|
Activity Number:
|
284
- Learning Individualized/Sub-Group Treatment Rules in Complex Settings
|
|
Type:
|
Topic-Contributed
|
|
Date/Time:
|
Wednesday, August 11, 2021 : 1:30 PM to 3:20 PM
|
|
Sponsor:
|
Health Policy Statistics Section
|
|
Abstract #317535
|
|
|
Title:
|
Personalizing Treatment Assignment Rules in Large Multi-Arm Experimental Settings
|
|
Author(s):
|
Rahul Ladhania* and Lyle Ungar
|
|
Companies:
|
University of Michigan and University of Pennsylvania
|
|
Keywords:
|
optimal assignment rule;
personalized treatments;
welfare maximization;
heterogeneous treatment effects;
causal inference;
machine learning
|
|
Abstract:
|
We learn personalized assignment rules from among many treatment arms from a large randomized controlled trial. We argue that a large number of treatment arms makes finding the best arm hard, while we can still achieve sizable welfare gains from personalization by direct optimization. We then document the performance of a forest-based assignment algorithm in a simulation exercise and apply it to a behavioral mega-study with more than 50 treatment arms, aimed at promoting the formation of lasting exercise habits.
|
Authors who are presenting talks have a * after their name.
