BIOS Seminar: Statistical Methods for Tobacco Research

David Vock, PhD, assistant professor in the Divison of Biostatistics in the School of Public Health at the University of Minnesota, will present this week’s seminar  titled, “Statistical Methods for Regulatory Tobacco Research.” Dr. Vock’s research focuses on statistical methods development for electronic health data with a particular focus on development of machine learning techniques to handle censored data. He also works on novel methods for causal inference and estimation of dynamic treatment regimes (DTRs), including methods for novel clinical trials (e.g., SMART designs) used to test DTRs. He collaborates on projects related to transplantation, smoking cessation, cardiovascular disease, and infectious disease.

Abstract

In 2009, Congress passed the Family Smoking Prevention and Tobacco Control Act which gave the Food and Drug Administration (FDA) the authority to limit the nicotine content of cigarettes to lower levels if such regulation would be likely to improve public health. The FDA has funded a variety of randomized controlled trials (RCTs) to investigate the effect of nicotine reduction on smoking behavior. However, there are a number of challenges with analyzing and interpreting data from these RCTs including (1) estimating the regulatory effect when noncompliance to randomized treatment assignment is reported with error; (2) understanding the effect in vulnerable subpopulations; and (3) synthesizing results from several medium-sized studies when there is treatment effect heterogeneity. In this talk, we present a suite of novel statistical methods to address these challenges. We apply our methods to data from recent clinical trials of very low nicotine content cigarettes conducted by the Center for the Evaluation of Nicotine Cigarettes (CENIC) at the University of Minnesota.

Joint work with Joseph S. Koopmeiners, Jeffrey A. Boatman, Ross L. Peterson, Ales Kotalik, and Chuyu Deng