Skip to Main Content

COVID-19 is an emerging, rapidly evolving situation.

What people with cancer should know:

Get the latest public health information from CDC:

Get the latest research information from NIH:

Principal Investigator
Zhiyue Huang
Position Title
About this CDAS Project
NLST (Learn more about this study)
Project ID
Initial CDAS Request Approval
Apr 16, 2020
Rank Tests for Time-to-Event Data Following Propensity Score Matching with Replacement
With the increasing accessibility of electric health data and raising costs and recognized limitation of traditional trials, how to use real-world data and observational studies to enhance the efficacy evidence and bridge the gap between trials and practice attracts interests again. To handle confounding, randomization remains as a key tool, while customized statistical methodologies may be applied in real-world data and observational studies. The propensity score matching (PSM) is one of the popular methods to handle confounding in observational studies.

However, the inference on time-to-event data by the PSM with replacement remains unsolved. There are two challenges. 1) the PSM with replacement is a non-smooth process; 2) how to deal with the censoring indicator in the time-to-event data. In this project, we aim to develop a class of statistical tests for the time-to-event data by the PSM with replacement.

1) develop a class of statistical tests for the time-to-event data by the PSM with replacement
2) Analyze the NLST data as a real example for the new proposed statistical methods. We try to compare the the PFS/OS of patients treated by different therapy in the NLST data.


Shanmei Liao, Beigene;
Yujie Zhong, School of Statistics and Management, Shanghai University of Finance and Economics