Conditional power calculations as an aid in the decision whether to continue a clinical trial

Area of Science:

• Biostatistics
• Clinical Trial Design
• Survival Analysis

Background:

• Clinical trials often involve sequential monitoring and interim analyses.
• Decisions to continue or stop trials are critical for resource allocation and patient safety.
• Stochastic curtailment is a statistical technique used in adaptive trial designs.

Purpose of the Study:

• To integrate stochastically curtailed testing into clinical trial continuation decisions.
• To evaluate the conditional power of statistical tests in ongoing comparative trials.
• To provide a framework for decision-making based on observed survival or event times.

Main Methods:

• Utilized a statistical model assuming constant hazard rates for two treatment arms.
• Developed methods to calculate the conditional power of a test at the decision point.
• Incorporated the concept of stochastically curtailed testing into the decision process.

Main Results:

• Demonstrated how to incorporate stochastic curtailment into the decision-making process for clinical trials.
• Showcased the examination of the power function conditional on the trial's progress.
• Provided a method for assessing the statistical evidence for continuing the trial.

Conclusions:

• Stochastic curtailment offers a valuable approach for adaptive clinical trial management.
• Conditional power analysis is essential for informed decisions regarding trial continuation.
• The proposed methods support efficient and statistically sound clinical trial conduct.