Adaptive group sequential design for phase II clinical trials: a Bayesian decision theoretic approach.

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Decision Theory

Background:

• Bayesian decision theoretic approaches (BDTAs) are established for phase II clinical trial design.
• Existing methods often lack comprehensive Bayesian analysis for multiple endpoints, particularly efficacy and toxicity.
• Monitoring toxicity is a critical objective in phase II trials.

Purpose of the Study:

• To propose an adaptive group sequential design for phase II clinical trials using a BDTA.
• To characterize efficacy and toxicity as correlated bivariate binary endpoints within a Bayesian framework.
• To allow for a trade-off between efficacy and toxicity endpoints during trial monitoring.

Main Methods:

• An adaptive group sequential design is developed, incorporating Bayesian decision theory.
• Efficacy and toxicity are modeled as correlated bivariate binary endpoints.
• A loss function is defined, balancing patient care costs and decision-making errors.
• Interim trial evaluations are conducted group sequentially with adaptive group sizes and look numbers.

Main Results:

• The performance of the Bayesian modeling and decision rules was evaluated via simulation across seven scenarios.
• Operating characteristics were assessed under various loss function parameters.
• The proposed method was illustrated using a real-world case study.

Conclusions:

• The developed adaptive group sequential design offers a robust Bayesian approach for phase II trials with multiple correlated endpoints.
• The method provides flexibility in managing the trade-off between efficacy and toxicity.
• This approach enhances decision-making during clinical trial conduct, as demonstrated in a pancreatic cancer trial.