DSpace Institution

DSpace Repository http://dspace.org

School of Public Health Thesis and Dissertations

2022-04

Changes in The Prevalence Rate of

Stunning, factors Associated with
Stunting and Changes in The
Prevalence Rate Among Under Five
Children in Ethiopia Using Ethiopian
Demographic Health Survey Data

Getenet, Dessie

http://ir.bdu.edu.et/handle/123456789/13755
Downloaded from DSpace Repository, DSpace Institution's institutional repository
 BAHIR DAR UNIVERSITY
COLLEGE OF MEDICINE AND HEALTH SCIENCES
SCHOOL OF PUBLIC HEALTH
DEPARTMENT OF EPIDEMIOLOGY AND BIOSTATISTICS

CHANGES IN THE PREVALENCE RATE OF STUNNING, FACTORS

ASSOCIATED WITH STUNTING AND CHANGES IN THE PREVALENCE RATE
AMONG UNDER FIVE CHILDREN IN ETHIOPIA USING ETHIOPIAN
DEMOGRAPHIC HEALTH SURVEY DATA.

BY

GETENET DESSIE (BSc, MSc)

APRIL, 2022

BAHIR DAR, ETHIOPIA

I
 Bahir Dar University
College of Medicine and Health Science
School of public health
Department of Epidemiology and Biostatistics

Full Name of Investigator Mr. Getenet Dessie (BSc, MSc )

Cell phone: +251-9-23-52- 27- 12
E-mail: [email protected]
Name of Advisors 1. Mr. Zelalem Mehari ( MSc, Assistant professor in
Biostatistics)
Cell phone: +251-9-11-05-89-18
E-mail- [email protected]
2. Mr. Gebeyaw Wudie (MPH, Assistant professor in
Epidemiology)
Cell phone: +251-9-20-05-45-99
E-mail- [email protected]

Full title of the proposal Changes in the prevalence rate of stunning, factors
associated with stunting and changes in the prevalence rate
among under five children in Ethiopia using Ethiopian
demographic health survey data

Duration of the study February 15/2021 to April 28, 2022 GC

Study area Ethiopia

April, 2022
Bahir Dar, Ethiopia

II
Acknowledgements
First and foremost, I would like to express my heartfelt appreciation to my advisors, Mr. Zelalem
Mehari and Mr. Gebeyaw Wudie, for their consistent and invaluable guidance and support
throughout the development of this thesis report.
Second, I would like to express my gratitude to the Measure DHS program for granting me to
download and use the Ethiopia demographic and health survey statistics.

I would also want to thank Mr. Berhanu Abebaw ,Mulat Tirfie and Dr.Gedafaw Abejie for their
invaluable assistance with the composition of my thesis.

Finally, I would like to express my gratitude to the Department of Epidemiology and

Biostatistics, School of Public Health, College of Medicine and Health Sciences at Bahir Dar
University's, for approving my thesis topic and assigning advisors.

I
Acronyms
ANC: Antenatal Care
DHS: Demographic and Health Survey
EAs: Enumeration Areas
EDHS: Ethiopian Demographic and Health Survey
SD: Standard Deviation
SSA: Sub Saharan Africa
WHO: World Health Organization

III
 Table of Contents
Contents Pages

Acknowledgements .......................................................................................................................... I
Acronyms ...................................................................................................................................... III
List of tables .................................................................................................................................. VI
List of figures ............................................................................................................................... VII
Abstract ...................................................................................................................................... VIII
1. Introduction ............................................................................................................................. 1
1.1. Background ...................................................................................................................... 1
1.2. Statement of the problem ..................................................................................................... 2
1.4. Literature review .............................................................................................................. 5
1.4.1. Magnitude and trends of stunting ............................................................................. 5
1.4.2. Factors associated with stunting ............................................................................... 6
2. Objectives .............................................................................................................................. 10
2.3. General objective............................................................................................................ 10
1.4. Specific objectives............................................................................................................. 10
3. Methods..................................................................................................................................... 10
3.1. Study settings ..................................................................................................................... 10
3.2. Study design and period ..................................................................................................... 10
3.4. Population........................................................................................................................... 11
3.4.1. Source Population ........................................................................................................ 11
3.4.2. Study Population.......................................................................................................... 11
3.5. Eligibility criteria ............................................................................................................... 11
3.5.1. Inclusion criteria .............................................................................................................. 11
3.5.1. Exclusion criteria............................................................................................................. 11
5.6. Sampling technique ............................................................................................................ 11
3.7. Sample size......................................................................................................................... 12
3.8. Data source and extraction ................................................................................................. 13
3.9. 1. Dependent variable ..................................................................................................... 13
3.9. 2. Independent variables ................................................................................................. 13
3.10. Operational definitions ..................................................................................................... 14
3.11. Data management and method of data analysis ............................................................... 14

IV
 3.12. Ethical Consideration ....................................................................................................... 16
4. Results ................................................................................................................................... 16
5. Discussion .............................................................................................................................. 40
6. Conclusion ............................................................................................................................. 47
7. Recommendation ................................................................................................................... 48
References ..................................................................................................................................... 49
Appendix 1. ................................................................................................................................... 61
Appendix 2:................................................................................................................................... 63
Appendix 3:................................................................................................................................... 65

V
List of tables
Table 1: Parameter estimates of household, maternal, and child characteristics: EDHS of 2005,
2011, 2016, and 2019 (weighted). ................................................................................................ 19
Table 2: A multivariate decomposition logistic regression analysis of different features in the
2005 -2011 EDHS and 2005 EDHS-2016 EDHS. ........................................................................ 27
Table 3: A multivariate decomposition analysis of different features in the 2005 -2019 EDHS
and 2011 EDHS-2016 EDHS........................................................................................................ 33
Table 4: A multivariate decomposition logistic regression analysis of different features in the
2011 EDHS-2016 EDHS. ............................................................................................................. 38

VI
List of figures
Figure 1: Showing the association between determinants and changes in the prevalence rate of
stunting among children under the age of five years Using EDHS data (2005-2019). .................. 9
Figure 2: Flow diagram shows the technique for selecting study participants for the final analysis
EDHS, 2005–2019, Ethiopia......................................................................................................... 12

VII
Abstract
Background: Childhood stunting is one of the most significant impediments to human
development. Aside from detecting stunting reduction, epidemiological data is insufficient to
determine whether the reduction was due by community behavioral change or community
decomposition change.
Objective: To assess changes in the prevalence rate of stunning, factors associated with
stunting and changes in the prevalence rate among under five children in Ethiopia.
Methods: A further analysis of the four consecutive Ethiopian demographic health surveys
(2005 – 2019) that were collected using a stratified two-stage cluster sampling method was
carried out. A total of 26048 children aged 0-5 years were included in the analysis. Multilevel
logistic regression model was fitted to identify individual and community level factors. The
models were compared using deviance information criteria (DIC). Multivariate decomposition
logistic regression analysis was also carried out to assess the role of compositional characteristics
and behavioral change for declined in stunting prevalence rate among under-five children in
Ethiopia. All through the analysis a p-value < 0.05 was determined to be statistically significant.

Results: Over the study period, the prevalence rate of stunting in under five children decreased
from 47 % to 37% in 2019. Differences in behavioral change among children under the age of
five account for 76.69% of the overall decline in stunting prevalence rate in the years 2005-2011,
86.53% in the years 2005-2016, 98.9% in the years 2005-2019, 70.34% in the years 2011-2016,
and 73.77% in the years 2011-2019. The remaining variation in stunting prevalence rate in each
combination was related to differences in child compositional characteristics. Behavioral
adjustments among breast feed children, diet diversity, place of delivery, ANC follow-up, and
region have all had a major effect on stunting prevalence rate. The wealth index, parenteral
education, child's age in months, length of breast feeding, and area were among the
compositional change factors.
Conclusion: A large percentage of children under the age of five remain stunted in Ethiopia.
Stunting was associated to alterations in the compositional and behavioral characteristics of
children. Stimulating existing nutritional measures, as well as raise the wealth index, will have a
significant impact to further reduce stunting on Ethiopian children under the age of five.
Keywords: Change; Stunting; under five children; Ethiopia;

VIII
1. Introduction
1.1. Background
A well-balanced, satisfying food is the foundation of a child's existence, health, and
development. Children who are well-fed are more likely to grow up to be healthy, industrious,
and eager to study throughout their lives. Under nutrition, by the same rationale, is disastrous,
potentially reducing brainpower and productivity (1, 2) and slowing economic downturn, which
can perpetuate a cycle of poverty and illness (3).

Stunting is defined as a reduction in height in relation to a child's age, which is frequently caused
by malnutrition, repeated illnesses, and/or a lack of social stimulation (4). Stunting is a condition
in which children's growth and development are stunted as a result of poor nutrition, frequent
infections, and insufficient psychosocial stimulation. If a child's height-for-age is more than two
standard deviations below the WHO Child Growth Standards median, they are considered
stunted (5, 6).

Stunting is a well-known indicator of a child's poor growth. Early stunting predicts lower
cognitive and educational performance in later childhood and adolescence, as well having major
educational and economic effects at the individual, household, and community levels. Stunting
was linked to a decline in schooling among children in Brazil, Guatemala, India, the Philippines,
and South Africa, according to recent longitudinal research, with stunted children completing
approximately one year less school than non-stunted children (7, 8).

Floods, droughts, frosts, earthquakes, and other natural disasters pose a risk of stunting in
children. Stunting is influenced by sociocultural, economic, and political factors such as equality,
education, and cultural standards, as well as work status and government regulations (9).
Stunting rates are similar in early childhood across low and high socioeconomic class
households, but they diverge dramatically between the sixth and twentieth months of life,
according to studies. This difference cannot be explained just by the presence or absence of
variables that can be modified by nutrition-specific interventions; determinants associated to
nutrition-sensitive therapies are also substantially influenced (10, 11). Childhood malnutrition
has a number of negative repercussions for a child's survival and long-term well-being. It also
has long-term implications for human capital, economic productivity, and overall national
development (12, 13).

1
1.2. Statement of the problem

Linear growth failure is the most common type of malnutrition worldwide (14), affecting an
estimated 165 million children under the age of five. Stunting has been designated as a major
public health issue, with aggressive objectives set to eliminate stunting prevalence by 40%
between 2010 and 2025 (14, 15). The global goal equates to a four percent reduction per year,
with the number of stunted children falling from 171 million in 2010 to around 100 million in
2025. However, with present rates of improvement, by 2025, there would be 127 million stunted
children, 27 million more than the aim and only a 26% reduction (16).

Malnutrition's repercussions are a major worry for Ethiopian authorities, with over 5.8 million
children under the age of five (38 percent) suffering from chronic malnutrition (12, 13). Poor
nutrition during pregnancy and early childhood might result in stunting. Stunted children may
never reach their full height potential, and their brains may never develop to their full cognitive
potential. These children start out in life with a significant disadvantage: they struggle in school,
earn less as adults, and encounter impediments to community participation (17, 18). Stunting,
along with other concomitant undernutrition issues such as fetal growth restriction, wasting, and
vitamin A and zinc deficiencies, as well as suboptimal nursing, were estimated to be the cause of
3.1 million child fatalities (45 percent of all child deaths) in 2011 (19). According to more recent
estimates, stunting and severe wasting account for one-third of all deaths among children under
the age of five (20).

Inadequate nutrition and repeated bouts of infection during the first 1000 days of a child's life
cause stunting, which is often permanent. It also has long-term consequences for individuals and
societies, such as impaired cognitive and physical development, decreased productive capacity,
poor health, and an increased risk of degenerative diseases like diabetes (21). Every year, one
million children die as a result of stunting. Stunting in infancy and early childhood has long-term
consequences for the children who survive, including impaired cognition and school
performance, reduced physical development, poor health, lost productivity, and low adult wages
for those who survive (22, 23).

Despite the fact that stunting is on the decline in Ethiopia, the prevalence remains high (12).
Ethiopia has one of the world's highest rates of stunted children under the age of five (24, 25).

2
The government has signed various global initiatives and set national commitments to eradicate
child malnutrition, including stunting (26, 27). It was discovered that the spatial patterns of
stunting and severe stunting are not random. Amhara, Benishangul Gumuz, Afar, Tigray, and
Oromia are among the country's stunted regions (28). Stunting prevalence in the United States
has decreased from 51% in 2000 to 32% in 2016. Regional differences, as well as pro-rich, pro-
urban, and pro-educated disparities, exist (29).

Changes in stunting during the last two decades can be attributed to a number of factors,
including improved total consumable agricultural output, increased number of health
professionals, parental education, maternal nutrition, economic improvement, and lower diarrhea
incidence. Stunting reduction is driven by a mother of high stature, living in a city, having a large
kid size, a mother without anemia, having a large child weight, higher agricultural productivity,
and improved sanitation and child care practices (29-31). Due to a lack of attention paid to these
factors, existing investment levels are insufficient to sustain the progress required to accomplish
these goals (32).

Effective nutrition strategies in Ethiopia, as elsewhere, require targeting on the basis of factors
for change to maximize decline in stunting. However, because Ethiopia is a developing country,
implementing this program will be difficult due to a lack of data on the pattern of stunting
through time and the determinant factors that influence stunting change. Aside from detecting
stunting reduction, epidemiological data is insufficient to determine whether the reduction was
due by behavioral change or community decomposition change. To overcome this visible
literature gap in the area, the current study aimed to assess change in stunting and its associated
factors among under-five children in Ethiopia.

3
1.3. Significance of the study

In order for under-five children to receive target full interventions, it is critical to identify
variables that cause change in stunting prevalence rate. This research could also assist decision-
makers by identifying gaps and assisting policymakers in selecting and implementing lessons
gained in order to achieve a successful plan to eradicate early life stunting in Ethiopia. The
current study's contribution would be in line with Ethiopia's 2030 sustainable development goal,
according to the federal minister of health, and would expand our understanding of factors that
cause stunting prevalence rate change in Ethiopia. The outcomes of this study would also
indirectly provide information to health managers about the efficacy of national nutritional
programs on reducing chronic malnutrition in children under the age of five. Furthermore, the
findings of this study could provide information to non-governmental organizations working on
malnutrition on the efficacy of their programs in reducing stunting and the challenges they face.
Finally, such research is critical in understanding the impact of population composition change
on stunting and intervening accordingly.

4
1.4. Literature review
1.4.1. Magnitude and trends of stunting

In 2011, an estimated 165 million children under the age of five, or 26% of all children
worldwide, were stunted, falling from an estimated 253 million in 1990. Stunting is still a major
public health concern among children under the age of five in Africa (36 percent in 2011) and
Asia (27 percent in 2011). Africa and Asia are home to more than 90% of the world's stunted
children (21).

The number of stunted children in upper-middle-income nations has decreased the most of any
income level (33). In 2000, there were 203.6 million stunted children in the world, or 33.1
percent of the population, and by 2020, that number is expected to drop to 149.2 million, or 22
percent of the population (33). Many countries, notably in South Asia, have been affected by
stunting as a result of nutrition transition procedures. Stunting was particularly common in
children under the age of five in Pakistan (44%) (34) and other nearby regional nations such as
Bangladesh (36%) (35) and Nepal (35.8%) (36). However, stunting has decreased in Nepal in
general, as well as in all categories and subgroups; from 2001 to 2016, stunting dropped by 18
and 10.7% in rural and urban areas, respectively, indicating that significant differences in
stunting remain in the country (36). This has also been demonstrated in India, where the
prevalence of stunting has decreased from 48.0 percent in 2006 to 38.4 percent in 2016 (37).

