Stunting in childhood: An overview of global burden, trends, determinants, and drivers of decline

Abstract

Background:
Progress has been made worldwide in reducing chronic undernutrition and rates of linear growth stunting in children under 5 y of age, although rates still remain high in many regions. Policies, programs, and interventions supporting maternal and child health and nutrition have the potential to improve child growth and development.

Objective:
This article synthesizes the available global evidence on the drivers of national declines in stunting prevalence and compares the relative effect of major drivers of stunting decline between countries.

Methods:
We conducted a systematic review of published peer-reviewed and gray literature analyzing the relation between changes in key determinants of child linear growth and contemporaneous changes in linear growth outcomes over time.

Results:
Among the basic determinants of stunting assessed within regression-decomposition analyses, improvement in asset index score was a consistent and strong driver of improved linear growth outcomes. Increased parental education was also a strong predictor of improved child growth. Of the underlying determinants of stunting, reduced rates of open defecation, improved sanitation infrastructure, and improved access to key maternal health services, including optimal antenatal care and delivery in a health facility or with a skilled birth attendant, all accounted for substantially improved child growth, although the magnitude of variation explained by each differed substantially between countries. At the immediate level, changes in several maternal characteristics predicted modest stunting reductions, including parity, interpregnancy interval, and maternal height.

Conclusions:
Unique sets of stunting determinants predicted stunting reduction within countries that have reduced stunting. Several common drivers emerge at the basic, underlying, and immediate levels, including improvements in maternal and paternal education, household socioeconomic status, sanitation conditions, maternal health services access, and family planning. Further data collection and in-depth mixed-methods research are required to strengthen recommendations for those countries where the stunting burden remains unacceptably high.

Full text

Stunting in childhood: an overview of global burden, trends,
determinants, and drivers of decline
Tyler Vaivada,1Nadia Akseer,1,2Selai Akseer,1Ahalya Somaskandan,1Marianne Stefopulos,1and Zulqar A Bhutta1,2,3
1Centre for Global Child Health, Hospital for Sick Children, Toronto, Canada; 2Dalla Lana School of Public Health, University of Toronto, Toronto, Canada;
and 3Center of Excellence in Women and Child Health, the Aga Khan University, Karachi, Pakistan
ABSTRACT
Background: Progress has been made worldwide in reducing
chronic undernutrition and rates of linear growth stunting in children
under 5 y of age, although rates still remain high in many regions.
Policies, programs, and interventions supporting maternal and child
health and nutrition have the potential to improve child growth and
development.
Objective: This article synthesizes the available global evidence on
the drivers of national declines in stunting prevalence and compares
the relative effect of major drivers of stunting decline between
countries.
Methods: We conducted a systematic review of published peer-
reviewed and gray literature analyzing the relation between changes
in key determinants of child linear growth and contemporaneous
changes in linear growth outcomes over time.
Results: Among the basic determinants of stunting assessed within
regression-decomposition analyses, improvement in asset index
score was a consistent and strong driver of improved linear growth
outcomes. Increased parental education was also a strong predictor of
improved child growth. Of the underlying determinants of stunting,
reduced rates of open defecation, improved sanitation infrastructure,
and improved access to key maternal health services, including
optimal antenatal care and delivery in a health facility or with
a skilled birth attendant, all accounted for substantially improved
child growth, although the magnitude of variation explained by each
differed substantially between countries. At the immediate level,
changes in several maternal characteristics predicted modest stunting
reductions, including parity, interpregnancy interval, and maternal
height.
Conclusions: Unique sets of stunting determinants predicted stunt-
ing reduction within countries that have reduced stunting. Several
common drivers emerge at the basic, underlying, and immediate
levels, including improvements in maternal and paternal education,
household socioeconomic status, sanitation conditions, maternal
health services access, and family planning. Further data collection
and in-depth mixed-methods research are required to strengthen rec-
ommendations for those countries where the stunting burden remains
unacceptably high. Am J Clin Nutr 2020;112(Suppl):777S–791S.
Keywords: child, infant, nutrition, height, length, linear growth,
stunting
Introduction
High rates of chronic malnutrition in young children persist
globally, a condition that is strongly linked to poverty. Maternal
malnutrition can start the process of linear growth faltering in
utero, contributing to intrauterine growth restriction and low
birth weight. Suboptimal feeding practices in infancy coupled
with a high burden of infectious diseases also predict poor child
growth. Linear growth stunting, dened as a height-for-age z
score (HAZ) ≥2 SDs below the median, is an easily recogniz-
able and quantiable physical indicator of chronic childhood
malnutrition.
Children whose growth is stunted are more likely to ex-
perience higher rates of mortality, morbidity, and suboptimal
This study was funded by a grant to the Centre for Global Child Health
from Gates Ventures. The funder had no role in the design, implementation,
analysis, or interpretation of the data.
Published in a supplement to The American Journal of Clinical Nutrition.
The Guest Editor for this supplement was Mark Manary, and has no
disclosures. The Supplement Coordinator for the supplement publication was
Nadia Akseer, Gates Ventures/Hospital for Sick Children, Toronto, Canada.
Supplement Coordinator disclosure: no conicts to disclose. The Stunting
Exemplars research Principal Investigator was Zulqar A Bhutta, Hospital for
Sick Children, Toronto,Canada. Principal Investigator disclosure: no conicts
to disclose. Publication costs for this supplement were defrayed in part by the
payment of page charges by Gates Ventures. The opinions expressed in this
publication are those of the authors and are not attributable to the sponsors or
the publisher, Editor, or Editorial Board of The American Journal of Clinical
Nutrition.
Supplemental Methods, Supplemental Tables 1–4, Supplemental Figures
1–3, and Supplemental Text are available from the “Supplementary data” link
in the online posting of the article and from the same link in the online table
of contents at https://academic.oup.com/ajcn/.
SA, AS, and MS contributed equally.
Data described in the manuscript, code book, and analytic code will be
made available upon request.
Address correspondence to ZAB (e-mail: zulqar.bhutta@sickkids.ca).
Abbreviations used: ANC4+, mother attended ≥4 antenatal care visits;
DHS, Demographic and Health Survey; HAZ, height-for-age zscore; LMIC,
low- and middle-income country; MICS, Multiple Indicator Cluster Survey.
Received December 20, 2019. Accepted for publication May 29, 2020.
First published online August 29, 2020; doi: https://doi.org/10.1093/ajcn/
nqaa159.
Am J Clin Nutr 2020;112(Suppl):777S–791S. Printed in USA. Copyright ©The Author(s) on behalf of the American Society for Nutrition 2020. This is an
Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits
unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 777S

