Learning When-to-Treat Policies
This work addresses dynamic decision-making challenges in fields like healthcare, offering a practical method for policy optimization, though it appears incremental as it builds on existing doubly robust learning techniques.
The paper tackles the problem of learning dynamic treatment rules from observational data, where decisions involve both whom to treat and when to start treatments, such as in medical settings. It introduces an 'advantage doubly robust' estimator that achieves promising empirical performance and provides theoretical regret bounds without needing structural assumptions.
Many applied decision-making problems have a dynamic component: The policymaker needs not only to choose whom to treat, but also when to start which treatment. For example, a medical doctor may choose between postponing treatment (watchful waiting) and prescribing one of several available treatments during the many visits from a patient. We develop an "advantage doubly robust" estimator for learning such dynamic treatment rules using observational data under the assumption of sequential ignorability. We prove welfare regret bounds that generalize results for doubly robust learning in the single-step setting, and show promising empirical performance in several different contexts. Our approach is practical for policy optimization, and does not need any structural (e.g., Markovian) assumptions.