poor maternal health and nutri- 1992 tion before, during, and after pregnancy. The prevalence of moderate and severe stunting Neonatal mortality 29 is 42% for children under five (MICS 2013–14), 2010 35 31 but the odds of being stunted are approxi- rate 2000 mately 12 percent higher for a child born to 1992 an adolescent mother than for a child born to a mother over age 20.3 In addition, a higher 0 50 100 150 200 250 300 percentage of children born to women with Number of deaths per 1,000 live births no education were stunted (47 percent) or No education Primary Secondary or higher severely stunted (19 percent), compared to ­ Source: Demographic and Health Surveys those born to women with secondary educa- tion or higher levels of education—where women with no education, compared with 3   Results of a multivariate regression model using 65 per 1,000 for babies born to women with 2013-14 MICS data on whether moderate stunting is at least a secondary education; the under-5 associated with mother’s adolescence (child’s birth mortality rates for these groups were 138 per between ages of 15 and 19), ethnicity, religion, region of residence, urban-rural residence, household wealth 1,000 and 94 per 1,000, respectively; and the quintile, mother’s level of educational attainment, and neonatal mortality rates for these groups were sex of the child. The full findings are available from the 29 per 1,000 and 31 per 1,000, respectively. authors upon request. 2 Figure 3. Stunting of Children in Malawi, by Mother’s Age, 2013–14 a. Moderate and severe stunting b. Severe stunting 50 50 % of children with height-for-age z score less % of children with height-for-age z score less than –3 standard deviations below normal than –2 standard deviations below normal 45 45 41 40.2 40 40 35 35 30 30 25 25 20 20 15.3 15 15 13.9 10 10 5 5 0 0 Mother 15–19 Mother 20+ Mothers, ages 15–19 Mothers, ages 20+ Source: MICS 2013–2014. 33 percent were stunted and 10 percent were prevalence of recent diarrhea and cough severely stunted. among children born to adolescent mothers (versus those born to older mothers) (p-value Children born to adolescent mothers < 0.05 for both), but no significant difference also face illness more often compared for fever. After controlling for important fac- to other children tors, analyses found that children of adoles- Overall, diarrhea prevalence has barely cent mothers are significantly more likely to changed since 1992 (from 22 percent ac- have experienced a recent case of diarrhea cording to the 1992 DHS, to 18 percent in the (approximately 17 percent higher odds than 2010 DHS, and 25 percent in the 2013–14 for children of older mothers), but found no MICS); cough prevalence rose considerably significant relationship for recent cough or during the 2000s before falling again (14 per- fever.5 cent in 1992, 27 percent in 2000, 19 percent Seeking medical treatment for diarrhea in 2004, 7 percent in 2010); and fever preva- and for cough decreased nationwide dur- lence has declined slightly, from 40 percent in ing the 2000s, but has since increased 1992 to 32 percent in 2012 (MICS). According (figure 4). In 2010, 62 percent of children to the 2013–14 MICS, a higher percentage with a recent case of diarrhea reportedly had of children born to adolescent mothers had received care at a health facility (80 percent fever, diarrhea, or cough than children born to received either medical care or ORS); among older women: 38 percent versus 37 percent; children of adolescent mothers with a recent 29 percent versus 24 percent; and 45 percent case of diarrhea, 66 percent had received versus 41 percent. These differences were medical care (80 percent received either statistically significantly for diarrhea and medical care or ORS). There are similar rates cough (but not for fever), even after control- of seeking care for recent cases of acute re- ling for important factors in multivariate spiratory infection (fever or cough): 61 per- analyses: children born to adolescent moth- cent for all mothers and 63 percent for ers saw 18–20 percent higher odds of recent adolescent mothers. diarrhea or fever than children born to older Multivariate results, however, show that mothers.4 Similarly, recent DHS (2004, 2010) there is no significant relationship between have also identified a significantly higher adolescent motherhood and seeking care for diarrhea or respiratory infection, suggesting 4   Results of a multivariate regression model using 2013–14 MICS data on cough in the past two weeks (or 5   Results of a multivariate regression model using a fever or diarrhea modeled separately) were associated pooled cross-sectional dataset on cough in the