Trend and Determinants of Infant Mortality in East Africa

No Thumbnail Available



Journal Title

Journal ISSN

Volume Title


Addis Ababa University


An understanding the risk factors related to infant mortality is to guide the development of focused and evidence-based health interventions to reduce infant mortality. This study aimed to see trend and identify detrminants of infant mortality in East African countries by Demographic Health Surveys. To achieve these objectives, multilevel logistic regres- sion model has been used. The results show that, overall infant mortality was decreasing over-time in East Africa. The average percentage change of infant mortality is signi_- cantly decrease by 49.3% in east Africa. Among the East African countries the lowest and highest percentage change of infant mortality observed in Rwanda and Mozambique, respectively. Multilevel analyses shows that child related factors: birth type, birth size, birth order, sex of child, anemia level, breastfed and factors related to mother: mother education, age of mother at _rst birth, religion, contraceptive use, marital status and _nally, factors related to household: family size, residence were found to be statistically signi_cant predictors of infant mortality in East Africa. Survey year's has also statis- tically signi_cant e_ect on infant mortality in East African countries. Infant mortality is high in all East African countries in which household members are four and more, also birth type is statistically signi_cant predictor of infant mortality in all East African countries. In order to reduce infant mortality in East Africa, awareness creation e_orts have to increase family planning, contraception, improve the education level of parents, encourage breastfeeding, providing availability of toilet and safe water. To this e_ect, the outcome of reducing infant mortality is achieved by considering those signi_cant factors related to infant birth.



Infant Mortality, Trend, Random Intercept Model, Random Slope Model, Intraclass Crrelation Coe_Cient, Marginal Quasi Likelihood(Mql), Penalized Quasi Like- lihood(PQL)