Department of Biostatistics
Environmental Statistics Seminar

2013 - 2014

Coordinators: Dr. Matthew Cefalu and Dr. Francesca Dominici

Schedule: Fridays, 12:30-2:00 p.m.
HSPH2, Room 426 (unless otherwise notified)

Contract All | Expand All
Seminar Description
This seminar focuses on statistical issues related to assessing environmental effects on human health and analyzing environmental data in general. Specific areas of interest include air pollution epidemiology, exposure assessment, teratology, fertility and reproduction, respiratory studies, and community-based research as well as general topics such as errors-in-variables models, missing data methods, hierarchical modeling, smoothing, and methods for correlated data such as longitudinal and spatial data analysis. The seminars are generally pitched at a level that encourages student participation. Students interested in receiving credit for attending the seminars may sign up with individual faculty members for some guided readings on a special topic. Please see Brent Coull for details.


September 20

Jack Cackler
Doctoral Student, Department of Biostatistics, Harvard University

"Spatiotemporal Methods for Disaster Risk Prediction and Management"
ABSTRACT: This paper focuses on two novel methods to increase inferential capability of health hazards posed by wildfires. The first is a Bayesian algorithm associating increases in particulate matter, which increases after wildfires, with increases in hospitalizations in counties in the Western United States. The second is an image smoothing algorithm that interpolates post-disaster low-quality images with pre-disaster high quality images to yield a fast an accurate post-disaster assessment of affected areas with limited resources. In conjunction, these two tools can greatly increase the capability to understand how wildfires affect the population and where they occur.

September 27

Mark Meyer
Doctoral Student, Department of Biostatistics, Harvard University

"Function-on-Function Mixed Models via Bayesian Wavelet Regression"
ABSTRACT: Medical and public health research increasingly involves the collection of more and more complex and high dimensional data. In particular, functional data---where the unit of observation is a curve or set of curves that are finely sampled over a grid---is frequently obtained. Moreover, researchers often sample multiple curves per person resulting in repeated functional measures. A common question is how to analyze the relationship between two functional variables. We propose a general function-on-function regression model for repeatedly sampled functional data. Both one and two sample settings are presented along with a Bayesian inference procedure. We will examine these models via simulation and a data analysis. Motivating data come from a neurological study examining how the brain processes various types of images. Subjects were taken from a pre-screening for a smoking cessation trial. Event-related potentials were then measured after presentation of neutral, positive, negative, and cigarette-related images. Our analysis will focus on the relationship between measurements from pairs of sensors during neutral and cigarette-related image presentation.

October 11

Matthew Cefalu, Ph.D.
Research Fellow, Department of Biostatistics, Harvard School of Public Health

"A Bayesian Analysis of Marginal Structural Models Estimating the Health Effects of Regulations under the Clean Air Act"
ABSTRACT: The US Environmental Protection Agency (EPA) devotes billions of dollars to regulatory processes designed to manage air quality in the US. However, there remains uncertainty surrounding the causal impact of specific regulatory decisions implemented by the EPA on both air pollution and public health. Focusing in on the non-attainment designation under the Clean Air Act, we propose a novel Bayesian approach to estimate parameters from a marginal structural model (MSM) relating non-attainment designation to reductions in air pollution. Our proposed Bayesian procedure for estimating causal effects with MSMs offers practical advantages over standard procedures that rely on inverse probability of treatment weights, including direct quantification of uncertainty, finite-sample performance when there is limited overlap in the propensity score distribution between treated and untreated observations, and automatic balance diagnostics. We present preliminary results indicating a small reduction in coarse particular matter can be attributed to the non-attainment designation under the Clean Air Act.

November 1

Laura F. White, Ph.D.
Associate Professor, Department of Biostatistics, Boston University
Adjunct Associate Professor, Department of Biostatistics, Harvard School of Public Health

"Assessing the Relationship Between Ambient Air Pollution and Diabetes and Hypertension in the Black Women's Health Study"
ABSTRACT: The Black Women's Health Study (BWHS) is an ongoing, prospective follow-up study of African American women in the United States that started in 1995, with approximately 59,000 African-American women. This cohort has been used to study numerous health issues. This talk will present current work to evaluate the relationship between ambient air pollution, particularly PM 2.5 and ozone levels on diabetes and hypertension incidence among women in the cohort. I will briefly describe measures of the pollutants that we are using in this analysis. I describe some of the challenging issues we have faced, such as spatial heterogeneity. We have used multiple modeling approaches to attempt to best characterize and understand the relationship between these pollutants and diabetes and hypertension. These include a random effects Cox modeling approach, a meta-analytic approach to Cox modeling to account for spatial heterogeneity, and traditional Cox modeling across the entire cohort.