past two with whether mother was an adolescent at the time of weeks (or fever or diarrhea modeled separately) were the child’s birth, woman’s educational attainment, associated with adolescence, woman’s educational household wealth quintile, religion, ethnicity, region of attainment, household wealth quintile, marital status, residence, urban/rural residence, child sex, and some religion, region of residence, urban/rural residence, child relevant behaviors (such as ownership of a bed net to sex, birth order, survey year and some relevant prevent fever, and improved water and sanitation for behaviors (such as use of a bed net to prevent fever, and diarrhea). The full findings are available from the authors improved water and sanitation for diarrhea). The full upon request. findings are available from the authors upon request. 3 Figure 4. Treatment-Seeking for Childhood Learning, Cognitive Development, and Illnesses in Malawi, 1992–2010, 2013–14 Achievement 80 Poor health and nutrition adversely affect learning, cognitive development, and recent diarrhea, or cough, who were taken to 70 70 67 % of children (under 5 years) who had 62 65 achievement. Among children enrolled in a health facility for treatment 60 54 Standard 1, 22 percent of them are 8 50 49 years old or older, which sets them on a challenging learning trajectory (EMIS 2014– 40 36 37 15). High repetition in early grades,8 stunting 28 27 30 or lack of maturity (parents may consider 20 children “too small” or “not ready” to start 10 Standard 1 and/or walk long distances to pri- mary school), lack of required documentation 0 to register in time for the school year, or lack S S S 4 S S of information among parents about the im- DH DH 3–1 DH DH DH 1 00 04 10 92 20 portance of starting school at the right age 20 20 20 19 Year of survey are some of the reasons for a delayed start in Diarrhea in the two weeks preceding survey schooling. Late entry has many negative im- Cough in the two weeks preceding survey plications, including higher likelihood of drop- ping out or repeating grades later. that factors other than age may be more im- Few Malawian children get preschool portant in motivating mothers to seek care.6 experience, as only 28 percent of children For example, there may be differences in the entering primary school have had preschool quality of care received by teenage mothers or experience (MoGCSW 2015); and only 39 per- in the timeliness with which they seek care, cent of children aged 36–59 months were at- which may in part explain poorer health out- tending early childhood education, with large comes for their children. Although they seek variations by socioeconomic characteristics treatment for their children’s illnesses, for (MICS 2014). The Ministry of Gender, Children, some sicknesses, adolescent mothers are less and Social Welfare (MoGCSW) reports that likely to take preventive steps to avoid illness. there are more than 11,000 early childhood For example, antenatal tetanus toxoid vacci- centers in Malawi that serve 1.4 million chil- nation protects the mother from contracting dren (2015). Over 8,000 of these preschools tetanus, prevents preterm birth, and protects are public community-based childcare cen- the newborn. In Malawi, it is relatively uncom- ters (CBCC), mostly catering to children in mon for adolescent mothers to receive ante- rural areas and opening for a few hours each natal tetanus toxoid vaccination. In 2010, only weekday. Most CBCCs are volunteer-run by 6 percent of adolescent mothers had been untrained caregivers and chaired by a com- vaccinated (versus 13 percent of women age munity-based organization; some also receive 20–34 at time of birth, and 18 percent of nongovernmental organization support. The women 35–49). Use of bed nets is also low quality and reliability of services to children for children of adolescent mothers. Only 52 vary greatly—a 2011 survey of CBCCs in 4 percent had slept under a bed net during the districts found only 53 percent of those previous night in 2013–14, compared to 53 listed in the government registry were opera- percent nationwide—and after adjusting for tional during unannounced visits (Neuman socioeconomic and demographic factors, chil- et al. 2014). CBCCs close for a variety of rea- 22% dren of adolescent mothers had 20 percent lower odds of sleeping under a bed net com- sons including unavailability of food, inade- quate shelter, and caregiver absenteeism. A The percentage of 1st graders pared to children of older mothers.7 baseline study of 199 CBCCs reported that who are 8 years or older. one-third of CBCC caregivers lacked Primary School Leaving Certificates and less than 6   Results of