Determinants of reduced child stunting in Cambodia: analysis of pooled data from three Demographic and Health Surveys
Nayu Ikeda a, Yuki Irie b & Kenji Shibuya c
a. AXA Department of Health and Human Security, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan.
b. Global Link Management, Inc., Tokyo, Japan.
c. Department of Global Health Policy, Graduate School of Medicine, the University of Tokyo, Tokyo, Japan.
Correspondence to Nayu Ikeda (e-mail: email@example.com).
(Submitted: 23 September 2012 – Revised version received: 21 December 2012 – Accepted: 21 January 2013 – Published online: 19 February 2013.)
Bulletin of the World Health Organization 2013;91:341-349. doi: http://dx.doi.org/10.2471/BLT.12.113381
Improving the nutritional status of children is a priority in child health. The global burden of undernutrition among children is enormous. Worldwide, more than 300 million children younger than 5 years are estimated to be chronically undernourished.1 Nutritional insufficiency is one of the most important factors contributing to illness among children in the world. In 2010, it accounted for 3.4% of the total global burden of disease.2 To reduce the burden of child undernutrition, it is important to implement appropriate policies and interventions targeting determinants such as education, economic status and empowerment of women.3
Cambodia has experienced rapid socioeconomic development during its democratic transition towards an open market economy since the international peace agreements were signed in 1991.4 Driven by strong economic activity in a period of peace and political stability, the country’s gross domestic product grew by an annual average of 9% during the 2000s5 and was continuing to grow by more than 6% annually in the early 2010s.6 However, nearly one third of the population still lives below the poverty line5 and many children suffer from hunger and poor health. Although mortality among Cambodian children under 5 years of age decreased at an accelerated rate during the 2000s, the decrease is not sufficient to achieve Millennium Development Goal (MDG) 4, which calls for a two-third reduction from 1990 to 2015.7 The nutritional status of Cambodian children has been one of the poorest in the world since at least the mid-1980s.1 Cambodia is not likely to meet MDG 1, which seeks to halve the proportion of people experiencing hunger,1 even though the prevalence of stunting among children younger than 5 years dropped from 50% in 2000 to 40% in 2010.8
Stunting is generally used as an indicator of long-term chronic nutritional deficiency. Child stunting has been extensively studied in developing countries, including Bangladesh,9 Brazil,10 Indonesia,9 Kenya11 and Mozambique.12 In Cambodia, several studies have analysed data from the Cambodian Demographic and Health Surveys (DHSs) conducted in 2000 and 2005 to explore the factors influencing child stunting. These studies have shown stunting was more common among Cambodian children who lived in poorer households,13 or whose mothers had little education14 or smoked tobacco.15 Another study also showed the absence of an association between child stunting among Cambodian children and mothers’ compliance with recommended breastfeeding and complementary feeding practices.16 However, no study has extended the analysis to include the Cambodian 2010 DHS, which covers data for the period of rapid economic growth, or examined how certain potential contributing factors may have affected the decline in stunting prevalence.
The present study aimed to explore what measures will further reduce child stunting in Cambodia, where many people still live in poverty despite rapid economic growth. Using pooled data from three Cambodian DHSs conducted in the period from 2000 to 2010, we investigated improvements in child stunting over the past decade to highlight areas in which effective policy intervention could further reduce stunting prevalence. This information will help policy-makers to organize effective programmes across various sectors to accelerate improvements in the nutrition and health of children in Cambodia.
We pooled data for children less than 5 years of age from the Cambodian DHS conducted in 2000, 2005 and 2010. In each survey, a nationally representative sample of households was obtained through multistage cluster sampling. The methods were detailed elsewhere.8,17,18 To summarize, in 2000 the sampling frame was 600 villages; in 2005 it was all of Cambodia’s 13 505 villages and in 2010 it was 28 764 enumeration areas. Provinces and groups of provinces were stratified into urban and rural areas. In the first stage, villages or census enumeration areas were selected, and, in 2000, a fixed number of segments or enumeration areas were further retained from each of the selected villages. Subsequently, a fixed number of households was selected from each of the selected clusters. All women of reproductive age (15–49 years) who were members of a sampled household or who had slept there the night before the survey were eligible for an interview. The DHS collected anthropometric data for women and for all children younger than 5 years from a 50% subsample of households. Trained personnel measured the recumbent length of children aged less than 24 months and the standing height of older children.8,17,18 Children were included in the study if they had no missing values for anthropometric measurements or other relevant variables. The final sample was composed of 10 366 children (3395 in 2000, 3418 in 2005 and 3553 in 2010).