778S Supplement
cognitive and motor development (1). Meta-analyses of 5
prospective cohort studies have shown that a unit increase
in HAZ for children ≤2 y was associated with a 0.22-SD
improvement in cognitive function later in childhood at 5–
11 y (2), illustrating the lingering effects of early-life chronic
malnutrition. This has serious implications for population health
and the fullment of the intellectual and economic potential
of low- and middle-income countries (LMICs). Despite these
associations, stunting has the potential to be misused as a
measure of population health, as poor nutritional status can
affect the health, growth, and development of children whose
linear growth falls above the HAZ cutoff (3). It is helpful
to conceptualize stunting as a robust indicator of a decient
environment, which has strong associations with adverse out-
comes in the short and long term, rather than the sole cause of
poor cognitive development or future risk for chronic diseases
(4).
There has been global progress on reducing rates of child
stunting in recent decades, but this progress has been uneven
(see Panel 1,Figures 1–4,Supplemental Figure 1). Some
particularly high-performing countries have reduced stunting
prevalence by >30 percentage points in the past 30 y, while
others have made negligible progress. It is crucial to examine
the key determinants and drivers of stunting reduction so that
individual countries can learn what works in order to imple-
ment targeted policies and programs. Countries that prioritize
the implementation and scale-up of evidence-based, nutrition-
sensitive, and nutrition-specic policies and programs stand to
make great improvements in human capital development and
economic productivity, as these initiatives generally have very
high benet–cost ratios (5). There is also a moral imperative to
act, as all children have the right to grow and develop optimally
in order to reach their full developmental potential. Targeted and
concerted action at the national level will be essential to achieve
the Sustainable Development Goals related to child health and
nutrition.
As an introduction to this supplement issue, this article
includes an overview of the epidemiology of stunting across
LMICs (Panel 1, Figures 1–4, Supplemental Figure 1) and
a summary of existing conceptual thinking around the major
determinants of chronic childhood malnutrition and stunting
(Panel 2). The main objective of this article is to synthesize
the available global evidence on the drivers of national declines
in stunting prevalence and compare the relative effect of major
drivers of stunting decline between LMICs. The remainder
of this article focuses on this objective. Specically, we
sought to synthesize the global evidence examining drivers
of reductions in child stunting over time. To this end, we
conducted a systematic review of published peer-reviewed and
gray literature that analyzed the relation between changes
in key determinants of child growth and contemporaneous
changes in growth outcomes over time. These theoretical
determinants, described in Panel 2 and Supplemental Figure
2, included contextual factors, interventions, policies, strategies,
programs, and other initiatives that may have accounted for
reductions in under-5 child stunting prevalence over time in
LMICs.
Panel 1:
Child Stunting Epidemiology
The changing global burden of childhood linear growth
stunting
Although stunting rates have been decreasing over the
past several decades, an estimated 21.3% (144 million) of
children under 5 y of age globally experienced stunted
growth in 2019 (6). Both regional and within-country dispar-
ities exist, with prevalences ranging from 34.5% in eastern
Africa to 4.5% in eastern Asia as of 2019 (6). Globally, there
were ∼109 million fewer children experiencing stunting
in 2019 compared with 1990. However, despite making
modest progress in reducing prevalence, due to substantial
population growth the total number of children experiencing
stunting in the African region has increased by ∼13.1
million since 1990 (see Figure 1). An estimated 17% of
mortality burden in children under 5 y is associated with
stunting (1). Compared with children with HAZ >−1,
children with HAZ between −2and−3 have a 118%
(HR: 2.18) and 138% (HR: 2.38) higher risk of dying from
pneumonia or diarrhea, respectively (7). Those children who
are severely stunted (HAZ ≤−3) are at even higher risk
(pneumonia mortality—HR: 6.39; diarrhea mortality—HR:
6.33) (7).
Those countries with the highest levels of stunting
prevalence are concentrated in South and Southeast Asia and
sub-Saharan Africa, as depicted on the map in Figure 2.A
chart of the most recent country-level estimates of stunting
prevalence worldwide can be found in Supplemental Figure
1. While all global regions have experienced decreases
in stunting prevalence since 1990, this progress has been
uneven. The regions of South Asia and East Asia and
the Pacic have seen the greatest improvements, reducing
stunting prevalence by ∼25 percentage points over the past
30 y (see Figure 3).
National trends in child stunting in top-performing
countries
Those countries that have achieved substantial reductions
in stunting prevalence over the past ∼30 y are geographi-
cally dispersed among several regions worldwide. Figure 4
depicts the trends in stunting prevalence in a sample of
13 of the best-performing countries globally, which were
selected based on consultations with experts. Although the
baseline prevalence and rate of reduction in stunting vary for
each of these countries throughout the period examined, one
consistent pattern emerges that characterizes several of these
top performers: an initial period of stagnation followed by a
consistent decline.
For example, between 1988 and 1993 Vietnam initially
experienced stagnation (∼61%) but saw a very steep
decline between 1993 and 1998, followed by relatively
consistent reductions until 2015 (∼25%). In Burkina
Faso, this initial plateau lasted until 2006, after which

Global overview of child stunting determinants 779S
FIGURE 1 Global and regional comparison of the total number of children aged 0–59 mo experiencing linear growth stunting in 1990 and 2019. Source
of data: UNICEF, World Bank Group joint malnutrition estimates, 2020 edition (6). Data not available for Europe and Central Asia.
dramatic and consistent reductions were seen. Although
data are not available for Nepal before 1995, since
then, Nepal (68.2%) and Bangladesh (65.8%) followed a
very similar and consistent pattern of decline until 2014,
reducing stunting prevalence by ∼30 percentage points.
An examination of the contributing factors to Peru’s own
steep decline between 2008 and 2016 (28.2–13.1%) is the
subject of an in-depth case study within this supplement
issue.
It is important to note that there exists considerable varia-
tion in both stunting burden and trends within countries. This
subnational variation is closely related to socioeconomic
and geographic disparities, including indicators such as
parental education, household wealth, and rural location.
These subnational inequities are also analyzed and discussed
in detail within each of the case-study articles within this
supplement issue.
Panel 2:
Determinants of Linear Growth in Childhood
Theoretical determinants of linear growth faltering in
young children
Policies and programs designed to alleviate childhood
undernutrition and growth faltering typically rely on tar-
geting a standard set of risk factors that represent the
immediate, underlying, and basic causes of stunting. The
main multilevel conceptual framework used by the global
nutrition community for the past 30 y is the UNICEF Un-
dernutrition Conceptual Framework (8), upon which several
variations have been based. One derivative developed by
the WHO called “Childhood Stunting: Context, Causes and
Consequences” summarizes 3 levels of factors associated
with stunting, and is a product of the Healthy Growth

780S Supplement
FIGURE 2 Stunting prevalence for children under 5 y based on the most recently available country-level estimates. Map based on longitude (generated)
and latitude (generated). Color shows sum of stunting prevalence. Details are shown for country. Source of data: UNICEF, WHO, World Bank Group joint
malnutrition estimates, 2020 edition (6). Data not available for Europe and Central Asia.
Project (9). Adapted versions of the UNICEF framework
were highlighted in the Lancet Series on Maternal and
Child Nutrition in 2008 (10), and expanded in 2013 (1)to
incorporate the theorized effects of both nutrition-sensitive
and nutrition-specic interventions.
Although there are many theoretical determinants of
stunting along the causal chain, only a subset has been
studied well enough to quantify the strength of the relation.
A recent comparative stunting risk-assessment analysis (11)
grouped risk factors into 5 clusters: maternal nutrition and
infection, teenage motherhood and short birth intervals, fetal
growth restriction and preterm birth, child nutrition and
infection, and environmental factors. With the exception
of zinc supplementation trials in zinc-decient children,
all of the effect sizes for stunting risk were derived from
meta-analyses of cohort studies or pooled analyses of
Demographic and Health Survey (DHS) data. The leading
global risk factors in terms of total number of attributable
stunting cases were identied as follows: fetal growth
restriction (dened as being born at term and small for
gestational age), unimproved sanitation, childhood diarrhea,
and maternal short stature.
Econometric analysis of underlying and basic determi-
nants using data from 116 countries between 1970 and
2012 (12) identied several drivers of stunting reduction,
including access to safe water, improved sanitation, gender
equality, women’s education, and nutritious food avail-
ability, with governance and income growth providing a
supporting environment. Another cross-country analysis
(13) of the developmental drivers of change in country-
level nutritional status also highlighted asset ownership,
health service access, maternal educational achievement,
and lower fertility. However, growth in the economy and
food production were key predictors only in countries
experiencing food insecurity, and infrastructure was found
to not be directly important to nutritional improvement.
A distinction must be made between analyses of the cross-
sectional associations of specic determinants and stunting
prevalence or mean HAZ and those that analyze the relative
contribution of drivers of change in measures of child growth
over time. Synthesizing the global evidence base detailing
the latter type of analysis is prioritized in this article.
Trends in indicators for key determinants in a set of top-
performing countries
Trends in key indicators for determinants of child stunting
are depicted in Supplemental Figure 2 for the 13 top-
performing countries described earlier. Overall, indicators
have generally been improving over the last 30 y for the
majority of top-performing countries, albeit unevenly. The
rate of progress also varies widely by country. Despite
these general improvements, at present there remain massive
disparities in literacy rates, access to safe water and basic
sanitation, and poverty rates between countries. Analyses of
the relative contribution of these determinants to stunting
reductions in a set of countries is described later in the
article.

Global overview of child stunting determinants 781S
FIGURE 3 Global and regional trends in stunting prevalence, 1990–2019. Source of data: UNICEF, WHO, World Bank Group joint malnutrition estimates,
2020 edition (6). Data not available for Europe and Central Asia.
Methods
Building on existing frameworks and a mapping of key
indicators and proxies from the global literature, we developed
an adapted conceptual framework (Figure 5) to aid in the
identication and interpretation of a variety of determinants of
child stunting. Standard systematic review methods were used
to identify and assess literature of interest. These included the
development and execution of a search strategy in 15 databases,
screening of titles and abstracts for relevance, followed by full-
text screening against inclusion criteria and categorization of
records. Additional studies were identied through gray literature
searches, hand-searching of review reference lists, and update
searches. Those records selected for inclusion were abstracted
using a standardized form and underwent methodological
quality appraisal. Abstracted data were then collated in tabular
format, organized by determinant category and country, and
narratively synthesized. The systematic review methods used are
summarized in Panel 3 and described in full in the Supplemental
Methods.
Panel 3:
Review Methods
Initial title and abstract screening of records was completed
by a team of reviewers and focused on sensitivity and
relevance. Studies were identied as potentially relevant
if they met the 3 following inclusion criteria: 1)aset
of participants that included children <5 y was analyzed,
2)≥1 anthropometric outcome was measured, and 3)the

