Related Experiment Videos
Seamlessly expanding a randomized phase II trial to phase III.
Lurdes Y T Inoue1, Peter F Thall, Donald A Berry
1Department of Biostatistics, University of Washington, Box 357232, Seattle, Washington 98195, USA. linoue@u.washington.edu
Biometrics
|December 24, 2002
Summary
This study introduces a novel sequential Bayesian design for clinical trials, integrating survival data and early events. This adaptive approach aims to reduce sample size and trial duration while maintaining statistical power for comparative studies.
Area of Science:
- Clinical Trials Methodology
- Biostatistics
- Medical Research Design
Background:
- Comparative clinical trials are essential for evaluating new treatments.
- Traditional designs may be inefficient in terms of sample size and duration.
- Integrating early event data with survival analysis offers potential for adaptive trial designs.
Purpose of the Study:
- To propose a novel sequential Bayesian phase II/III design for comparative clinical trials.
- To incorporate both survival time and discrete early events into the trial design.
- To evaluate the efficiency of this design compared to conventional methods.
Main Methods:
- A sequential Bayesian phase II/III design is developed using a parametric mixture model.
- Patients are randomized throughout phase II, with sequential decisions based on predictive probabilities.
- Phase III is seamlessly integrated into phase II by adding more centers.
Main Results:
- Simulation studies demonstrate that the proposed design maintains overall statistical size and power.
- The Bayesian sequential design typically requires a substantially smaller sample size.
- The adaptive nature of the design leads to a shorter overall trial duration.
Conclusions:
- The proposed sequential Bayesian phase II/III design is an efficient alternative for comparative clinical trials.
- This adaptive design offers significant advantages in reducing sample size and trial duration.
- The method shows promise for optimizing clinical trial conduct, as exemplified in non-small-cell lung cancer research.