Jef Leroy, Kaleab Baye
WEBINAR
Thought-Provoking Perspectives on Stunting
Co-Organized by the National Information Platform for Nutrition (NIPN), the Ethiopian Public Health Institute (EPHI) and IFPRI
NOV 18, 2020 - 03:30 PM TO 05:00 PM EAT
2. What does stunting really mean?
Jef L Leroy, Edward Frongillo
International Food Policy Research Institute
Webinar NIPN Ethiopia
November 2020
3. Definitions ….
-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
height-for-age Z-score
All children suffer from growth retardation~60%
Stunted
4. Motivation
• Last decade:
• Unprecedented increase in attention to undernutrition;
• Reducing child stunting has become a global development objective.
• The strong focus on stunting:
• Enabled successful advocacy for nutrition;
• But also… led to confusion and misunderstanding about meaning of
stunting among researchers, donors, and agencies active in nutrition.
• My concern:
• “Overselling” the importance of stunting;
• Drawing attention away from where the focus should be;
• “Underdelivering” on stunting may cause the world to lose interest.
• Cause of this confusion: poor understanding of the difference
between growth as an outcome vs. a marker.
5. Outcome vs. marker – Esio Trot
Mr Hoppy
• shy old man;
• secretly in love with Mrs
Silver.
Mrs Silver
• has a small pet tortoise (Alfie);
• wishes she knew how to make
Alfie grow.
7. Growth in Esio Trot: theory of change
I say “Esio
trot”
Alphie’s
growth ↑
She says
“Esio trot”
Mrs
Silver’s
trust in
me↑
Mrs
Silver’s
love for
me ↑
Mrs Silver’s theory of
change
I swap
tortoise
Mr Hoppy’s theory of
change
Alphie’s
growth
Alphie’s growth:
outcome of interest
in and of itself
Alphie’s growth:
marker of trust
9. What we are writing…
Recent comprehensive review (McGovern, IJE 2017):
Over half of the 68 papers on linear growth or height made these types direct causal claims
11. The world has heard us…
Causal view is now strongly embedded among
agencies and donors active in nutrition:
• Widespread adoption of the view that stunting leads to
developmental delays, lower levels of schooling,
reduced earnings, and chronic disease risk.
• Linear growth retardation and stunting have become a
primary development objective.
• General belief that eliminating stunting will
automatically lead to meaningful benefits in a large
number of other domains.
14. What does the evidence
say?
Linear growth
retardation
Child
development
Schooling
Work capacity
Chronic
disease
Birth outcomes
of F1
Encephalop.
disproportion
Productivity,
earnings
Mortality,
morbidity
Mortality,
morbidity,
stunting of F1
Mortality,
morbidity
1
2
3
4
5
[cause]
16. Just so it is clear….
The evidence on causal effects relates
to linear growth retardation,
*not* to undernutrition
17. Summary of evidence
Linear growth
retardation
Child
development
Schooling
Work capacity
Chronic
disease
Birth outcomes
of F1
Encephalop.
disproportion
Productivity,
earnings
Mortality,
morbidity
Mortality,
morbidity,
stunting of F1
“causes
”
Mortality,
morbidity
1
2
3
4
5
“causes
”
“causes
”
“causes
”
“causes
”
√
√
18. Child
development
Schooling
Work capacity
Physiological
functioning
Birth outcomes
of F1
Encephalop.
disproportion
Productivity,
earnings
Mortality,
morbidity
Mortality,
morbidity,
stunting of F1
Linear growth
retardation
Mortality,
morbidity
[cause]
Deficient
environment
[cause]
[cause]
[markerofpast]
[markeroffuture]
New framework: outcome vs. marker
21. Just semantics?
“It does not matter. Whatever is good to reduce
stunting is also good for the other outcomes”
Three reasons to disagree:
• Improving linear growth is often not necessary:
We do not have to fix growth completely to have significant and
*meaningful* impacts on important outcomes.
• Improving linear growth is often not sufficient:
A well-growing child with anemia in an environment lacking
stimulation is unlikely to develop properly.