782S Supplement
FIGURE 4 Stunting prevalence, top-performing countries. Source of data: UNICEF, WHO, World Bank Group joint malnutrition estimates, 2020 edition
(6).
association between ≥1 stunting determinants and child
growth outcomes was examined.
Subsequently, the full text of records was retrieved and
reviewed, inclusion criteria were applied, and tags were
assigned to the studies using a predened algorithm, which
was used to categorize included articles based on their
study design. For the purposes of the current review, only
the subset of studies examining the drivers of stunting
decline or improvements in child growth outcomes at the
national level were considered for full data abstraction.
These studies contained analyses of multiple national cross-
sectional surveys (e.g., DHS). For this subset of included
studies, the prior categorization exercise was reassessed by
a second reviewer to conrm eligibility for data abstraction.
At this stage, the reference lists of reviews identied during
the eligibility screening process were hand-searched for
additional relevant studies for inclusion.
From the set of included national-level studies, quantita-
tive and qualitative data were extracted, and methodological
quality was appraised by the review team in duplicate. A
standardized abstraction form was generated, which was
designed to collect data on study characteristics, target pop-
ulation, outcome data, intervention/policy/program char-
acteristics, and analysis methods. The estimates extracted
included percentage contributions from decomposition anal-
yses, regression coefcients, ORs, and RRs. In order to
assess the quality of included studies based on their study
design, we produced a tailored quality appraisal tool. We

Global overview of child stunting determinants 783S
FIGURE 5 Conceptual framework of child stunting determinants. Determinants include those identied during the review process, and are based on those
originally described in the UNICEF Undernutrition Conceptual Framework (8) and 2013 Lancet Maternal and Child Nutrition Series framework (1).
used a star rating system to assess quality across 4 domains:
study design, sample selection, data sources, and statistical
analyses measures. Abstracted data and quality appraisal
ratings were matched between ≥2 reviewers, and any
disagreements were resolved through discussion reaching a
consensus.
Following the completion of data extraction, study
variables were categorized into groups and subgroups based
on the conceptual framework. The determinants/covariates
were then mapped according to their conceptual domain
grouping and subgrouping, and further study information
was collated to assist with narrative synthesis.
Ethics statement
As this was a systematic review of publicly available literature,
ethical review was not required.
Results
Study selection
After database searches were executed (19 June 2018) and
records exported and de-duplicated, a total of 16059 titles and
abstracts were screened within Covidence, from which 2141
records were identied as potentially relevant. Full texts were
retrieved and then screened against broad inclusion and exclusion
criteria, which yielded a total of 1156 studies. Concurrently, all
1156 studies were then assigned a set of “tags” based on study
design for further categorization. For the purposes of this work,
the subset of studies that were assigned all of the following tags
were included and abstracted: 1) national-level or multinational-
level, 2) quantitative analysis, and 3) analysis of trends over time
(including ≥2 time points).
A total of 55 studies identied from the original indexed
literature were assigned this set of tags and were eligible for
inclusion in this systematic review. The reference lists of those

784S Supplement
studies tagged as “reviews” within Covidence were also screened,
yielding an additional 4 studies not previously identied. Further
gray literature searching done in February 2019 yielded 12
additional studies eligible for inclusion, and rapid catch-up
searches for indexed literature done in May and August 2019
yielded an additional 6 and 12 studies, respectively. Thus, data
included in this review were abstracted from a grand total of 89
discrete studies (see Supplemental Figure 3 for a review ow
diagram).
Study characteristics and quality appraisal
The complete list of included studies (14–102), their character-
istics, and quality appraisal scores can be found in Supplemental
Table 1. The quality appraisal of included studies did not identify
any meaningful differences in their individual methodological
quality, nor raise any signicant concerns that would affect the
interpretation or synthesis of this set of observational studies.
Additionally, groups and subgroups of determinants analyzed
within the included studies are summarized in Supplemental
Table 2.
Synthesis of results from analyses of the drivers of improved
child linear growth
The included studies analyzed data on the determinants of
child growth and drivers of stunting reduction from >70 countries
worldwide. A total 11 of studies (29,30,33,40,46,48–51,55,
99) contained data from national-level regression-decomposition
analyses of change in HAZ and stunting prevalence in 14
countries (see Table 1). These studies best address our research
question, and the following sections focus on synthesizing
key ndings from robust models across this set of studies,
organized by groupings based on the basic, underlying, and
immediate determinants of stunting. Other included studies
using different methodologies are also described to augment the
results.
In a majority of these regression-decomposition studies (33,
40,46,48–51) multivariable regression and Oaxaca-Blinder
decomposition methods were used to examine how different
determinants predicted change in nutrition status. Multivari-
able linear regression and linear probability modeling were
used to examine associations between HAZ and covariates of
interest based on data collected regularly through DHS, the
Multiple Indicator Cluster Survey (MICS), and other nationally
representative surveys. The Oaxaca-Blinder decomposition is
complementary to this initial regression analysis using the same
individual/household-level data and ecological variables to assess
predictors of HAZ or stunting change within a country between 2
survey time points at national or subnational levels. Some studies
used extensions of Oaxaca-Blinder methods to incorporate
dummy and nominal variables into the decomposition analysis
(103,104) or account for logit and probit models (105).
Econometric analysis (89), quantile regression-decomposition
(102), and calculation of the relative contribution to decreases in
stunting prevalence (55) were other methods used by authors of
included studies.
Supplemental Table 3 summarizes the effect estimates from
those studies that analyzed the associations between a variety
of key indicators and the risk of stunting across multiple
years. Estimates from decomposition analyses of changes in
the population-level inequality of stunting at the national level
are summarized in Supplemental Table 4 and described in the
Supplemental Text.
Basic Stunting Determinants
Asset index.
Household income is an important measure of a household’s
capacity to afford important elements related to improved
nutrition such as food, water, sanitation, and medical care (12).
Compared with other determinants, improvements in asset index
consistently predicted some of the greatest improvements in
HAZ across the countries analyzed. Of the total HAZ change
observed in Cambodia (55) and Pakistan (49), 42% and 33%
were attributed to asset index scores, respectively—the largest
values analyzed. Similarly, improvements in asset index drove an
estimated 25% of total HAZ change in Bangladesh (50).
Parental education.
Maternal education is associated with decreased odds of
stunting due to improvements in child health and care, and
enhanced uptake and benets from health interventions (1).
Higher levels of paternal education are also associated with
reduced odds of child stunting (106). Improvements in maternal
educational attainment predicted 17% of the total HAZ change
in Pakistan (49), between 11% and 14% in Nepal (33,49–
51), 10% in Guinea (29) and India (49), and 7% in Cambodia
(55). Improvements in paternal education generally appeared
to explain less HAZ change than maternal education, with the
exceptions of Cambodia and Guinea. Increases in combined
measures of parental education were estimated to predict 30%
of the HAZ change in India (50).
Underlying Stunting Determinants
Open defecation and sanitation.
Environmental enteropathy and repeated diarrhea due to
environmental fecal contamination and ingestion by young
children—often related to widespread open-defecation practices
or improper feces disposal—are theorized to increase the risk of
stunting through reduced nutrient absorption and inammation
(107–110). Reductions in open defecation accounted for 17% of
the total HAZ change in Pakistan (49), 10–14% in Nepal (49–
51), 8% in Ethiopia (50), and 7–10% in India (49,50). Similarly,
improved sanitation infrastructure was found to be an important
predictor of HAZ change in Cambodia (12%) (55), Guinea (18%)
(29), and Nepal (7%) (33).
Access to improved water sources.
The presence of a piped water source in the yard of a
house is associated with water-related safe hygiene practices in
mothers (111) and represents a pathway associated with diarrhea
reduction (112). Improved access to safe water source predicted
7% of the change in HAZ in rural Paraguay (40) and 6% in
Senegal (50).