To assess the chronic nutritional status of children, we used height-for-age z-scores following the Child Growth Standards of the World Health Organization (WHO).19 The height-for-age z-score, as defined by WHO, expresses a child’s height in terms of the number of standard deviations above or below the median height of healthy children in the same age group or in a reference group. We classified children as being stunted if they had a height-for-age z-score of less than −2.
To explore factors that may have contributed to the reduction in child stunting during the past 10 years, we conducted a two-step analysis: (i) we examined the association between stunting in individual children and various socioeconomic characteristics and public health determinants, and (ii) we quantified the extent to which a change in each associated determinant decreased the prevalence of child stunting.
In the first part of the analysis, we used multivariate hierarchical logistic regression to model the stunting of children as a function of their socioeconomic characteristics. In this pooled analysis, we included survey year as a fixed effect and adjusted for any correlation among mothers having several children because 41.3% of the sample had siblings aged less than 5 years. We used only covariates that could be consistently measured across the three surveys, which were: the child’s sex; the child’s age in months (continuous variable); type of birth (single or multiple); initiation of breastfeeding within 1 hour of birth (dichotomous variable); preceding birth interval; maternal height (in centimeters squared); maternal body mass index (a proxy for maternal nutritional status); maternal and paternal educational level; maternal use of tobacco at the time of the survey (dichotomous variable); type of residence (urban or rural); number of household members; household wealth status score; access to improved sources of water during the rainy season; access to improved sanitation facilities; the occurrence of diarrhoea in the two weeks preceding the survey; and mother’s birth during the 1975–1978 Khmer Rouge regime (as an indicator variable). In constructing the indicators on water and sanitation, we followed the classifications proposed by the Joint Monitoring Programme for Water Supply and Sanitation (run by WHO and the United Nation’s Children Fund).20 The wealth index scores originally provided by the Cambodian DHS were estimated separately and were not comparable across survey years. Therefore, to compute household wealth index scores on a single common scale across surveys, we pooled data sets for 42 077 households (12 195 in 2000, 14 224 in 2005 and 15 658 in 2010) and conducted a principal component analysis on household possession of 12 durable assets and housing materials.
To measure the relative contribution of different factors to the decline in the prevalence of child stunting between 2000 and 2010, we computed a base value for stunting prevalence by using β coefficients estimated from the regression and the mean values of the explanatory variables in 2000. We then changed the value of each variable to its mean in 2010 while keeping all other covariates constant at their 2000 mean levels. We took any difference from the base value to be the effect of each factor on stunting prevalence. The contributions were not mutually exclusive across factors because the change in means over time could be correlated, particularly with wealth. We used the same approach to estimate the changes in the stunting prevalence in 2010 that would have been achieved by counterfactual improvements in socioeconomic status and public health, as represented by sanitation, spacing between births, smoking prevalence, infection and maternal nutrition. We took into account the complex sampling method (i.e. stratification, clustering and sample weights) in estimating summary statistics for the variables at the population level. We used Stata version 12 (StataCorp LP, College Station, United States of America) for all statistical analyses.
The prevalence of stunting in the study population decreased from 49.3% (95% confidence interval, CI: 47.1–51.5) in 2000 to 39.0% (95% CI: 36.7–41.2) in 2010. Broken down by age group, in 2000 the stunting prevalence was 35.6% (95% CI: 32.7–38.7) among children aged 0–23 months and 57.8% (95% CI: 55.1–60.5) among those aged 24–59 months; in 2010 it was 27.5% (95% CI: 24.7–30.4) and 46.7% (95% CI: 43.8–49.6), respectively. Table 1 lists the point estimates and 95% CIs for the explanatory variables at the national level by survey year. A substantial increase was observed in several public health variables, including early initiation of breastfeeding, improved sources of drinking water and improved sanitation facilities. Fig. 1 illustrates trends in the distribution of household wealth index scores across survey years. The fraction of wealthier households increased over time, while the distribution was still skewed to the right, with the majority of households below the mean value and a long tail at the higher end of the distribution in 2010.
