Bayesian Random-Effects Meta-Analysis of Aggregate Data on Clinical Events.

Area of Science:

• Biostatistics
• Clinical Trial Analysis
• Epidemiology

Background:

• Meta-analysis is crucial for assessing evidence on rare events, particularly adverse effects in clinical trials.
• Survival methods are optimal for handling adverse event data.
• Existing common-effect models process hazard ratio data but don't fully address study heterogeneity.

Purpose of the Study:

• To extend existing meta-analysis models for rare events.
• To develop a Bayesian random-effects approach accommodating study heterogeneity.
• To provide a robust method for analyzing adverse effects in clinical trials.

Main Methods:

• Developed a Bayesian random-effects meta-analysis model.
• Extended Holzhauer's common-effect model.
• Utilized hazard ratios for data analysis.
• Performed sensitivity analyses and Monte Carlo simulations.

Main Results:

• The proposed Bayesian random-effects model effectively handles heterogeneity in rare event data.
• The model is applicable to realistic clinical trial datasets.
• Simulations and sensitivity analyses support the model's validity and robustness.

Conclusions:

• The novel Bayesian random-effects model offers an advanced approach for meta-analysis of rare adverse events.
• This method enhances the accurate assessment of intervention effects when dealing with sparse data.
• The model provides a more comprehensive understanding of treatment-related risks in clinical research.