A Bayesian Optimal Adaptive Clinical Trial Design for Sequentially Integrated Therapies

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacology

Background:

• Complex diseases often require sequential, integrated treatments targeting specific aspects.
• Alcohol-associated hepatitis (AH) exemplifies this, needing acute inflammation management followed by alcohol use disorder (AUD) interventions.
• Designing trials for sequential therapies with multiple drugs is challenging.

Purpose of the Study:

• To introduce a novel Bayesian model-based trial design for sequentially integrated therapies.
• To address the challenges in optimizing treatment combinations and study designs for complex diseases.
• To enhance trial efficiency and statistical power in sequential treatment studies.

Main Methods:

• Development of the Bayesian Optimal Adaptive Design for Sequentially Integrated Therapies (BIT) design.
• Incorporation of Bayesian modeling to link outcomes across treatment phases.
• Inclusion of adaptive stopping rules for futility and superiority at interim analyses.
• Application of the BIT design to a simulated clinical trial for severe AH treatment.

Main Results:

• The BIT design effectively connects outcomes from acute and subsequent treatment phases.
• The design enhances efficiency through multiple interim analyses and adaptive stopping rules.
• Simulation studies confirm excellent operating characteristics, including experiment-wise type I error rate control.
• The BIT design demonstrates improved statistical power and optimized sample size determination.

Conclusions:

• The BIT design offers an efficient and statistically sound framework for clinical trials involving sequentially integrated therapies.
• This adaptive design is broadly applicable to complex diseases beyond alcohol-associated hepatitis.
• The proposed methodology facilitates the optimization of multi-phase treatment strategies.
• Implementation of the BIT design can lead to more powerful and efficient clinical trials.