Global overview of child stunting determinants 785S
TABLE 1 Summary of changes in stunting prevalence and HAZ statistically explained by changes in stunting determinant indicators within regression-decomposition analyses.
Bangladesh Cambodia Ethiopia Guinea India Kenya Liberia Namibia Nepal Pakistan Paraguay Rwanda Senegal Zambia
Headey
2015b
(48)
Headey
2016
(45)
Headey
2017
(50)
World
Bank
2005
(99)
Ikeda
2013 (55)
Buisman
2019 *
(30)
Headey
2017
(50)
Headey
2014
(46)
Boccanfuso
2013 (29)
Headey
2016
(45)
Headey
2017**
(50)
Buisman
2019 *
(30)
Buisman
2019 *
(30)
Buisman
2019 *
(30)
Cunningham
2017 * (33)
Headey
2015a
(51)
Headey
2016
(45)
Headey
2017
(50)
Headey
2016
(45)
Ervin 2019
(1997‐2012) (40)
Buisman
2019 *
(30)
Headey
2017
(50)
Headey
2017
(50)
Category Determinant/Indicator 1997‐
2011
1997‐
2011
1997‐
2014
1996‐
2000
2000‐
2010
2005‐
2011
2000‐
2011
2000‐
2011 1999‐2005 1993‐
2006
1993‐
2006
2008‐
2014
2007‐
2013
2006‐
2013 1996‐2011 2001‐
2011
1996‐
2011
1996‐
2011
1991‐
2013 Rural Urban
2010‐
2014
1993‐
2011
2002‐
2014
Basic
Causes
Household
Socioeconomic
Status
Household Income 8% 3%
Livestock
17% 532% 37% 29%
Asset Index 13% 23% 25% 42% 9% 23% 12% 26% 28% 33% 20% 9%
Wealth 8%
Occupaon 15%
Literacy
Maternal Educaon 8% 12%
13%
6% 7%
4%
10% 10%
30%
11% 12% 14%
11%
17% 18% 25%
9%
Paternal Educaon 5% 5% 10% 3% 80% 3% 3% 2% 6% 10% 1% 2%
Region Rural Residence 2%
Underlying
Causes
Water,
Sanitaon,
and Hygiene
Open Defecaon 6% 8% 3% 8% 10% 7% 10% 14% 12% 17%
Safe Water (piped
water, tube well) ‐25% 0% ‐2% ‐1% 7% 10% 6% 0%
Improved Sanitaon 1% 12% 18% 7% 10% 14%
Unhealthy
Household
Environment
Household Size 4% 118%
Bed nets 35%
Health
Services
Skilled Birth Aendant
25% 52% 29%
Mother received 4+
antenatal care visits 4%
5% 7% 3%
3%
7% 10%
3% 4%
16% 13% 40% 34% 2%
Place of Birth/
Delivered at Health
Facility
4% 11% 10% 17% 15%
Vaccinaon N/A 6% 4% 3% 3%
Feeding
Pracces Breaseeding ‐85% 3% 1%
Immediate
Causes
Maternal
Characteriscs
Parity
10%
6%
3% 4% 7%
‐5%
Interpregnancy
Interval 3% 8%
28% 29%
4% 13% 13%
24%
3%
Birth Order 4% 14% 10% 9%
Maternal Age 1%
Maternal Height 5% 3% 4% 10% 1%
Maternal BMI 2%
Infecous
comorbidies Diarrhea 1%
Overall Variance (%) explained by
model†53% 63% 57% 32% 94% 29% 22% 15% 122% 57% 58% 23% 55% 13% 58% 79% 82% 70% 123% 104% 192% 77% 66% 48%
Buisman 2019: Maternal Risk (birth order, birth interval >24 months, mothers taller than 150 cm, mother’sage at birth)
Ervin 2019: ln(income), delayed vaccines, child breastfed at birth, ln(birth interval)
Ikeda 2013: Outcome is stunting prevalence, all other studies included outcome is HAZ
∗0–23 months
∗∗0–47 months
∗∗∗0–10 years
†The total variance is the variance calculated by the study authors. Some models have adjusted for other covariatesthat have not been included in this table.

786S Supplement
A total of 40 included studies explored the association between
childhood stunting and improved water sources (15,21,28–30,
33,37,39–41,43,46–51,54,55,58,59,62,66,68–70,74,76,
77,79,83,87,89,90,93,96,98,99,101,102). These studies
used a variety of methods and examined associations with several
indicators measuring access to clean water, including presence
of improved sources, unimproved sources, and physical distance
to water sources. There was variability in the signicance and
magnitude of the relation between improved sources of water and
stunting.
Optimal antenatal care coverage and place of birth.
High antenatal care coverage within a population is necessary
to optimize maternal health and nutrition, as well as fetal growth
and development. Evidence from a study of available health
services in several LMICs demonstrated that a mother attending
≥4 antenatal care visits (ANC4+) with ≥1 visit with a skilled
medical professional has been associated with a reduced risk of
stunting (113). In addition, improved access to health care and
skilled birth attendance at a health facility is associated with
increased HAZ scores in children (114).
The extent to which improved antenatal care coverage pre-
dicted changes in child growth varied widely across the countries
of interest. A combined measure of increases in coverage in
ANC4+and health facility births or skilled birth attendance
accounted for 40%, 34%, and 29% of the change in HAZ
in Pakistan (49), Senegal (50), and Rwanda (30), respectively.
Associations between child growth and ANC4+, facility birth,
or skilled attendance were analyzed in 14 different studies (19,
22,23,30,33,46,48–51,74,75,83,93).
Bed nets.
The largest predictor of stunting reduction and HAZ change in
Zambia (35%) was the change in the proportion of households
with bed nets (114), likely due in part to reductions in maternal
malaria risk in the population and improved birth outcomes (115).
Vaccination coverage.
High childhood vaccination coverage is an indicator of
a functional health system. Improved vaccination coverage
predicted between 4% (51) and 6% (33) of HAZ change in Nepal
and3%inParaguay(40). A total of 11 studies (14,22,33,40,48,
51,59,68,93,99,102) analyzed the relation between vaccination
coverage and stunting.
Breastfeeding practices.
In addition to being an optimal nutrition source, exclusive
breastfeeding for the rst 6 mo of life followed by continued
breastfeeding for 2 y has a protective effect against diarrhea-
related morbidity and mortality by reducing exposure to water-
borne pathogens (116). Being breastfed at birth predicted 3% of
the change in HAZ in rural Paraguay (40). There were 14 studies
(21,23,29,40,43,47,55,74,81,82,87,93,98,102)that
explored the relation between improved breastfeeding practices
and childhood stunting prevalence. Most of these analyses
revealed signicant associations between ever breastfeeding,
breastfeeding duration, and child growth, although a handful
of countries displayed nonsignicant relations, including Brazil,
Dominican Republic, Honduras, Peru, and Sri Lanka.
Complementary feeding practices and food security.
A total of 3 studies analyzed the associations between dietary
intake and child growth outcomes including complementary food
selection (82), actual micronutrient intake (25), and consumption
of nonhuman milk (97). There were 4 studies that assessed
the association between indicators of food insecurity and child
growth (20,62,82,100). However, none of these analyses were
dynamic since they did not assess the relative predicted HAZ
change over an interval.
Immediate Stunting Determinants
Fertility.
Family planning improves birth spacing and is important
in preventing high-risk pregnancies among younger and older
mothers, as well as women who have experienced closely
spaced births (117). A longer time interval between births has
been associated with lowered odds of stunting and reduced
susceptibility to unfavorable outcomes for infants and children
(118). Family-planning interventions may also reduce the number
of children ever born to a mother, also known as parity (117).
The association between fertility and stunting can be linked to
the former’s effect on preceding birth intervals (119), as longer
birth intervals are thought to increase the amount of “nutrition-
specic resources” available to individual children (118).
Declines in parity accounted for <7% of the observed HAZ
change in the countries assessed (49,50). While interpregnancy
interval predicted 13% of HAZ change in Paraguay (40) and 8%
of HAZ change in Cambodia (55), this value was only 3–4%
across other countries analyzed (48–51).
A district-level multilevel ecological analysis in Peru (120)
did not nd a signicant association between total fertility rate
and stunting, whereas a pooled multicountry study (121)en-
compassing 23 countries found a signicant association between
fertility rate (births per 1000 women) and stunting. Three studies
(22,59,70) examined the relation between childhood stunting
and access to family planning including modern contraceptive
use.
Maternal height.
In a cross-country analysis of several LMICs, maternal height
was found to be negatively correlated with stunting in infants and
children, highlighting the importance of maternal nutrition and
early-life factors on maternal growth and the effect on offspring
(122).
There was considerable variability among the prediction
estimates across countries analyzed. The largest estimates of
maternal height predicting HAZ change were seen in Nepal,
with values of 4% (33) and 10% (51) of HAZ change explained
provided in separate analyses, while in Bangladesh these values
were 3–5% (48,50). In Rwanda, 24% was explained by a
combined measure of maternal age, height, and interpregnancy
interval (30).

