Association Between PM2.5 and Lung Cancer Incidence Rates (United States 1999-2013 Ecological Study with Generalized Linear Mixed Model)

Nakanishi, Masahiro

Full Metadata

Overview

Year:: 2017
Contributor:: Nakanishi, Masahiro (creator); Sullivan, Adam (Advisor); Zheng, Tongzhang (Advisor); Galárraga, Omar (Reader); Sandstede, Björn (Advisor); Brown University. Department of Applied Mathematics (sponsor)
Genre:: theses
Subject:: PM2.5; Lung Cancer; Multilevel models (Statistics); Generalized Linear Mixed Model
Extent:: viii, 50 p.
DOI: https://doi.org/10.7301/Z0Q23XQ0

Files

Description

Abstract:: Abstract Introduction: Cancer is a leading cause of death worldwide, accounting for 8.8 million deaths in 2015, and among which Lung Cancer consists 1.69 million deaths (19.2%). [WHO] Particulate Matter was classified as IARC Group 1 carcinogens in 2013 by the International Agency for Research on Cancer (IARC), specialized cancer agency of the World Health Organization. [Press Release] The public interests in the potential detrimental health effect of air pollution, especially those casued by Particulate Matters (PM2.5, PM10, etc), are soaring in the recent years. There have been rising number of studies conducted across the globe to report the association between PM2.5 and Lung Cancer Mortality Rates, but very few have successfully studied the association between PM2.5 and Lung Cancer Incidence Rates as suggested by [Vinikoor-Imler et al]. Hence, it is of immense interest to us to assess the association between PM2.5 and Lung Cancer Incidence Rates. Methods: Lung Cancer Incidence Rates, Tobacco, and Healthcare datasets obtained from CDC and PM2.5 dataset obtained from EPA were aggregated into one large longitudinal dataset (panel dataset or cross-sectional time-series dataset) with Lung Cancer Incidence Rates, PM2.5, Tobacco, and Healthcare being the cross-sectional data and Year 1999- 2013 (15 years) being the time-series data. Generalized Linear Mixed Model with Poisson Distribution Log Link was extensively employed in this study to estimate the association between PM2.5 and Lung Cancer Incidence Rates. Results: Through our analysis with Generalized Linear Mixed Model using Poisson Distribution Log Link, we obtained the Relative Risk of 1.183 (95% CI: 1.181, 1.186) as the association between PM2.5 and Lung Cancer Incidence Rates on State level. Our Modeled-Relative Risk’s 95% CI overlapped with 10 out of 14 studies’ 95% CIs quoted in the Systematic Review and Meta Analysis published by Harma et al. After obtaining the fixed and random effects of the covariates on Lung Cancer Incidence Rates for each of the 50 States and D.C., we projected the estimates onto the U.S. map to see if there is any geographical difference. Interestingly enough, even after alleviating the negative correlation between the intercepts and the slopes by centering the predictors, we observed a generally higher effect of PM2.5 on the East Coast and a generally higher effect of Tobacco and Healthcare on the West Coast.
Notes:: Thesis (Sc. M.)--Brown University, 2017

Content

Access Conditions

Rights: In Copyright
Restrictions on Use: Collection is open for research.

Citation

Nakanishi, Masahiro, "Association Between PM2.5 and Lung Cancer Incidence Rates (United States 1999-2013 Ecological Study with Generalized Linear Mixed Model)" (2017). Applied Mathematics Theses and Dissertations. Brown Digital Repository. Brown University Library. https://doi.org/10.7301/Z0Q23XQ0

Relations

Collection:

Applied Mathematics Theses and Dissertations

Theses and Dissertations for the Applied Mathematics department.
...