Joint Modeling of Quality of Life and Survival Using a Bayesian Approach in a Retrospective Time Scale

Area of Science:

• Biostatistics
• Palliative Care Research
• Longitudinal Data Analysis

Background:

• Quality of Life (QoL) is a key outcome in palliative care trials.
• The "truncation by death" problem poses a significant challenge, potentially biasing QoL analyses due to unobserved data post-mortem.
• Accurate estimation of QoL-treatment associations is crucial, especially with high mortality rates in studies.

Purpose of the Study:

• To propose a novel Bayesian joint modeling framework to simultaneously analyze longitudinal Quality of Life (QoL) trajectories and survival outcomes.
• To address the "truncation by death" issue by modeling QoL retrospectively relative to the time of death.
• To incorporate both individual and cluster-level dependencies into the joint model.

Main Methods:

• Developed a Bayesian joint model incorporating cluster-level random effects.
• Modeled longitudinal QoL using penalized regression splines for flexible trajectories.
• Employed a proportional hazards frailty model with a Weibull baseline for survival, linked via subject and cluster random effects.
• Utilized Markov Chain Monte Carlo (MCMC) sampling for model estimation.

Main Results:

• A comprehensive simulation study evaluated the method's performance across various cluster numbers.
• The novel methodology was successfully applied to real-world data from the Reducing End of Life Symptoms with Touch (REST) study.
• The proposed framework effectively handles dependencies and flexible trajectories in joint QoL and survival modeling.

Conclusions:

• The proposed Bayesian joint modeling framework offers a robust solution to the "truncation by death" problem in palliative care research.
• This approach allows for more accurate estimation of treatment effects on QoL by accounting for survival outcomes and complex data dependencies.
• The methodology provides a valuable tool for analyzing longitudinal QoL and survival data in clinical trials with high mortality rates.