EcoSta 2022: Start Registration
View Submission - EcoSta2022
A0740
Title: Transporting randomized trial results to estimate counterfactual survival functions in target populations Authors:  Youngjoo Cho - Konkuk University (Korea, South) [presenting]
Zhiqiang Cao - Shenzhen Technology University (China)
Fan Li - Yale University (United States)
Abstract: Generalizability and transportability have been studied extensively for uncensored data. However, scarce literature focuses on survival data with censoring. Motivated by controversial results from two clinical trials of blood pressure, we study the transportability of survival outcome findings from randomized clinical trials to an external target population. Based on four assumptions, we propose inverse probability weighting estimators and doubly robust estimators and show that when both the sampling score model and censoring model are correctly specified, the proposed estimators are consistent. Furthermore, the doubly robust estimators are still consistent if the survival time model is correct no matter sampling score model and censoring model are misspecified or not. We derive the influence functions of the proposed estimators and conduct simulation studies to examine their finite-sample performances. We finally apply our proposed estimators to assess the transportability of survival difference between treatment and control groups found in ACCORD-BP trial to the adults with Diabetes mellitus in the U.S. population.