Zoom Meeting Passcode: N0Uz24

“Empower RCT analysis with integrated information from RWD”

Shu Yang, Ph.D. (NCSU)

Randomized clinical trials (RCT) have been the gold standard for treatment effect evaluation; however, the RCT sample may be small, lack generalizability to a target population, and limited in the patient diversity. On the other hand, large real-world data (RWD) are becoming increasingly available for research purposes; however, RWD analysis may suffer from biases due to confounding. In this talk, we leverage the complementary features of RCT and RWD to improve treatment effect evaluation. First, we propose calibration weighting estimators that improve the generalizability of the RCT-based estimator by leveraging the representativeness of the RWD study sample.  Second, we develop various integrative strategies that borrow RWD to improve HTE (heterogeneity of treatment effect) estimation, while ensuring that any confounding biases present in the RWD do not leak into the proposed estimator. We apply our proposed methods to estimate the effects of adjuvant chemotherapy in early-stage resected non–small-cell lung cancer integrating data from an RCT and a sample from the National Cancer Database.