Activity Number:
|
271
- Recent Statistical Advances of Handling Nonproportional Hazards in Study Design and Data Interpretation
|
Type:
|
Topic Contributed
|
Date/Time:
|
Tuesday, August 9, 2022 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Biopharmaceutical Section
|
Abstract #323328
|
|
Title:
|
Adaptive Extension of Group-Sequential Time-to-Event Trials in the Presence of Non-Proportional Hazards
|
Author(s):
|
Yue Shentu* and Yu Li and Xieyang Jia
|
Companies:
|
Daiichi Sankyo Inc. and Merck & Co and Merck & Co
|
Keywords:
|
adaptive design;
group-sequential trial;
non-proportional hazards;
Bayesian modeling;
probability of success
|
Abstract:
|
With potential delayed separation or crossing of survival curves, the power of time-to-event trials depends not only on number of events, but also on sufficient duration of follow-up. A trial design that ensures both the target number of events and minimum follow-up time may help mitigate the risk of power loss due to non-proportional hazards. Since the pattern of non-proportional hazards is uncertain at the time of trial design, the pre-specified minimum follow-up time is often based on heuristics. We propose to use a Bayesian piecewise exponential model to calculate the conditional probability of success at the penultimate interim analysis in a group-sequential trial, and adaptively extend the follow-up time for the final analysis when necessary. The resulting combination test statistic at the final analysis follows the same rejection boundary as the original group-sequential trial in a straightforward manner. The performance of this adaptive design is assessed through simulations and demonstrated in a trial that could have benefited from such adaptation.
|
Authors who are presenting talks have a * after their name.