Global overview of child stunting determinants 787S
Low birth weight.
Being born with a low birth weight (<2500 g) can be an
indicator of fetal growth restriction in utero, a process that can
contribute to linear growth faltering. National analyses from
Bangladesh (99), Malawi (37), Sri Lanka (82), and Uganda
(101) examined the relation between a child’s low birth weight
and stunting as an outcome. In all studies, improved birth
outcomes (i.e., increased birth weight or reduced low birth
weight) were signicantly associated with improved measures of
child growth. However, only the Sri Lanka study used actual birth
weight measured in the hospital, while others used a categorical
subjective measure of relative size at birth.
Dietary diversity.
Dietary diversity scores are used as an indicator of diet quality
and density of micronutrients and macronutrients required for
optimal growth and development. Inadequate dietary diversity
is associated with increased odds of childhood stunting (123).
One national study from Sri Lanka using multivariate regression
did not nd a signicant relation between dietary diversity and
child stunting (82), while a multinational study using logistic
regression found that this relation was only signicant in India
(64).
Diarrhea.
Diarrhea incidence has been found to be associated with
stunting in young children, although ndings have been incon-
sistent (124) and effect sizes are generally small (107). Diarrhea
itself may not represent a direct cause of growth faltering, but
rather, indicate enteric inammation and dysfunction. Recent
ndings from the Etiology, Risk Factors and Interactions of
Enteric Infections and Malnutrition and the Consequences for
Child Health and Development (MAL-ED) birth cohort study
(125) revealed that children with enteric pathogens had enteric
inammation and reduced linear growth, even when diarrhea was
not present. Another recent cohort study (126) from Bangladesh
found that diarrhea caused by certain pathogens was associated
with linear growth but not all-cause diarrhea.
Reductions in diarrhea frequency predicted only 1% of HAZ
change in Cambodia (55). Diarrhea was signicantly associated
with odds of stunting in Cambodia (55), Bangladesh (41), Malawi
(37,75), and Uganda (101), although the effect size varied. The
relation between the incidence of diarrhea and growth outcomes
in children was examined in 11 studies in total (14,21,35,37,
41,55,62,66,75,97,98).
Discussion
Summary of evidence
Due to the very high heterogeneity and observational nature
of the data within the included studies, the aim of this review is
to identify broad patterns from existing national-level analyses
examining how determinants of stunting can predict child
growth outcomes. While the adapted tool used to assess the
methodological quality of included studies did not identify
meaningful differences in study quality, it is important to consider
the percentage of statistically explained HAZ change described
in studies with regression-decomposition analyses in the context
of the total variance explained (see bottom row of Table 1),
which can serve as an indicator of the strength of the model. For
example, the models produced through regression-decomposition
analyses for Ethiopia, Kenya, and Namibia have relatively
low total variance explained. Rather than attempt to interpret
individual estimates, this discussion highlights relatively large
predicted values that arise consistently across multiple countries.
Among the basic determinants of stunting assessed, improve-
ments in asset index score within households appeared to have
the strongest explanatory power within national-level regression-
decomposition analyses of the drivers of stunting reduction. This
was especially true for several South Asian countries, Senegal,
and Cambodia. Increasing parental educational levels was also
found to be a consistently strong predictor of improvements in
child growth outcomes.
Of the underlying determinants of stunting, reduction in the
prevalence of open defecation and improved sanitation infra-
structure were relatively important drivers of HAZ improvement
in Cambodia, Guinea, India, Nepal, and Pakistan. Independent
and combined measures of access to key maternal health services,
including optimal antenatal care coverage and delivery in a health
facility or with a skilled birth attendant, also accounted for
substantially improved child growth, although the magnitude of
variation explained differed substantially between countries.
Due to the unavailability of robust data collection for nutrition-
specic factors within DHS and MICS datasets, there was
less variety in the indicators representing the most immediate
determinants of stunting, including dietary intake and birth
outcomes. Several maternal characteristics predicted modest
stunting reduction across the countries analyzed, including parity,
interpregnancy interval, and maternal height.
Given the nature of these analyses, it is important to consider
possible nonlinearity in some of the associations between
determinants and child growth outcomes. For example, there
appears to be a nonlinear relation between the prevalence of open
defecation and mean HAZ scores within populations (48). This
means that a 20 percentage point decrease in open-defecation
prevalence from 80% to 60% compared with 30% to 10% may
have very different impacts on child growth. This can potentially
explain why reductions in open-defecation rates in Bangladesh
predicted relatively less improvement in HAZ compared with
other countries considered and suggests diminishing returns.
These ndings are generally aligned with those from existing
econometric analyses (12,13) of the key drivers of stunting
decline over the past few decades, including improvements in
household asset index, parental education, health service access
(ANC4+), and sanitation infrastructure. However, clear gaps in
the evidence include those determinants where data availability
and subsequent analyses were scarce, the most glaring of which
are the lack of analyses on how dietary intake and diversity
predict changes in nutrition status.
Limitations
Due to the observational nature of the survey data discussed
in the included studies, making causal inferences from the
prediction values produced by regression-decomposition tech-
niques is not possible. Additionally, many analyses had a very

788S Supplement
high proportion of unexplained variance or generated models
that explained >100% of variation in HAZ change in a given
country. This suggests that there may be other potential drivers
of stunting reduction that have yet to be theorized, measured,
or analyzed—some of which could be particularly important to
the unique stories of stunting reduction in individual countries.
The risk of omitted variable bias is a potential issue for the
analyses of observational data, and a high percentage prediction
value may represent a strong association but does not suggest a
reduced risk of confounding. There are also potential limitations
related to the datasets available, as we were not able to assess
the quality of the stunting determinant indicator variables, nor
the anthropometric data quality. Nevertheless, the regression-
decomposition approach is relatively agnostic in its assessment
of multiple stunting determinants at the national or subnational
level, lending comprehensiveness and rigor to these analyses of
observational data.
Not all LMICs were represented among the analyses dis-
cussed, and therefore this is not a globally exhaustive synthesis of
the drivers of national stunting reduction. While there was good
South Asian region representation, there was a particular lack of
regression-decomposition analyses from countries in the African
region. Sparse data from fragile and conict settings limited our
assessment of the determinants of stunting in these contexts.
Future national-level explorations of the determinants of stunting
may reveal additional important drivers of reduction. Despite
having conducted thorough database searches, there remains the
possibility of incomplete retrieval of studies that would have been
eligible for inclusion and may have affected the interpretation of
the overall results.
Rationale for examining exemplars in stunting reduction
with in-depth country case studies using mixed methods
Despite apparent progress on stunting reduction worldwide,
regional trends do not illustrate the large variations in the rate
of stunting reduction at the national level. Some countries have
made excellent progress, while others lag behind. In order to
rene our understanding of the drivers of changes in childhood
linear growth faltering and generate meaningful and granular
recommendations that countries can act upon, it is necessary
to unpack the contributing factors surrounding these national
variations in decline. In particular, it is helpful to focus analyses
on periods of rapid national reductions in stunting prevalence
in order to effectively determine the factors that accounted for
these steep declines. This necessarily involves assessing which
programs and policies have successfully predicted changes in
coverage of key indicators.
Quantitative analyses of national survey data can provide
an indication of which sectors were important to the national
stunting-reduction story. However, in-depth country case studies
using both quantitative and qualitative methods—folding in
higher-resolution data on key indicators at the subnational level—
can provide a more comprehensive and nuanced picture of the
drivers of stunting reduction. This is especially important for
examining within-country inequities in stunting reduction, which
can be just as wide as the variation between countries in a given
region. The other articles in this supplement issue describe the
methods and results of in-depth case studies in 5 countries that
have made exemplary progress in stunting reduction despite only
modest economic growth.
Conclusions
There are unique sets of stunting determinants that have
predicted stunting reduction among countries that have reduced
stunting, although there are several common drivers at the basic,
underlying, and immediate level. Determinants identied to be
particularly impactful include improvements in maternal and
paternal education, household socioeconomic status, sanitation
conditions, maternal health services access, and family planning.
There is a need to conduct in-depth, retrospective, and mixed-
methods case studies of determinants of stunting decline over
multiple decades in order to overcome the limitations inherent
in the existing literature and analyses of national survey data.
The authors’ responsibilities were as follows—TV, NA, and ZAB:
designed the research; MS, AS, SA, and TV: conducted the research; MS, AS,
SA, and TV: analyzed data; TV, MS, and AS: drafted the manuscript; NA and
ZAB: critically revised the manuscript; ZAB: had primary responsibility for
the nal content; and all authors: read and approved the nal manuscript. The
authors report no conicts of interest.
References
1. Black RE, Victora CG, Walker SP, Bhutta ZA, Christian P, De Onis
M, Ezzati M, Grantham-McGregor S, Katz J, Martorell R. Maternal
and child undernutrition and overweight in low-income and middle-
income countries. Lancet North Am Ed 2013;382(9890):427–51.
2. Sudfeld CR, Charles McCoy D, Danaei G, Fink G, Ezzati M,
Andrews KG, Fawzi WW. Linear growth and child development
in low- and middle-income countries: a meta-analysis. Pediatrics
2015;135(5):e1266–e75.
3. Perumal N, Bassani DG, Roth DE. Use and misuse of stunting as a
measure of child health. J Nutr 2018;148(3):311–15.
4. Leroy JL, Frongillo EA. Perspective: what does stunting really mean?
A critical review of the evidence. Adv Nutr 2019;10(2):196–204.
5. Hoddinott J, Alderman H, Behrman JR, Haddad L, Horton S. The
economic rationale for investing in stunting reduction. Matern Child
Nutr 2013;9:69–82.
6. UNICEF/WHO/World Bank. Joint child malnutrition estimates,
March 2020 edition. 2020 [cited 2020 May 05]. Available from: https:
//data.unicef.org/topic/nutrition/malnutrition/.
7. Olon I, McDonald CM, Ezzati M, Flaxman S, Black RE, Fawzi
WW, Cauleld LE, Danaei G, for the Nutrition Impact Model
Study (anthropometry cohort pooling). Associations of suboptimal
growth with all-cause and cause-specic mortality in children under
ve years: a pooled analysis of ten prospective studies. PLoS One
2013;8(5):e64636.
8. UNICEF. Strategy for improved nutrition of children and women in
developing countries. New York: United Nations Children’s Fund;
1990.
9. Stewart CP, Iannotti L, Dewey KG, Michaelsen KF, Onyango AW.
Contextualising complementary feeding in a broader framework for
stunting prevention. Matern Child Nutr 2013;9:27–45.
10. Black RE, Allen LH, Bhutta ZA, Cauleld LE, De Onis M, Ezzati
M, Mathers C, Rivera J. Maternal and child undernutrition: global
and regional exposures and health consequences. Lancet North Am
Ed 2008;371(9608):243–60.
11. Danaei G, Andrews KG, Sudfeld CR, Fink G, McCoy DC, Peet
E, Sania A, Fawzi MCS, Ezzati M, Fawzi WW. Risk factors for
childhood stunting in 137 developing countries: a comparative risk
assessment analysis at global, regional, and country levels. PLoS Med
2016;13(11):e1002164.
12. Smith LC, Haddad L. Reducing child undernutrition: past drivers and
priorities for the post-MDG era. World Dev 2015;68:180–204.
13. Headey DD. Developmental drivers of nutritional change: a cross-
country analysis. World Dev 2013;42:76–88.