multivariate regression models, using both 2013–14 MICS and pooled cross-sectional datasets (modeled separately) on cough in the past two weeks (or fever or diarrhea modeled separately) were associated fever or diarrhea modeled separately), were associated with whether mother was an adolescent at the time of with adolescence, woman’s educational attainment, the child’s birth, woman’s educational attainment, household wealth quintile, marital status, religion, region household wealth quintile, religion, ethnicity, region of of residence, urban/rural residence, child sex, birth order, residence, urban/rural residence, child sex, and and survey year (for pooled analysis). The full findings household bednet ownership. The full findings are are available from the authors upon request. available from the authors upon request. 7   Results of a multivariate regression model using 8   In 2014–15, 1 in 4 children in Standard 1 were repeaters 2013–14 MICS data on cough in the past two weeks (or (EMIS). 4 40 percent had received any training on early Accreditation Guidelines (2012); the National childhood development (World Bank 2015b). Nutrition Policy and Strategic Plan (2009) The recent 2014 MICS ECDI9 shows that and the National ECD Curriculum all aim to only 60 percent of children aged 35–59 ensure Malawian children are given a fair months were on track in their literacy-­ numeracy, physical, social-emotional, and chance to survive, grow, develop, and partici- pate. The Growth and Development Strategy 17% learning domains, though there were large II (MGDS II) also emphasizes the importance The percentage of children socioeconomic differences. A World Bank of child development for building human capi- age 36–59 months who are (2015b) study found that child development tal and productivity, and there are multi-sec- developmentally on track for outcomes tend to be better for mothers toral Population and Youth Policies. These literacy and numeracy. with more formal education. Controlling for policies incorporate some evidence-based height and age, maternal primary school approaches including positioning ECD at the education was associated with a 1.2 point intersection of health and early education, increase in children’s receptive vocabulary emphasizing ways to improve quality of edu- scores and maternal secondary school edu- cation across levels (including ECE), promot- cation was associated with a 3.2 point in- ing breastfeeding and community-based crease in scores.10 The MICS also showed management of childhood illness, and the that only 1.2 percent of children lived in provision of psychosocial care for children – households that had 3 or more books for but important gaps remain, including mater- the child, though about 30 percent lived in nal education about early learning and households where they had some sort of stimulation, as well as preparatory programs support for learning through engagement and policies about parental leave. with adults in four or more activities. This set of ECD-relevant policies has seen wide support from the government and the Policy Framework in Malawi community. They have created new institu- Malawi’s Growth and Development tions and structures, and the multitude of Strategy II (MGDS II) places great emphasis NGOs and partners in this space see many in the role of child development in building opportunities for partnership and cooperative human capital and productivity. Malawi has implementation. There have been implemen- a multisector approach to ECD, centered in tation challenges, however; some laws have the MoGCSW, with implementation at the been enacted but not implemented, and the district level by district social welfare officers. policy space analysis based on stakeholder In recent years there has been great progress interviews found that there are insufficient in ECD laws, policy, and guidelines develop- resources for policy implementation – both in ment, including the National Policy on ECD terms of budgetary shortages and staffing (2006); National Social Support Policy (2012), shortages, as well as poor coordination which has a link with other policies including among the many implementing partners. The ECD; National Strategic Plan for ECD (2009- budget currently allocated to the operating 2014), National ECD Operational and costs of the country’s more than 11,000 pub- lic CBCCs and private ECD centers is inade- quate. Although there are monitoring and 9   ECDI is an overall the index calculated as the percentage of children who are on track in at least 3 of evaluation plans, there are concerns that 4 areas- Literacy-numeracy tests naming of letters of these may not function well across