Training Program in Environmental Health Statistics



Current Fellows (2012-2013)

Jack Cackler is an incoming first year student who will be taking core courses in the Biostatistics Ph.D. program.

Joseph Antonelli, and Patrick Staples are second year students taking core courses in the Biostatistics Ph.D. program. During summer 2012, they participated in data analysis projects in environmental health research, working jointly with faculty in Biostatistics and Environmental Health.

Joseph Antonelli was awarded a grant to study environmental statistics in Cyprus during Winter Session. His summer 2012 research project focused on flexible nonparametric Bayes methods for logistic random effects models. He will take his qualifying exams in January 2013.

Patrick Staples' summer research project was conducted under the supervision of Dr. Christopher Barr, Brent Coull, and Jim Shine on spatial models, particularly Voronoi tessellations, for detection of Superfund hazardous waste site contamination hot spots. This work represents a collaboration within the Harvard Superfund Research Program. Mr. Staples plans to take his qualifying exams in January 2013.

Stacey Ackerman-Alexeeff (dissertation advisor Brent Coull), Elizabeth Smoot (dissertation advisor Sebastien Hanuese) , Tristan Hayeck (academic advisor Brent Coull), and Matt Cefalu (dissertation advisor Francesca Dominici) are third year students.

Stacey Ackerman-Alexeeff has been involved in two research projects to date. The first deals with measurement error in air pollution epidemiology. Ms. Ackerman-Alexeeff is working to formulate measurement error problems as mixtures of Berkson and classical errors and develop new methodology to adjust for bias and variance. The second project focuses on testing for genetic pathway effects. She is working to extend kernel machine regression testing for genetic pathway effects to longitudinal designs and for gene-environment interaction testing. In March 2011, Ms. Ackerman-Alexeeff presented a poster at the International Biometric Society ENAR Conference entitled, "Testing for the Effect of a Genetic Pathway in Longitudinal Data: Kernel Machine Regression". She also made a presentation at the International Biometric Society ENAR Conference entitled, "Methods for Spatially-varying Measurement Error in Air Pollution Epidemiology" that was held in March, 2012 in Washington, DC. She collaborated with Drs. Coull, Dominici, and Schwartz on the submission of a methodologic R01 to NIEHS in July 2011.

Matthew Cefalu's summer project (between his first and second years) focused on the health effects of long-term air pollution. His thesis research will be on developing methods for causal inference in environmental health studies, with an emphasis on Bayesian and semi-parametric methods. He is currently heading an epidemiological study of the health effects of long-term air pollution exposure among the elderly in New England.

Tristan Hayeck is interested in Statistical genetics methods, specifically, working on a liability threshold model for a new family association statistic applied to diseases such diabetes and coronary heart disease. Last summer he worked with Professor Alkes Price on a Liability Threshold Model analyzing Jackson Heart Study for patients with type 2 diabetes and Coronany heart disease. The goal of the analysis is to improve estimation of family association statistics by modeling the pedigrees based on inferences made on genetic data. He developed a C++ model, where he implemented a multivariate Gibbs sampler to model the family interactions.

Elizabeth Smoot is working on building on the hybrid design for combining ecological and case control data proposed by Haneuse and Wakefield. A main obstacle to the applicability of the hybrid design is its large computational burden. One potential solution to this obstacle is to use an alternative decomposition of the likelihood (into case-control and ecological sans case control) and then exploit the independence of case-control and ecological-sans-case control datasets by approximating the ecological portion of the likelihood; a second potential solution she is studying is to use post-stratification methods. Ms. Smoot completed core coursework in biostatistics and coursework for a cognate in lung function and pollution.

Mark Meyer (dissertation advisor Brent Coull) is a fourth year student in the Biostatistics Ph.D. program. His research focuses on developing semi-parametric regression techniques for longitudinal data. He is working on functional models with functional responses and covariates; he intends to extend this work with inferential methods. His projects have focused on Bayesian methods for longitudinal data, incorporating smoothing techniques, and on developing an R package for fitting penalized Bayesian Distributed Lag Models. A second project examines the effects of airplane cabin pressure on heart variability.

Andrew Correia (dissertation advisor Francesca Dominici) is a fifth year student in the Biostatistics Ph.D. program. His research is centered on developing statistical methodology to assess the relationship between PM2.5 and life expectancy in the United States. His most recent work has extended and confirmed the analysis of Pope et al (NEJM, 2009), showing a positive association between reductions in PM2.5 and gains in life expectancy across 545 U.S. counties for the period 2000 - 2007. The effect appears to be greatest in densely populated counties, highly urban counties, and counties with small land area, suggesting the possibility of differential exposure misclassification. In March 2011 Andrew presented a poster at the International Biometric Society ENAR Conference entitled, "The Effect of Air Pollution Control on Life Expectancy in the United States: A population based analysis of major metropolitan areas."

Jennifer Bobb is a postdoctoral research fellow in the Department of Epidemiology working under the direction of Drs. Francesca Dominici and Brent Coull. Dr. Bobb's research focuses on computationally efficient methods for modeling (1) the joint health effects of pollution mixtures, and (2) the uncertainty associated with climate change. Specifically, she is developing Bayesian formulations of kernel regression machines for high-dimensional pollution exposures, while also developing model averaging approaches in climate change settings.