Global overview of child stunting determinants 789S
14. Adesugba M, Edeh H, Mavrotas G. Child nutritional status, welfare
and health in Nigerian households. IFPRI—Discussion Papers; 2018
(1776). Washington, DC: International Food Policy Research Institute;
2018.
15. Adhikari RP, Shrestha ML, Acharya A, Upadhaya N. Determinants of
stunting among children aged 0–59 months in Nepal: ndings from
Nepal Demographic and Health Survey, 2006, 2011, and 2016. BMC
Nutr 2019;5(1):37.
16. Adjaye-Gbewonyo K, Vollmer S, Avendano M, Harttgen K.
Agricultural trade policies and child nutrition in low- and
middle-income countries: a cross-national analysis. Global Health
2019;15(1):21.
17. Akombi BJ, Agho KE, Renzaho AM, Hall JJ, Merom DR. Trends
in socioeconomic inequalities in child undernutrition: evidence from
Nigeria Demographic and Health Survey (2003–2013). PLoS One
2019;14(2):e0211883.
18. Alfani F, Dabalen A, Fisker P, Molini V. Vulnerability to Malnutrition
in the West African Sahel. Washington, DC: World Bank Group; 2015.
Contract No.: 7171.
19. Almasi A, Saeidi S, Zangeneh A, Ziapour A, Choobtashani M,
Saeidi F, Salahshoor MR, Azar FEF. Investigation of some factors
affecting stunting and wasting among the under-ve children in
eastern Mediterranean region. Int J Pediatr Mashhad 2019;7(7):
9759–72.
20. Amaral MM, Herrin WE, Gulere GB. Using the Uganda National
Panel Survey to analyze the effect of staple food consumption
on undernourishment in Ugandan children. BMC Public Health
2018;18(1):32.
21. Amarante V, Figueroa N, Ullman H. Inequalities in the reduction of
child stunting over time in Latin America: evidence from the DHS
2000–2010. Oxford Development Studies 2018;46(4):519–35.
22. Ambel AA, Andrews C, Bakilana AM, Foster EM, Khan Q, Wang
H. Examining changes in maternal and child health inequalities in
Ethiopia. Int J Equity Health 2017;16(1):152.
23. Angdembe MR, Dulal BP, Bhattarai K, Karn S. Trends and predictors
of inequality in childhood stunting in Nepal from 1996 to 2016. Int J
Equity Health 2019;18(1):42.
24. Bagmar SH, Khudri M. Tracking changes and identifying
determinants of child malnutrition status over the past decade in
Bangladesh. Pakistan J Nutr 2015;14(12):964–71.
25. Ballew C, Khan LK, Kaufmann R, Mokdad A, Miller DT, Gunter EW.
Blood lead concentration and children’s anthropometric dimensions
in the Third National Health and Nutrition Examination Survey
(NHANES III), 1988–1994. J Pediatr 1999;134(5):623–30.
26. Barankanira E, Molinari N, Msellati P, Laurent C, Bork KA. Stunting
among children under 3 years of age in Cote d’Ivoire: spatial and
temporal variations between 1994 and 2011. Public Health Nutr
2017;20(9):1627–39.
27. Biadgilign S, Ayenew HY, Shumetie A, Chitekwe S, Tolla A, Haile
D, Gebreyesus SH, Deribew A, Gebre B. Good governance, public
health expenditures, urbanization and child undernutrition nexus in
Ethiopia: an ecological analysis. BMC Health Serv Res 2019;19(1):
40.
28. Biadgilign S, Shumetie A, Yesigat H. Does economic growth
reduce childhood undernutrition in Ethiopia? PLoS One
2016;11(8):e0160050.
29. Boccanfuso D, Bruce O. A new avenue for understanding
the nutritional status of children in Guinea. Eur J Dev Res
2013;25(5):714–36.
30. Buisman LR, Van de Poel E, O’Donnell O, van Doorslaer EKA. What
explains the fall in child stunting in sub-Saharan Africa? SSM Popul
Health 2019;8:100384.
31. Christiaensen L, Alderman H. Child malnutrition in Ethiopia: can
maternal knowledge augment the role of income? Econ Dev Cultural
Change 2004;52(2):287–312.
32. Corsi DJ, Gaffey MF, Bassani DG, Subramanian SV. No female
disadvantage in anthropometric status among children in India:
analysis of the 1992–1993 and 2005–2006 Indian National Family
Health Surveys. J South Asian Dev 2015;10(2):119–47.
33. Cunningham K, Headey D, Singh A, Karmacharya C, Rana PP.
Maternal and child nutrition in Nepal: examining drivers of progress
from the mid-1990s to 2010s. Global Food Security 2017;13:30–7.
34. da Silva ICM, Franca GV, Barros AJD, Amouzou A, Krasevec
J, Victora CG. Socioeconomic inequalities persist despite declining
stunting prevalence in low- and middle-income countries. J Nutr
2018;148(2):254–8.
35. Darrouzet-Nardi AF. Nonviolent civil insecurity is negatively
associated with subsequent height-for-age in children aged <5 y born
between 1998 and 2014 in rural areas of Africa. Am J Clin Nutr
2017;105(2):485–93.
36. Das Gupta M, Gragnolati M, Ivaschenko O, Lokshin M. Improving
child nutrition outcomes in india: can the integrated child development
services program be more effective? Washington, DC: World Bank
Group; 2005. Contract No.: No. 3647.
37. Doctor HV, Nkhana-Salimu S. Trends and determinants of
child growth indicators in malawi and implications for the
Sustainable Development Goals. AIMS Public Health 2017;4(6):
590–614.
38. Efevbera Y, Bhabha J, Farmer PE, Fink G. Girl child marriage as a
risk factor for early childhood development and stunting. Soc Sci Med
2017;185:91–101.
39. World Bank. Stunting reduction in Sub-Saharan Africa. Washington,
DC: World Bank Group; 2017.
40. Ervin PA, Bubak V. Closing the rural-urban gap in child malnutrition:
evidence from Paraguay, 1997–2012. Econ Hum Biol 2019;32:
1–10.
41. Goyal N, Canning D. Exposure to ambient ne particulate air pollution
in utero as a risk factor for child stunting in Bangladesh. Int J Environ
Res Public Health 2017;15(1):23.
42. Greffeuille V, Sophonneary P, Laillou A, Gauthier L, Hong R, Hong R,
Poirot E, Dijkhuizen M, Wieringa F, Berger J. Persistent inequalities
in child undernutrition in Cambodia from 2000 until today. Nutrients
2016;8(5):16.
43. Hangoma P, Aakvik A, Robberstad B. Explaining changes in child
health inequality in the run up to the 2015 Millennium Development
Goals (MDGs): the case of Zambia. PLoS One 2017;12(2):
e0170995.
44. Harttgen K, Klasen S, Vollmer S. Economic growth and child
undernutrition in sub-Saharan Africa. Popul Dev Rev 2013;39(3):397–
412.
45. Hasan MT, Soares Magalhaes RJ, Williams GM, Mamun AA. The
role of maternal education in the 15-year trajectory of malnutrition
in children under 5 years of age in Bangladesh. Matern Child Nutr
2016;12(4):929–39.
46. Headey D. An analysis of trends and determinants of child
undernutrition in Ethiopia, 2000-2011. Washington, DC: International
Food Policy Research Institute; 2014. Contract No.: 70.
47. Headey D, Hirvonen K, Hoddinott J. Animal sourced foods and child
stunting. Am J Agric Econ 2018;100(5):1302–19.
48. Headey D, Hoddinott J, Ali D, Tesfaye R, Dereje M. The other
Asian enigma: explaining the rapid reduction of undernutrition in
Bangladesh. World Dev 2015;66:749–61.
49. Headey D, Hoddinott J, Park S. Drivers of nutritional change in four
South Asian countries: a dynamic observational analysis. Matern Child
Nutr 2016;12(Suppl 1):210–18.
50. Headey D, Hoddinott J, Park S. Accounting for nutritional changes in
six success stories: a regression-decomposition approach. Glob Food
Secur-AgricPolicy 2017;13:12–20.
51. Headey DD, Hoddinott J. Understanding the rapid reduction
of undernutrition in Nepal, 2001–2011. PLoS One
2015;10(12):e0145738.
52. Hoffman D, Cacciola T, Barrios P, Simon J. Temporal changes and
determinants of childhood nutritional status in Kenya and Zambia. J
Health Popul Nutr 2017;36(1):27.
53. Huda TM, Hayes A, El Arifeen S, Dibley MJ. Social determinants of
inequalities in child undernutrition in Bangladesh: a decomposition
analysis. Matern Child Nutr 2018;14(1):e12440.
54. Huicho L, Huayanay-Espinoza CA, Herrera-Perez E, Segura ER, Nino
de Guzman J, Rivera-Ch M, Barros AJ. Factors behind the success
story of under-ve stunting in Peru: a district ecological multilevel
analysis. BMC Pediatr 2017;17(1):29.
55. Ikeda N, Irie Y, Shibuya K. Determinants of reduced child stunting in
Cambodia: analysis of pooled data from three demographic and health
surveys. Bull World Health Organ 2013;91(5):341–9.
56. Jahagirdar D, Harper S, Heymann J, Swaminathan H, Mukherji A,
Nandi A. The effect of paid maternity leave on early childhood
growth in low-income and middle-income countries. BMJ Glob Health
2017;2(3):e000294.

