|
Andrea Verhulst , French Institute for Demographic Studies (INED)
José Henrique Monteiro da Silva, University of Pennsylvania
Julio E. Romero Prieto, London School of Hygiene and Tropical Medicine (LSHTM)
Michel Guillot, University of Pennsylvania
In LMICs, incomplete vital registration data are entirely dismissed to estimate under-5 mortality. In this paper, we propose a model for estimating the under-five mortality rate from this type of incomplete information. The model relies on age regularities between age 0 and 5 observed in the Under-Five Mortality Database (U5MD) providing distributions of deaths by detailed age, including daily, weekly, and monthly breakdowns. The premise of this approach is that underregistration does not affect all ages equally. Some ages are more likely to be underreported than others, especially the first months after birth. The strategy consists then in correcting the mortality for very specific age (such as neonatal age) based on the remaining age ranges considered more robust. The goal of this paper is to use the broadest possible set of countries to provide guidelines with global applicability.
Presented in Session 113. Modelling and Exploring Infant Mortality