Table 1. Characteristics of children included in the analysis, by year, based on data obtained from the Cambodia Demographic and Health Surveys of 2000, 2005 and 2010
Fig. 1. Distribution of household wealth index scores, by year, estimated from Cambodian Demographic and Health Survey data for 2000, 2005 and 2010
After adjustment for other confounding factors, child stunting showed a statistically significant association with a child’s sex and age, type of birth, maternal height, a maternal body mass index of < 18.5 kg/m2, preceding birth interval (except for firstborn children), number of household members, household wealth index scores, improved sanitation facilities, maternal and paternal educational level, maternal use of tobacco at the time of the survey, episode of diarrhoea in the child in the two weeks immediately before the survey, and mother’s birth during the Khmer Rouge regime (Table 2).
Table 2. Variables associated with child stunting, based on data from the Cambodian Demographic and Health Surveys of 2000, 2005 and 2010
Table 3 shows the extent to which various factors contributed to the decrease in child stunting prevalence between 2000 and 2010. Of all the independent variables included in the model, increased household wealth index score was the one that made the largest contribution, followed by increased access to improved sanitation facilities, improved maternal and paternal educational level, longer interval between preceding birth and birth of surveyed child and decreased prevalence of maternal tobacco use.
Table 3. Estimated contribution of each explanatory variable, in percentage points, to the stunting prevalence among children younger than 5 years from 2000 to 2010, based on data from the Cambodian Demographic and Health Surveys of 2000–2010
The prevalence of child stunting in 2010 would have been 3.4 percentage points (95% CI: 0.8–6.1) higher if the coverage of improved sanitation facilities had extended to all households; 1.8 percentage points (95% CI: 1.1–2.4) higher if the interval between the birth of every surveyed child and that of the preceding child had been at least 36 months, and 0.4 percentage points (95% CI: 0.1–0.6) if maternal tobacco use had been eradicated.
To our knowledge, this is the first study to quantify the contribution made by various factors to the decrease in the prevalence of stunting among Cambodian children less than 5 years of age during the 2000s, a period of peace and rapid economic growth. We found that increased household wealth made the largest contribution, followed by improvements in sanitation facilities, parental education and birth spacing and by a decrease in maternal tobacco use.
It is widely accepted that when economies grow and poverty is reduced, child nutrition improves owing to greater access to food, improved maternal and child care and better public health services.21 However, the evidence surrounding the relationship between improvements in the economy and childhood nutrition is inconclusive.22,23 Findings have been heterogeneous, perhaps because of differences in study design and unobserved confounders, such as the policies and circumstances that exist in individual countries. In this regard, our study contributes to the evidence that economic growth exerts a positive influence on child nutrition. It also points to what previous studies have shown, namely, that economic development alone is not enough to improve the nutritional status of children. Equitable income allocation and investment in public health and education programmes are needed to promote a nourishing diet and to keep children healthy.1,22,23 In our study, substantial economic disparity between households was observed: a large fraction of households ranked low on the nationwide wealth scale. Equitable economic growth is needed to attain sustained reductions in the prevalence of child stunting in Cambodia.
The other major determinants in our model contributed in approximately equal measure to the reduction in stunting prevalence. The rate of access to improved sanitation facilities increased from less than 10% to more than 25% between 2000 and 2010. This affected the prevalence of stunting, particularly among children aged 24–59 months (results not shown), perhaps because hygienic toilets prevent young children growing out of diapers from coming into contact with faeces and hence reduce the risk of infections transmitted by the faecal-oral route. 24-26 Since more than 70% of Cambodian children still live in households without improved sanitary facilities, strengthening strategies to increase the provision of hygienic toilets is essential for the health of preschool children.