Although the baseline prevalence and rate of reduction in stunting differ for each of these
countries, stunting has been steadily declining across the time period studied, with an initial
period of stagnation followed by a steady fall (38, 39). For example, between 1988 and 1993
Vietnam initially experienced stagnation (∼61%) but has seen a very steep decline between 1993
and 1998, followed by relatively consistent reductions until 2015 (∼25%). In Nepal before 1995,
since then (68.2%) (40, 41) and Bangladesh (65.8%) followed a very similar and consistent
pattern of decline until 2014,reducing stunting prevalence approximately by 30 percentage points
(42, 43). An examination of the contributing factors to Peru’s own steep decline between 2008
and 2016 (28.2–13.1%) (39). Similarly in some African countries such as BurkinaFaso, this
initial plateau lasted until 2006, after which dramatic and consistent reductions were seen (39,
44).

5
In the Sub-Saharan area, a strong decrease trend in stunting prevalence, often larger than 0.5
pp/yr, was also seen. Despite this decreased trend, stunting prevalence rates remained high in
most countries towards the conclusion of the period, particularly in Nigeria (37 percent in 2013),
Mali (38 percent in 2012), Tanzania (41 percent in 2010), Malawi (47 percent in 2010), and
Mozambique (42 percent in 2011) (45). Another trend analysis in Tanzania revealed a
considerable reduction in stunting among children under the age of five by 30 percent over a 25-
year period, although one in every three children under the age of five remains stunted (46).

Between 2000 and 2014, Ethiopia saw a considerable decrease in the prevalence of stunting, with
a 31% decrease in stunting (47, 48). Overall, 40.2% of children under the age of five were
stunted, with 19.1 percent severely stunted. Stunting was found to be most common in children
between 24-35 months (49%). Children under the age of six months had the lowest rate (14.6%)
(49). Bogale, T.Y. et al. reported a significant prevalence of stunting in Ethiopia, with a
magnitude of 57 percent, with 3.5 percent severely stunted, 27.3 percent moderately stunted, and
the remaining 26.4 percent mildly stunted, and a relatively high burden of severely stunted
among males (50).

1.4.2. Factors associated with stunting

1.4.2.1. Socio-demographic and economic factors

Improvement in asset index scores was a constant and substantial driver of enhanced linear
growth outcomes, according to a large scale study followed by regression-decomposition
analysis. Parental education was also found to be a substantial predictor of increased child
development. Open defecation, inadequate sanitation, and poor infrastructure have all had a
negative impact on child development (39). Temperature, per capita agricultural production at
the district level, wealth index, and parental education levels were all independent predictors of
childhood stunting in India (51). A single study in South Jakarta found a strong link between
family factors and nutritional parenting and the frequency of stunting in children under the age of
five (52). Mother's education, occupation, household income, and environmental factors such as
rural or urban location and sanitation were consistently indicated as determining factors that
closely connected to stunting in studies in Sub-Saharan Africa (53-55).

6
Similarly, high levels of child undernutrition in developing nations are caused by poverty, food
insecurity, ignorance, and poor hygiene and sanitation (20). Another study found that caregivers'
age (years), greatest level of education (primary, secondary, or higher), number of children in
their homes, and location (urban or rural) were all important predictors in stunting (46).
Household food security, region, and wealth index have all been established as independent
determinants for severe stunting in Ethiopia (28, 56). Children from male-headed households
with moms and fathers who had completed higher education had reduced probabilities of being
stunted than their contemporaries, according to another comprehensive survey in Ethiopia.
Stunting is also linked with the worst wealth strata, a lack of improved sanitation facilities, and a
lack of housing (49, 57).

1.4.2.2. Maternal characteristics

Stunting was more common in children from advanced maternal age, polygamous families, and
mothers who had several unions than in children from monogamous families and moms who had
only one union. In addition, children born in multiple births were more likely to be stunted than
singleton children. Stunting was also linked to age at first childbirth (years) and children who
had a 24-month gap between births (46, 54). Low mother height and a history of not taking
deworming medicine during pregnancy were also found to be predictors of stunting in a
Rwandan population health survey (55). Improved access to critical maternal health services,
such as optimal prenatal care and delivery in a health facility or with a qualified birth attendant,
on the other hand, all contributed to significant increases in infant growth. Several maternal
factors, such as parity, inter-pregnancy interval, and maternal height, also indicated moderate
stunting reductions (39).

1.4.2.3. Child characteristics

According to previous research, the rate of stunted children in SSA countries was higher among
men than females and among rural children than their urban counterparts. A linear association
exists between children's age and stunting, with a 2% increase in stunted children for every one
month increase in child age (54). Similarly, a prior study in Rwanda identified low child weight
at birth as a risk factor (55). Children born with low birth weight had a higher prevalence of
stunting than those born with normal or high birth weight, according to Sunguya, B.F.et al. In

7
1990, over 60% of children born with a low birth weight experienced stunting compared to 46
and 40% of normal and high birth weight children, respectively. In 2015, 48 percent of infants
born with a low birth weight became stunted, compared to 31 percent of children born with a
normal birth weight (46). From Ethiopian perspective, a kid's gender, multiple births, and the
age of the child were all independent predictors of severe stunting (28, 56). Another large scale
study in Ethiopia found that being male and having a short birth interval were the factors that
enhanced the likelihood of stunting at the individual level. Furthermore, children between the
ages of 24 and 35 months were more likely to be stunted than children under the age of one year
(58).

1.4.2.4. Child caring practices

The high levels of child undernutrition in underdeveloped nations were caused by ignorance and
a lack of adequate newborn and young child feeding practices (20). According to the Sunguya,
B.F.et al study, children who were inconsistently nursed or never breastfed were more likely to
be stunted than those who were breastfed for less than six months (46). A lack of dietary
diversification is also linked to an increased risk of childhood stunting (59). Single data suggests
that children who began supplemental meals before or after the recommended 6-month period
were more likely to be stunted (60).

1.4.2.5. Child morbidity

From prior published literatures, anemia was identified as an independent factor for severe type
of stunting among children (28, 56, 58). Infectious diseases have also contributed to high levels
of child malnutrition in developing countries (20). Stunting was caused by a variety of factors,
one of which being diarrhea (59). Despite the fact that children with upper respiratory tract
infections show signs of malnutrition, there is no significant difference in height-for-age (61).
Another conclusion was that neither the occurrence nor the duration of upper respiratory
infections were linked to increased height (62).

8
 Conceptual framework
A graphic representation of the conceptual framework is shown below, which depicts the
relationship between independent and dependent variables and was constructed after analyzing
several literatures. As shown in Figure 1, the dependent variable (change in the prevalence rate
of stunting) is thought to be influenced by socio-demographic and economic factors, mother
features, and child caring behaviors, child characteristics, and child morbidity.

Socio-demographic and economic factors Maternal characteristics mother’s age,

Residence, region, wealth index, source of number of under-five children, children
drinking water, toilet facilities, parent living with mother, age at first child birth,
education and occupation, head of household
mothers height, deworming during
pregnancy , antenatal care and place of
delivery
Child caring practices

Early initiation of breast feeding

Stunting and
Exclusive breastfeeding. its change Child characteristics
Duration of breast feeding
Sex, residence, age,
Bottle feeding birth weight, birth
order, and birth type
Diversity feeding

Child morbidity

Diarrheal diseases

Respiratory tract infection

Anemia

Figure 1: Showing the association between determinants and changes in the prevalence rate of
stunting among children under the age of five years Using EDHS data (2005-2019).

9
2. Objectives
2.3. General objective
1.3. To assess changes in the prevalence rate of stunning, factors associated with stunting and
changes in the prevalence rate among under five children in Ethiopia using Ethiopian
demographic health survey data from 2005-2019.
1.4. Specific objectives
 To describe change in the prevalence rate of stunting among under-five children in
Ethiopia using EDHS data from 2005-2019
 To identify factors associated with stunting among under-five children in Ethiopia using
EDHS data from 2005-2019
 To identify factors associated with change in the prevalence rate of stunting among
under-five children in Ethiopia using EDHS data from 2005-2019

3. Methods
3.1. Study settings
The study used data from national, population-based, cross-sectional surveys; EDHS: 2005,
2011, 2016, and 2019, a global effort supported by the US Agency for International
Development to collect nationally representative demographic and health data. Ethiopia is
structured into 11 administrative regions, each of which is subdivided into zones, with each zone
further subdivided into districts. The districts are subdivided further into kebeles. Every five
years, the EDHS conduct a national and subnational representative household survey. Ethiopia is
Africa's second most populated country, situated between 3 and 15 degrees north latitude and 33
and 48 degrees east longitude in the horn of Africa (63) . Around 13 million children under the
age of five live in Ethiopia, accounting for roughly 16% of the country's total population of 96
million (64).
3.2. Study design and period
The cross-sectional data from the Ethiopian Demography and Health Surveys (EDHSs), which
are collected every five years, were used in this secondary data analysis.

10
3.4. Population

3.4.1. Source Population

All Ethiopian children under the age of five who were living in the country at the time of the
demographic health surveys.
3.4.2. Study Population
All under five years children that had been included in the four consecutive EDHS datasets
(2005-2019) were included.

3.5. Eligibility criteria

3.5.1. Inclusion criteria
All Ethiopian under five years children who were included in the four years of EDHS data
(2005-2019) were included.

3.5.1. Exclusion criteria

Children whose data were incomplete, flagged cases and de facto resident were not included in
this study.

5.6. Sampling technique

A stratified, two-stage cluster sampling technique was utilized in all four DHSs (2005, 2011,
2016, and 2019). Children 0-59 months in the selected house-holds were measured for height for
age in the first three DHS data collections. Each region was stratified into urban and rural areas
to achieve stratification. The Somali region was separated into two parts, the first of which
consisted of the first three zones, and the second of which consisted of the three additional zones
that were added latter. Because the Addis Ababa region is fully urban, 23 sampling strata were
constructed. In addition, the 2019 EMDHS sample was stratified and selected in two stages.
There were 21 sampling strata in each region, which were divided into urban and rural areas. In
two steps, EA samples were selected independently in each stratum. At each of the lower
administrative levels, implicit stratification and proportionate allocation were used by sorting the
sampling frame within each sampling stratum before sample selection. Participants were chosen
based on administrative units at various levels and a probability proportional to size selection
method at the first stage of the sampling study. In all DHS reports, the sampling technique is
described in detail (65-68).

11
3.7. Sample size
The EDHS data set was accessed after asking measure of DHS for permission through the
project title ―change in stunting and its associated factors among under five children in Ethiopia
using Ethiopian demographic health survey data from 2005-2019‖. A total of 29267 children
under the age of five were screened for height-for-age. In 2005, 4,586 individuals were
examined, followed by 10,282 in 2011, 9462 in 2016, and 4937 in 2019. After taking into
account exclusion criteria, 3,476 participants from the 2005EDHS, 9,013 from the 2011EDHS,
8,567 from the 2016EDHS, and 4,992 from the 2019MEDHS were included in the final analysis
(Figure 2).

Each survey yielded a number of children;

2005EDHS= 9,861
Identification

2011EDHS= 11,654
2016EDHS= 10,641
2019MEDHS= 5,753

Reasons for children to be excluded

 Incompleteness
Screening

 Flagged cases
 Not dejure resident

Number of children included in each

survey after taking into account exclusion
criteria;
2005EDHS= 3,476
2011EDHS= 9,013
Included

2016EDHS= 8,567
2019MEDHS= 4,992

Final total sample size =26048

Figure 2: Flow diagram shows the technique for selecting study participants for the final analysis
EDHS, 2005–2019, Ethiopia.
12
3.8. Data source and extraction
The Ethiopia demographic and health survey datasets from 2005 to 2019 were requested and
downloaded from the Measure DHS program website, which are freely available at
(https://dhsprogram.com/data/dataset_admin/login_main.cfm) to all registered users. The
recommended dataset type for under-five children was chosen after analyzing and examining the
details of the EDHS data structure and dataset types. Stunting data and potential independent
variables were extracted in this manner.

3.9. Variables of the study

3.9. 1. Dependent variable

Stunting and its change (―stunting = 1, not stunting = 0‖)

3.9. 2. Independent variables

Socio-demographic and economic factors

Residence, region, wealth index, source of drinking water, toilet facilities, parent education and
occupation, head of household

Maternal characteristics

Mother’s age, number of under-five children, children living with mother, age at first child birth,
mother’s height, deworming during pregnancy, antenatal care and place of delivery.

Child caring practices

Early initiation of breast feeding, exclusive breastfeeding, duration of breast feeding, bottle
feeding, and diversity feeding.

Child characteristics

Sex, residence, age, birth weight, birth order, and birth type

Child morbidity

Diarrheal diseases, respiratory tract infection and anemia

13
3.10. Operational definitions
Stunting in under five children: Stunting was defined as having a height for age that is less
than 2.0 SD HAZ-scores below the WHO reference population's median height for age, as
defined by the World Health Organization (WHO) Height-for-age z-scores as Child Growth
Standards (69).

Early initiation of breast feeding: Provision of mother's breast milk to infants within one hour
of birth to ensures that the infant receives the colostrum, or ―first milk‖, which is rich in
protective factors (70).

Exclusive breastfeeding: The term "exclusive breastfeeding‖ refers to a baby who solely
receives breast milk. For the first six months of life, no other liquids or solids are provided – not
even water – with the exception of oral rehydration solution or drops/syrups of vitamins,
minerals, or medicines (71).

Diversity feeding: The proportion of infants and young children aged 6 to 59 months who
received at least four out of seven food groups in the previous 24-hours (grain, legumes, dairy
products, egg, meat, fruits, and vegetables) recommended by the World Health Organization
(72).

Improved latrine: Latrine facilities are those that are designed to keep excreta away from
human touch and include flush/pour flush toilets connected to piped sewer systems, septic tanks,
or pit latrines; pit latrines with slabs (including ventilated pit latrines); and composting toilets
(73).
Open defecation: Human feces disposal in fields, forests, bushes, open bodies of water, beaches,
and other open spaces, or with solid waste (73).
Improved drinking water sources: Are those that have the potential to deliver safe water by
nature of their design and construction, and include: piped water, boreholes or tubewells,
protected dug wells, protected springs, rainwater, and packaged or delivered water (73).
3.11. Data management and method of data analysis

The outcome variable was coded as a binary variable ("stunting" = 1 and "not stunting" = 0),
similar to a prior study (74), and the data analysis was executed using STATA 15. To account
for the uneven probability of selection between the strata that were geographically specified,

14
sample weights were used. The methodology of EDHS final reports contains a full explanation
of the weighting procedure (65-68). First, descriptive statistics and trends in stunting were
examined across all surveys using recoded background variables. Second, a Multilevel logistic
regression models were fitted to find predictors of stunting at individual and community levels,
taking into account the hierarchical nature of the four-year EDHS data, which included 26048
children aged 0–5 years nested inside each year's enumeration areas. Four models were fitted to
compare and select the best fit model: the first model (model I), also known as the null model,
was fitted as a baseline model without any predictor variables, the second model (model II) was
fitted with individual level variables, the third model (Model III) was fitted with community
level (region & residence) variables, and the final model (model IV) was fitted with both
individual and community level variables. Then the models were compared using deviance
information criteria (DIC), and the final best fit model (model IV) was selected as the model
with the smallest DIC value (Appendix 3) (75). For measures of association, adjusted odds ratio
with 95% confidence intervals was used to declare statistical significance. For measures of
variation (random effects), Intra-class correlation coefficient (ICC) were utilized. Akaike's
information criterion and Bayesian information criterion were also used to take assess how well
the model fits the data (Appendix 3). Lower scores in both criteria were considered to choose
the best model. The standard error was used to identify multicollinearity. We have added and
removed one of the variables and compare the standard error and parameter estimates to see if
multicollinearity introduces bias into the standard errors of the parameter estimates.