• Improving linear growth is often not cost-efficient:
Addressing outcomes directly is more efficient than trying to
address them indirectly through linear growth.
22. Just semantics?
“It does not matter. Whatever is good to reduce
stunting is also good for the other outcomes”
Three reasons to disagree:
• Improving linear growth is often not necessary:
We do not have to fix growth completely to have significant and
*meaningful* impacts on important outcomes.
• Improving linear growth is often not sufficient:
A well-growing child with anemia in an environment lacking
stimulation is unlikely to develop properly.
• Improving linear growth is often not cost-efficient:
Addressing outcomes directly is more efficient than trying to
address them indirectly through linear growth.
23. Just semantics?
“It does not matter. Whatever is good to reduce stunting
is also good for the other outcomes”
Three reasons to disagree:
• Improving linear growth is often not necessary:
We do not have to fix growth completely to have significant and
*meaningful* impacts on important outcomes.
• Improving linear growth is often not sufficient:
A well-growing child with anemia in an environment lacking stimulation
is unlikely to develop properly.
• Improving linear growth is often not cost-efficient:
Addressing outcomes directly is more efficient than trying to address
them indirectly through linear growth.
e.g.: obstructed labor can be prevented (nearly) entirely with good
operative delivery;
24. Just semantics? (2)
• Single focus on stunting unnecessarily downplays the
importance of interventions that “only” improve …
• infant & young feeding practices;
• dietary adequacy;
• water, hygiene, and sanitation practices.
• Single focus on stunting may undermine efforts to engage
other sectors in nutrition-sensitive programs:
• Agriculture sector: more likely to “sign up” for improvements in diet
than for linear growth.
• Social protection: more likely to commit to reducing poverty then to
reducing stunting.
• Sense of fatalism: “little or nothing that can be done after 2
years of age”
• Growth retardation continues beyond 2 years
• Interventions in other domains (e.g. ECD) can still make a difference
25. Proposed way forward
• Be specific
• Undernutrition ≠ stunting; stunting ≠ linear growth
retardation
• Be explicit about reason for focus on stunting:
• Population assessment? To count those affected? Program
design and evaluation?
• In programs, is it used as a marker of another outcome (and
why is that outcome not addressed directly?) or is it an
outcome of immediate interest (and why was it chosen as an
outcome?)
• Population assessment: ok
Good marker for population assessment (comparing
regions, assessing changes over time)
26. Proposed way forward (cont’d)
• Counting cases: limitations
• Use of stunting to count the number of children affected has
limitations (Perumal, 2018)
• Other exercises using stunting to count those affected (e.g.
cost per case of stunting averted) will result in inaccurate
estimates
• Programs and impact evaluation: growth not a good
primary outcome
• Stunting linked causally to two outcomes which can be more
effectively addressed through direct interventions
• Rather, focus on what matters *and* can sensibly “respond”
to the intervention:
e.g.: morbidity, dietary intake, food security, child development,
immune function, nutrient status, body composition, …
27. Thank you!
‘Why, Mr Hoppy!... Of course I’ll mary you!’
‘It’s all due to Alfie’, Mrs Silver said, slightly breathless.
‘Good old Alfie’, Mr Hoppy said. ‘We’ll keep him for
ever.’
29. Evaluation of Linear Growth at
Higher Altitudes
Kaleab Baye (Associate Prof.; PhD)
Center for Food Science and Nutrition, Addis Ababa University
Kalle Hirvonen (PhD; Senior Researcher)
International Food Policy Research Institute
30. Background
SDG Target
Progress is routinely measured using anthropometric indices (i.e. HAZ)
that compare child height to the WHO growth standards
31. Background
Key finding: all children have the
same growth potential when their
physiologic needs are met and the
environments support healthy
growth
development of a prescriptive
growth standard for children less
than 5 years of age
Excluded sites
>1,500 masl
Is the application of the WHO child growth
standard in high altitude settings justified?
