Bayesian hierarchical modeling in interim futility analysis for two parallel clinical trials

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacoeconomics

Background:

• Interim analysis is increasingly used in confirmatory clinical trials, particularly twin studies, to meet regulatory requirements like FDA guidance.
• Interim futility analysis helps mitigate risks by stopping trials unlikely to show treatment efficacy, preventing wasted resources.
• A key challenge in twin studies is deciding whether interim analysis should use individual or pooled data, as true treatment effects are unknown.

Purpose of the Study:

• To develop a novel Bayesian hierarchical modeling method for interim analysis in twin clinical trials.
• To enable dynamic data borrowing between parallel twin studies.
• To compare the performance of the new method against traditional separate and pooled analyses.

Main Methods:

• Developed a Bayesian hierarchical model allowing dynamic data borrowing between twin studies.
• Evaluated the impact of various heterogeneity hyperparameters on the Bayesian model's performance.
• Applied the developed method to a case study for comparing predictive powers.

Main Results:

• The proposed Bayesian method demonstrated favorable characteristics compared to separate and pooled analysis strategies.
• The study visualized the critical impact of heterogeneity hyperparameters on the Bayesian model.
• A data-driven suggestion for selecting heterogeneity hyperparameters was provided, independent of prior knowledge.

Conclusions:

• The Bayesian hierarchical model with dynamic data borrowing offers an efficient approach for interim analysis in twin clinical trials.
• The method provides a robust framework for handling uncertainty in treatment effects between studies.
• This approach enhances risk mitigation and resource optimization in complex clinical trial designs.