Third, a multivariate decomposition logistic regression analysis was used to determine the
contribution of each covariate to the observed change in stunting prevalence rate. The influence
of changes in population structure in terms of children's characteristics on percentage of stunting
over time was investigated using random-effects generalized least square regression. At a P level
of less than 0.05, any statistical test was refereed as significant. The decomposition approach
divides the total drop in stunting into two parts: the endowments component, which can be
attributed to changes in the composition or prevalence of a set of indicators, and the effect
portion, which can be assigned to changes in the effect of these indicators (referred to as the
coefficient portion) (76). The formula is given by

ΔYi−j= (Xi−Xj) βi+Xj (βi−βj), i≠j,

15
Where i,j= 2005, 2011, 2016, and 2019,

ΔY is the difference in mean prediction of stunting between year I and year j, given that of
different characteristics X.

β is the estimated regression coefficients.

(Xi−Xj) βi represents the difference due to endowment between the ith and jth years.

Xj (βi−βj) represents the difference due to coefficients between the ith and jth years.

Finally, stunting was regressed over time to check if the difference was significant.

3.12. Ethical Consideration

After being registered and submitting a request with short specified study objectives, the data
from 2005-2019 Ethiopia demographic and health surveys were used for this study with
permission from the Measure DHS program. Ethical approval was obtained from the ethical
review Committee of the school of Public Health, college of medicine and health science of the
Bahir Dar University. The information was solely utilized for this study and could not be shared
with other researchers. The ethical difficulties have been detailed in the Ethiopia DHS final
report, which is available on the on: https://dhsprogram.com/publications/index.cfm.

4. Results
4.1 Socio-demographic and socio- economic characteristics

Data from 3,476 under five children in 2005, 9,013 in 2011, 8,567 in 2016, and 4,992 under five
children in the 2019 EDHS were utilized in this analysis. Females made up about 51% of
children under the age of five. The majority of the children were under the age of two in all of
the surveys included in the analysis. The vast majority of those polled (68%) reside in rural
areas. The Oromo region had the highest number of children (12-20%), followed by Amhara (9-
13%), and SNNPR (11.80-18.79 %). Open defecation was used by 41.55-65.54 % of the
individuals who responded. Antenatal care follow-up has increased from 33.79 % in 2005 to
74.64 % in the 2019 EDHS surveys. Over the course of the study, the percentage of children in
the poorest wealth index category increased over the survey period, and the proportion has
become larger than in other categories (see Appendix 1).

16
4.2. Trend of stunting
Over the study period (2005–2019), the prevalence of stunting fell from 47% in 2005 to 37% in
2019. The difference was statistically significant, with a beta coefficient of -.767184 and a 95%
CI of:-1.23, -0.30 with a P value of 0.019. With a six percent decline, the survey period 2011–
2016 saw the highest decline. From 2005 to 2019, the rate of reduction in stunting varies
depending on the child's attributes. Females had the largest decrease (12.4%) in the specified
period, compared to males who had the lowest (7 %). Despite the fact that stunting was relatively
low among urban residents, the decline was higher (7%) among children from rural settlements
across the study period. Breastfed children had a decreased trend by 11.85% during the study
period, which was higher than the other groups. Over the study period, the Amahara region has
shown the largest reduction in stunting (15.88%), followed by SNNPR (15.41%). In the
Benishangul Gumez region, however, the percentage of stunted children has increased over time
(see Appendix 2).

4.3. Multilevel logistic regression analysis to identify Factors associated with

stunting
Wealth index, source of drinking water, toilet facilities, parent education, occupation, mother’s
age, number of under-five children in the house , living status of the child , age of mother at first
child birth, mothers height, child’s age in month ,antenatal care visit, place of delivery, early
initiation of breast feeding, exclusive breastfeeding, duration of breast feeding, bottle feeding
,diversity feeding, sex of child , age of mother, birth order, birth type, diarrheal diseases ,cough
,short rapid breathing , and anemia were fitted in model II as Individual level factors. Model III
has been fitted with residence and region as second level variables. The final model showed that
in the 2016 EDHS survey, women with higher education had a reduced likelihood of stunting
than mothers with no education (AOR = 0.42, 95% CI: 0.22, 0.85). In both the 2011 and 2019
EDHS datasets, the number of children under the age of five demonstrated a statistically
significant association with stunting. In the 2011 and 2016 EDHSs, when the number of under-
five children in the house increased by one, the likelihood of stunting increased by 1.2 and 1.40
times, respectively, with AOR = 1.2, 95% CI: 1.02, 1.4 and AOR = 1.40, 95% CI: 1.11, 1.69. In
two of the studies examined, the respondent's height was a statistically significant factor in
stunting. In the 2005 and 2016 EDHSs, a centimeter increase in respondent height reduces

17
stunting by 0.6 percent, with AOR = 0.994, 95 % CI: 0.992, 0.996 and AOR=0.994, 95% CI:
0.992, 0.995, respectively. As children's ages increased by one month, their chances of being
stunted increased by a percentage ranging from 3-7% in all surveys. In the 2005 and 2011
EDHSs, diarrhea increased the likelihood of stunting by 1.4 times, with AOR= 1.4, 95% CI:
1.06, 1.82 and AOR= 1.4, 95% percent CI: 1.10, 1.80, respectively. In the 2016 EDHS, bottle
feeding was the sole significant factor, and when comparing children who were bottle fed to
children who were not bottle fed, the probability of becoming stunted fell by 30% (AOR= 0.70,
95% CI:0.53, 0.92) (Table 1).

The wealth index has a substantial relationship with stunting in this study. Children from the
poorest and poorer categories were 1.73 and 1.79 times more likely to be stunted than children
from the richest household, according to a multilevel logistic analysis of the 2005 EDHS, with
AOR= 1.73, 95% CI: 1.13, 2.64 and AOR=1.79, 95% CI: 1.19, 2.70, respectively. Similarly,
according to the 2011 EDHS, children from the poorest and poorer families were 2.3 and 2.29
times more likely to be stunted than children from the richest families, with AOR=2.3, 95% CI:
1.4, 3. 5 and AOR= 2.29, 95 % CI: 1.48,3.60, respectively. Females were less likely than male
children to be stunted, according to the 2011 and 2019 EDHS (AOR = 0.77, 95% CI = 0.65, 0.92
and AOR = 0.67, 95 % CI = 0.49, 0.92, respectively). In 2011 and 2019, EDHS, we found that
increasing the birth order by one unit reduced the likelihood of stunting by 7% (AOR =0.93,
95% CI: 0.90, 0.96 and AOR =0.93, 95% CI: 0.87, 0.998, respectively) (Table 1).

Only region has a significant association with stunting when considering community-level
determinants. In the 2005 EDHS, children from the Amhara region had a significantly higher
likelihood of stunting than children from Tigray (AOR=1.68, 95% CI: 1.16, 2.43). Children from
the Amhara region, on the other hand, had a lower likelihood of stunting in 2011 than children
from Tigray (AOR= 0.71, 95 percent CI: 0.52, 0.96). The two surveys analysis (2011&2019)
revealed that children from Oromia region have a lower likelihood of stunting than children from
Tigray (AOR= 0.50, 95% CI: 0.36, 0.70 and 0.43, 95% CI: 0.23, 0.79, respectively) (Table 1).

Children from Somalia have a lower likelihood of stunting than children from Tigray region,
according to the analysis of 2011 and 2019 EDHS, with (AOR= 0.46, 95% CI: 0.30, 0.72 and
AOR=0.29, 95% CI: 0.14, 0.54 respectively). Only the analysis of 2011 EDHS surveys showed a
strong link between residing in the Beneshangul Gumez regions and stunting. When compared to

18
 children in Tigray Region, living in Benishhangul Gumez reduces the likelihood of stunting by
32% (AOR=0.68, 95% CI: 0.48, 0.97). There are two types of findings in the current study for
under-five children in SNNPR. In 2005, under-five children in the SNNPR had a 50% higher
likelihood of being stunted than under-five children in the Tigray region. Under-five children
from SNNPR, on the other hand, were 0.51 and 0.43 times less likely to be stunted in 2011 and
2019 EDHS data analysis, with (AOR=0.51, 95 % CI: 0.37, 0.70 and AOR=0.43, 95% CI: 0.24,
0.78, respectively). Children from Gambela, Harari, Adis Abeba, and Dire Dewa had a lower
likelihood of stunting than children from Tigray region, with (AOR= 0.25, 95%CI: 0.16, 0.39,
AOR=0.40, 95%CI: 0.26, 0.62, AOR=0.44, 95% CI: 0.25, 0.76, and AOR=0.56, 95% CI: 0.37,
0.84, respectively). Similarly, according to the 2019 EDHS data, under five children in the
Gambela and Adis Abeba regions had a lower likelihood of stunting than under five children in
Tigray (AOR= 0.27, 95% CI: 0.15,0.50 and AOR= 0.35, 95% CI: 0.23,0.68 respectively) (Table
1).

Table 1: Parameter estimates of household, maternal, and child characteristics: EDHS of 2005,
2011, 2016, and 2019 (weighted).

AOR
Characteristics 2005 2011 2016 2019
Model IV AOR Model IV AOR Model IV AOR Model IV AOR
(95%CI) (95%CI) (95%CI) (95%CI)
highest No 1 1 1 1
educational education
level Primary 0.90(0.68,1.2) 1.1 (0.88, 1.30) 0.93(0.77, 1.14) 1.02(0.71,1.48)
Secondary 0.51(0.24,1.08) 0.79 (0.41,1.54) 0.78(0.51,1.20) 0.71(0.34,1.48)
Higher 0.39(0.12,1.3) 0.91 (0.37, 2.25) 0.42(0.22, 0.85)** 2(0.38, 10.54)
Toilet Open 0.77(0.16,3.68) 1.11(0.80,1.54) 1.18 (0.86,1.62\) 0.75( 0.35,1.61)
defecation
Unimprove 0.63 (0.13,3.06) 1.03 (0.65, 1.25) 0.98(0.73, 1.32) 0.75( 0.35,1.61)
d
Improved 1 1 1 1
Number of under five children 1.06(0.90, 1.26) 1.2 (1.02, 1.41)* 1.1(0.96, 1.26) 1.40(1.11,
1.69)*
Height of respondent 0.994(0.992 1(1,1.01) 0.994(0.992, -
0.996)*** 0.995)***
ANC follow Yes 1 1 1 1
up No 1.21(0.95, 1.53) 1.18 (0.99, 1.40) 1.03(0.84, 1.25) 0.75(0.46,1.22)
place of Home 1 1 1 1
delivery Health 0.54(0.24, 1.23) 0.73(0.41,1.3) 0.83(0.61, 1.14) 1.17(0.78,1.75)
institution
Others 0.8(0.25,2.54) 2.13(0.51,8.71) 1.51(0.78, 2.91) 0.
20(0.031,1.26)

19
Diarrhea No 1 -
Yes 1.40(1.06, 1.82)** 1. 4(1.1,1.8)** 1.27(0.94,1.70) -
Bottle feeding (Yes ) 1.33(0.85, 2.09) 0.92 (0.67, 1.26) 0.70(0.53, 0.92)* 0.92(0.66,1.28)
Wealth index Poorest 1.73(1.13, 2.64)* 2.3(1.4, 3. 5)** 1.54(1.03,2.32)* 4.85(1.44,
16.36)*
Poorer 1.79(1.19, 2.70)*** 2.29(1.48,3.60) 1.35(0.97,1.90) 3.37(1.10,
*** 10.37)*
Middle 1.48(1.01, 2.18)* 2.24(1.42, 3.53) 1.03(0.72,1.47) 3.16(0.996,10.0
** 0)
Richer 1.68(1.16, 2.51)* 1. 9(1.20, 2.93)** 1.17(0.85,1.61) 2.39(0.84,6.80)
Richest 1 1 1 1
Age of child in month 1.05(1.03, 1.06)*** 1.06(1.05,1.07)** 1.03(1.02, 1.04)*** 1.07(1.04,
* 1.09)***
Sex of child Male - 1 1 1
Female 0.83(0.67,1.04) 0.77 (0.65, 0.84(0.71, 0.999)* 0.67(0.49,
0.92)** 0.92)**
Birth order 0.97(0.94,1.01) 0.93 0.97(0.94,1.01) 0.93(0.87,
(0.90,0.96)*** 0.998)**
Dietary diversity Yes - 1 1 1
No - 1.63(1.20,2.23)** 1.53(1.10,2.14)* 0.72(0.41, 1.26)
Region Tigray 1 1 1 1
Affar 1.24(0.75, 2.06) 0.84 (0.59,1.19) 0.99 (0.69,1.42) 0.70 (0.42,
1.19)
Amhara 1.78(1.15, 2.77)* 0.71 (0.52, 0.96)* 1.31(0.98, 1.75) 0.73(0.39,1.38)
Oromiya 1.07(0.76,1.5) 0.50(0.36,0.70)** 0.98 (0.72,1.34) 0.43(0.23,0.79)
* **
Somali 1.28(0.8,2.04) 0.46 (0.30, 1.03(0.73, 1.45) 0.
0.72)** 29(0.14,0.54)**
*
Benishangu 0.99(0.6,1.63) 0.68 (0.48, 0.97)* 1.08(0.78,1.51) 0. 6(0.31,1.15)
l-gumuz
SNNP 1.55(0.98 ,2.45) 0.51(0.37, 0. 1.06(0.78, 1.44) 0.
70)*** 43(0.24,0.78)**
Gambela 0.96(0.51,1.81) 0.25 0.70(0.47, 1.04) 0.27(0.15,0.50)
(0.16,0.39)*** ***
Harari 1.55(0.86,2.79) 0.40 (0.26, 1.08(0.76,1.53) 0.89(0.44, 1.80)
0.62)***
Adis 1.02(0.47,2.24) 0.44 (0.25, 0.71(0.42,1.19) 0.35(0.23,0.68)
Abeba 0.76)*** *
Dire Dawa 0.73(0.42,1.25) 0.56 (0.37, 1.44 (0. 90, 2.33) 0.61(0.33,1.12)
0.84)*
Key: ∗pvalue<0.05; ∗∗pvalue<0.01; ∗∗pvalue<0.001

4.4. Multivariate decomposition logistic regression analysis of 2005 -20011 &

2005-2016 EDHS
Difference due to characteristics (endowment)
The overall change in the prevalence rate of stunting between 2005 and 2011 was three percent,
according to a multivariate decomposed logistic regression analysis. Differences in the

20
compositional characteristics of children accounted for 25.31% of the overall change in in the
prevalence rate of stunting. According to the multivariate decomposition logistic regression
study from 2005 to 2016, 13.47% of the overall change was related to change in child
compositional characteristics. In multivariate decomposed logistic regression analysis of the
2005-2011 and 2005-2016 EDHS datasets, among the compositional change factors the
following ; wealth index, parenteral education, child's age in months, duration of breast feeding,
and region all had a statistically significant contribution to change in the prevalence rate of
stunting (Table 2).