November 22

Michael Greenstone, Ph.D.
Professor of Environmental Economics, Department of Economics, Massachusetts Institute of Technology and Director of the Hamilton Project

"Evidence on the Impact of Sustained Exposure to Air Pollution on Life Expectancy from China's Huai River Policy "
ABSTRACT: This paper's findings suggest that an arbitrary Chinese policy that greatly increases total suspended particulates (TSPs) air pollution is causing the 500 million residents of Northern China to lose more than 2.5 billion life years of life expectancy. The quasi-experimental empirical approach is based on China's Huai River policy, which provided free winter heating via the provision of coal for boilers in cities north of the Huai River but denied heat to the south. Using a regression discontinuity design based on distance from the Huai River, we find that ambient concentrations of TSPs are about 184 g/m3 [95% confidence interval (CI): 61, 307] or 55% higher in the north. Further, the results indicate that life expectancies are about 5.5 y (95% CI: 0.8, 10.2) lower in the north owing to an increased incidence of cardiorespiratory mortality. More generally, the analysis suggests that long-term exposure to an additional 100 g/m3 of TSPs is associated with a reduction in life expectancy at birth of about 3.0 y (95% CI: 0.4, 5.6).

December 6

Forrest Crawford, Ph.D.
Assistant Professor, Department of Biostatistics, Yale School of Public Health, and the Department of Ecology and Evolutionary Biology, Yale University

"Markov Models for Dependent Binary Responses in Epidemiology"
ABSTRACT: Epidemiologists often attempt to determine whether a disease runs in families or has a heritable/genetic/household component, using only observational or administrative datasets. Dependency of responses within households complicates statistical analysis, and traditional methods can produce results whose epidemiological interpretation is unclear. We propose a class of Markov counting processes for analyzing correlated binary data and establish a correspondence between these models and sums of dependent Bernoulli random variables. Our approach generalizes previous models for correlated outcomes, admits easily interpretable parameterizations, and incorporates ascertainment bias in a natural way. This allows simultaneous estimation of disease risk due to household and extra-household sources. We show how to incorporate covariates in a regression setting and apply our method to problems in respiratory disease epidemiology.

February 7

Guojun He, Ph.D.
Research Fellow, China Initiative and Department of Global Health and Population, Harvard School of Public Health

"The Effect of Air Pollution on Cardiovascular Mortality: Evidence from the 2008 Beijing Olympic Games"
ABSTRACT: Exogenous air pollution variations induced by the 2008 Beijing Olympic Games provide a natural experiment to estimate the health effects of air pollution. This study finds that air pollution has a significant effect on cardiovascular mortality in China. A 10 ug/m^3 (10 percent) decrease in PM10 mean concentrations decreases monthly cardiovascular mortality by 13.6 percent, implying that more than 67,000 premature cardiovascular deaths could be avoided each year by a 10 percent reduction in PM10 concentrations. The estimates are robust to a variety of specifications.

April 4

Christine Choirat, Ph.D.
Research Associate, Institute for Quantitative Social Science (IQSS), Harvard University
Associate Professor of Quantitative Methods, School of Economics and Business Administration, University of Navarra, Spain

"Accountability of the Health Benefits of Air Pollution Regulatory Policies: An Exploratory Analysis of the Acid Rain Program Data"
ABSTRACT: The US Environmental Protection Agency (EPA) devotes resources on the order of billions of dollars to scientific and regulatory processes designed to manage air quality in the US. However, despite evident link between exposure to particulate air pollution and health, uncertainty regarding the public-health impact of specific regulatory decisions remains.

The regulatory environment surrounding air pollution control policies demands a new type of epidemiological evidence. Whereas the field of air pollution epidemiology has historically informed policies with estimates of exposure-response relationships between pollution exposure and health outcomes, these estimates alone cannot support current debates surrounding the actual health impacts of past and current air quality control strategies. We argue that directly evaluating specific control strategies is a distinct endeavor from estimating exposure-response relationships, and that increased emphasis on analyses designed to estimate causal effects of well-defined regulatory interventions could provide a new dimension to the epidemiological evidence supporting policy decisions.

This aim of the present work is to conduct accountability assessment of actions taken to control polluting emissions from power plants across the US. We have compiled an unprecedented data source for conducting accountability assessments of power plant controls undertaken as part of the Acid Rain Program. These data include continuous-time emissions for over 4000 power generating units across approximately 1300 power plants in the US, as well as data on controls implemented (e.g., scrubbers), fuel shifts (e.g., switch from coal to natural gas), and other actions to control emissions. Our ultimate objective is to link these data to ambient pollution data and to Medicare data to conduct accountability assessment.

April 18

Michelle L. Bell, Ph.D.
Professor of Environmental Health, Yale University

"Airborne Particles and Pregnancy Outcomes: Current Evidence and Remaining Challenges"
ABSTRACT: None Given

May 2

To Be Announced


"Talk Title TBD"
ABSTRACT: None Given



Back to HSPH Biostatistics Maintained by the Biostatistics Webmaster
Last Update: April 11, 2014