Group-Sequential Designs With an Externally-Driven Change of Primary Endpoint.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Epidemiology

Background:

• Adaptive clinical trial designs are valuable in emerging disease settings with high uncertainty.
• Uncertainty in disease natural history, effect sizes, and population can necessitate changing primary endpoints during a trial.

Purpose of the Study:

• To develop statistical methods for controlling type I error rates when a primary endpoint is changed mid-adaptive trial.
• To address the challenge of interim analyses based on an initial primary endpoint when a new primary endpoint is adopted.

Main Methods:

• Modification of group-sequential methods to adjust for interim analyses of an initial primary endpoint.
• Simulation studies to evaluate the type I error rate control under different correlation scenarios between initial and new primary endpoints.
• Illustration using a simulated remdesivir clinical trial for COVID-19.

Main Results:

• The proposed group-sequential method effectively controls the type I error rate for the new primary endpoint, regardless of the treatment effect on the initial endpoint.
• Type I error rate is adequately controlled even when the correlation between the initial and new primary endpoints is estimated from trial data.

Conclusions:

• Modified group-sequential methods provide a robust framework for adaptive trials with primary endpoint changes.
• This approach ensures statistical validity in clinical trials facing evolving uncertainties, as demonstrated in the COVID-19 pandemic context.