Children from families with no or limited education were more likely to be stunted than those
from families with higher education. Parents primary school coverage grew from 16.60% in
2005 to 25.3% in 2011 and 25.73% in 2016 (Appendix 1), resulting in a negative significant
compositional contribution to a 14.80% and 8.78% reduction in stunting prevalence rate,
respectively. From 2005 to 2011 and 2016 EDHS, the proportion of women with no education
decreased from 76.41% to 69.87% and 64.01%, respectively (Appendix 1), resulting in a 14.95%
and 9.25 % increase change in the prevalence rate of stunting. The likelihood of becoming
stunted decreased as respondents' height increased. Between 2005 and the two most recent
EDHS surveys (2011&2016), community compositional changes in respondent height had a
positive contribution to change in stunting prevalence rate by 43.41 % and 21.86 %, respectively
(Table 2).

During the data collection period, under-five children who were on breastfeeding were more
likely to be stunted than children who were Ever breastfed but not during the study period. As a
result, the drop in the number of breast-fed children between surveys contributed to a positive
improvement in the prevalence rate of stunting of 2.26% between 2005 and 2011, and 1.88%
between 2005 and 2016 EDHS. Children under the age of five in the poorest household wealth
index category were more likely to be stunted than children in the highest household wealth
index category. Between 2005 and 2011, increasing the proportion of children in the poorest
wealth index group reduced change in prevalence rate of stunting by 0.31%. The present study
relates a 5.31% drop change in the prevalence rate of stunting between 2005 and 2011 and a
5.30% decline between 2005 and 2016 to mean age compositional differences among under-five
children. Between 2005 and 2011, change the composition of under five children in the Amhara

21
region and Dire Dewa city contributed to a 0.05% increase and a 0.07% decrease change in the
prevalence rate of stunting, respectively. In addition, between 2005 and 2016, changes in the
composition of children in the Amhara region and Dire Dewa city contributed to a 0.46%
increase and a 0.06% reduction in the prevalence of stunting, respectively (Table 2).

Difference due to effects of coefficient (Coefficients)

Controlling the effects of change in compositional features, behavioral changes among children
under the age of five who were on breast feeding during the survey time increased change in the
prevalence rate of stunting by 178.2%, according to a multivariate decomposed logistic
regression analysis conducted between 2005 and 2011. Change in the prevalence rate of stunting
progressed by 211.32% as a result of age-related behavioral changes in children from young to
old. According to a multivariate decomposed logistic regression analysis conducted between
2005 and 2016, the change in stunting prevalence rate increased by 62.98 % with behavioral
changes among children who were breast-feeding during the survey period (Table 2).

22
Table 2: A multivariate decomposition logistic regression analysis of different features in the 2005 -2011 EDHS and 2005 EDHS-2016
EDHS.

2005 EDHS-2011 EDHS 2005 EDHS-2016 EDHS

Characteristics Endowments Coefficients Endowments Coefficients
Coef with 95% CI Perce Coef with 95% Percent Coef with 95% Perce Coef with 95% Perce
nt CI CI nt CI nt
Residence (urban as ref) -0.0001(-0.0004, 0.07 0.047(-0.0995, -64.65 -0.0003(-0.001, 0.21 0.062(-0.075, 0.13
Rural 0.00022) 0.1938) 0.0004) 0.199)

ANC (no as ref) -0.00013(-0.0062, -0.05 -0.0006(-0.051, 0.79 -0.00003(- 0.02 -0.030(-0.077, 0.18
Yes 0.006) 0.049) 0.0067, 0.007) 0.018)
highest No -0.011(-0.0214, 14.95 0.100(-0.1713, -136.92 -0.012(-0.024, 9.25 0.038(-0.217, -
educational education 0.0004)* 0.371) 0.00053)* 0.294) 28.90
level primary 0.011(0.0002, - 0.0242(-0.0362, -33.16 0.012(0.0005, -8.78 0.009(-0.047, -7.11
0.0214)* 14.80 0.08453) 0.023)* 0.066)
secondary 0.0005(-0.00012, -0.70 0.00813(-0.009, -11.16 0.001(-0.0002, -0.57 0.0036(-0.010, -2.73
0.00114) 0.025) 0.002) 0.018)
Higher 1 1 1 1 1 1
Height of respondent -0.032(-0.0524, 43.41 0.0358(-0.799, -49.12 -0.029(-0.048, 21.86 -0.148(-0.924, 111.8
0.011)** 0.871) 0.01)** 0.627) 4
Duration Ever 1 1 1 1 1 1 1 1
of breastfed, not
breastfeed currently
ing Never 0.0006(-0.0002, -0.79 0.0003(-0.00262, -0.34 0.0005(-0.0003, -0.35 -0.0004(- 0.31
breastfed 0.00133) 0.00311) 0.0012) 0.0033, 0.0024)
still -0.0017(-0.003, 2.26 -0.130(-0.2046, 178.19 -0.003(-0.005, 1.88 -0.084(-0.151, 62.98
breastfeeding 0.00034)* 0.0552)* 0.001)* 0.0167)*
water source(improved as -0.0034(-0.0105, 4.64 0.0018(-0.0085, -2.45 -0.004(-0.0114, 2.95 -0.002(-0.011, 1.28
ref) 0.0037) 0.0121) 0.0036) 0.008)
unimproved
Diarrhea Yes -0.0014(-0.0036, 1.96 -0.0055(-0.0228, 7.52 -0.0014(-0.003, 1.03 0.006(-0.010, -4.61

27
 0.0007) 0.012) 0.0006) 0.022)
Wealth poorest 0.00023(0.00001, -0.31 -0.0022(-0.032, 3.01 -0.00001(- 0.000 -0.006(-0.034, 4.31
index 0.00044)* 0.028) 0.0001,0.0001) 1 0.023)
poorer 0.00051(-0.0001, -0.69 -0.0107(-0.040, 14.61 0.0006(-0.0002, -0.43 -0.011(-0.038, 8.29
0.0011) 0.019) 0.0013) 0.016)
middle 0.00013(-0.0006, -0.18 -0.00833(-0.035, 11.43 0.00003(-0.0002, -0.02 -0.017(-0.043, 12.62
0.00083) 0.0185) 0.0003) 0.010)
richer -0.0001(-0.0003, 0.11 -0.014(-0.041, 19.23 -0.0001(- 0.06 -0.015(-0.040, 11.47
0.00014) 0.013) 0.00031, 0.0002) 0.010)
Richest 1 1 1 1 1 1 1 1
Age of child in month 0.004(0.0015, -5.31 -0.154(-0.223, 211.32 0.007(0.003, -5.30 -0.044(-0.10, 33.36
0.0063)* 0.085)*** 0.011)** 0.011)
Sex of child (male as ref ) -0.0004(-0.00082, 0.48 0.0064(-0.023, -8.76 -0.0003(-0.0006, 0.22 -0.005(-0.032, 3.40
0.00012) 0.036) 0.00005) 0.023)
Region Tigray 0.0002(-0.00004, -0.28 0.0042(-0.01083, -5.73 0.0003(-0.0001, -0.23 0.008(-0.005, -6.29
0.00045) 0.01918) 0.001) 0.021)
Affar 0.00003(-0.00002, -0.05 0.0001(-0.0018, -0.07 -0.0000(- 0.00 0.0003(-0.002, -0.19
0.00008) 0.002) 0.0001,0.001) 0.002)
Amhara -0.00004(-0.00007, 0.05 -0.002(-0.0487, 2.54 -0.001(-0.0012, 0.46 0.009(-0.034, -6.79
0.00001)* 0.04501) 0.0001)* 0.052)
Oromiya 0.0014(-0.0005, -1.94 -0.00701(-0.084, 9.62 0.002(-0.0005, -1.52 0.024(-0.044, -
0.0033) 0.070) 0.005) 0.091) 17.87
Somali 0.00003(-0.00003, -0.05 -0.002(-0.0091, 2.74 -0.0001(-0.0001, 0.04 -0.0017(-0.009, 1.27
0.0001) 0.0051) 0.0001) 0.005)
Benishangu 0.00002(-0.00000, -0.03 0.0007(-0.00162, -0.91 0.00002(- -0.02 0.0012(-0.001, -0.91
l-gumuz 0.00005) 0.00294) 0.00001, 0.003)
0.00005)
SNNP -0.0008(-0.0018, 1.06 -0.0065(- 8.90 -0.001(-0.0024, 0.83 0.004(-0.038, -4.51
0.00021) 0.05442, 0.0415) 0.0002) 0.050)
Gambela 0.00001(-0.00001, -0.00 0.00014(-0.0005, -0.19 -0.00001(- 0.00 -0.0001(-0.001, 0.00
0.00001) 0.0007) 0.00001, 0.0006)

28
 0.00001)
Harari 0.00001(-0.00001, -0.01 -0.0001(-0.0005, 0.11 0.00001(- -0.01 -0.00001(- 0.00
0.00002) 0.00032) 0.00001, 0.0004,
0.00002) 0.00034)
Addis 1 0.00 1 1 1 1 1 0.00
Ababa
Dire Dawa 0.00005(0.00001, -0.07 0.00025(- -0.35 0.0001(0.00001, -0.06 0.0005(- -0.40
0.0001)* 0.00043, 0.00014)* 0.00003,
0.00093) 0.0011)
Anemia Severe 0.00001(-0.00004, 0.809 0.00323(- -4.43 -0.00001(- 0.01 0.0021(-0.0015, -1.59
0.00005) 0.00024, 0.00005, 0.006)
0.00670) 0.00002)
Moderate -0.00102(-0.0023, 0.104 0.00927(- -12.72 -0.0013(-0.0025, 0.95 0.011(0.0011, -8.13
0.00021) 0.00091, 0.0195) 0.00003) 0.021)*
Mildanemi 0.0008(0.00001, 0.049 0.0155(0.00063, -21.27 0.0006(-0.00003, -0.44 0.014(0.00001, -
a 0.0015) 0.0304) 0.0012) 0.027)* 10.32
not anemic 1 1 1 1 1 1 1 1
Key: pvalue<0.05; ∗∗pvalue<0.01; ∗∗pvalue<0.001
∗

29
4.5. Multivariate decomposition logistic regression analysis of 2005 -2019 &
2011-2016 EDHS
Difference due to characteristics (endowment)
From 2005 to 2019, a multivariate decomposition logistic regression analysis revealed that
98.9% of the overall reduction in stunting prevalence rate was due to children's behavioral
changes. None of the compositional variables have a significant relationship with stunting
prevalence rate change. However, a multivariate decomposition logistic regression analysis of
2011-2016 revealed that alterations in compositional characteristics of under five children
accounted for 29.66% of the entire change. Among the compositional change factors parents
education, height of mothers, children’s age, sex of children in month, anemia and region had a
statistically significant contribution on change in stunting. Children from family who had
primary education were more likely to be stunted than children from family who had higher
education. The coverage of parent’s primary education was increased from 25.53% in the year
2011 to 25.73% in 2016 (Table 1) that had a negative significant compositional contribution to
the decline in the prevalence rate of stunting by 10.05 (Table 3).

The likelihood of becoming stunted decreased as respondents' height increased. Between the
2011 and 2016 EDHS surveys, community compositional changes in respondents' height
(increased mean height) had increased reduction in the prevalence rate of stunting by 13.23%. In
a multilevel logistic analysis, female children were less likely than male children to be stunted.
Between 2005 and 2016, the proportion of female children included in the study increased by
0.1%. Stunting decreased by 4.28% as a result of this compositional alteration. Changes in the
composition of children in the Amhara region between surveys contributed to a 4.71% reduction
in the prevalence rate of stunting between 2011 and 2016. However, raising the proportion of
children with moderate anemia from 16.81% to 23.51% (table 1) had a negative impact on
stunting prevalence rate reduction by 7.26 % (Table 3).

Difference due to effects of coefficient (Coefficients)

Controlling the roles of change in compositional characteristics, behavioral changes among

women toward ANC follow up enhanced decline in the prevalence rate of stunting by 49.94 %,
according to a multivariate decomposed logistic regression analysis from 2005 to 2019. Also,

31
changing the behavior of children who were breastfed during the study period resulted in a
72.9% increase change in the prevalence rate of stunting. Stunting had been decreased by 37.7%
as a result of age-related behavioral changes in under-five children from young to old. Stunting
increased by 8.34% as a result of behavioral changes among children in the Tigray region as
compared with Adis Abeba. Stunting could have been decreased by 119.3% with behavioral
changes among children from middle-income families, according to a multivariate
decomposition logistic regression analysis conducted between 2011 and 2016. Change in the
prevalence rate of stunting was reduced by 233.14% and 109.63%, respectively, due to
behavioral modifications among children in the Oromia and SNNP regions. A good practice of
eating solid, semi-solid, or soft foods one or more times on the day before data collection time
resulted with a 119.97% increase in stunting prevalence rate change (Table 3).

