Adjusting Survival Estimates for Treatment Switching: New Guidance and the Role of Real-World Evidence

Speaker(s)

Discussion Leader: Shannon Cope, MSc, Precision AQ, VANCOUVER, Canada
Discussants: Nicholas Latimer, PhD, MSc, Sheffield Centre for Health and Related Research (SCHARR), Delta Hat Ltd, Nottingham, UK; University of Sheffield, Sheffield, Great Britain; Jeroen P Jansen, PhD, Center for Translational and Policy Research on Precision Medicine (TRANSPERS), University of California – San Francisco, San Francisco, CA, USA; Harlan Campbell, PhD, Department of Statistics, University of British Columbia, Vancouver, BC, Canada

Presentation Documents

PURPOSE:

NICE's Decision Support Unit has published a new technical support document about treatment switching (when trial participants receiving the control treatment crossover to the experimental treatment). The TSD 24 (April, 2024) offers comprehensive reporting guidelines and specific recommendations. We review the various crossover adjustment methods recommended (e.g., rank-preserving structural failure time models (RPSFT), inverse probability of censoring weighting (IPCW), two-stage estimation (TSE)) and highlight recent statistical developments. We then explain the reasoning behind the TSD's recommendations and review key takeaways from a review of recent NICE submissions.

Given that real-world evidence (RWE) is increasingly used to support applications for regulatory and HTA approvals, the TSD 24 discusses the potential of using external data for crossover adjustment. We review examples of how external data has been used for crossover adjustment in a number of recent studies (e.g., used to validate different crossover adjustment methods, and used to create an external control arm without any crossover contamination). Finally, we discuss a newly proposed method to estimate counterfactual survival by leveraging RWE to supplement RCT data in an efficient and unbiased manner. This approach may prove particularly valuable given the increasing availability of routine electronic health record (EHR) data.

DESCRIPTION:

Workshop attendees will be asked (real-time polling) to consider the impact of the latest NICE guidance, if RWE has potential to improve cross-over adjustment, and if different sources of bias can be adequately addressed.

Ms. Cope will chair the session and begin by summarizing the various statistical methodologies often used for cross-over adjustment (10 min). Dr. Latimer will then review the recently published TSD 24 guidance (12 min). Dr. Jansen will review specific examples of how RWE has been considered for cross-over adjustment (12 min) and Dr. Campbell will conclude by considering newly proposed methods for integrating external data (12 min).

Code

312

Topic

Study Approaches