A Bayesian Model for Estimating under-5 Mortality Schedules in Small Areas

Jose Henrique Monteiro da Silva , University of Pennsylvania

The probability of dying under age 5, also known as the Under-5 Mortality Rate (U5MR) is a key indicator of a population’s health and well-being. Recently, scholars have moved forward in understanding under-5 mortality by trying to model it in more depth, by addressing the detailed age schedule of under-5 mortality instead of focusing on the single U5MR measure. Previous works show that under-5 mortality age schedule has a regular shape, similarly to the overall mortality age schedule. Hence, this regularity favors the use of relational models to estimate mortality patterns in a data-sparse context (e.g., small areas). As more and more countries experience a decline in under-5 mortality, the stochasticity of deaths might incur in very irregular and sometimes unreasonable age-specific morality curves. Therefore, new methods are needed to address these issues and measure the true underlying mortality curve. By taking advantage of the regularity of the mortality schedule below age 5, in this work I propose a methodology for estimating under-5 mortality curves in small areas or data-scarce contexts. I extend a previous Bayesian strategy used for modeling the overall mortality curve in small areas to 22 age groups below age 5.

See paper

 Presented in Session 52. Flash Session Advances in Subnational and Small-Area Population Analysis