A Bayesian quasi-likelihood design for identifying the minimum effective dose and maximum utility dose in...

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacometrics

Background:

• Traditional dose-ranging studies prioritize dose-efficacy assessment and minimum effective dose identification.
• There is a growing need to optimize drug dosage based on a comprehensive benefit-risk assessment.

Purpose of the Study:

• To propose a Bayesian quasi-likelihood dose-ranging design for simultaneously identifying the minimum effective dose and maximum utility dose.
• To optimize the benefit-risk tradeoff by jointly considering safety and efficacy endpoints.
• To establish proof-of-concept for dose-response relationships.

Main Methods:

• Utilizes a Bayesian quasi-likelihood approach with a beta-binomial model for binary toxicity.
• Employs quasi-likelihood for efficacy endpoints (binary, ordinal, continuous) without parametric assumptions.
• Incorporates a utility function for benefit-risk tradeoff and adaptive patient allocation.
• Employs a group-sequential design with interim dose dropping (toxicity/futility) and posterior probability criteria for proof-of-concept.

Main Results:

• The proposed design demonstrates robustness and competitive performance in simulations compared to existing methods.
• Successfully establishes proof-of-concept and identifies both minimum effective dose and maximum utility dose.
• The design effectively optimizes the benefit-risk tradeoff by integrating safety and efficacy data.

Conclusions:

• The Bayesian quasi-likelihood dose-ranging design offers a superior approach for optimizing drug dosage by balancing efficacy and safety.
• It provides a robust framework for identifying key dose metrics (MED, MUD) and establishing proof-of-concept.
• This design enhances clinical trial efficiency and drug development decision-making through adaptive strategies.