U.S. flag An official website of the United States government
  1. Home
  2. Science & Research
  3. Science and Research Special Topics
  4. Advancing Regulatory Science
  5. Principal stratification methods and software for intercurrent events in clinical trials
  1. Advancing Regulatory Science

Principal stratification methods and software for intercurrent events in clinical trials

CERSI Collaborators: Triangle CERSI, Duke University: Fan Li, PhD; Laine Thomas, PhD; Anqi Zhao, PhD; Susan Halabi, PhD

FDA Collaborators: Yuan-Li Shen, Dr.P.H; Pallavi Mishra-Kalyani, PhD; Shu Wang, PhD.; Xiaoxue Li, PhD.; Joyce Cheng, PhD

Project Start Date: September 8, 2023

Regulatory Science Challenge

Events that occur post randomization in randomized control trials, known as intercurrent events, can alter the course of the randomized clinical trials and jeopardize comparative effectiveness evaluation and consequently decision making in regulatory science. The standard approach of intention-to-treat analysis ignores intercurrent events and thus preserves the trial validity based on randomization, but it fails to capture treatment effect heterogeneity and the complex causal mechanism. The 2018 ICH E9(R1) addendum suggests principal stratification as an alternative approach to handle intercurrent events, but significant gaps exist between the theory and practice of principal stratification in regulatory science. In particular, there is a lack of transparent and accessible analytical methods, practical guidelines, and software of principal stratification in the context of regulatory science.

Project Description and Goals

This project aims to develop a suite of transparent and accessible analysis tools, software and educational material for applying the principal stratification method to analyze intercurrent events in clinical trials. Investigators will focus on two prevalent types of intercurrent events: (1) nonadherence to assigned treatment, including treatment switching and discontinuation and (2) truncation of the target outcome by a terminal event. For each type, investigators will develop estimand, computational, visualization, and sensitivity analysis tools, with a special emphasis on time-to-event outcomes. They will also develop a companion R package and tutorials with illustrations of clinical trials in oncology and other diseases. The results of this study will impact clinical trials in two ways: (1) produce new methodological tools for addressing a pressing and prevalent complication in clinical trials, (2) provide associated open-source software and educational material to disseminate the methodology to regulatory agencies, health researchers, and industry. Investigators also plan to develop scientific publications describing the outcomes of this research and discuss it at public forums.

Back to Top