A Tutorial on Optimal Dynamic Treatment Regimes

Area of Science:

• Biostatistics
• Causal Inference
• Precision Medicine

Background:

• Dynamic Treatment Regimes (DTRs) are adaptive strategies for sequential treatment decisions.
• Optimizing DTRs is crucial for improving average population outcomes in precision medicine.

Purpose of the Study:

• To provide a systematic and accessible introduction to optimal DTRs.
• To cover the formal definition, causal inference framework, and identification assumptions for DTRs.
• To review statistical models and estimation methods for learning optimal DTRs from data.

Main Methods:

• Formal definition of DTRs within causal inference.
• Discussion of identification assumptions linking causal effects to observed data.
• Overview of statistical models and estimation techniques for optimal DTR learning.

Main Results:

• The tutorial systematically presents the theoretical and practical aspects of optimal DTRs.
• It details methods for learning optimal DTRs from both simulated and real-world data.

Conclusions:

• This tutorial serves as a comprehensive guide for researchers interested in optimal DTRs.
• It equips readers with the foundational knowledge and methods to apply DTRs in precision medicine research.