A group sequential test for survival trials: an alternative to rank-based procedures

Area of Science:

• Biostatistics
• Clinical Trial Design
• Survival Analysis

Background:

• Interim monitoring is crucial for ethical and efficient survival trials.
• The proportional hazards assumption is a common limitation in standard survival analysis methods.
• Existing group sequential methods may not be optimal when proportional hazards do not hold.

Purpose of the Study:

• To develop and present a novel method for interim monitoring in survival trials.
• To extend existing test statistics for cumulative weighted differences to a sequential setting.
• To provide an interpretable measure of treatment effect that does not rely on the proportional hazards assumption.

Main Methods:

• Extension of Pepe and Fleming's cumulative weighted difference test statistics to the sequential setting.
• Utilizing an appropriate weight function to derive the test statistic.
• Illustration with a real clinical trial design and operating characteristics studied via simulation.

Main Results:

• The proposed method provides a viable alternative to group sequential linear rank tests.
• The test statistic serves as an estimator for the cumulative weighted difference in survival probabilities.
• The method is demonstrated to be applicable and interpretable, particularly when proportional hazards fail.

Conclusions:

• The developed method offers robust interim monitoring for survival trials.
• It provides an interpretable measure of treatment difference, enhancing clinical trial analysis.
• This approach is valuable when the proportional hazards assumption is not met.