brea

brea implements a Bayesian discrete-time proportional hazards framework for analyzing recurrent discrete-time event data, enabling semiparametric modeling of time-varying covariates and MCMC-based inference for interval-recorded events such as patient-reported outcomes recorded in whole months.

Key Features:

• Bayesian discrete-time proportional hazards: Implements a Bayesian discrete-time proportional hazards model for recurrent events, supporting competing risks, frailties (random effects), and tied outcomes.
• Flexible baseline hazards: Allows flexible specification of baseline hazards to capture complex time effects without restrictive parametric forms.
• Semiparametric covariate modeling: Employs a generalized additive model (GAM)-style semiparametric approach to incorporate time-varying covariates.
• MCMC-based inference: Uses Markov chain Monte Carlo (MCMC) algorithms for parameter estimation and uncertainty quantification.

Scientific Applications:

• Epidemiology and clinical research: Applied to studies with discrete-time recurrent events such as patient-reported outcomes and other interval-recorded event data.
• Illicit drug use cessation: Used to analyze time to cessation recorded in whole months, accommodating tied outcomes and recurrent events.
• Clinical trials (anesthesia methods): Applied in trial settings involving anesthesia administration methods where recurrent and interval-recorded events and random effects are relevant.

Methodology:

Uses a Bayesian discrete-time proportional hazards model tailored for recurrent event data with semiparametric GAM-style covariate terms, flexible baseline hazard specification, explicit handling of competing risks and frailties, and inference via Markov chain Monte Carlo (MCMC).