Competing risks analysis of under-five child mortality in Bangladesh

Abstract

This study implements an analysis of under five mortality to the Bangladesh Demographic and Health Survey, 2011 data from a competing risks perspective. Kaplan - Meier overall survival curves along with log-rank test p-values are employed to determine the prospective covariates for the Cox proportional hazard (PH) model. Later the typical Cox PH model is used to model the causespecific hazard for the two competing causes, namely disease and non-disease on the selected covariates. It is revealed that mothers primary and secondary education, birth order of the index child in case of disease model; mothers higher education, wealth index of the index child in case of non-disease model and mothers aged greater than 30 years at the time of giving birth, home as a place of child delivery in both model played significant role in decreasing under five child mortality.

Bangladesh J. Sci. Res. 27(1): 27-38, June-2014