Post on 29-Dec-2015
Effect of early childhood stunting on schooling among poor urban households in
Nairobi, Kenya
Maurice Mutisya
APHRC
Introduction
• Interest in early childhood development (Fall et al.)
– Nutrition key for the development of a child (Jyoti,
Frangillo, & Jones, 2005).
– Effects accrue to later life - low academic achievement
and lower earnings (Save the Children, 2013)
• Recent focus – growth during the first 1000 days
– Moving beyond mortality to cognitive development
– Challenge – lack of data – longitudinal studies!
Objective
Use existing longitudinal study to
determine the effect of stunting and food
security on achievement among children
aged between 5 &8 years in Korogocho
and Viwandani.
Design• Quantitative survey – Integrated:
– MCH- longitudinal study 2007 and 2010
• Secondary data on stunting during first 1000 days
– Cross-sectional survey (primary data) – achievement
– NUHDSS – Food security information
• Assessment
– Literacy and Numeracy basic skills - % score
Sampling
• Loss to follow up
– Migration
• Response rate of
71%
• Biased sample –
non-random
loss?
7425 Total recruitment (IDS) 2007-10
<5 years7425
5+ years7425
Lost1235
Lost2340
2056 1794
Migrations
121
1673
Stunting Info. not enough/available
Normal618 (36.9)
Stunted1055 (63.1)
1189
Normal436 (36.7)
Stunted753 (63.3)
182 302 Migration; no Respondent
Migration; no Respondent
Study Setting and Target population
• Nairobi Urban Demographic Surveillance System (NUHDSS)
– Two study sites – Korogocho and Viwandani
– ……. about 70,000 individuals in 23000 HHs
• Target Population
– Households participating in MCH study
– Inclusion – currently active in the NUHDSS
– Child aged between 5 and 8 years
Measurements – Key variablesOutcome Z-score on literacy and numeracy Basic skills
Stunting Height for Age during the first 2 years
Food security Items on food uncertainty, worry,
child and adult hunger
Wealth Index PCA - Assets and amenities
Background information
Variable Percentage
% Stunting 63.3%
Not stunted 36.7
Temporal stunt 24.5
Stunted 38.8
% food secure 39.2
% Poorest (wealth index) 34.7
% joined school 97.4
Numeracy z-scores
Not stunted
Temorary
Stunted
Secure
Moderate
Severely
Poorest
Middle
Least poor
Stun
ting
Food
Sec
urit
yW
ealt
h
-0.35-0.25
-0.15
-0.04999999999999990.05
0.150.25
0.35
Numeracy Z-Scores
Literacy Z-scores
Not stunted
Temorary
Stunted
Secure
Moderate
Severely
Poorest
Middle
Least poor
Stun
ting
Food
Sec
urit
yW
ealt
h
Literacy Z Scores
SEM Model
Achievement
Food security
Stunting
Gender
ADHD
Wealth Index
Father Educ.
Mother Educ.
School years
HH size
HomeworkSch. type
Mother age
Orphan hood
HH: Household; ADHD = Attention Deficit Hyperactivity Disorder
SEM: NumeracySEM: Numeracy - direct and indirect effects
Variable Direct Effects Indirect Effects Total EffectsStunting
Not stunted - -Temporal -0.123*** -0.123***Stunted -0.183*** -0.183***
Food SecuritySecure - - -Moderate -0.035 -0.003 -0.037Insecure -0.018 -0.008 -0.025
Wealth IndexPoorest - - -Middle 0.122** 0.009 0.131**Least poor 0.096 0.019 0.115*
Controlling for other covariates; ***P<0.01; **P<0.05; *P<0.1
SEM: LiteracySEM: Numeracy - direct and indirect effects
Variable Direct Effects Indirect Effects Total EffectsStunting
Not stunted - -Temporal -0.181*** -0.181***Stunted -0.293*** -0.293***
Food SecuritySecureModerate -0.028 -0.028Insecure -0.068 -0.068
Wealth IndexPoorestMiddle 0.072 0.020 0.093*Least poor 0.209** 0.044** 0.253**
Controlling for other covariates; ***P<0.01; **P<0.05; *P<0.1
Conclusion• Opportunity provided by longitudinal studies
• Early childhood stunting associated with literacy and
numeracy achievement
– Poor health
– Cognitive development?
• Food security not related with achievement
– Food security related to nutrition (Reis, 2012; Saaka &
Osman, 2013)
– Effect – could be attenuated by nutrition