Although formal education has improved gradually in Cambodia, there is substantial room for further advances which would in turn lead to further reductions in child stunting. Currently, most parents of children under 5 years of age in Cambodia have never enrolled in secondary education. Our finding of an association between maternal and paternal educational level and childhood stunting is consistent with the results of other studies conducted in developing countries, including Cambodia.9,12-14 Moreover, our additional analyses suggest that maternal educational level is inversely associated with stunting in children aged 24–59 months (results not shown). Educated parents may be better equipped to offer their children good care than parents of low educational attainment. It is important that the Cambodian government pursue policies aimed at improving access to education among the poor in Cambodia.
According to our findings, an increase in preceding birth interval made a modest contribution to the decreased prevalence of stunting in Cambodian children. The progress observed in birth spacing in Cambodia is reflected in the substantial decline in child mortality in the country during the past 10 years. A short interval between births can have an adverse effect on child nutrition by causing intrauterine growth retardation or undermining the quality of child care.27
A reduction in the maternal use of tobacco also made a small contribution to the drop in stunting prevalence in Cambodia in the past decade. In our study, maternal smoking and tobacco chewing was significantly associated with child stunting at the individual level, as shown by previous studies in developing countries.9,15,28 The use of tobacco can cause growth retardation in utero29 and divert money away from food and towards tobacco, particularly among the poor.30 Child growth in Cambodia can be improved by increasing public awareness about the dangers of tobacco use and implementing interventions to combat the notion, propagated primarily by older women smokers, that smoking reduces stress during pregnancy.31
Like previous studies in developing countries, ours detected an association between maternal height and child stunting.32,33 Short maternal stature is associated with intrauterine growth retardation and low birth weight, which are in turn predictors of infant death and impaired child growth.2 Moreover, birth of the mother during the Khmer Rouge famine of 1975–1978 also showed a positive association with stunting in the offspring. This suggests the presence of a transgenerational link in which exposure to hunger during the Khmer Rouge regime influences a mother’s stature, which predicts, in turn, the growth of her child. A previous study also showed that intrauterine exposure to the Dutch famine of 1944–1945 was associated with shorter offspring length at birth.34 Further research is needed to explore the life-course effects on child growth of maternal exposure in utero to risk factors present during the Cambodian famine.
The present study has several limitations. First, it did not address causal effects, since it was based on data from cross-sectional surveys. Second, children who died before the survey were not included in the sample, although some may have been severely undernourished. Third, a few key variables could not be included in the model because of differences in survey design and questions. For example, exclusive breastfeeding was excluded from the model because, since the Cambodian DHS only includes questions about current breastfeeding, no information on the early diet of children older than 6 months was available. However, our subgroup analysis of children younger than 6 months suggests the absence of a statistically significant association exists between exclusive breastfeeding and early postnatal stunting (results not shown). This is consistent with previous findings.16 Similarly, we were unable to include complementary feeding status in the model because the list of food items was different across the surveys. However, our sub-group analysis in children aged 6–23 months suggests that less frequent feedings are not associated with stunting (results not shown). Acute respiratory infection was also excluded from our model because the 2000 DHS included no question about the presence of a cough in connection with a respiratory problem. Although these variables were missing from the analysis, our model did consider all those covariates linked to child stunting that were treated homogeneously across the three surveys.
In conclusion, the decrease in the prevalence of stunting among Cambodian children between 2000 and 2010 was partly attributable to improvements in household wealth and in some public health measures. Further reductions in the prevalence of child stunting could be attained through more equitable economic growth, greater access to hygienic toilets and improved educational opportunities. Interventions addressing birth spacing and tobacco use are also needed. Policy-makers need to prioritize these measures and effectively coordinate multisectoral programmes to improve the diet of Cambodian children.
This study was supported in part by the Grant-In-Aid for Scientific Research from the Japan Ministry of Health, Labour and Welfare, grant number H21-Chikyukibo-Ippan-002.
Yuki Irie and Nayu Ikeda contributed equally to the authorship of this paper.
- Stevens GA, Finucane MM, Paciorek CJ, Flaxman SR, White RA, Donner AJ, et al., Nutrition Impact Model Study Group (Child Growth), et al. Trends in mild, moderate, and severe stunting and underweight, and progress towards MDG 1 in 141 developing countries: a systematic analysis of population representative data. Lancet 2012; 380: 824-34 http://dx.doi.org/10.1016/S0140-6736(12)60647-3 pmid: 22770478.
