causalBETA: An R Package for Bayesian Semiparametric Causal Inference with Event-Time Outcomes

Area of Science:

• Causal inference
• Biostatistics
• Computational statistics

Background:

• Randomized trials are ideal but often infeasible for causal inference.
• Observational studies require causal inference techniques to adjust for confounding.
• Bayesian methods offer advantages like prior smoothing, flexible modeling, and full uncertainty quantification.

Purpose of the Study:

• To address the implementation gap in Bayesian causal inference.
• To introduce causalBETA, an open-source R package for Bayesian event-time analysis.
• To connect statistical causal inference formulas with practical software implementation.

Main Methods:

• Development of the causalBETA R package.
• Utilizing Bayesian semiparametric models for event-time outcomes.
• Leveraging Stan for efficient Bayesian posterior computation.

Main Results:

• The causalBETA package offers a user-friendly interface for Bayesian causal inference.
• Syntax is compatible with existing R survival analysis packages.
• Custom S3 objects facilitate results visualization and summarization.

Conclusions:

• causalBETA lowers the barrier to using advanced Bayesian causal inference methods.
• The package enables robust estimation of causal effects on event-time outcomes.
• Provides methodological details, data demonstration, and computational guidance for users.