Spatial Modeling of Child Mortality in Nepal

Brian Chin, University of Pennsylvania
Livia Montana, Harvard University

Nepal has achieved remarkable reduction in child mortality over the last 15 years, yet the rates are still among the highest in the world. About one of every 11 children dies before reaching age five. Using the Nepal Demographic and Health Surveys from 2001 and 2006 and geographic data, this paper describes the spatial pattern of child mortality by modeling child mortality as a function of individual and geographic covariates in a multivariate generalized additive model. The analysis considers household and individual predictors of mortality, as well as contextual indicators of access to health services and the biophysical and built environments. Results show a strong east-west spatial pattern in child mortality which remains after accounting for individual and household level covariates. This study demonstrates a new approach to analyzing child mortality from DHS data. The results can inform policies to address health inequalities and improve access to underutilized health services.