Bayesian survival analysis in clinical trials: What methods are used in practice?

Area of Science:

• Biostatistics
• Clinical Trial Design
• Survival Analysis

Background:

• Bayesian statistics offer an alternative to frequentist methods in clinical trials, particularly for rare diseases.
• Time-to-event endpoints are common, but Bayesian survival trials present unique modeling complexities.
• This review critically examines the application and reporting of Bayesian methods in survival trials.

Purpose of the Study:

• To systematically review the use and reporting of Bayesian methods in survival clinical trials.
• To identify trends in the application of Bayesian approaches for trial design, analysis, and monitoring.
• To assess the quality of reporting for Bayesian survival analyses.

Main Methods:

• Systematic review of clinical trials using Bayesian survival analyses via PubMed and Web of Science.
• Inclusion of original trial reports and re-analyses; exclusion of meta-analyses and methodological papers.
• Full-text search of pre-selected journal repositories.

Main Results:

• 28 articles met inclusion criteria; most were oncology (25), randomized controlled (21), phase III (13) trials, with half in rare diseases (13).
• Bayesian methods were used for monitoring (14) and final analysis (14); regression models were fitted in 12 articles.
• Prior distributions were often incompletely reported (20 articles); informative priors or external data were rarely used (4 trials).

Conclusions:

• Bayesian survival analysis is infrequently implemented, with limited incorporation of external data into priors.
• There is a significant need to enhance the quality of reporting for Bayesian methods in survival trials.
• Extending reporting guidelines like the Consolidated Standards of Reporting Trials is recommended for Bayesian clinical trials.