Rob Strawderman is the Donald M Foster MD Distinguished Professor in Biostatistics and Chair of the Department of Biostatistics and Computational Biology at the University of Rochester. His principal research interests include semiparametric methods for missing and censored data and statistical learning methods for risk and outcome prediction; he has numerous other research interests, including inference and variable selection in the areas of dynamic treatment regimes and causal inference in mediation analysis and for recurrent events.
Rob earned his bachelor’s degree in mathematics from Rutgers University (1988); he subsequently earned both a masters (1990) and doctoral (1992) degree in biostatistics from Harvard University. He has held previous tenured faculty appointments at Cornell (2000-2012) and at the University of Michigan (1992-2000). Honors include being named Fellow of both the American Statistical Association (2006) and the Institute of Mathematical Statistics (2012); he was also the recipient of the Distinguished Alumni Award from the Department of Biostatistics at Harvard University (2008). Rob’s significant track record of professional service includes service as an associate editor for the Journal of the American Statistical Association (1997-2017) and the Electronic Journal of Statistics (2007-2013).
Talk: Robust Q-learning for Dynamic Treatment Regimes
Abstract: Q-learning is a regression-based approach that is widely used to formalize the development of an optimal dynamic treatment strategy; however, it is also highly sensitive to the specification of finite-dimensional working models used to estimate ``treatment free’’ nuisance parameters. Misspecification of these working models can lead to serious bias due to residual confounding, and may result in treatment strategies that are sub-optimal. We propose a robust Q-learning approach that allows estimating such nuisance parameters using data-adaptive techniques. Methodology, asymptotics and simulations will be summarized and we highlight the utility of the proposed methods through simulation. Time permitting, data from the ``Extending Treatment Effectiveness of Naltrexone'' multistage randomized trial will be used to illustrate the proposed methods.