the many the alphabet, reading simple words, and recognizing partners. Specifically for child health, the single digit numbers. Physical tests small motor skills multi-sectoral National Nutrition Policy has and asks about recent illness. Social-emotional tests good support although it would benefit from how the child interacts with other children, their attention span, and behavior issues. Learning tests a visible policy champion. While there are following directions and executing a task independently. committees and coordinating bodies which For literacy-numeracy and social-emotional, a child is have raised the visibility of this topic (al- considered on track if two of the three components in though visibility and priority remain low, ac- those areas are correct. Sixty percent of children age cording to some stakeholders), policy 36–59 months who are developmentally on track in literacy-numeracy, physical, social-emotional, and implementation is under-resourced in terms learning domains; however only 17 percent of children of budget and human resources, particularly age 36-59 months were developmentally on track for for activities at the community level. Similar literacy and numeracy. overall results were found by the World Bank 10   Measured by the Peabody Picture Vocabulary Test. SABER report (2015a) – which rated Malawi’s The mean score for mothers without a Primary School Leaving Certificate was 24.4. Maternal education has policy environment, implementation, and been associated with PPVT scores for preschoolers in monitoring and assurance of ECD policy as Cambodia, Ecuador, Madagascar, and Mozambique. “emergent,” with some of the laws enacted but 5 not implemented, and while established poli- To improve outcomes of children of teen- cies provide a solid basis for implementation age mothers, services need to be tailored to there is a need to strengthen them through reach them. ECD policies and programs need an overarching law specifically aimed at ECD to be of better quality to address the many policy and delivery; in addition it was found challenges facing adolescents and their chil- that lack of a corresponding budget is an ob- dren. For Malawi to reap its full demographic stacle to implementation of policies. dividend, it will be necessary to ensure im- proved health and early childhood develop- Key Findings and Conclusions ment outcomes—particularly given the large Teenage pregnancies have negative conse- number of children born to adolescent moth- quences for the next generation: ers currently and in the coming decades. •• Children of adolescent mothers are at References higher risk of mortality, poor nutrition, EMIS. 2015. MoEST Statistics. Ministry of Education and and onset of illness than are children of Science and Technology. Lilongwe. Malawi. older mothers. National Statistical Office (NSO). 2015. “Malawi MDG •• Teenage mothers are less likely to take Endline Survey 2014.” NSO, Zomba, Malawi. Demographic and Health Surveys. 2015. Macro preventive measures for ensuring better International. Maryland. health of their children. Ministry of Gender, Children and Social Welfare, Republic •• Poor health and nutrition adversely affect of Malawi. 2015. “2015 Annual Report for Integrated cognitive and socioemotional development Early Childhood Development. Mmera Mpoyamba.” and learning, putting children of teenage Government of Malawi, Lilongwe. Neuman, M., C. McConnell, and K. Foster. 2014. mothers at higher risk of poor development “From Early Childhood Development Policy to outcomes. Sustainability: The Fragility of Community-Based •• The number of children reached with ECD Childcare Services in Malawi.” Int. Journal of Early services continues to fall far short of goals, Childhood 46 (1): 81–99. particularly in rural areas and there was World Bank. 2015a. “Malawi. ECD. Systems Approach for Better Education Results. SABER Report.” World Bank, need to increase the budget allocated Washington, D.C. to operating costs of the country’s more ———. 2015b. “Protecting Early Childhood Development in than 11,000public CBCCs and private ECD Malawi: Baseline Report.” World Bank, Washington, D.C. centers. © 2016 International Bank for Reconstruction and Development / The World Bank. Some rights reserved. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. This work is subject to a CC BY 3.0 IGO license (https://creativecommons.org/licenses/by/3.0/ igo). The World Bank does not necessarily own each component of the content. It is your responsibility to determine whether permission is needed for reuse and to obtain permission from the copyright owner. If you have questions, email pubrights@ worldbank.org. SKU K8736 6