A Spatial Survival Model For Risk Factors Of Under-Five Mortality In Kenya

Abstract

Child mortality refers to death of a child between 0 and ve years old. Kenya like many Sub-Saharan countries faces a burden of child mortality, and has made e orts towards reducing the same. The main The main objective of the study is to evaluate the spatial variation in under- ve mortality in Kenya using spatial survival methods. Methods Data on child mortality was collected through the Demographic Health Survey 2014. The survey collected information on children, demographic indicators related to the mother and child, and various social and economic attributes. Intrinsic Conditional Autoregressive Models were tted to account for spatial dependence and clustering to estimate the hazard at county level, together with a cox-proportional hazard model to estimate the risk factors associated with child mortality. Results The spatial cox proportional hazard model was identi ed as the best t based of the Deviance Information Criterion (DIC). There exists a spatial structure on the hazard of death in the Kenyan counties. Counties with the highest hazard of death include counties around central Kenya (Laikipia, Nyandarua, Nyeri, Kiambu, Machakos and Makueni), although most counties have similar hazards. The lowest hazard is found in Western Kenya counties and Nyanza. Sex of the child, sex of household head, age of respondent at rst birth, level of education, and whether a child is in a multiple birth are signi cant risk factors of child mortality. Conclusion This study brings out the spatial disparities that exist in the country on child mortality in Kenya. The speci c counties have mortality rates that are county-speci c, with neighbouring counties having similar hazards for death of a child. is important therefore for interventions to take into consideration the e ect of where a child is born(county) to reduce mortality.