Semiparametric Bayesian Density Estimation with Disparate Data Sources: A Meta-Analysis of Global Childhood Undernutrition

• Activity Number: 583
• Type: Invited
• Date/Time: Wednesday, August 12, 2015 : 2:00 PM to 3:50 PM
• Sponsor: JASA, Applications and Case Studies
• Author(s): Mariel M. Finucane* and Christopher J. Paciorek and Gretchen A. Stevens and Majid Ezzati
• Companies: Mathematica Policy Research and UC Berkeley and World Health Organization and Imperial College London
• Keywords: Complex survey data ; Hierarchical modeling ; Missing data ; Normal mixture model ; Stick-breaking prior

Abstract

Undernutrition, quantified using height-for-age z-scores, is an important contributor to childhood morbidity and mortality. Since all levels of mild, moderate, and severe undernutrition are of clinical and public health importance, it is of interest to estimate the shape of the z-scores' distributions. We present a finite mixture model that uses data on 4.3 million children to make annual estimates of these distributions for children in each of the world's 141 low- and middle-income countries. We incorporate both individual-level data, as well as aggregated summary statistics from studies whose individual-level data was unavailable. We place a hierarchical Bayesian model on the mixture weights, which allows for nonlinear changes in time, and borrows strength in time, in covariates, and within and across regional country clusters to make estimates where data are uncertain or missing. This work addresses three important problems that often arise in the field of global health monitoring. First, data are always incomplete. Second, different data sources commonly use different reporting metrics. Last, distributions, and especially their tails, are often of substantive interest.