Indirect Estimation of the Neonatal Mortality: A Model Life Table Approach Using Prior Distributions

Julio E. Romero Prieto , London School of Hygiene and Tropical Medicine (LSHTM)
Andrea Verhulst, University of Pennsylvania
Michel Guillot, University of Pennsylvania

This paper proposes a method for the indirect estimation of the neonatal mortality. The model was estimated from the most reliable records of the Global Age Patterns of Under-5 Mortality database, a newly collected database of national distributions of deaths by detailed age and calendar year. It is a general model able to evaluate and correct both vital record and survey estimates of mortality. The model predicts the mortality rates at early ages, by weeks and months of life, for a given combination of parameters related to the level and pattern of mortality. Hence, the indirect estimation of the neonatal mortality is to find the optimal combination of parameters, matching or fitting the model to the most reliable probabilities of dying observed at post-neonatal ages. Results shows that neonatal mortality can be estimated from post-neonatal mortality and improves significantly if a prior distribution of the level of mortality is available.

See extended abstract

 Presented in Session 30. Life Tables and Applications