- Murray CJL, Vos T, Lozano R, Naghavi M, Flaxman AD, Michaud C, et al., et al. Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012; 380: 2197-223.
- Bhutta ZA, Ahmed T, Black RE, Cousens S, Dewey K, Giugliani E, et al., Maternal and Child Undernutrition Study Group, et al. What works? Interventions for maternal and child undernutrition and survival. Lancet 2008; 371: 417-40 http://dx.doi.org/10.1016/S0140-6736(07)61693-6 pmid: 18206226.
- Sharing growth: equity and development in Cambodia. Washington: The World Bank; 2007. Available from: http://siteresources.worldbank.org/INTCAMBODIA/Resources/293755-1181597206916/E&D_Full-Report.pdf [accessed 12 February 2013].
- World development indicators 2012. Washington: The World Bank; 2012. Available from: http://data.worldbank.org/news/world-development-indicators-2012-now-available [accessed 12 February 2013].
- World dataBank [Internet]. Washington: The World Bank; 2013. Available from: http://databank.worldbank.org/data/home.aspx [accessed 12 February 2013].
- Lozano R, Wang H, Foreman KJ, Rajaratnam JK, Naghavi M, Marcus JR, et al., et al. Progress towards Millennium Development Goals 4 and 5 on maternal and child mortality: an updated systematic analysis. Lancet 2011; 378: 1139-65 http://dx.doi.org/10.1016/S0140-6736(11)61337-8 pmid: 21937100.
- Cambodia Demographic and Health Survey 2010. Phnom Penh & Calverton: National Institute of Statistics & ICF Macro; 2011.
- Semba RD, de Pee S, Sun K, Sari M, Akhter N, Bloem MW. Effect of parental formal education on risk of child stunting in Indonesia and Bangladesh: a cross-sectional study. Lancet 2008; 371: 322-8 http://dx.doi.org/10.1016/S0140-6736(08)60169-5 pmid: 18294999.
- Monteiro CA, Benicio MH, Conde WL, Konno S, Lovadino AL, Barros AJD, et al., et al. Narrowing socioeconomic inequality in child stunting: the Brazilian experience, 1974–2007. Bull World Health Organ 2010; 88: 305-11 http://dx.doi.org/10.2471/BLT.09.069195 pmid: 20431795.
- Masibo PK, Makoka D. Trends and determinants of undernutrition among young Kenyan children: Kenya Demographic and Health Survey; 1993, 1998, 2003 and 2008–2009. Public Health Nutr 2012; 15: 1715-27 http://dx.doi.org/10.1017/S1368980012002856 pmid: 22694984.
- Burchi F. Child nutrition in Mozambique in 2003: the role of mother’s schooling and nutrition knowledge. Econ Hum Biol 2010; 8: 331-45 http://dx.doi.org/10.1016/j.ehb.2010.05.010 pmid: 20646971.
- Hong R, Mishra V. Effect of wealth inequality on chronic under-nutrition in Cambodian children. J Health Popul Nutr 2006; 24: 89-99 pmid: 16796155.
- Miller JE, Rodgers YV. Mother's education and children's nutritional status: new evidence from Cambodia. Asian Dev Rev 2009; 26: 131-65.
- Kyu HH, Georgiades K, Boyle MH. Maternal smoking, biofuel smoke exposure and child height-for-age in seven developing countries. Int J Epidemiol 2009; 38: 1342-50 http://dx.doi.org/10.1093/ije/dyp253 pmid: 19622677.
- Marriott BP, White AJ, Hadden L, Davies JC, Wallingford JC. How well are infant and young child World Health Organization (WHO) feeding indicators associated with growth outcomes? An example from Cambodia. Matern Child Nutr 2010; 6: 358-73 http://dx.doi.org/10.1111/j.1740-8709.2009.00217.x pmid: 21050390.
- Cambodia Demographic and Health Survey 2005. Phnom Penh & Calverton: National Institute of Statistics & ICF Macro; 2006.
- Cambodia Demographic and Health Survey 2000. Phnom Penh & Calverton: National Institute of Statistics & ICF Macro; 2001.