32
Table 3: A multivariate decomposition analysis of different features in the 2005 -2019 EDHS and 2011 EDHS-2016 EDHS.

2005 EDHS-2019 EDHS 2011 EDHS-2016 EDHS

Characteristics Endowments Coefficients Endowments Coefficients
Coef with 95% CI Perc Coef with 95% CI Percen Coef with 95% Perce Coef with Percent
ent t CI nt 95% CI
Residence (urban as ref) -0.0004(-0.009, 0.27 0.081(-0.055, -52.21 0.0002(-0.001, -0.44 0.077(-0.0812, -164.63
Rural 0.008) 0.218) 0.0014) 0.235)

ANC (no as ref) 0.002(-0.034, 0.037) -1.19 -0.078(-0.137, 49.94 0.0063(-0.004, - -0.0313(- 66.90
Yes 0.019)* 0.016) 13.49 0.077, 0.015)
highest No -0.0024(-0.051, 1.56 0.016(-0.214, -10.05 -0.0094(-0.0193, 20.17 0.128(-0.120, -274.51
educational education 0.046) 0.245) 0.0005) 0.376)
level primary 0.002(-0.030, 0.034) -1.04 0.0023(-0.048, -1.45 0.005(0.00052, - 0.065(-0.037, -139.87
0.052) 0.0093)* 10.47 0.169)
secondary -0.0001(-0.0024, 0.07 -0.003(-0.016, 1.78 0.0037(-0.0005, -7.91 0.0048(-0.007, -10.28
0.002) 0.011) 0.008) 0.017)
Higher 1 1 1 1 1 1 1
Height of respondent -0.0062(-0.009, 13.23 0.277(-0.730, -593.33
0.0032)*** 1.284)
Duration Ever 1 1 1 1 1 1 1 1
of breastfed, not
breastfeed currently
ing Never 0.0001(-0.002, -0.07 -0.0003(-0.003, 0.21 -0.0005(-0.0034, 0.99 0.0006(-0.001, -1.19
breastfed 0.0021) 0.0023) 0.0025) 0.002)
still -0.0003(-0.0067, 0.20 -0.112(-0.185, 72.09 -0.0016(-0.0051, 3.33 -0.10(-0.250, 213.12
breastfeeding 0.006) 0.039)** 0.0020) 0.051)
water source(improved as -0.001(-0.020, 0.61 -0.0056(-0.016, 0.0003(-0.005, -0.67 -0.010(-0.052, 20.82
ref) 0.018) 0.005) 3.56 0.0053) 0.032)
unimproved
Wealth poorest -0.0001(-0.0017, 0.05 0.017(-0.016, -10.83 -0.0001(-0.0002, 0.19 -0.031(-0.075, 66.06
index 0.0015) 0.050) 0.0001) 0.013)
poorer 0.0001(-0.0014, -0.05 0.0025(-0.029, -1.60 -0.0001(-0.0005, 0.14 -0.039(- 84.26
0.0016) 0.037) 0.00033) 0.0822,
33
 0.0034)
middle -0.00003(-0.0007, 0.02 0.014(-0.016, -8.95 -0.0002(-0.0005, 0.34 -0.056(-0.100, 119.33
0.00061) 0.044) 0.0002) 0.013)*
richer -0.0001(-0.0017, 0.05 -0.001(-0.029, 0.41 -0.00003(-0.001, 0.06 -0.027(-0.060, 57.42
0.0016) 0.028) 0.0010) 0.005)
Richest 1 1 1 1 1 1 1 1
Age of child in month 0.0013(-0.025, -0.85 -0.059(-0.118, 37.70 0.0041(0.0031, -8.86 0.044(-0.023, -94.28
0.028) 0.00002)* 0.0052)*** 0.111)
Sex of child (male as ref ) -0.0001(-0.0018, 0.05 -0.022(-0.053, 13.96 -0.002(-0.0036, 4.28 -0.005(-0.038, 9.75
0.0016) 0.010) 0.0004)* 0.027)
Region Tigray 0.00014(-0.0026, -0.09 0.013(0.0003, -8.34 0.0006(-0.0005, -1.25 0.003(-0.0101, -6.92
0.0028) 0.026)* 0.0020) 0.017)
Affar 0.00003(-0.0006, -0.02 0.0011(-0.001, -0.71 -0.00001(- 0.00 0.00002(- -0.03
0.00064) 0.003) 0.0001, 0.0001) 0.002, 0.002)
Amhara -0.0001(-0.002, 0.06 -0.005(-0.049, 2.88 -0.002(-0.004, 4.71 0.026(-0.019, -55.20
0.002) 0.040) 0.00042)* 0.071)
Oromiya 0.0001(-0.001, -0.03 0.0036(-0.065, -2.32 0.0004(-0.0001, -0.76 0.110(0.004, -233.14
0.001) 0.073) 0.001) 0.214)*
Somali -0.00002(-0.0004, 0.01 -0.0024(-0.011, 1.51 -0.0001(-0.0014, 0.28 0.0012(-0.005, -2.57
0.0004) 0.006) 0.0011) 0.007)
Benishan 0.00001(-0.0003, -0.01 0.0012(-0.0008, -0.77 -0.0001(- 0.21 0.0024(- -5.09
gul- 0.00031) 0.003) 0.00022, 0.0003,
gumuz 0.00003) 0.0051)
SNNP -0.0001(-0.002, 0.06 -0.0018(-0.046, 1.12 -0.0008(-0.002, 1.80 0.051(0.0014, -109.63
0.002) 0.043) 0.00022) 0.101)*
Gambela -0.00001(-0.0003, 0.01 -0.00024(-0.0008, 0.15 -0.00001(- 0.01 0.001(-0.0001, -1.43
0.0002) 0.0004) 0.0001, 0.0001) 0.001)
Harari 0.00001(-0.0001, -0.00 0.0001(-0.0003, -0.05 -0.00001(- 0.01 0.0004(- -0.78
0.0001) 0.0005) 0.00002, 0.0001, 0.001)
0.00002)
Addis 1 1 1 1 1 1 1 1
Ababa
Dire 0.00001(-0.00003, -0.00 0.0002(-0.0004, -0.12 -0.00001(- 0.01 0.0004(- -0.86

34
 Dawa 0.00003) 0.001) 0.00002, 0.00033,
0.00001) 0.001)
Anemia Severe - - - 0.0007(-0.0011, -1.38 -0.0003(- 0.68
0.0024) 0.0025, 0.002)
Moderate - - - 0.0034(0.0011, -7.26 0.003(-0.003, -5.61
0.0057)** 0.008)
Mildane - - - 0.0012(-0.0025, -2.49 0.0014(- -3.09
mia 0.005) 0.0114,
0.0143)
not - - - 1 1 1 1 1
anemic
place of Home 1 1 1 1 1 1 1 1
delivery Health -0.0005(-0.0114, 0.33 0.005(-0.002, -3.26 -0.012(-0.025, 25.22 0.0035(-0.011, -7.54
institutio 0.010) 0.013) 0.0012) 0.0178)
n
Others -0.0001(-0.003, 0.08 -0.001(-0.002, 0.45 0.0008(-0.001, -1.62 0.0001(-0.002, -0.13
0.002) 0.0004) 0.0025) 0.002)
Dietery no - - - - 1 0.00 0.00000 0.00
diversity Yes - - - - -0.0015(-0.0056, 3.17 0.0002(-0.002, -0.39
0.0026) 0.002)
number None - - - - 1 0.00 0.00000 0.00
of times ate >=1 - - - - -0.0024(-0.004, 5.19 -0.056(-0.110, 119.97
solid, 0.001) 0.002)*
semi-solid
or soft
food
yesterday
Key: ∗pvalue<0.05; ∗∗pvalue<0.01; ∗∗pvalue<0.001

35
4.6. Multivariate decomposition logistic regression analysis of 2011-2019
EDHS
Difference due to characteristics (endowment)
The multivariate decomposition logistic regression analysis of 2011-2019 found that
decomposition changes in children's characteristics account for 26.23% of the overall change in
the prevalence rate of stunting. The remaining 73.77% was attributed to children's behavioral
changes. Wealth index, parenteral education, home water source, child's age in month, child's
sex, place of delivery, and region all exhibited a statistically significant compositional influence
on change in the prevalence rate of stunting when variables were decomposed. Keeping all
other behavioral variables equal, children from families with no education were more likely to
be stunted than children from families with a higher level of education. Between the surveys, a
reduction in the share of mothers with no education (Table 1) contributed to a 31% increase
change in the prevalence rate of stunting. In contrast, increased the proportion of mothers with a
primary education between the comparison periods, had reduced decline in stunting by 13.97%
(Table 4).

Stunting was more common in children from the poorest, poorer, middle, and richer families
than in children from the richest families. Between the surveys, changes in family composition
among poorer, medium, and richer households increased change in the prevalence rate of
stunting by 0.13%, 3.11%, and 2.28%, respectively, but changes in family composition among
the poorest families decreased change in stunting prevalence rate by 1.12%. The shift in age
structure among children included in the 2011 and 2019 EDHS accelerated the pace of change
in stunting prevalence rate by 4.23%. More female children were included in the 2019 EDHS,
which resulted in a 0.32 percent increase in stunting prevalence rate change. Between the
surveys, change the proportion of children in Tigray, Affar, and Harari declined change in
stunting prevalence rate by 0.76%, 0.79%, and 0.12%, respectively. However, between the
surveys, changes in the composition of children in the Amhara region and changes in the
composition of the place of delivery increased the change in the prevalence rate of stunting by
2.5% and 2.35%, respectively. Finally, decrease in coverage of an unimproved water source
from 40.06 % to 37.54% accounts for 8.57% declined of change in the prevalence rate of

37
stunting as compared with increase in coverage of improved water source from 36.61%
to37.23% (Table 4).

Difference due to effects of coefficient (Coefficients)

The multivariate decomposition logistic regression analysis of 2011-2019 found that behavioral
changes toward duration of breast feeding, dietary diversity, age in month, place of delivery,
and region had a significant effect on change in the prevalence rate of stunting between surveys
when keeping compositional characteristics constant. During the study period, changing the
behavior of breast-fed children resulted in a 57.2 percent rise change in the prevalence rate of
stunting. Stunting had been dropped by 134.53% as a result of age-related behavioral changes in
under-five children. Behavioral modifications in children from Tigray and Harari regions have
detrimental effect on change in stunting by 15.01% and 0.61%, respectively. It was able to
boost the change in stunting prevalence rate by 36.92% between the survey years by changing
children's behavior toward dietary diversification. Finally, behavioral adjustments toward other
places of delivery resulted in a 2.29% increase in stunting prevalence rate change (Table 4).
Table 4: A multivariate decomposition logistic regression analysis of different features in the
2011 EDHS-2016 EDHS.

2011 EDHS-2019 EDHS

Characteristics Endowments Coefficients
Coef with 95% CI Percent. Coef with 95% CI Percent.
Residence (urban as ref) -0.005(-0.013, 0.003) 7.07 0.026(-0.080, 0.132) -37.18
Rural

highest No -0.022(-0.039, 31.00 -0.002(-0.160, 0.160) 2.35

educational education 0.0041)*
level Primary 0.010(0.001, 0.020)* -13.97 -0.0092(-0.070, 13.32
0.052)
secondary -0.001(-0.007, 0.005) 1.31 -0.002(-0.006, 0.003) 2.94
Higher
Duration Ever breastfed, 1 1 1 1
of not currently
breastfeed Never 0.0001(-0.003, 0.003) -0.07 0.0012(-0.0013, -1.77
ing breastfed 0.004)
still -0.001(-0.002, 0.001) 0.74 -0.040(-0.079, 57.20
breastfeeding 0.001)*
water source(improved as 0.006(0.00001, -8.57 -0.024(-0.055, 0.007) 34.64
ref) 0.012)*
unimproved

38
Wealth Poorest 0.001(0.00033, -1.12 0.001(-0.030, 0.032) -1.41
index 0.00122)**
Poorer -0.0001(-0.0002, 0.13 -0.008(-0.040, 0.022) 12.16
0.00001)*
Middle -0.0022(-0.004, 3.11 -0.001(-0.030, 0.027) 1.05
0.001)*
Richer -0.0016(-0.0031, 2.28 -0.007(-0.033, 0.020) 9.51
0.00001)*
Richest 1 1
Age of child in month -0.001(-0.0044, 4.23 -0.093(-0.170, 134.53
0.0015)*** 0.020)*
Sex of child (male as ref ) -0.00022(-0.0004, 0.32 -0.013(-0.040, 0.012) 18.97
0.0001)**
Region Tigray 0.0005(0.0003, -0.76 0.010(0.00011, -15.01
0.0010)*** 0.021)*
Affar 0.0006(0.0001, -0.79 0.0002(-0.0013, -0.34
0.001)* 0.002)
Amhara -0.002(-0.0033, 2.50 0.012(-0.022, 0.045) -17.17
0.00014)*
Oromiya -0.0021(-0.0052, 3.02 0.023(-0.044, 0.090) -33.24
0.001)
Somali -0.0004(-0.005, 0.62 0.001(-0.0034, 0.005) -1.26
0.004)
Benishang 0.00001(-0.00001, -0.02 0.0006(-0.0012, -0.90
ul-gumuz 0.00003) 0.0025)
SNNP -0.0007(-0.0016, 0.94 0.013(-0.021, 0.047) -18.83
0.00033)
Gambela -0.0001(-0.0002, 0.10 0.0002(-0.0004, -0.24
0.00004) 0.001)
Harari 0.0001(0.00002, -0.12 0.0004(0.0001, -0.61
0.00014)** 0.001)*
Addis 1 1 1 1
Ababa
Dire 0.0001(-0.0001, -0.10 0.0001(-0.0004, -0.12
Dawa 0.00023) 0.001)
place of Home 1 1 1 1
delivery Health 0.007(-0.0062, 0.020) -10.14 0.004(-0.005, 0.012) -5.35
institution
Others -0.002(-0.003, 2.35 -0.002(-0.003, 2.29
0.0004)* 0.0004)*
Dietary No 1 1 1 0.00
diversity Yes 0.0004(-0.0013, -0.61 -0.026(-0.050, 36.92
0.0022) 0.003)*
Key: pvalue<0.05; pvalue<0.01; ∗∗pvalue<0.001
∗ ∗∗

39
5. Discussion
The aim of this study was to estimate changes in the prevalence rate of stunning, factors
associated with stunting and changes in the prevalence rate among Ethiopian children under the
age of five from 2005 to 2019. In the previous 15 years, Ethiopia has increased the number of
national policies and large-scale health, nutrition, and food security programs (77), resulting in a
steady drop in stunting from 47% in 2005 to 37% in 2019, which is comparable to other
countries (33). Stunting, on the other hand, continues to be a serious problem in Ethiopia due to
a variety of circumstances. The risk of stunting was higher in under-five children whose
mothers had no education than in those whose mothers had a higher education level, according
to a multilevel logistic regression analysis of the 2016 EDHS survey. A multivariate
decomposition analysis of the EDHS from 2005 to 2011 and 2005 to 2016 revealed that
lowering the number of mothers without education contributed to a reduction in stunting. There
is a considerable link between maternal education and children's nutrition, according to earlier
studies. Children born to educated mothers are less likely to be stunted than children born to
uneducated mothers (78, 79). Women's higher education is a critical component in improving a
family's socioeconomic level (80), and excellent socioeconomic status influences predictors of
stunting such as reproductive factors, feeding patterns, and health-care utilization (81). Our
findings, as well as those of earlier studies, have major policy implications because they imply
that by boosting mother's formal education, Ethiopia could ameliorate the impact of stunting on
children and lessen the country's high stunting-related ill-health among children.
When compared to children under the age of five from mothers with higher education, the
coverage of parent’s primary education rose from previous to recent surveys; yet, it had a
negative significant compositional impact to the decline in the prevalence rate of stunting. This
finding is consistent with a large-scale study conducted across three African countries, which
found that women's primary education had no significant impact on child stunting (82). These
findings suggest that educating women at the primary school level may not be adequate to
reduce stunting to the levels desired, and that policies to keep mothers in school beyond primary
school should be prioritized in order to reduce the number of stunted children in the country.

Behavioral changes toward mother's education did not indicate a significant relation to stunting
reduction in any of the analyses. Although a variety of initiatives, such as the Sustainable
Development Goals (SDGs), are emerging in Ethiopia to support and encourage women's
40
empowerment, reaching this goal has not been straightforward and has been hampered by
persisting regional inequities. In most regions, community attitudes on women's engagement in
development, women's access to and management of productive resources, and gender-based
equalities in training and education are unsatisfactory (83). The current finding implies that
focusing on women's perceptions and attitudes, as well as boosting women's education, may
have a good impact on bringing about the desired behavioral change in the community
regarding child nutrition.

In a multilevel logistic regression analysis, maternal height was found to be inversely related to
the prevalence of childhood stunting. Between 2005 and 2011, as well as 2005 and 2011, and
2011 and 2016, community compositional changes in responder height had a favorable impact
on change in the prevalence rate of stunting. Aside from genetics, environmental factors
including maternal nutrition, feeding patterns, and nutritional quality and quantity can all
influence growth of children before to the age of two (23). In addition, a variety of
socioeconomic issues, ranging from general conditions to poor feeding practices, which may
result in low maternal stature, may have an impact on early childhood growth and development
(84). In addition to heredity and shared environmental factors, the biological significance of the
mother milieu during pregnancy and lactation could have explained the link between maternal
height and early life stunting (23). It is plausible to claim that stunting is a cyclical process in
which women who were stunted as children have stunted offspring, producing an
intergenerational cycle of poverty and diminished human capital that is difficult to break (85).
As a result, policies and tactics that consider mothers and have been implemented over a short,
medium, or long period of time may have the intended effect on stunting in children's early
lives.

