Skip to Main Content

An official website of the United States government

Principal Investigator
Xia Junwen
Renmin University of China
Position Title
About this CDAS Project
PLCO (Learn more about this study)
Project ID
Initial CDAS Request Approval
Sep 6, 2022
An instrumental-variable method on estimating optimal treatment regime by maximizing t-year survival function
In the past decades, the literature on optimal treatment regime growed fast in the context of randomized experiment or observational study with the foundational assumption of no unmeasured confounding. However, The assumption sometimes is too restrict to be applied for the noncompliance in randomized experiment or impossible ability to collect all confounders in observational study. The violence of the assumption of no unmeasured confouding would necessarily yield bias on estimation of optimal treatment regime and further wrong assignment of clinical treatment. To tackle the problem it brings, an ubiquitous method in the economy and epidemiology called Instrumental Variable (IV) was introduced to the content of optimal treatment regime by (Cui and Tchetgen Tchetgen, 2021). In clinical obser- vational study, censoring data is common and survival function is of the interest. Moreover, the proposed identification estimotor in (Cui and Tchetgen Tchetgen, 2021) cannot be directly applied under the circumstance of survival time. Thus, in this manuscript, we advance an IV based Kaplan-Meier estimator and further an estimation of optimal treatment regime. As far as we know, it is the first time a nonparametric survival time function estimator is proposed within the IV framework.

1. Providing a new IV based Kaplan-Meier estimator
2. Estimating the optimal treatment regime by maximizing our new estimator
3. Applying our new method to the PLCO Data


Zhang Jingxiao, Renmin University of China;
Zhan Zishu, Renmin University of China.

Related Publications