- WHO Multicentre Growth Reference Study Group. WHO Child Growth Standards: length/height-for-age, weight-for-age, weight-for-length, weight-for-height and body mass index-for-age: methods and development. Geneva: World Health Organization; 2006.
- Progress on drinking water and sanitation: 2012 update. New York: United Nations Children’s Fund & World Health Organization; 2012.
- Smith Lisa C, Haddad L. How potent is economic growth in reducing undernutrition? What are the pathways of impact? New cross-country evidence. Econ Dev Cult Change 2002; 51: 55-76 http://dx.doi.org/10.1086/345313.
- Haddad L, Alderman H, Appleton S, Song L, Yohannes Y. Reducing child malnutrition: How far does income growth take us? World Bank Econ Rev 2003; 17: 107-31 http://dx.doi.org/10.1093/wber/lhg012.
- Subramanyam MA, Kawachi I, Berkman LF, Subramanian SV. Is economic growth associated with reduction in child undernutrition in India? PLoS Med 2011; 8: e1000424 http://dx.doi.org/10.1371/journal.pmed.1000424 pmid: 21408084.
- Water, sanitation & hygiene: strategy overview. Seattle: Bill & Melinda Gates Foundation; 2012. Available from: http://www.gatesfoundation.org/watersanitationhygiene/Documents/wsh-strategy-overview.pdf [accessed 12 February 2013].
- Humphrey JH. Child undernutrition, tropical enteropathy, toilets and handwashing. Lancet 2009; 374: 1032-5 http://dx.doi.org/10.1016/S0140-6736(09)60950-8 pmid: 19766883.
- Mara D, Lane J, Scott B, Trouba D. Sanitation and health. PLoS Med 2010; 7: e1000363 http://dx.doi.org/10.1371/journal.pmed.1000363 pmid: 21125018.
- Dewey KG, Cohen RJ. Does birth spacing affect maternal or child nutritional status? A systematic literature review. Matern Child Nutr 2007; 3: 151-73 http://dx.doi.org/10.1111/j.1740-8709.2007.00092.x pmid: 17539885.
- Best CM, Sun K, de Pee S, Bloem MW, Stallkamp G, Semba RD. Parental tobacco use is associated with increased risk of child malnutrition in Bangladesh. Nutrition 2007; 23: 731-8 http://dx.doi.org/10.1016/j.nut.2007.06.014 pmid: 17664060.
- Strauss RS. Effects of the intrauterine environment on childhood growth. Br Med Bull 1997; 53: 81-95 http://dx.doi.org/10.1093/oxfordjournals.bmb.a011608 pmid: 9158286.
- Block S, Webb P. Up in smoke: Tobacco use, expenditure on food, and child malnutrition in developing countries. Econ Dev Cult Change 2009; 58: 1-23 http://dx.doi.org/10.1086/605207.
- Singh PN, Yel D, Sin S, Khieng S, Lopez J, Job J, et al., et al. Tobacco use among adults in Cambodia: evidence for a tobacco epidemic among women. Bull World Health Organ 2009; 87: 905-12 http://dx.doi.org/10.2471/BLT.08.058917 pmid: 20454481.
- Ozaltin E, Hill K, Subramanian SV. Association of maternal stature with offspring mortality, underweight and stunting in low- to middle-income countries. JAMA 2010; 303: 1507-16 http://dx.doi.org/10.1001/jama.2010.450 pmid: 20407060.
- Subramanian SV, Ackerson LK, Davey Smith G, John NA. Association of maternal height with child mortality, anthropometric failure, and anemia in India. JAMA 2009; 301: 1691-701 http://dx.doi.org/10.1001/jama.2009.548 pmid: 19383960.
- Painter RC, Osmond C, Gluckman P, Hanson M, Phillips DI, Roseboom TJ. Transgenerational effects of prenatal exposure to the Dutch famine on neonatal adiposity and health in later life. BJOG 2008; 115: 1243-9 http://dx.doi.org/10.1111/j.1471-0528.2008.01822.x pmid: 18715409.
- Rutstein SO, Johnson K. The DHS wealth index. Calverton: ORC Macro; 2004 (DHS Comparative Reports No. 6).