Joint Dispersion Model with a Flexible Link

TL;DR

This paper introduces a flexible joint modeling framework for longitudinal and survival data that accounts for heteroscedasticity and time-varying relationships, improving understanding of disease progression in HIV/AIDS.

Contribution

It proposes a novel Bayesian joint model with a dispersion component and time-varying effects, relaxing common variance assumptions and enhancing model flexibility.

Findings

Outperforms existing joint models in predictive accuracy.

Highlights the importance of heteroscedasticity in disease modeling.

Demonstrates the impact of CD4 count stability on survival analysis.

Abstract

The objective is to model longitudinal and survival data jointly taking into account the dependence between the two responses in a real HIV/AIDS dataset using a shared parameter approach inside a Bayesian framework. We propose a linear mixed effects dispersion model to adjust the CD4 longitudinal biomarker data with a between-individual heterogeneity in the mean and variance. In doing so we are relaxing the usual assumption of a common variance for the longitudinal residuals. A hazard regression model is considered in addition to model the time since HIV/AIDS diagnostic until failure, being the coefficients, accounting for the linking between the longitudinal and survival processes, time-varying. This flexibility is specified using Penalized Splines and allows the relationship to vary in time. Because heteroscedasticity may be related with the survival, the standard deviation is considered as a covariate in the hazard model, thus enabling to study the effect of the CD4 counts' stability on the survival. The proposed framework outperforms the most used joint models, highlighting the importance in correctly taking account the individual heterogeneity for the measurement errors variance and the evolution of the disease over time in bringing new insights to better understand this biomarker-survival relation.