Bayesian analysis of paired survival data using a bivariate exponential distribution.

Area of Science:

• Biostatistics
• Statistical Modeling
• Survival Analysis

Background:

• Paired survival data analysis is crucial in various fields.
• Moran's bivariate exponential model offers unique advantages over existing models.
• Computational challenges have limited the application of Moran's model.

Purpose of the Study:

• To develop a feasible Bayesian computational method for Moran's bivariate exponential model.
• To address the lack of a closed-form likelihood function in Moran's model.
• To introduce a model checking procedure for enhanced reliability.

Main Methods:

• Introduced a latent variable to overcome computational hurdles in Bayesian analysis.
• Employed a predictive Bayesian P-value for robust model checking.
• Utilized Moran's bivariate exponential model for paired survival data.

Main Results:

• Successfully implemented a Bayesian analysis for Moran's model.
• The latent variable approach simplified complex Bayesian computations.
• The predictive Bayesian P-value provided an effective model checking tool.

Conclusions:

• The proposed Bayesian method enhances the applicability of Moran's bivariate exponential model.
• This approach facilitates more reliable analysis of paired survival data.
• The study offers a practical solution for previously intractable statistical problems.