790S Supplement
57. Joe W, Rajaram R, Subramanian SV. Understanding the null-to-small
association between increased macroeconomic growth and reducing
child undernutrition in India: role of development expenditures and
poverty alleviation. Matern Child Nutr 2016;12(Suppl 1):196–209.
58. Joshi N, Bolorhon B, Narula I, Zhu S, Manaseki-Holland S. Social
and environmental determinants of child health in Mongolia across
years of rapid economic growth: 2000–2010. Int J Equity Health
2017;16(1):189.
59. Kabubo-Mariara J, Ndenge GK, Mwabu DK. Determinants of
children’s nutritional status in Kenya:evidence from Demographic and
Health Surveys. J African Econ 2009;18(3):363–87.
60. Khatun W, Rasheed S, Alam A, Huda TM, Dibley MJ. Assessing the
intergenerational linkage between short maternal stature and under-
ve stunting and wasting in Bangladesh. Nutrients 2019;11(8):07.
61. Kien VD, Lee HY, Nam YS, Oh J, Giang KB, Van Minh H. Trends in
socioeconomic inequalities in child malnutrition in Vietnam: ndings
from the Multiple Indicator Cluster Surveys, 2000–2011. Glob Health
Action 2016;9:29263.
62. Kinyoki DK, Berkley JA, Moloney GM, Kandala NB, Noor AM.
Predictors of the risk of malnutrition among children under the
age of 5 years in Somalia. Public Health Nutr 2015;18(17):
3125–33.
63. Kov P, Smets S, Spears D, Vyas S. Growing taller among toilets:
evidence from changes in sanitation and child height in Cambodia,
2005–2010. Amston, CT: Research Institute for Compassionate
Economics Working Paper; 2013.
64. Krishna A, Mejia-Guevara I, McGovern M, Aguayo VM,
Subramanian SV. Trends in inequalities in child stunting in South
Asia. Matern Child Nutr 2018;14(Suppl 4):e12517.
65. Lovo S, Veronesi M. Crop diversication and child health: empirical
evidence from Tanzania. Ecol Econ 2019;158:168–79.
66. 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(9):1715–27.
67. Matanda DJ, Mittelmark MB, Kigaru DM. Child undernutrition in
Kenya: trend analyses from 1993 to 2008–09. BMC Pediatr 2014;
14:5.
68. Milman A, Frongillo EA, de Onis M, Hwang JY. Differential
improvement among countries in child stunting is associated
with long-term development and specic interventions. J Nutr
2005;135(6):1415–22.
69. Mohsena M, Goto R, Mascie-Taylor CG. Socioeconomic and
demographic variation in nutritional status of under-ve Bangladeshi
children and trend over the twelve-year period 1996–2007. J Biosoc
Sci 2017;49(2):222–38.
70. Monteiro CA, Benicio MH, Conde WL, Konno S, Lovadino AL,
Barros AJ, Victora CG. Narrowing socioeconomic inequality in child
stunting: the Brazilian experience, 1974–2007. Bull World Health
Organ 2010;88(4):305–11.
71. Monteiro CA, Benicio MH, Konno SC, Silva AC, Lima AL, Conde
WL. Causes for the decline in child under-nutrition in Brazil, 1996–
2007. Rev Saude Publica 2009;43(1):35–43.
72. Nepali S, Simkhada P, Davies I. Trends and inequalities in stunting
in Nepal: a secondary data analysis of four Nepal demographic health
surveys from 2001 to 2016. BMC Nutrition 2019;5(1):19.
73. Nie P, Rammohan A, Gwozdz W, Sousa-Poza A. Changes in child
nutrition in India: a decomposition approach. Int J Environ Res Public
Health 2019;16(10):22.
74. Nisar YB, Dibley MJ, Aguayo VM. Iron-folic acid supplementation
during pregnancy reduces the risk of stunting in children less than
2 years of age: a retrospective cohort study from Nepal. Nutrients
2016;8(2):67.
75. Ntenda PAM, Chuang YC. Analysis of individual-level and
community-level effects on childhood undernutrition in Malawi.
Pediatr Neonatol 2018;59(4):380–9.
76. O’Donnell O, Nicolas AL, Van Doorslaer E. Growing richer and taller:
explaining change in the distribution of child nutritional status during
Vietnam’s economic boom. J Dev Econ 2009;88(1):45–58.
77. Otterbach S, Rogan M. Exploring spatial differences in the risk of child
stunting: evidence from a South African national panel survey. J Rural
Stud 2019;65:65–78.
78. Pierce H. Increasing health facility deliveries in Cambodia and its
inuence on child health. Int J Equity Health 2019;18(1):67.
79. Ponce N, Shimkhada R, Raub A, Daoud A, Nandi A, Richter L,
Heymann J. The association of minimum wage change on child
nutritional status in LMICs: a quasi-experimental multi-country study.
Glob Public Health 2018;13(9):1307–21.
80. Rabassa M, Skouas E, Jacoby H. Weather and child health in rural
Nigeria. Washington, DC: World Bank; 2012. Report No.: 1813–9450;
Contract No.: 6214.
81. Rabbani A, Khan A, Yusuf S, Adams A. Trends and determinants of
inequities in childhood stunting in Bangladesh from 1996/7 to 2014.
Int J Equity Health 2016;15(1):186.
82. Rannan-Eliya RP, Hossain SM, Anuranga C, Wickramasinghe R,
Jayatissa R, Abeykoon AT. Trends and determinants of childhood
stunting and underweight in Sri Lanka. Ceylon Med J 2013;58(1):10–
18.
83. Rashad A, Sharaf M. Does economic growth reduce child
malnutrition in Egypt? New evidence from National Demographic
and Health Survey. Working Papers—Department of Economics,
University of Alberta; 2015. Alberta, Canada: University of Alberta;
2015.
84. Restrepo-Mendez MC, Barros AJ, Black RE, Victora CG. Time
trends in socio-economic inequalities in stunting prevalence: analyses
of repeated national surveys. Public Health Nutr 2015;18(12):
2097–104.
85. Rieger M, Trommlerova SK, Ban R, Jeffers K, Hutmacher M.
Temporal stability of child growth associations in Demographic and
Health Surveys in 25 countries. SSM Popul Health 2019;7:100352.
86. Saha UR, Chattapadhayay A, Richardus JH. Trends, prevalence and
determinants of childhood chronic undernutrition in regional divisions
of Bangladesh: evidence from demographic health surveys, 2011 and
2014. PLoS One 2019;14(8):e0220062.
87. Smith T, Shively G. Multilevel analysis of individual, household, and
community factors inuencing child growth in Nepal. BMC Pediatr
2019;19(1):91.
88. Sobrino M, Gutierrez C, Alarcon J, Davila M, Cunha AJ. Birth interval
and stunting in children under ve years of age in Peru (1996–2014).
Child Care Health Dev 2017;43(1):97–103.
89. Spears D. How much international variation in child height can
sanitation explain? Washington, DC: World Bank Group; 2013.
Contract No.: WPS 6351.
90. Stifel D, Alderman H. The “Glass of Milk” subsidy program and
malnutrition in Peru. World Bank Econ Rev 2006;20(3):421–48.
91. Subramanyam MA, Kawachi I, Berkman LF, Subramanian
SV. Socioeconomic inequalities in childhood undernutrition
in India: analyzing trends between 1992 and 2005. PLoS One
2010;5(6):e11392.
92. Subramanyam MA, Kawachi I, Berkman LF, Subramanian SV. Is
economic growth associated with reduction in child undernutrition in
India? PLoS Med 2011;8(3):e1000424.
93. Sunil TS, Sagna M. Decomposition of childhood malnutrition in
Cambodia. Matern Child Nutr 2015;11(4):973–86.
94. Surkan PJ, Ettinger AK, Ahmed S, Minkovitz CS, Strobino D. Impact
of maternal depressive symptoms on growth of preschool- and school-
aged children. Pediatrics 2012;130(4):e847–55.
95. Urke HB, Mittelmark MB, Valdivia M. Trends in stunting and
overweight in Peruvian pre-schoolers from 1991 to 2011: ndings
from the Demographic and Health Surveys. Public Health Nutr
2014;17(11):2407–18.
96. Wagstaff A, van Doorslaer E, Watanabe N. On decomposing
the causes of health sector inequalities with an application to
malnutrition inequalities in Vietnam. J Econometrics 2003;112(1):
207–23.
97. Woodruff BA, Wirth JP, Bailes A, Matji J, Timmer A, Rohner F.
Determinants of stunting reduction in Ethiopia 2000-2011. Matern
Child Nutr 2017;13(2):e12307.
98. Woodruff BA, Wirth JP, Ngnie-Teta I, Beauliere JM, Mamady D,
Ayoya MA, Rohner F. Determinants of stunting, wasting, and anemia
in Guinean preschool-age children: an analysis of DHS data from
1999, 2005, and 2012. Food Nutr Bull 2018;39(1):39–53.
99. World Bank. Maintaining momentum to 2015: an impact evaluation
of interventions to improve maternal and child health and nutrition
in Bangladesh. Washington, DC: World Bank, Operations Evaluation
Department; 2005.
100. Wu L, Yang Z, Yin SA, Zhu M, Gao H. The relationship between
socioeconomic development and malnutrition in children younger than

