Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 208 - Personalized and Precision Medicine
Type: Contributed
Date/Time: Tuesday, August 10, 2021 : 1:30 PM to 3:20 PM
Sponsor: Biometrics Section
Abstract #318516
Title: Transportability of Causal Inference Under Probabilistic Dynamic Treatment Regimes for Organ Transplantation
Author(s): Grace R Lyden* and David M Vock
Companies: University of Minnesota School of Public Health and University of Minnesota School of Public Health
Keywords: Causal inference; Transplant; Dynamic treatment regimes; Survival analysis; Inverse probability weights; Bayes theorem
Abstract:

Patients who are eligible for two possible organ transplant strategies with differing expected wait times (e.g. deceased-donor kidney-pancreas vs living-donor kidney) face a difficult decision: Is the strategy with the longer wait time worth the wait? This question can be answered by comparing survival under two dynamic treatment regimes (DTRs). Existing DTR methods fail, however, when the goal is to inform decision-making in patients today, perhaps at a particular center, where the distribution of wait times and organ quality might differ from that of a national, historical database. We introduce the concept of a generalized representative intervention: a random DTR that assigns treatment according to the distribution observed in the target population. We propose a class of consistent, asymptotically normal, weighted survival estimators, which can handle multi-dimensional and possibly continuous time-varying treatments. Estimable weights are derived using Bayes' theorem to handle continuous components (e.g., organ quality) that circumvent the need to model densities. We apply our method to the Scientific Registry of Transplant Recipients database.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2021 program