Correlates of Spatial Differences in under-five mortality in Nairobi’s informal Settlements

Abstract

Child mortality in Kenya is often associated with individual level factors including socio-economic status, nutritional status, and poor access to health services. Geographical factors are less documented as important predictors of child mortality in the country. Using a Bayesian geo-additive survival model, this paper examines the factors associated with child mortality in two Nairobi slums, Korogocho and Viwandani, accounting for spatial random effects. It uses longitudinal data for the period 2006-2011 involving 30339 children aged below five years from the Nairobi Urban Health and Demographic Surveillance System implemented in the two slums. In addition to determinants such as mother's education and age, size of the household and ethnicity, our findings show a certain spatial structure in child mortality risk, with differences between some villages in Viwandani, while no spatial variations were observed in Korogocho. The results call for specific efforts from policymakers to refine child health interventions in Nairobi's urban slums.