Global overview of child stunting determinants 791S
5 years in China during the period 1990 to 2010. Asia Pac J Clin Nutr
2015;24(4):665–73.
101. Yang YY, Kaddu G, Ngendahimana D, Barkoukis H, Freedman D,
Lubaale YA, Mupere E, Bakaki PM. Trends and determinants of
stunting among under-5s: evidence from the 1995, 2001, 2006 and
2011 Uganda Demographic and Health Surveys. Public Health Nutr
2018;21(16):2915–28.
102. Zanello G, Srinivasan CS, Shankar B. What explains Cambodia’s
success in reducing child stunting—2000-2014? PLoS One
2016;11(9):e0162668.
103. Fortin NM. The gender wage gap among young adults in the United
States: the importance of money versus people. J Hum Resour
2008;43(4):884–918.
104. Yun MS. A simple solution to the identication problem in detailed
wage decompositions. Econ Inq 2005;43(4):766–72.
105. Fairlie RW. An extension of the Blinder-Oaxaca decomposition
technique to logit and probit models. JEM 2005;30(4):
305–16.
106. Vollmer S, Bommer C, Krishna A, Harttgen K, Subramanian S. The
association of parental education with childhood undernutrition in
low-and middle-income countries: comparing the role of paternal and
maternal education. Int J Epidemiol 2016;46(1):312–23.
107. Checkley W, Buckley G, Gilman R, Assis A, Guerrant R, Morris S,
Molbak K, Valentiner-Branth P, Lanata C, Black R. Multi-country
analysis of the effects of diarrhoea on childhood stunting. Int J
Epidemiol 2008;37(4):816–30.
108. Humphrey JH. Child undernutrition, tropical enteropathy,
toilets, and handwashing. Lancet North Am Ed 2009;374(9694):
1032–5.
109. Budge S, Parker AH, Hutchings PT, Garbutt C. Environmental
enteric dysfunction and child stunting. Nutr Rev 2019;77(4):
240–53.
110. Tickell KD, Atlas HE, Walson JL. Environmental enteric dysfunction:
a review of potential mechanisms, consequences and management
strategies. BMC Med 2019;17(1):181.
111. Curtis V, Kanki B, Mertens T, Traore E, Diallo I, Tall F, Cousens S.
Potties, pits and pipes: explaining hygiene behaviour in Burkina Faso.
Soc Sci Med 1995;41(3):383–93.
112. Cairncross S, Valdmanis V. Water supply, sanitation and hygiene
promotion (chapter 41). In: Jamison D, Breman J, Measham A,
Alleyne G, Claeson M, Evans DB, Jha P, Mills A, Musgrove
P,, et al.editors. Disease control priorities in developing countries.
Washington, DC: The World Bank; 2006. p. 771–92.
113. Kuhnt J, Vollmer S. Antenatal care services and its implications
for vital and health outcomes of children: evidence from 193
surveys in 69 low-income and middle-income countries. BMJ Open
2017;7(11):e017122.
114. Headey D, Hoddinott J, Park S. Accounting for nutritional changes in
six success stories: a regression-decomposition approach. Global Food
Security 2017;13:12–20.
115. Eisele TP, Larsen DA, Anglewicz PA, Keating J, Yukich J, Bennett
A, Hutchinson P, Steketee RW. Malaria prevention in pregnancy,
birthweight, and neonatal mortality: a meta-analysis of 32 national
cross-sectional datasets in Africa. Lancet Infect Dis 2012;12(12):942–
9.
116. Lamberti LM, Walker CLF, Noiman A, Victora C, Black RE.
Breastfeeding and the risk for diarrhea morbidity and mortality. BMC
Public Health 2011;11(3):S15.
117. Smith R, Ashford L, Gribble J, Clifton D. Family planning saves lives.
Washington, DC: Population Reference Bureau; 2009.
118. Gribble JN, Murray NJ, Menotti EP. Reconsidering childhood
undernutrition: can birth spacing make a difference? An analysis of
the 2002–2003 El Salvador National Family Health Survey. Matern
Child Nutr 2009;5(1):49–63.
119. Rutstein SO. Further evidence of the effectsof preceding birth intervals
on neonatal infant and under-ve-years mortality and nutritional
status in developing countries: evidence from the Demographic and
Health Surveys 2008. DHS Working Paper 41. Calverton, MD: DHS;
2008.
120. Huicho L, Huayanay-Espinoza CA, Herrera-Perez E, Segura ER, de
Guzman JN, Rivera-Ch M, Barros AJ. Factors behind the success story
of under-ve stunting in Peru: a district ecological multilevel analysis.
BMC Pediatr 2017;17(1):29.
121. Ponce N, Shimkhada R, Raub A, Daoud A, Nandi A, Richter L,
Heymann J. The association of minimum wage change on child
nutritional status in LMICs: a quasi-experimental multi-country study.
Global Public Health 2018;13(9):1307–21.
122. Özaltin E, Hill K, Subramanian S. Association of maternal stature
with offspring mortality, underweight, and stunting in low-to middle-
income countries. JAMA 2010;303(15):1507–16.
123. Arimond M, Ruel MT. Dietary diversity is associated with child
nutritional status: evidence from 11 demographic and health surveys.
J Nutr 2004;134(10):2579–85.
124. Briend A. Is diarrhoea a major cause of malnutrition among the under-
ves in developing countries? A review of available evidence. Eur J
Clin Nutr 1990;44(9):611–28.
125. Kosek MN, Ahmed T, Bhutta Z, Cauleld L, Guerrant R, Houpt
E, Kang G, Kosek M, Lee G, Lima A. Causal pathways from
enteropathogens to environmental enteropathy: ndings from the
MAL-ED birth cohort study. EBioMedicine 2017;18:109–17.
126. Schnee AE, Haque R, Taniuchi M, Uddin MJ, Alam MM, Liu J,
Rogawski ET, Kirkpatrick B, Houpt ER, Petri WA Jr. Identication
of etiology-specic diarrhea associated with linear growth faltering in
Bangladeshi infants. Am J Epidemiol 2018;187(10):2210–18.