Behavioral changes among children who were on breast feeding during the survey time had a
positive contribution for change in stunting among children compared to children who were on
breast feeding previously but not currently, according to the multivariate decomposition logistic
regression analysis of 2005-2011, 2005-2016, and 2011-2019. The National Nutrition Strategy
of the Federal Democratic Republic of Ethiopia, which has been implemented over the last few
decades, focuses on mainstreaming and strengthening nutrition activities through community-
based nutrition programs that help to reduce food insecurity and unbalanced nutrient

41
consumption. Community-based health and agriculture extension programs, health service
delivery, education, and gender programs all received more attention. The community-based
nutrition program also includes growth monitoring and promotion for all children under the age
of two, as well as caregiver counseling (93, 94). Thus, the encouraging drop in stunting
observed due to behavioral changes among children who were breast-fed during data collection
could be attributed to the implementation of a community-based nutrition program. This
indicates that further enhancing the program will provide very promising results in terms of
eradicating stunting among Ethiopian children.

Between 2005 and 2011, the number of children in the poorest wealth index category increased,
and change in stunting among children decreased, according to the EDHS. At the same time,
between 2011 and 2019, the proportion of children from lower, middle, and upper-middle-class
families decreased, resulting in a more rapidly fall in child stunting. Previous findings that
attempted to investigate the effects of economic growth on undernutrition in Ethiopia have
similarly confirmed the direct effect of economic growth on stunting (95). According to a
study, a 10% rise in GDP per capita reduces the frequency of child stunting by 2.7 percent. In
this regard, the average cost of stunting in poor nations has been estimated to be around 13.5
percent of GDP per capita (96). According to published literature, the link between the
prevalence of stunting and economic growth is stronger among children from low-income
nations (97, 98), implying that the household's financial level is the foundation for all nutritional
interventions implemented in disadvantaged areas. These findings may serve as a reminder to
Ethiopian policymakers to place a greater emphasis on policies that promote economic growth
as well as nutrition-related programs.

Similarly, a multivariate decomposition logistic regression assessment of 2011-2016 revealed

that behavioral changes among children from middle-income families were associated with an
increase change in stunting. Ethiopia's Growth and Transformation Plan (GTP) is a five-year
development plan that runs from 2010/11 to 2014/15 (99). In both the health and agriculture
sectors, community-based service delivery systems have been made available throughout this
time period to assure decentralized and democratized public services. Health extension workers,
in particular, play a critical role in strengthening and accelerating social and behavioral changes
in children's eating habits, both in rural and urban regions (93). As a result, the decrease in

42
stunting owing to behavioral changes among children from middle-income homes could be the
outcome of programs established during this time period, which could be a useful lesson in
achieving the country's aim.

The risk of stunting grew as the child's age climbed month by month. Changes in age structure
among children in the EDHS from 2005 to 2011, 2005 to 2016, and 2011 to 2019 showed an
increase in stunting. To achieve optimal growth in children, the amount and frequency with
which they are fed should be increased: 2–3 meals per day for infants aged 6–8 months, 3–4
meals per day for infants aged 9–23 months, plus 1–2 additional snacks as needed (100).
However, findings from national representative data showed that the frequency of infant and
child feeding practices dropped as the child's age increased by one unit (101). In most locations,
young child feeding practice is also inadequate, and providing children the minimum
appropriate diet variety does not grow with age (102). Short birth intervals, poor weaning and
dietary practices, reduced and inappropriate breast and supplemental feeding techniques, food
insecurity and poverty, as well as the nutritional status of mothers or caregivers, are all common
contributors. As a child grows older, he or she is more likely to drink unclean water and eat
improperly prepared food, both of which can expose youngsters to a number of infections and
diarrheal disorders, leading to chronic malnutrition (103). All of these things could be
contributing factors to the child's inability to achieve optimal growth as they get older.

The multivariate decomposition logistic regression analysis of the EDHS for the years 2005-
2011, 2005-2019, and 2011-2019 revealed that age-related behavioral changes among children
from young to old age played a favorable role in stunting reduction. Since 2004, Ethiopia's
Federal Ministry of Health's Family Health Department has adopted a national policy to
improve baby and child feeding practices, with the goal of gradually increasing food
consistency and diversity as newborns grow older, while responding to their needs and skills
(104). Ethiopia made significant progress in extending community-based primary health care
delivered by health extension workers as a result of these programs. Because of their influence
on eating decisions and access to mass media, Alive & Thrive launched a radio and television
campaign aimed largely at men to reinforce and expand the impact of community interventions
and to reach individuals outside of program areas. Each television and radio broadcast
emphasized the importance of male involvement in infant feeding. (105). The decrease in

43
stunting may be due to changes in parental behavior towards newborn and young child feeding
practices, which may have been affected by radio and television programs used as
communication tools under Ethiopia's Growth and Nutrition program (106). This means that,
while the economic and political hurdles to improving Ethiopia’s nutritional status are
enormous and appears insurmountable; strengthening the existing nutritional interventions can
make more of a difference in reducing early life growth failure.

From 2011 to 2019, the proportion of female children included in the study increased. Stunting
among children under the age of five has decreased significantly as a result of this
compositional change. Male children under the age of five in Sub-Saharan Africa are more
likely than girls to be stunted (107). Gender differences in mortality and morbidity could
explain this. Even though there is no clear understanding of early childhood health inequalities,
epidemiological research consistently show that boys have higher mortality and morbidity than
females (108). Other potential determinants, such as social role valorization of daughters and
nutritional discrimination, have not been widely investigated in Ethiopia, implying the need for
more exploratory research in the area.

In all combinations of survey data, a decrease in the proportion of children from the Amhara
region increased the change in stunting. Amhara region has the country's third highest rate of
monetary poverty, as well as the greatest disparity between rural and urban communities (109,
110). Many households can only generate enough food to meet their nutritional needs for about
six months of the year (111). The Amhara region's children are similarly worse off than the
national average in terms of basic necessities and services (112). There are significant gaps in
health care professionals' knowledge and abilities, facility readiness, administration and
leadership, and the availability of crucial supplies in various parts of the region. Maternal and
newborn health services are still underutilized, and maternal and newborn care is of poor quality
(112, 113). This could all be contributing to the high rate of early life growth failure in the
Amhara region, implying the need for a variety of interventions to guarantee children have
access to both meals and health services that would effectively meet their multifaceted needs for
growth and development.

Between 2011 and 2019, the proportion of children in the Tigray region who were stunted
dropped. Despite considerable improvements in access to health care services in the region,

44
producing an acceptable amount of food is extremely challenging due to a scarcity of suitable
farmland. Eight nine percent of the population earns less than £2 a day, while the bulk of the
population produces less than half of their annual minimum food requirements (114). However,
there is successful execution of a health extension program and the expansion of health-care
facilities. (115). In the Tigray area, total universal health service coverage is nearly comparable
to Addis Abeba, and is complemented by a high level of facility delivery and children
vaccination (116, 117). All of these could not aid communities in developing appropriate child
feeding practices behavior as it illustrated in analysis of EDHS between 2005 and 2019,
behavioral changes among children in the Tigray region contributed significantly to change in
stunting negatively compared to Addis Abeba.

Similarly, between 2011 and 2019, the proportion of children in the Harari and Affar regions
who were stunted increased. Harari has had a significant decrease in monetary poverty in recent
decades, beginning in 2004/05. The region’s overall monetary poverty rate has dropped to 7%,
the lowest in the country. Similarly to monetary poverty, the number of people living in food
poverty has considerably dropped (118). According to reports, children in the Harari region are
less likely than the national average to be deprived of a greater number of fundamental
requirements and rights (119). Despite the fact that many people in the Afar region face chronic
food poverty, over 90% of the Afar community relies on a pastoralist subsistence strategy (118).
As a result, there is a relatively high culture of feeding children animal products with great
nutritional content to counteract stunting (120). As a result, when additional children from the
two locations are sampled between the surveys, a relatively low degree of stunting is
anticipated. In contrast, between 2005 and 2016, a fall in the proportion of children from less
risky areas such as Dire Dewa city had a deleterious effect on stunting reduction.

From 2011 to 2016, behavioral modifications among children in the Oromia and SNNP regions
were responsible for a significant drop in stunting change. Between 2011 and 2019, this was
also true in the Harari region. This could be linked to the political resistance that existed in
southern part of Ethiopia, particularly in the Ormia region. Oromia has underperformed on
maternity and child health care over this time period. During the study period, good practices
such as facility delivery, ANC, and postnatal care that might change mothers' behavior toward
child feeding practice were the lowest in all regions (121). In the pastoralist areas of Oromia,

45
the execution of the Health Extension Programme (HEP) has also been hampered (122).
Similarly, despite its economic success, SNNPR has Ethiopia's highest multi-dimensional child
deprivation (MCD) rate (123). Despite the fact that Ethiopia has experienced significant poverty
reduction in these areas, coordination for the development of good child feeding habits is
lacking due to a lack of awareness, frequent turnover of focal persons and management, a lack
of accountability and responsibility, and a lack of nutrition structures in each specific area (123,
124).

Anemia and a lack of improved water sources, both of which are well-known causes of chronic
malnutrition, are also negative drivers of stunting change in our study (125). On the other hand,
behavioral changes toward nutritional diversity, ANC follow-up, place of delivery, and eating
solid, semi-solid, or soft foods have all been linked to a reduction in childhood stunting.
Ethiopia could reduce the burden of early life growth failure by increasing access to improved
water sources, maternal and child care, and well-structured patient education programs to
increase self-awareness and a positive attitude toward maternal care and child feeding practice,
according to this finding.

46
6. Conclusion

Despite the fact that several projects to eliminate stunting have been implemented in Ethiopia, a
significant number of children remain stunted. Compositional features of children, including
wealth index, parental education, child's age in months, sex of child, duration of breast feeding,
anemia, unimproved water supply, and region, all had a statistically significant impact on
stunting change. Due to changes in coefficients; dietary diversity, ANC follow-up, place of
delivery, eating solid, semi-solid, or soft meals, and age all exhibited a significant association
with change in stunting.

47
7. Recommendation
 The Ethiopian Ministry of Health should maintain its present efforts to improve dietary
diversity, ANC follow-up, institutional delivery, and the feeding of solid, semi-solid, or soft
foods to children above the age of six months.
 The Ethiopian government and the ministry of health should place a specific emphasis on
impoverished areas, such as the Amhara region.
 Ethiopia's government should step up its efforts to raise the wealth index.
 These findings imply that nongovernmental organizations aiming to combat malnutrition
should include a structured economic empowerment program in addition to their regular
health activities.
 In this study, maternal height has an impact on children's linear growth over time. Genetic
and non-genetic variables are likely to play a role. Despite the fact that other aspects could
be investigated by our study, the genetic component is beyond the scope of this study, and a
gentical investigation undertaken by genetic epidemiologists would be beneficial.

48
References
1. Price S. Understanding the importance to health of a balanced diet. Nursing times.
2005;101(1):30-1.
2. Lim S. Eating a balanced diet: a healthy life through a balanced diet in the age of
longevity. Journal of obesity & metabolic syndrome. 2018;27(1):39.
3. Acosta A, Fanzo J. Fighting maternal and child malnutrition: analysing the political and
institutional determinants of delivering a national multisectoral response in six countries.
Synthesis paper Brighton, UK: Institute for Development Studies. 2012.
4. World Health Organization. Reducing stunting in children: equity considerations for
achieving the Global Nutrition Targets 2025. 2018.
5. Sferruzzi-Perri A, Vaughan O, Forhead Al FA. IN A NUTSHELL.
growth.198935070:94.
6. World Health Organization. Nutrition landscape information system ( NLiS): results of a
user survey. World Health Organization, 2019.
7. Black RE, Victora CG, Walker SP, Bhutta ZA, Christian P, De Onis M, et al. Maternal
and child undernutrition and overweight in low-income and middle-income countries.
The lancet. 2013;382(9890):427-51.
8. Black RE, Allen LH, Bhutta ZA, Caulfield LE, De Onis M, Ezzati M, et al. Maternal
and child undernutrition: global and regional exposures and health consequences. The
lancet. 2008;371(9608):243-60.
9. WFP/Nyani Quarmyne, Social and Economic Impact of Child Undernutrition on
Ghana’s Long-Term Development. Available on
:https://docs.wfp.org/api/documents/WFP-0000119786/download/ . Extracted on
8/9/2021. .
10. Bommer C, Vollmer S, Subramanian S. How socioeconomic status moderates the
stunting-age relationship in low-income and middle-income countries. BMJ global
health. 2019;4(1):e001175.
11. Jonah CM, Sambu WC, May JD. A comparative analysis of socioeconomic inequities in
stunting: a case of three middle-income African countries. Archives of Public Health.
2018;76(1):1-15.
12. Indicators K. Mini Demographic and Health Survey. 2019.

49
13. Bach A, Gregor E, Sridhar S, Fekadu H, Fawzi W. Multisectoral Integration of
Nutrition, Health, and Agriculture: Implementation Lessons From Ethiopia. Food and
nutrition bulletin. 2020;41(2):275-92.
14. De Onis M, Branca F. Childhood stunting: a global perspective. Maternal & child
nutrition. 2016;12:12-26.
15. UNICEF, WHO, World Bank. Levels and Trends in Child Malnutrition. Joint Child
Malnutrition Estimates. New York, NY: United Nations International Children’s Fund;
Geneva: WHO; Washington, DC: World Bank, 2012. Available from:
http://www.who.int/nutgrowthdb/estimates/en/.
16. de Onis M, Dewey KG, Borghi E, Onyango AW, Blössner M, Daelmans B, et al. The
World Health Organization's global target for reducing childhood stunting by 2025:
rationale and proposed actions. Maternal & child nutrition. 2013 Sep;9 Suppl 2(Suppl
2):6-26. PubMed PMID: 24074315. Pubmed Central PMCID: PMC6860845. Epub
2013/10/23. eng.
17. Unicef. Levels and trends in child malnutrition. eSocialSciences, 2018.
18. World Health Organization. Levels and trends in child malnutrition: key findings of the
2019 edition. World Health Organization, 2019.
19. Black RE, Victora CG, Walker SP, Bhutta ZA, Christian P, de Onis M, et al. Maternal
and child undernutrition and overweight in low-income and middle-income countries.
Lancet (London, England). 2013 Aug 3;382(9890):427-51. PubMed PMID: 23746772.
Epub 2013/06/12. eng.
20. Ahmed T, Hossain M, Sanin KI. Global burden of maternal and child undernutrition and
micronutrient deficiencies. Annals of nutrition & metabolism. 2012;61 Suppl 1:8-17.
PubMed PMID: 23343943. Epub 2013/02/21. eng.
21. World Health Organization. Global nutrition targets 2025: Stunting policy brief. World
Health Organization, 2014.
22. UNICEF. Stop Stunting: Progress Report 2013–2015. Unicef; 2015.
23. Martorell R, Zongrone A. Intergenerational influences on child growth and
undernutrition. Paediatric and perinatal epidemiology. 2012;26:302-14.
24. IFPRI. Global Nutrition Report 2016: From Promise to Impact: Ending Malnutrition by
2030. International Food Policy Research Institute (IFPRI). 2016.

