Intra- and Inter-Regional Variation and Risk Factors of Childhood Stunting in Low-Income and Middle-Income Eastern African Countries

Abstract

Stunting in children continues to persist in many developing countries, particularly in sub-Saharan Africa. An analysis of the four countries enables a comparative analysis and opportunities for learning what works as far as tackling child stunting in the Sub-Saharan Africa region. The main objective of the study was to analyse the intra-and inter-regional variations and risk factors associated with childhood stunting in four East African countries (Kenya, Tanzania, Malawi and Zambia). Using the Demographic and Health Survey (DHS) and Multiple Indicator Cluster Survey (MICS), four specific objectives were examined. In the first objective, trend analysis in stunting was carried out in each country, across a selection of variables that were identified based on previous findings on what influences child under- nutrition. Then country-specific multilevel binary logistic models were fitted to estimate factors associated with stunting at child, mother, household, and area levels. In all four countries, there were a decreasing trend with varying regression slopes, between and across inequality variables. Both shared and country-specific risk factors were identified. Shared risk factors included age of child, mother’s education, mother's age, and wealth index. Clustering at community and district levels were also evident. The second objective identified the most effective model and key predictors for stunting prediction using machine learning (ML) techniques. Based on best model, XGBoost, the most important predictors of stunting status differed in the four countries, but included child, maternal, and household factors. The third objective aimed to investigate the context and identify the vulnerabilities that are associated with child stunting. A multilevel latent class analysis approach was used to identify clusters of childhood stunting. We identified three clusters, which reflected poor health in the first class, poor socio-economic conditions and less diverse diets in the second class, and mixed poor health in the third class. The last objective is to estimate direct and indirect effects, while examining the practicability of the UNICEF conceptual model on determinants of child malnutrition. Overall, results show significant indirect effects between maternal education and wealth index, maternal education and breastfeeding, breastfeeding and dietary score. A direct positive significant association between wealth index, and education level on stunting, while there was a direct negative association between occupation, dietary diversity, and breastfeeding on stunting. Indirect effects were established between the wealth index on stunting in Malawi, and of maternal education on stunting in Tanzania and Zambia.
In conclusion, we observe that regardless of the income-level of the country, risk factors of stunting remain similar, while others remain unique to a country. Urban-rural disparities persist in Tanzania and Malawi, while they are narrowing in Kenya and Zambia. On the other hand, the rich-poor differences are increasing in Malawi. Our research has revealed that stunting is a complex issue and several factors contribute to it at different levels, including child, maternal and household levels. It is essential to focus on preventive measures to reduce stunting, and these measures should be tailored to the specific context. Providing essential nutrition, counselling parents on giving priority to children's food, promoting optimal duration of breastfeeding, and preventing diarrheal diseases among children are some of the key preventive measures that need to be emphasized. To effectively tackle stunting, a multi-sectoral approach is required, which involves reducing socioeconomic inequalities between urban and rural areas, removing barriers to women's access to education, improving access to safe water and sanitation services, and promoting diverse diets.

Date of Award	22 Apr 2025
Original language	English
Awarding Institution	University of Portsmouth
Supervisor	Ngianga Ii Kandala (Supervisor) & Sasee Pallikadavath (Supervisor)