|
Activity Number:
|
293
- Causality for Complex Data
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 11, 2021 : 3:30 PM to 5:20 PM
|
|
Sponsor:
|
IMS
|
|
Abstract #315541
|
|
|
Title:
|
Assessing Causal Inference Using Adaptively Collected Data
|
|
Author(s):
|
Susan Murphy*
|
|
Companies:
|
Harvard University
|
|
Keywords:
|
causal inference;
sequential decision making;
reinforcement learning;
digital health
|
|
Abstract:
|
Increasingly treatments are being delivered by online stochastic algorithms that take into account prior responses by the individual or similar individuals. Since these algorithms result in sequentially randomized treatments with known probabilities, one might think that the analysis of the treatment effects would be straightforward. However in some scenarios the algorithms can produce randomization probabilities that do not converge even in large samples. Further reconciling the definition of a causal effect with desired scientific goals can be tricky. This talk will discuss our efforts in combating these challenges in digital health.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2021 program
|
