Generalized triple outcome decision-making in basket trials.

Area of Science:

• Clinical Trials
• Biostatistics
• Drug Development

Background:

• The go/no-go decision in Phase II/Ib clinical trials is crucial for drug development.
• Traditional binary hypothesis testing is being replaced by Triple Outcome Decision-Making (TODeM) for greater flexibility.
• Existing Bayesian methods for information borrowing are limited to binary decision frameworks.

Purpose of the Study:

• To propose a novel decision-making framework, Generalized Triple Outcome Decision-Making (G-TODeM).
• To adapt Bayesian methods with information borrowing for the TODeM framework.
• To improve decision-making flexibility in clinical trials by incorporating cross-cohort data.

Main Methods:

• Development of the Generalized Triple Outcome Decision-Making (G-TODeM) framework.
• Application of Bayesian methods, such as EXNEX and MUCE, for information borrowing across cohorts.
• Integration of these Bayesian approaches within the TODeM structure.

Main Results:

• The G-TODeM framework successfully applies Bayesian information borrowing to the TODeM structure.
• Information borrowing from other cohorts effectively shrinks the 'consider' zone for the inference cohort.
• This leads to more nuanced and data-driven go/no-go decisions.

Conclusions:

• G-TODeM offers a robust method for enhancing clinical trial decision-making.
• The framework improves upon existing methods by allowing flexible integration of Bayesian statistics and TODeM.
• This approach facilitates more efficient and informative clinical trial progression.