Implementing Personalized Medicine: Estimation of Optimal Dynamic Treatment Regimes
A dynamic treatment regime is a list of sequential decision rules that formalizes clinical practice.  Each rule corresponds to a key treatment decision point in the course of patient’s disease or disorder at which clinicians would select a treatment for the patient from among the available options based on information on the patient to that point.  Given data from a clinical trial or observational study, the statistical problem is to estimate the optimal dynamic treatment regime, that leading to the most favorable outcome if followed by all patients in the population.  Approaches to this problem may be formulated based on the potential outcomes framework described in Marie Davidian’s talk, which immediately precedes this one.  We focus first in detail on the case of a single treatment decision and discuss various methods for estimation or the optimal regime, including an approach based on recasting this problem from a classification perspective.  Extension to multiple decision points is then discussed.  This work is joint with Baqun Zhang, Phillip Schulte, Eric B. Laber, and Marie Davidian.

Bio: Anastasios (Butch) Tsiatis is Gertrude M. Cox Distinguished Professor in the Department of Statistics at North Carolina State University and Adjunct Professor of Biostatistics and Bioinformatics at Duke University.  He is a Fellow of the American Statistical Association and the Institute of Mathematical Statistics, and is an elected member of the International Statistical Institute.   He is a leading expert in clinical trials, survival analysis, causal inference, dynamic treatment regimes, group sequential trials, and semiparametric inference, and he is author of the 2006 book “Semiparametric Theory and Missing Data” on the latter topic.  His methodological research has been funded continuously on grants from the National Institutes of Health since 1983, and he is recipient of a (Method to Extend Research In Time (MERIT) award from National Institute of Allergy and Infectious Diseases.   Butch has served as Co-Editor of the journal Biostatistics since 2009.