50
25. World Health Organization. World health statistics 2016: monitoring health for the
SDGs sustainable development goals: World Health Organization; 2016.
26. FDRE. ―SEQOTA‖ Declaration: Implementation Plan (2016-2030). 2016.
27. FMOH. Federal DemocraticRepublic of Ethiopia Ministry of Health. HSTP: Health
Sector Transformation Plan (2015/16-2019/20).
28. Hailu BA, Bogale GG, Beyene J. Spatial heterogeneity and factors influencing stunting
and severe stunting among under-5 children in Ethiopia: spatial and multilevel analysis.
Scientific Reports. 2020 2020/10/02;10(1):16427.
29. Tasic H, Akseer N, Gebreyesus SH, Ataullahjan A, Brar S, Confreda E, et al. Drivers of
stunting reduction in Ethiopia: a country case study. The American journal of clinical
nutrition. 2020;112(Supplement_2):875S-93S.
30. Mengesha HG, Vatanparast H, Feng C, Petrucka P. Modeling the predictors of stunting
in Ethiopia: analysis of 2016 Ethiopian demographic health survey data (EDHS). BMC
Nutrition. 2020 2020/09/22;6(1):52.
31. Abeway S, Gebremichael B, Murugan R, Assefa M, Adinew YM. Stunting and Its
Determinants among Children Aged 6–59 Months in Northern Ethiopia: A Cross-
Sectional Study. Journal of Nutrition and Metabolism. 2018 2018/06/25;2018:1078480.
32. Brief AE-IP. Reducing Stunting in Ethiopia:―From Promise to Impact‖. 2019.
33. Organization WH. Levels and trends in child malnutrition: UNICEF/WHO/The World
Bank Group joint child malnutrition estimates: key findings of the 2021 edition. Levels
and trends in child malnutrition: UNICEF/WHO/The World Bank Group joint child
malnutrition estimates: key findings of the 2021 edition2021.
34. Khan S, Zaheer S, Safdar NF. Determinants of stunting, underweight and wasting
among children < 5 years of age: evidence from 2012-2013 Pakistan demographic and
health survey. BMC public health. 2019 2019/04/01;19(1):358.
35. Das S, Gulshan J. Different forms of malnutrition among under five children in
Bangladesh: a cross sectional study on prevalence and determinants. BMC Nutrition.
2017 Dec;3(1):1-2.
36. 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 2019/03/04;5(1):19.

51
37. Menon P, Headey D, Avula R, Nguyen PH. Understanding the geographical burden of
stunting in India: A regression-decomposition analysis of district-level data from 2015-
16. Maternal & child nutrition. 2018 Oct;14(4):e12620. PubMed PMID: 29797455.
Pubmed Central PMCID: PMC6175441. Epub 2018/05/26. eng.
38. 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.
39. Vaivada T, Akseer N, Akseer S, Somaskandan A, Stefopulos M, Bhutta ZA. Stunting in
Childhood: an overview of global burden, trends, determinants, and drivers of decline.
The American journal of clinical nutrition. 2020;112(Supplement_2):777S-91S.
40. 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.
41. Headey D, Hoddinott J. Understanding the rapid reduction of undernutrition in Nepal.
2001.
42. Goyal N, Canning D. Exposure to ambient fine particulate air pollution in utero as a risk
factor for child stunting in Bangladesh. International journal of environmental research
and public health. 2018;15(1):22.
43. Headey D, Hoddinott J, Ali D, Tesfaye R, Dereje M. The other Asian enigma:
explaining the rapid reduction of undernutrition in Bangladesh. World Development.
2015;66:749-61.
44. Ervin PA, Bubak V. Closing the rural-urban gap in child malnutrition: evidence from
Paraguay, 1997–2012. Economics & Human Biology. 2019;32:1-10.
45. Humbwavali JB, Giugliani C, Silva I, Duncan BB. Temporal trends in the nutritional
status of women and children under five years of age in sub-Saharan African countries:
ecological study. Sao Paulo medical journal = Revista paulista de medicina. 2018 Sep-
Oct;136(5):454-63. PubMed PMID: 30570097. Epub 2018/12/21. eng.
46. Sunguya BF, Zhu S, Mpembeni R, Huang J. Trends in prevalence and determinants of
stunting in Tanzania: an analysis of Tanzania demographic health surveys (1991–2016).
Nutrition journal. 2019 2019/12/10;18(1):85.

52
47. CSA-Ethiopia I. International: Ethiopia Demographic and Health Survey 2011. Central
Statistical Agency of Ethiopia and ICF International Addis Ababa, Ethiopia and
Calverton, Maryland, USA. 2012.
48. Macro ORC. Central Statistical Agency. Ethiopia demographic and health survey. 2005.
49. Gebreyesus SH, Mariam DH, Woldehanna T, Lindtjørn B. Local spatial clustering of
stunting and wasting among children under the age of 5 years: implications for
intervention strategies. Public health nutrition. 2016;19(8):1417-27.
50. Bogale TY, Bala ET, Tadesse M, Asamoah BO. Prevalence and associated factors for
stunting among 6–12 years old school age children from rural community of Humbo
district, Southern Ethiopia. BMC public health. 2018 2018/05/24;18(1):653.
51. Bharti R, Dhillon P, Narzary PK. A spatial analysis of childhood stunting and its
contextual correlates in India. Clinical Epidemiology and Global Health. 2019;7(3):488-
95.
52. Utami RA, Setiawan A, Fitriyani P. Identifying causal risk factors for stunting in
children under five years of age in South Jakarta, Indonesia. Enfermeria clinica.
2019;29:606-11.
53. Keino S, Plasqui G, Ettyang G, van den Borne B. Determinants of stunting and
overweight among young children and adolescents in sub-Saharan Africa. Food and
nutrition bulletin. 2014 Jun;35(2):167-78. PubMed PMID: 25076764. Epub 2014/08/01.
eng.
54. Sanni Yaya, Olanrewaju Oladimeji, Emmanuel Kolawole Odusina, Ghose Bishwajit,
Household structure, maternal characteristics and children’s stunting in sub-Saharan
Africa: evidence from 35 countries, International Health, 2020;, ihz105,
https://doi.org/10.1093/inthealth/ihz105.
55. Nshimyiryo A, Hedt-Gauthier B, Mutaganzwa C, Kirk CM, Beck K, Ndayisaba A, et al.
Risk factors for stunting among children under five years: a cross-sectional population-
based study in Rwanda using the 2015 Demographic and Health Survey. BMC public
health. 2019 2019/02/11;19(1):175.
56. Hagos S, Hailemariam D, WoldeHanna T, Lindtjørn B. Spatial heterogeneity and risk
factors for stunting among children under age five in Ethiopia: a Bayesian geo-statistical
model. PloS one. 2017;12(2):e0170785.

53
57. Haile D, Azage M, Mola T, Rainey R. Exploring spatial variations and factors
associated with childhood stunting in Ethiopia: spatial and multilevel analysis. BMC
pediatrics. 2016;16(1):1-14.
58. Haile D, Azage M, Mola T, Rainey R. Exploring spatial variations and factors
associated with childhood stunting in Ethiopia: spatial and multilevel analysis. BMC
pediatrics. 2016 2016/04/15;16(1):49.
59. Arimond M, Ruel MT. Dietary diversity is associated with child nutritional status:
evidence from 11 demographic and health surveys. The Journal of nutrition.
2004;134(10):2579-85.
60. Tessema M, Belachew T, Ersino G. Feeding patterns and stunting during early
childhood in rural communities of Sidama, South Ethiopia. Pan Afr Med J. 2013;14:75.
PubMed PMID: 23646211. Pubmed Central PMCID: PMC3641921. Epub 2013/05/07.
eng.
61. Chalabi D. Acute respiratory infection and malnutrition among children below 5 years
of age in Erbil governorate, Iraq. EMHJ-Eastern Mediterranean Health Journal, 19 (1),
66-70, 2013. 2013.
62. Torres AM, Peterson KE, Souza ACTd, Orav EJ, Hughes M, Chen LC. Association of
diarrhoea and upper respiratory infections with weight and height gains in Bangladeshi
children aged 5 to 11 years. Bulletin of the World Health Organization. 2000;78:1316-
23.
63. De Meillon B. The Anophelini of the Ethiopian geographical region. Publications of the
South African Institute for Medical Research. 1947;10(49).
64. UNICEF-Children in Ethiopia,2018. Available on
https://www.unicef.org/ethiopia/children-ethiopia.
65. Central Statistical Agency .Ethiopia Demographic and Health Survey . Addis Ababa,
Ethiopia , September 2006
66. Federal Ministry of Health. Ethiopia Mini Demographic and Health Survey 2019. Key
Indicators , Ethiopian Public Health Institute Addis Ababa. The DHS Program ICF
Rockville, Maryland, USA,July 2019 .
67. Central Statistical Agency - CSA/Ethiopia and ICF, Ethiopia Demographic and
Health Survey 2016. 2017, CSA and ICF: Addis Ababa, Ethiopia.

54
68. Central Statistical Agency - CSA/Ethiopia and ICF, Ethiopia. Ethiopia
Demographic and Health Survey 2011 . Available on :
https://dhsprogram.com/publications/publication-fr255-dhs-final-reports.cfm. Extracted
on 8/10/2021.
69. World Health Organization. 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: World Health Organization; 2006.
70. World Health Organization. Early initiation of breastfeeding to promote exclusive
breastfeeding. World Health Organization www who
int/elena/titles/early_breastfeeding/en/,/entity/elena/titles/early_breastfeeding/en/index
html. 2019.
71. World health organization. Exclusive breastfeeding for optimal growth, development
and health of infants. e-Library of Evidence for Nutrition Actions (eLENA) ,1 February
2019 09:29 CET.Available on
https://www.who.int/elena/titles/exclusive_breastfeeding/en/. .
72. Marriott BP, White A, Hadden L, Davies JC, Wallingford JC. World Health
Organization (WHO) infant and young child feeding indicators: associations with
growth measures in 14 low‐ income countries. Maternal & child nutrition.
2012;8(3):354-70.
73. Kvarnström E, McConville J, Bracken P, Johansson M, Fogde M. The sanitation ladder–
a need for a revamp?. Journal of Water, Sanitation and Hygiene for Development. 2011
Mar;1(1):3-12.
74. WHO. Childhood Stunting: Contexts, Causes and Consequences; WHO: Geneva,
Switzerland, 2017.
75. Akaike H. A new look at the statistical model identification. IEEE transactions on
automatic control. 1974;19(6):716-23.
76. Powers DA, Yoshioka H, Yun M-S. mvdcmp: Multivariate decomposition for nonlinear
response models. The Stata Journal. 2011;11(4):556-76.
77. Ayele S, Zegeye EA, Nisbett N. Multi-Sectoral Nutrition Policy and Programme Design,
Coordination and Implementation in Ethiopia. 2020.

55
78. Abuya B, Onsomu E, Kimani J, Moore D. Influence of maternal education on child
immunization and stunting in Kenya. Maternal and child health journal.
2011;15(8):1389-99.
79. Kabubo-Mariara J, Ndenge GK, Mwabu DK. Determinants of children's nutritional
status in Kenya: evidence from demographic and health surveys. Journal of African
Economies. 2009;18(3):363-87.
80. Desai S, Alva S. Maternal education and child health: is there a strong causal
relationship? Demography. 1998 Feb;35(1):71-81. PubMed PMID: 9512911. Epub
1998/03/26. eng.
81. Makoka D. The impact of maternal education on child nutrition: evidence from Malawi,
Tanzania, and Zimbabwe: ICF International; 2013.
82. Makoka D, Masibo PK. Is there a threshold level of maternal education sufficient to
reduce child undernutrition? Evidence from Malawi, Tanzania and Zimbabwe. BMC
pediatrics. 2015;15(1):1-10.
83. Elezaj E, Ramful N, Cebotari V, Neubourg CD. Gender Equality, Women’s
Empowerment and Child Wellbeing in Ethiopia. UNICEF, 2019.
84. Kim R, Mejia-Guevara I, Corsi DJ, Aguayo VM, Subramanian S. Relative importance
of 13 correlates of child stunting in South Asia: insights from nationally representative
data from Afghanistan, Bangladesh, India, Nepal, and Pakistan. Social Science &
Medicine. 2017;187:144-54.
85. Prendergast AJ, Humphrey JH. The stunting syndrome in developing countries.
Paediatrics and international child health. 2014;34(4):250-65.
86. Target WGN. Stunting Policy Brief. World Heal Organ. 2012;14(9):1-8.
87. Dewey KG, Adu‐ Afarwuah S. Systematic review of the efficacy and effectiveness of
complementary feeding interventions in developing countries. Maternal & child
nutrition. 2008;4:24-85.
88. Federal Ministry of Health, Family Health Department, Ethiopia. Ethiopian National
Strategy on Infant and Young Child Feeding,April,2004:Available at
:https://motherchildnutrition.org/nutrition-protection-promotion/pdf/mcn-national-
strategy-for-infant-and-young-child-feeding-ethiopia.pdf.Ectracted on 04/04/2022.

56
89. Ahmed KY, Page A, Arora A, Ogbo FA. Trends and factors associated with
complementary feeding practices in Ethiopia from 2005 to 2016. Maternal & child
nutrition. 2020;16(2):e12926.
90. Dafursa K, Gebremedhin S. Dietary diversity among children aged 6–23 months in
Aleta Wondo District, Southern Ethiopia. Journal of nutrition and metabolism.
2019;2019.
91. Hirvonen K, Wolle A, Laillou A, Vinci V, Chitekwe S, Baye K. Understanding delays in
the introduction of complementary foods in rural Ethiopia. Maternal & child nutrition.
2021:e13247.
92. Masuke R, Msuya SE, Mahande JM, Diarz EJ, Stray-Pedersen B, Jahanpour O, et al.
Effect of inappropriate complementary feeding practices on the nutritional status of
children aged 6-24 months in urban Moshi, Northern Tanzania: Cohort study. PloS one.
2021;16(5):e0250562.
93. Democratic F. Republic of Ethiopia: National Nutrition Program: 2015. 2015.
94. Ethiopia FMoHotFDRo. National Nutrition Strategy. The Federal Democratic Republic
of Ethiopia Addis Ababa, Ethiopia; 2008.
95. Biadgilign S, Shumetie A, Yesigat H. Does economic growth reduce childhood
undernutrition in Ethiopia? PloS one. 2016;11(8):e0160050.
96. Mary S. How much does economic growth contribute to child stunting reductions?
Economies. 2018;6(4):55.
97. Yaya S, Uthman OA, Kunnuji M, Navaneetham K, Akinyemi JO, Kananura RM, et al.
Does economic growth reduce childhood stunting? A multicountry analysis of 89
demographic and health surveys in sub-Saharan Africa. BMJ global health.
2020;5(1):e002042.
98. McGovern ME, Krishna A, Aguayo VM, Subramanian S. A review of the evidence
linking child stunting to economic outcomes. International journal of epidemiology.
2017;46(4):1171-91.
99. Shapiro BI, Gebru G, Desta S, Negassa A, Negussie K, Aboset G, et al. Ethiopia
livestock master plan: Roadmaps for growth and transformation. 2015.