32. Background
Characteristics of earlier studies
- Inconsistent findings
- Single country, often small sample
size
- Studies across low altitude range
(e.g. 3000-4000 masl)
- Potential confounding factors not
accounted
33. Background
• How many people live at high altitude?
In 2010, 842 million people in the global population (~12%) lived ≥1500 masl, with 67% in Asia and Africa
34. Objectives
• Given that stunting is now used as an indicator to
track progress on global nutrition targets and the
SDGs, it is necessary to revisit the association
between altitude and child growth
• We aimed to re-evaluate this relationship and
answer the following specific questions:
– Does residence at higher altitude increase the risk of
linear growth faltering?
– Does the prescriptive WHO growth standard equally
apply to children residing at higher altitudes?
35. Methods
• Association b/n altitude and child linear growth
was assessed by analyzing 133 nationally
representative DHS surveys conducted between
1992 and 2018
– ~1 M height measurements from 59 LMICs
– 96,552 clusters
– elevations ranging from -372 to 5951 masl
– common covariates including diet, disease, and
underlying causes like education, household wealth,
WASH, etc.
36. Methods
1. Using linear regression methods, we quantified the
child height deficit associated with altitude after
adjusting for confounding factors
2. We run multivariable regression models to assess
how altitude is associated with immediate causes of
malnutrition (diet & diseases)
3. Restricted the sample to children who resided in ideal
home environments and re-assessed the r/n between
altitude and child health
37. Methods
• Child growth trajectories above/below 1500 masl
were compared
• Robustness checks
– Climate factors
– Maternal height
– Residence in the same location since conception
– Omitting low altitude countries
– Excluding each country individually
– Using stunting as an outcome
39. Results
A 1000-masl ↑in altitude ~ a 0.163-unit (95%CI, −0.205 to −0.120 units) ↓in HAZ after
adjusting for common risk factors
Altitude-HAZ relationship
41. Results
“Ideal home environment “: a home
environment that allow children
achieve their full genetic potential
(Karra et al., 2016)
Criteria used to define “ideal home
environment”:
1. Singleton
2. Access to safe water and
sanitation
3. HH with finished floors, owning a
TV, and a car
4. Born to highly educated mothers
(> 13 yrs. of schooling)
5. Children that received BCG and
DTP vaccinations
44. Discussion
• Just shorter height or more than that?
– Only a small proportion are likely to benefit from
genetic adaptations related to residence in high
altitude over multiple generations
– Hypoxia induced IUGR is associated with cognitive
impairments (Miller et al., 2017; Moore et al., 2011)
– Abnormal cardiac functions (Patterson& Zhang, 2010)
– Increased risk of chronic disease in later life
(Moore et al., 2011; Barker 2006)
45. Discussion
• Is the magnitude of the growth deficit
biologically relevant?
La Paz, 3600 masl- 0.6 SD Addis Ababa, 2355 masl- 0.4 SD
46. Discussion
Limitations of the study
- This is an observational study, which cannot allow
causal inferences to be made
- Generating experimental evidence on this topic is
unfeasible
instead, we attempted to control for as
many confounding factors as possible
Our findings are also supported by mechanistic and animal model
studies (Moore et al., 2011)
47. Summary and way forward
1. Children residing at higher altitudes were, on average,
born shorter and remained on a lower growth trajectory
than children residing at lower altitudes
2. The altitude mediated growth deficits were biologically
significant and applied even for children living in ideal
home environments
3. The perinatal period is the most vulnerable period
4. Given the possible functional deficit associated with
altitude mediated IUGR, we do NOT recommend a
downward adjustment of the WHO growth standards
48. 5. Specific health care guidance and monitoring are needed
for the management of pregnancies and early child
development at higher altitudes
6. Residence at higher altitude is one additional risk factor
for linear growth faltering
7. Failing to address altitude-mediated growth deficits may
result in a significant proportion of the world population
NOT meeting the SDGs and the WHA nutrition targets
Summary and way forward
redistributing the oxygenated blood to vital organs (brain, heart and adrenals) and restrict, ‘non-essential’ somatic movements and would attempt to increase the heart rate to obtain more oxygenated blood from the placenta