Date of Award

2019

Embargo Period

8-1-2024

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Public Health Sciences

College

College of Graduate Studies

First Advisor

Andrew B. Lawson

Second Advisor

John Pearce

Third Advisor

Charlton Strange

Fourth Advisor

Brian Neelon

Fifth Advisor

Bo Cai

Sixth Advisor

Lucas Neas

Abstract

Background: Air pollution is associated with adverse health outcomes ranging from increased respiratory incidence to increased mortality; however, the health impacts from exposure to multiple pollutants remain unclear. Large gaps in knowledge remain for developing flexible models that address the decomposition of chemical mixtures in relation to health outcomes. In particular, application of complex fusion models, which combine observed and modeled data, to areas with sparse monitoring networks with multiple chemical components is under-developed. Objective: The overarching objective of this proposal is to improve health effects studies of air pollution by improving predictive capabilities of multipollutant exposure characterizations across space and time. Approach: This project focuses on the development of methods for improved estimation of pollutant concentrations when only sparse monitor networks are found. (Aim 1) Particularly, a multivariate air pollutant statistical model to predict spatiotemporally resolved concentration fields for multiple pollutants is developed and evaluated. (Aim 2) Following on from that work a Bayesian latent grouping model for chemical mixtures is proposed and tested on existing air quality data and health registry data. Simulated evaluation with known mixtures classes and their health effects is also proposed. (Aim 3) Finally we propose to apply the developed methods to the analysis of COPD exacerbations resulting in hospitalization in South Carolina. We will also develop an implementation of the software from this work to allow greater public access to the methodology derived. Impact: These methods utilize only widely available data resources, meaning that the improved predictive accuracy of sparsely monitored pollutant concentrations can benefit future studies by improving estimation of health effects and saving resources needed for supplemental monitoring campaigns. Further, the characterization of air pollution as mixtures provides a more realistic understanding of the health impact of ambient air quality. Finally, the case study demonstrates the feasibility and ease of application of our methods (by using the software we develop) for future researchers, providing an example to be replicated and/or emulated by others.

Rights

All rights reserved. Copyright is held by the author.

Download

Share

COinS