57
100. Victora CG, Bahl R, Barros AJ, França GV, Horton S, Krasevec J, et al. Breastfeeding
in the 21st century: epidemiology, mechanisms, and lifelong effect. The Lancet.
2016;387(10017):475-90.
101. Geda NR, Feng CX, Janzen B, Lepnurm R, Henry CJ, Whiting SJ. Infant and young
child feeding practices in Ethiopia: analysis of socioeconomic disparities based on
nationally representative data. Archives of Public Health. 2021;79(1):1-10.
102. Tassew AA, Tekle DY, Belachew AB, Adhena BM. Factors affecting feeding 6–23
months age children according to minimum acceptable diet in Ethiopia: A multilevel
analysis of the Ethiopian Demographic Health Survey. PloS one. 2019;14(2):e0203098.
103. Gebrehiwot SW, Abera G, Tesfay K, Tilahun W. Short birth interval and associated
factors among women of child bearing age in northern Ethiopia, 2016. BMC women's
health. 2019;19(1):1-9.
104. Federal Ministry of Health Family Health Department Ethiopia .National strategy for
infant and young child feeding.April 2004.Available on
:https://extranet.who.int/nutrition/gina/sites/default/filesstore/ETH%202004%20Nationa
l%20Strategy%20for%20Infant%20and%20Young%20Child%20Feeding.pdf. Extracted
on Novemebr 11/16/2021.
105. Alive and thrive .Improving child feeding practices to prevent malnutrition: Alive &
Thrive’s approach and results in Ethiopia. Available at:
https://www.aliveandthrive.org/sites/default/files/attachments/Ethiopia-country-
brief_web-version.pdf.
106. Workicho A, Biadgilgn S, Kershaw M, Gizaw R, Stickland J, Assefa W, et al. Social
and behaviour change communication to improve child feeding practices in Ethiopia.
Maternal & child nutrition. 2021:e13231.
107. Wamani H, Åstrøm AN, Peterson S, Tumwine JK, Tylleskär T. Boys are more stunted
than girls in sub-Saharan Africa: a meta-analysis of 16 demographic and health surveys.
BMC pediatrics. 2007;7(1):1-10.
108. Alkema L, Chao F, You D, Pedersen J, Sawyer CC. National, regional, and global sex
ratios of infant, child, and under-5 mortality and identification of countries with outlying
ratios: a systematic assessment. The Lancet Global Health. 2014;2(9):e521-e30.

58
109. CSA, LSMS—Integrated Surveys on Agriculture, Ethiopia Socioeconomic Survey
(ESS) 2015/16, 2017, p. 49.
110. Eradicating NEsPT. Poverty: An Interim Report on 2015/16 poverty analysis study.
Addis Ababa. 2017.
111. National Planning Commission. September 2017. Ethiopia’s Progress towards
Eradicating Poverty፡ an Interim Report on 2015/16 Poverty Analysis Study. .

112. UNICEF. SITUATION ANALYSIS OF CHILDREN AND WOMEN: AMHARA

REGION. 2019.
113. Bank W. Maintaining the Momentum While Addressing Service Quality and Equity: A
Diagnostic of Water Supply, Sanitation, Hygiene, and Poverty in Ethiopia: World Bank;
2017.
114. Van der Veen A, Gebrehiwot T. Effect of policy interventions on food security in
Tigray, Northern Ethiopia. Ecology and Society. 2011;16(1).
115. Sebastian MS, Lemma H. Efficiency of the health extension programme in Tigray,
Ethiopia: a data envelopment analysis. BMC Int Health Hum Rights. 2010 Jun 14;10:16.
PubMed PMID: 20546597. Pubmed Central PMCID: PMC2893104. Epub 2010/06/16.
eng.
116. Assefa Y, Hill PS, Gilks CF, Admassu M, Tesfaye D, Van Damme W. Primary health
care contributions to universal health coverage, Ethiopia. Bulletin of the World Health
Organization. 2020;98(12):894.
117. Eregata GT, Hailu A, Memirie ST, Norheim OF. Measuring progress towards universal
health coverage: national and subnational analysis in Ethiopia. BMJ global health.
2019;4(6):e001843.
118. Federal Democratic Republic of Ethiopia (FDRE), National Planning Commission,
Ethiopia’s Progress Towards Eradicating Poverty: An interim report on 2015/16 poverty
analysis study, 2017, p. 21.
119. CSA and UNICEF, MCD in Ethiopia, First National Estimates, 2018, p. 101.
120. UNICEF. Situation Analysis of Children and Women: Afar Region. 2019.
121. Regional Health Bureau, 2016, EmONC assessment. Note that the EmONC assessment
reported 77 per cent of women delivering in a health facility in 2016.

59
122. Federal Ministry of Health, HSTP 2015/16–2019/20. Mid-Term Review (VOL. I).
Comprehensive Report. DRAFT REPORT, 2018, p. 34.
123. UNICEF. SITUATION ANALYSIS OF CHILDREN AND WOMEN: SOUTHERN
NATIONS, NATIONALITIES,AND PEOPLE. 2019.
124. Ethiopia U. National Situation Analysis of Children and Women in Ethiopia. UNICEF
Research Reports. 2019.
125. Mohammed SH, Larijani B, Esmaillzadeh A. Concurrent anemia and stunting in young
children: prevalence, dietary and non-dietary associated factors. Nutrition journal.
2019;18(1):1-10.
126. Soriano JB, Abajobir AA, Abate KH, Abera SF, Agrawal A, Ahmed MB, et al. Global,
regional, and national deaths, prevalence, disability-adjusted life years, and years lived
with disability for chronic obstructive pulmonary disease and asthma, 1990–2015: a
systematic analysis for the Global Burden of Disease Study 2015. The Lancet
Respiratory Medicine. 2017;5(9):691-706.

60
Appendix 1.
Supportive table 1: The unweighted percent distribution of children under the age of five, as
well as mother and household characteristics, from the Ethiopian Demographic and Health
Surveys of 2005, 2011, 2016, and 2019.
Characteristics 2005(n= 2011(n= 2016(n=8,567) 2019(n=
3,476) 9,013) 4,992)
Number (%) Number (%) Number (%) Number (%)
Sex male 1,764(50.75) 4,597 (51.00) 4,361(50.90) 2,556 ( 51.20)
Female 1,712(49.25) 4,416 (49.00) 4,206(49.10) 2,436(48.80)
Age of child =<2 2,011(57.85) 5,225( 57.97) 5,195 (60.64) 2,989(59.88)
>2 1,465(42.15) 3,788 (42.03) 3,372 (39.36) 2,003(40.12)
Residence urban 432(12.43) 1,407(15.61) 1,553(18.13) 1,142(22.88)
Rural 3,044( 87.57) 7,606 (84.39) 7,014(81.87) 3,850( 77.12)
Region Tigray 399(11.48) 1,017(11.28) 898(10.48) 425(8.51)
Affar 171(4.92) 850(9.43) 807(9.42) 559(11.20)
Amhara 463(13.32) 961(10.66) 853(9.96) 458(9.17)
Oromiya 727(20.91) 1,438(15.95) 1,335(15.58) 628(12.58)
Somali 191(5.49) 698(7.74) 1,135(13.25) 537(10.76)
Benishangul- 274(7.88) 786(8.72) 713(8.32) 448(8.97)
gumuz
SNNP 653(18.79) 1,310(14.53) 1,076(12.56) 589(11.80)
Gambela 162(4.66) 644(7.15) 550(6.42) 377(7.55)
Harari 165(4.75) 458(5.08) 442(5.16) 375(7.51)
Addis Ababa 125(3.60) 300(3.33) 381(4.45) 254(5.09)
Dire Dawa 146(4.20) 551(6.11) 377(4.40) 342(6.85)
Respondents No education 2,656(76.41) 6,297(69.87) 5,484(64.01) 2,756(55.21)
highest
primary 577(16.60) 2,301(25.53) 2,204(25.73) 1,552(31.09)
educational
secondary 216(6.21) 282(3.13) 575(6.71) 422(8.45)
level
Higher 27(0.78) 133(1.48) 304(3.55) 262(5.25)
Water source Improved 3,044(87.57) 4,523 (50.18) 5,156(60.18) 3,105(62.20)

61
 Unimproved 432(12.43) 4,490(49.82) 3,411(39.82) 1,887(37.80)
Type of Open 2,278(65.54) 4,635(51.43) 3,738(43.63) 2,074 (41.55)
toilet defecation
facility Unimproved 862(24.80) 2,973 (32.99) 3,382(39.48) 1,885(37.76)
Improved 332( 9.55) 1,393(15.46) 1,412(16.48) 1,018(20.39)
Others 4 (0.12) 12(0.13) 35(0.41) 15 (0.30)
Early Yes 2,403(70.04) 5,010 (55.59) 6,047(70.58) 2,707(73.54)
initiation
No 1,028(29.96) 4,003(44.41) 2,520(29.42) 974(26.46)
Duration of ever 1,773(51.01) 4,841( 53.71) 4,499(52.52) 2,679( 53.67)
breastfeeding breastfed,
not currently
never breast 61(1.75) 141(1.56) 329 (3.84) 209(4.19)
fed
still 1,642(47.24) 4,031(44.72) 3,739 (43.64) 2,104(42.15)
breastfeeding
ANC follow No 1,540(66.21) 3,423 (55.40) 1,982(33.27) 895( 25.36)
up Yes 786(33.79) 2,756 (44.60 3,975(66.73) 2,634( 74.64)
)
Wealth index Poorest 876(25.20) 2,806 (31.13) 3,091(36.08) 1,670(33.45)
Poorer 658(18.93) 1,666 (18.48) 1,497(17.47) 882(17.67)
Middle 659(18.96) 1,493 (16.56) 1,237(14.44) 704(14.10)
Richer 626(18.01) 1,486 (16.49) 1,087(12.69) 655(13.12)
Richest 657(18.90) 1,562 (17.33) 1,655(19.32) 1,081(21.65)
Anemia sever 56 (1.61) 106 (1.18) 138 (1.64) -
Moderate 323 (9.3) 562 (6.24) 824 (9.8) -
Mild 673 (19.4) 1,515(16.81) 1,982 (23.51) -
Not anemic 2,424 (69.7) 6,830 (75.78) 5,486 (65.08) -
Mean age of children in 29.95 29.64 28.66 28.96
month

62
Appendix 2:

Supportive table 2: Trends of stunting among under-five children by children characteristics,

2005, 2011, 2016 EDHS and 2019 mini EDHS (Weighted).

Characteristics 2005(n= 2011(n= 2016(n=8,567) 2019(n=

3,476) 9,013) 4,992)
Percentage Percentage Percentage Percentage
Prevalence of stunting 47.35 44 38.09 37
sex male 47.2 46.13 40.65 40.2
Female 46 42.5 35.40 33.6
Age of child =<2 42 38.51 34.05 33.81
>2 55 52.16 44.38 42.53
Residence urban 30 31.5 26.52 25.8
Rural 48 46.2 39.50 41
Region Tigray 41 51 39 49
Affar 41.14 50. 2 41 43.4
Amhara 57.20 52 46 41.32
Oromiya 41.2 41 37 36
Somali 45.2 33 27.1 30.7
Benishangul- 39.7 48.6 42.80 41
gumuz
SNNPR 51.61 44 39 36.2
Gambela 29.3 27.3 23.73 18
Harari 38.6 30 31.62 35
Addis Ababa 18.12 22.26 15 .2 14
Dire Dawa 31.36 36 40.21 25
Mothers No education 49.1 46.6 41.15 42
education
primary 39.8 42 35.35 35.41
secondary 26.59 20.02 21.46 18.68

63
 Higher 15.81 17.62 17.37 17.0
Water source Improved 47.41 42.29 36.61 37.23
Unimproved 46.94 46.13 40.06 37.54
Type of Open 48.28 47.26 43.22 42.50
toilet facility defecation
Unimproved 46.12 44.28 36.56 36.45
Improved 42.85 34.13 27.56 30.74
Others 78.54 71.59 28.14 18.52
Duration of ever 50.59 49.81 41.65 41.54
breastfeeding breastfed,
not currently
never breast 39.52 36.83 42.94 38.88
fed
still 44.44 38.72 34.09 32.59
breastfeeding
ANC follow No 47.56 46.14 34.19 35.50
up Yes 40.80 36.56 39.36 34.50
Place of Home 48.56 45.91 40.84 40.16
delivery
Health 24.10 29.16 30.54 34.80
institution
Others 36.79 61.06 41.65 16.2
Wealth index Poorest 48 49 43.96 42
Poorer 54.0 47 42.85 42
Middle 46 46.34 37.42 40
Richer 46 45 35.03 35
Richest 35 29 25.57 23.6

64
Appendix 3:

Supportive table 3: Intra-class correlation coefficient(ICC), Akaike's information criterion(126)

, Bayesian information criterion(BIC) and Deviance information criteria (DIC) of for the four
models in each survey year of analysis,Ethiopia,2022
Years Criteria’s Models Remark
of
Model-I Model-II Model-III Model-IV
surveys
2005 ICC 9.8% 6.2 6.3 5.3 The Model-
IV has been
AIC 6112.233 3956.289 5995.82 3947.856
chosen for
BIC 6099.43 4178.513 6079.043 4236.145 interpretation

DIC 6095.4 3882.3 5969.8 3851

2011 ICC 4.1% 4.1% 3.2% 2.6% The Model-
IV has been
AIC 14298.81 9062.682 14073.36 8986.356
chosen for
BIC 14313.31 9344.752 14167.59 9344.103 interpretation

DIC 14294.8 8980.7 14047.4 8882.4

2016 ICC 6.5 3.6% 3.7% 3.3% The Model-
IV has been
AIC 11770.51 7430.823 11546.21 7429.339
chosen for
BIC 11784.72 7659.858 11638.53 7732.474 interpretation

DIC 11766.5 7362.8 11520.2 7339.4

2019 ICC 12% 7% 6.8% 2.7% The Model-
IV has been
AIC 6555.229 2946.049 6471.117 2913
chosen for
BIC 6568.31 3098.182 6549.603 3129.497 interpretation

DIC 6551.2 2894.05 6447.2 2839

65
 Declaration form
I, the undersigned, declare that this is my original work, that it has never been presented
before at this or any other university, and that all resources and materials used in the
research have been fully recognized.

Principal investigator
Name: Getenet Dessie
Signature: _____________
Date: _________________

Advisors
Advisor 1
Name: ________________________
Signature: _______________________
Date: ____________________________

Advisor 2
Name: _________________________
Signature: _______________________
Date: _____________________________

66
FINAL APPROVAL OF RESEARCH THESIS REPORT SHEET

BAHIRDAR UNIVERSITY,

COLLEGE OF MEDICINE AND HEALTH SCIENCES,

SCHOOL OF PUBLIC HEALTH,

DEPARTMENT OF EPIDEMIOLOGY AND BIOSTATISTICS

I hereby certify that I have examined this thesis report entitled ―Changes in the prevalence rate of
stunning, factors associated with stunting and changes in the prevalence rate among under five children
in Ethiopia using Ethiopian demographic health survey data‖, reported By Getenet Dessie. We
recommend and approved the thesis report for a degree of ―Master of Public Health in
Epidemiology‖.

Board of Examiners

External examiner

Name _________________________________________Signature _________Date ________

Internal examiner

Mr. Getachew Hailu (Assistant Professor in Epidemiology) Signature _________ Date ________

Department head

Name ________________________ sig ____________ date _______________

67