EAPSI: Statistical Methodology for Estimating Health Effects of Exposures to Complex Metal Mixtures, with a Focus on Identifying Time Windows of Susceptibility to Exposures

Liu Shelley H, Boston MA

Investigators

Abstract

This award supports research aimed at developing a novel statistical methodology to be applied to environmental health research. The statistical approach will investigate how time-varying exposures to complex metal mixtures affects child health. The existing scientific literature shows that exposures to metal mixtures during early life can impact cognitive function. Furthermore, there may be certain developmental time windows during which there is an increased vulnerability to metal mixture exposures. However, there is a lack of statistical approaches to accurately determine how the health impact of metal mixture exposures depends on exposure timing. Therefore, this project will develop a novel, flexible statistical methodology to address these concerns. The research will be conducted in collaboration with Dr. Matt Wand, Distinguished Professor of Statistics and world leader in variational approximate Bayesian inference, at the University of Technology Sydney in Sydney, Australia. Fast approximation approaches will be implemented in the statistical model to ensure computational efficiency. This novel statistical approach could be used to analyze a variety of environmental health studies. It will help identify the sensitive time windows of exposure to metal mixtures during early life. This information could guide the development of regulatory policies to help the most vulnerable populations. This project will implement variational approximations to a Bayesian hierarchical model. The Bayesian model estimates how the effects of mixture exposures change with the exposure window, and accounts for complex non-linear and non-additive time-varying mixture effects. By developing computationally efficient Bayesian inference in a large-data setting, the method will be scalable to longitudinal studies with large number of subjects and time points. This NSF EAPSI award supports the research of a U.S. graduate student and is funded in collaboration with the Australian Academy of Science.