Determinants of Under-Five Child Mortality in Ethiopia: a Spatial and Multilevel Count Regression Analysis
No Thumbnail Available
Date
2024-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
Introduction:- The under-five mortality rate serves as a significant indicator of child health and overall development within countries, highlighting the quality of life within a given population.
Tragically, millions of children below the age of five lose their lives each year. In Ethiopia, the burden of under-five deaths remains unevenly distributed.
Objective:- This study aims to examine the spatial distribution of under-five mortality in Ethiopia and identify factors impacting number of under-five child deaths using spatial analysis methods and multilevel count regression models.
Materials and Method:- the data was retrieved from the EMDHS 2019, and the total unweighted number of 5753 women were used in this study. Various count models were considered and the multilevel zero inflated negative binomial model happened to be the best model to fit under-five child mortality data. IRR (incidence rate ratio) has been used for interpretations. Hotspot and cold spot, as well as kriging interpolation analysis were done for describing the spatial variation of under-five child mortality.
Result:- Spatial patterns in the number of under-five deaths were observed. The global Moran’s I value (I = 0.416267, z score =5.410983, P value < 0.001) revealed a substantial group of under-five child mortality throughout the country. Under-five death had a statistically significant regional clustering pattern among Ethiopian children. The multilevel zero inflated negative binomial model fit results revealed that the expected number of under-five child mortality in female headed households is 1.27 times higher than the expected number of under-five deaths in male headed households (IRR=1.27). Also, the expected number of under-five deaths in rural areas is 0.6356 times the expected number of under-five deaths in urban areas. Likewise, when water is accessible on the premises, the number of under-five deaths is 0.05779 times lower compared to where obtaining water takes more than 30 minutes. And the expected number of under-five deaths when the mother uses contraceptive methods decreased by a factor of 0.79 (IRR=0.79) compared to the expected number of under-five deaths when the mother is not using contraception. Moreover, the expected number of under-five deaths for a birth order of 2nd and 3rd, 4th and 5th, and 6th and above is 2.10, 4.0812 and 9.73 times the expected number of under-five deaths for a first birth respectively. Moreover, multiple birth increases the incidence rate of under-five mortality (IRR=1.5595).
The spatial analysis verified hot and cold spot areas of under-five mortality among under-five children in Ethiopia
Conclusion:- The spatial analysis verified hot(high risky) and cold (less risky) spot of under-five mortality among under-five children in Ethiopia. The zero inflated negative binomial model revealed that under-five deaths per mother differ in age at first birth time, current contraceptive use, birth order, and type of birth. Place of residence, time to get water, household size, sex of household head and age of household head are statistically significant factors. Policy makers should develop targeted policies that consider the specific challenges and needs of different regions to address geographical variations in under-five mortality.
Description
Keywords
Under-Five Mortality, Count Models, Spatial Analysis, Multilevel Analysis