This special topics course was last offered in Spring 2023.

Course instructor: Marie Davidian, davidian@ncsu.edu

Course meetings: TBA

Tentative guest lecturer: Shannon Holloway (developer of the R package DynTxRegime)

Course Syllabus

Course objective: This course will provide a comprehensive introduction to methodology for data-based development and evaluation of dynamic treatment regimes. A dynamic treatment regime is a set of sequential decision rules, each corresponding to a key point in a disease or disorder process at which a decision on the next treatment action must be made. Each rule takes patient information to that point as input and returns the treatment s/he should receive from among the available options, thus tailoring treatment decisions to a patient’s individual characteristics. Dynamic treatment regimes formalize how clinicians make decisions in practice by synthesizing evolving information on a patient and are thus of considerable importance in precision medicine. Dynamic treatment regimes are also relevant in other contexts in which sequential decisions on interventions or policies must be made, as in education, engineering, economics and finance, and resource management. Of critical importance is the notion of an optimal treatment regime, one that, if used to select treatments for the patient population, would lead to the most beneficial outcome on average.

Methods for estimation of dynamic treatment regimes and in particular optimal treatment regimes from data will be motivated and developed through a formal time-dependent causal inference framework. The gold standard study design for developing and evaluating regimes is the sequential multiple assignment randomized trial (SMART), considerations for which will be discussed. Inference for optimal treatment regimes is a nonstandard statistical problem and is thus notoriously difficult; an introduction to this challenge will be presented. Examples throughout the course will be drawn from cancer and other chronic disease research and research in the behavioral, educational, and other sciences.

Use of the comprehensive R package DynTxRegime to implement many of the methods discussed in the lectures will be introduced.

Students completing this course will have a foundation in causal inference and fundamental results and methods for dynamic treatment regimes that will provide the basis for study of the rapidly evolving literature on dynamic treatment regimes and data-driven sequential decision-making in precision medicine.

Course text: Lecture notes prepared by the instructor. The notes incorporate content from the book
Dynamic Treatment Regimes: Statistical Methods for Precision Medicine by Tsiatis, Davidian, Holloway, and Laber, which will serve as an optional reference.

Course prerequisites: ST 702, Statistical Theory II, and ST 705, Linear Models and Variance Components (or equivalents); programming skills