The determinants of childhood diarrhea and acute respiratory infection: testing the importance of community-level determinants in Eswatini.
Date
2020
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Background: It has been noted that regardless of the observed decline in childhood mortality in
developing countries, the mortality rate is still extremely high. Approximately 30 countries in
Africa and Asia need to triple their current standard of reduction of childhood mortality to realise
the sustainable development goal (SDG) number 3 of ensuring healthy lives and wellbeing for all
age groups. The study investigated the following research questions: 1. what is the prevalence of
childhood diarrhoea and ARIs? 2. What are the individual, household and community-level
factors associated with childhood diarrhoea and ARIs? 3. To what extent do community-level
factors explain variations in childhood diarrhoea and ARIs? 4. Which clusters or communities
have a higher burden of childhood diarrhoea and ARIs?
Methodology: The study used a pooled secondary data analysis approach utilising data from the
2010 and 2014 Eswatini Multiple Indicator Cluster Surveys (MICS) that were cross-sectional.
The study utilised 5340 children aged below five years as the unit of analysis. The outcome
variables of the study were child diarrhoea and acute respiratory infection. The single-level
multivariate logistic regression, multilevel multivariate logistic regression and Quantum GIS
were used to answer the study questions. STATA 13 was used for the analyses.
Results: The study found that the magnitude of diarrhoea in Eswatini was 16.2% among underfive
children. In the multilevel model, child age, current height for age, toilet facility region of
residence overall community diarrhoea, were found to be important variables. For instance,
children aged 6-11 months and 12-23 months were 434% and 290% respectively, more likely to
have diarrhoea relative to those aged less than 6 months. Children with a normal and above
normal current height for age had 91% and 93% respectively increased odds of having diarrhoea
compared to those with a growth deficit. Children from households with no toilet facility were
283% more likely to have diarrhoea compared to those from households with a flush toilet.
Residents in the Hohho region were 159% more likely to have diarrhoea compared to children
who resided in the Lubombo region. Children from communities that had medium and high
overall community diarrhoea were 26% and 877% more likely to have diarrhoea compared to
children from communities with low overall community diarrhoea. The results showed a
substantial variation of diarrhoea across communities. The individual-level factors explained
5.6% of diarrhoea across communities; household factors explained 65.3% of the variation while
community-level factors explained 85% of the variation even though it was not significant.
The Mapping analysis revealed that the severity of diarrhoea was most pronounced in the
Manzini region and the Shiselweni region when compared to the Hhohho region. A total of 31%
of the constituencies had a high prevalence of diarrhoea between 18.4-28.1% in the Manzini
region. A consideration of the Shiselweni region revealed that 50% of the communities had a
high prevalence of diarrhoea of between 18.4-28.1%
The study found that the magnitude of ARI in Eswatini was 20.9% among under-five children.
In the multilevel model, maternal age, household wealth index, shared toilet with neighbours,
and region of residence were important factors that explained the variation of ARI across
communities. Individual factors explained about 76.05 % as shown by the PVC; household level
factors explained about 94% of the variance, and community-level factors explained about
93.6% of the variation of child ARI across communities.
The study mapped the prevalence of ARI across communities (clusters) and found that the
Hhohho region had four constituencies with a high prevalence of ARI, Motshane, Mbabane
South, Mbabane East, Nkhaba and Tiphisini. In the Manzini region, five constituencies had a
higher burden of ARI, namely Lamgabhi, Kwaluseni, Ntontozi, Mafutseni and Ludzeludze. In
the Shiswelweni region, the burden of diarrhoea was higher among five constituencies, namely
Zombodze, Mbangweni, Kubuta, Ngudzeni and Sigwe. In the Lubombo region, the burden of
ARI was higher at Lubulini, Hlane, and Mhlume.
Conclusion: The study demonstrated that the prevalence of diarrhoea and ARI are still very high
and a persistent public health problem in Eswatini. The causes of the high magnitude of
diarrhoea and ARI vary by individual, household and community factors. Policies that aim to
ensure reduction in child morbidity from diarrhoea and ARI in Eswatini include strategies and
programmes that rectify characteristics of the community contexts which mainly in the socially
and economically disadvantaged communities and regions of Eswatini.
Description
Doctoral Degree. University of KwaZulu-Natal, Durban.