SPATIAL MODELLING OF UNDER-FIVE MORTALITY AND ASSOCIATED RISK FACTORS IN ETHIOPIA ADMINISTRATIVE ZONES

Aweke Mitku, University of Kwazulu-Natal
Endeshaw Amuka, Bahir Dar University
Melkamu Zeru, Bahir Dar University

Under-five mortality rate is one of the benchmarks towards the Millennium Development Goals' (MDGs') goal of reducing child mortality. Thus, this study aimed to identify the risk factors by incorporating spatial variation of under-five children among Ethiopian administrative zones. This study used the Ethiopian Demographic and Health survey data, which were collected using a two-stage sampling method. The spatial regression model was employed to account for spatial effect in 65 administrative zones in Ethiopia. The general nesting spatial and spatial Durbin error regression model was a good fit for the data with the smallest AIC and BIC compared with other spatial models. The estimate of the spatial covariates, ?, and ? were -0.1039* and 0.1060*, which is statistically significant(p<=0.05). According to the model results, the increase in the percentage of children who were not vaccinated(0.1680*, p<=0.05), and unemployment status of women(0.7034**, p<=0.01), women who use media exposure(-0.9284*, p<=0.05), Children not currently breastfeeding(0.6654*, p<=0.05), and ANC visit at least 4 times(-0.7962*, p<=0.05) contributed to the statistically significant effect on the under-five mortality in Ethiopia administrative zones. The under-five mortality risk factors are spatially differ across 65 administrative zones in Ethiopia. There exists spatial variation in under-five mortality among in zones, which are affected by spatial neighborhoods of the zones.

Keywords: Mortality and Longevity, Spatial Demography, Geo-referenced/geo-coded data

See extended abstract.