“From Patients to Policy: Epidemiologic Perspectives on Implementation Science for HIV”
Daniel Westreich, PhD, Assistant Professor, Department of Epidemiology
University North Carolina Chapel Hill
“Targeted Maximum Likelihood in Pharmacoepidemiology”
Robert Platt, PhD, Biostatistics and Occupational Health
McGill University, Montreal Quebec
“Analysis of Complex Data Under Biased Sampling”
Yong Chen, PhD
University of Pennsylvania
"The Effects of Local Police Surges on Crime and Arrests in New York City"
John MacDonald, PhD
Professor of Criminology and Sociology, University of Pennsylvania, School of Arts and Sciences
"Bayesian Methods for Multiple Mediators: Principal Stratification and Causal Mediation Analysis of Power Plant Emission Controls"
Corwin Zigler, PhD Assistant Professor of Biostatistics Department of Biostatistics Harvard T.H. Chan School of Public Health
Abstract: A major feature of air quality regulatory policy in the US is to incentivize the installation of flue-gas-desulfurization scrubbers on power plant smokestacks. One goal of these policies is to reduce emissions that are precursors to the formation of PM2.5 in the atmosphere, which is known to be associated with adverse health outcomes. However, the presumed relationships between scrubbers, emissions, and ambient PM2.5 have never been estimated or empirically verified amid the realities of actual regulatory implementation. The goal of this talk is to develop new statistical methods to quantify these causal relationships. We frame this problem as one of mediation analysis to evaluate the extent to which the causal effect of a scrubber on ambient PM2.5 is mediated through causal effects on power plant emissions. Since power plants emit various pollutants including sulfur dioxide (SO2), nitrous oxides (NOx) and carbon dioxide (CO2), we develop new statistical methods for settings with multiple intermediate mediating factors that are measured contemporaneously, may interact with one another, and may exhibit joint mediating effects. Specifically, we propose new methods leveraging two related frameworks for causal inference in the presence of mediating variables: principal stratification and causal mediation analysis. We define principal effects based on multiple mediators, and also introduce a new decomposition of the total effect of a scrubber on ambient PM2.5 into the natural direct effect and natural indirect effects for all mediating emissions jointly, each pair of emissions, and each emission individually. Both approaches are anchored to the exact same models for the observed data, which we specify with flexible Bayesian nonparametric techniques. We first provide assumptions for identifiability of principal causal effects, then augment these with two additional assumptions required to conduct a genuine mediation analysis relying on natural direct and indirect effects. The principal stratification and causal mediation analyses are interpreted in tandem to provide the first comprehensive empirical investigation of the presumed causal pathways that motivate a variety of air quality control strategies that aim to reduce harmful emissions from power plants.
"Empirical likelihood ratio test for a mean change point model with a linear trend followed by an abrupt change"
Wei Ning, PhD Associate Professor of Statistics Department of Mathematics and Statistics Bowling Green State University
Abstract: In this talk, a change point model with the mean being constant up to some unknown point, and increasing linearly to another unknown point, then dropping back to the original level will be introduced. A nonparametric method based on the empirical likelihood test is proposed to detect and estimate the locations of change points. Under some mild conditions, the asymptotic null distribution of an empirical likelihood ratio test statistic is shown to have the extreme distribution. The consistency of the test is also proved. Simulations of the powers of the test indicate that it performs well under different assumptions of the data distribution. Applications on real data will be given to illustrate the testing procedure.
"GMM Logistic Regression Models with Time-dependent Covariates"
Jeffrey Wilson, PhD Associate Professor of Statistics and Biostatistics Department of Economics W.P. Carey School of Business Arizona State University
Abstract: When analyzing longitudinal data it is essential to account both for the correlation inherent from the repeated measures of the responses as well as the correlation realized on account of the feedback created between the responses at a particular time and the predictors at other times. However, since it is essential to include all the appropriate moment conditions as you solve for the regression coefficients, we explore an approach using a generalized method of moments for estimating the coefficients in such data. This approach makes use of all the valid moment conditions necessary with each time-dependent and time-independent covariate. We use continuously updating generalized method of moments (CUGMM) in obtaining estimates. We fit the generalized method of moments (GMM) logistic regression model with time-dependent covariates using SAS MACRO. We used p-values adjusted for multiple correlated tests to determine the appropriate moment conditions for determining the regression coefficients. We illustrate by looking at re-hospitalization data taken from a Medicare database.
"Young Driver Safety: Graduated Driver Licensing, Distracted Driving and Beyond"
Motao Zhu, MD, MS, PhD
Associate Professor, Department of Epidemiology
West Virginia University, School of Public Health
SPECIAL SEMINAR "Identifying the Average Treatment Effect in a Two Threshold Model"
Arthur Lewbel, PhD Barbara A. and Patrick E. Roche Professor of Economics Department of Economics Boston College
Abstract: Assume individuals are treated if a latent variable, containing a continuous instrument, lies between two thresholds. We place no functional form restrictions on the latent errors. Here unconfoundedness does not hold and identification at infinity is not possible. Yet we still show nonparametric point identification of the average treatment effect, and we provide an associated estimator. We apply our model to reinvestigate the inverted-U relationship between competition and innovation, estimating the impact of moderate competitiveness on innovation without the distributional assumptions required by previous analyses. We find no evidence of an inverted-U in US data.
"The spatiotemporal relationship between alcohol outlets and violence before and after privatization: Does it really matter?”
Loni Philip Tabb, PhD Assistant Professor Department of Epidemiology and Biostatistics Drexel University
Abstract: Alcohol-related violence is a well-documented public health concern, where various individual- and community-level factors contribute to this relationship. These factors range from age, race/ethnicity, and income to vacant units, retail space, public transportation, and even risky retailer density. Various geographic scales have been considered, including census block groups, census tracts, zip codes, and neighborhoods. Even though these scales vary in size, their general findings suggest that areas that have more alcohol outlets or greater alcohol outlet density tend to have higher rates of violence. While many studies focus on the cross-sectional nature of this association between alcohol and violence, longitudinal studies capture the complex attributes of both alcohol and violence while allowing for a more causal assessment.
We focused on the impact of a significant policy change at the local level on the relationship between alcohol and violence over time, where alcohol was captured by off-premise and on-premise alcohol outlets and violence was characterized as both non-aggravated and aggravated assaults. In 2012, Initiative 1183 (I-1183) was passed in the state of Washington, which privatized wholesale distribution and retail sales of liquor. Using Bayesian hierarchical models, we were able to measure and map this spatio temporal relationship between alcohol and violence before and after the implementation of I-1183. Our findings suggest that the impact of this policy varies, and characterizing violence by severity in addition to operationalizing alcohol by type of outlet is key.