Sudden cardiac death (SCD) is a leading cause of death in the U.S. Patients at elevated risk of SCD are primarily treated with an implantable cardioverter-defibrillator (ICD), which can prevent cardiovascular death but often causes severe side effects, such as reduced quality of life from shock-induced pain. Decisions about ICD treatment involve complex personal trade-offs across multiple health events, including mortality and quality of life. While prediction tools could help patients and providers weigh these trade-offs, they focus on univariate outcomes, treating other clinical endpoints as inputs, so trade-offs cannot be directly informed. To address this, we propose a novel general framework that jointly models a terminal event and a longitudinal marker as a bivariate process over discrete time, for settings where prediction is the primary goal. Discretizing study time lets the framework capture the dynamic interplay between outcomes while avoiding implicit extrapolation beyond truncation by a terminal event, such as death. The framework flexibly accommodates phenomena arising in applied contexts, including global time-invariant and local time-dependent dependence structures between the terminal event and marker, and latent association via a shared frailty term. Estimation proceeds via the Bayesian paradigm, facilitating uncertainty quantification and yielding patient-specific joint posterior predictions for time to terminal event and future marker trajectory. We introduce this bivariate discrete-time framework with a focus on its flexibility, provide guidance on choosing the discretization of study time and on Bayesian model construction and selection, and discuss insights gained from its novel output, joint posterior predictions. Finally, we demonstrate its clinical applicability using data from the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) on SCD and ICD therapy.

翻译：暂无翻译