Air Quality and Health Effects: Systematic Investigation
Open Access
- Author:
- You, Cheng
- Graduate Program:
- Statistics
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- November 25, 2018
- Committee Members:
- Dennis K.J. Lin, Dissertation Advisor/Co-Advisor
Dennis K.J. Lin, Committee Chair/Co-Chair
Ephraim M. Hanks, Committee Member
Yanyuan Ma, Committee Member
Fuqing Zhang, Outside Member - Keywords:
- Time Series
Detrending
Smoothing
Spline
Kernel
Regression
Case Crossover
Short-term Change
Long-term Trend
Acute Effect
Chronic Effect
Air Quality
Environmental Epidemiology - Abstract:
- To examine whether contemporary datasets support the cross-association between air quality and human health, a systematic statistical investigation is conducted. Our major contributions are proposing a novel detrending method on time series data to address the statistical instability of cross-association detection, investigating the method properties and selecting the parameters. In addition, two prevailing detection methods, time series regression and case-crossover analysis, are studied and compared side by side to demonstrate their performances and consistency. This thesis is mainly threefold. First, the problem and research objectives are elaborated. Second, the new detrending methodology is proposed along with its properties and parameter selection; the two common detection methods are studied and compared. Third, the general procedures are presented and the conclusions are made along with discussions. It is found that the current spline detrending methods may induce any cross-correlation, from negative to positive, while the proposed time series smoothers provide a consistent solution to identify the correct short-term effects. The proposed methods' properties, parameter selection, and diagnosis can also ensure the new opportunity in studying the cross-association among multiple time series data. In environmental epidemiology, especially the air quality studies, it is often encountered that multiple time series data with a certain long-term trend, including seasonality, cannot be fully adjusted by the observed covariates. The long-term trend is difficult to separate from abnormal short-term signals of interest. The newly proposed detrending method addresses how to estimate the long-term trend in order to recover short-term signals. Our case study demonstrates that the current spline detrending methods can result in significant positive and negative cross-correlations from the same dataset, depending on how the smoothing parameters are chosen. To circumvent this dilemma, three classes of time series smoothers are proposed to detrend time series data. These smoothers do not require fine-tuning of parameters and can be applied to recover short-term signals. The properties of these smoothers are shown with both a case study using a fully crossed factorial design and a simulation study using datasets generated from the original dataset. General guidelines are provided on how to discover short-term signals from time series with a long-term trend. The benefit of this research is that a problem is identified and characteristics of possible solutions are determined. Furthermore, since the London Great Smog of 1952 was estimated to have caused the acute deaths of over 4000 people, scientists have studied the relationship between air quality and human mortality. Currently, the cross-association between air quality and acute deaths is usually taken as evidence for causality. As global air quality has markedly improved since 1952, do contemporary datasets support this view? A large dataset, eight air basins in California, is assembled to examine the possible cross association of ozone and PM2.5 with acute deaths after removing seasonal and weather effects. A regression-corrected, case-crossover analysis for all non-accidental deaths age 75 and older of different causes was conducted. A stepwise time series regression was used to examine three causes of deaths. After seasonal and weather adjustments, there was essentially no predictive power of ozone or PM2.5 for acute deaths. The case-crossover analysis produced odds ratio very close to 1.000, which means no acute effect. The very narrow confidence limits indicated good statistical power. The recent air quality in California was investigated in both time-stratified, symmetric, bi-directional case-crossover and time series regression methods and both give consistent results. There is no statistically significant cross association between either ozone or PM2.5 and acute mortality. In the absence of an association, the causality is in question.