Bayesian multivariate joint modeling of competing risks with unit-bounded and ordinal longitudinal information
J. Gutiérrez Botella, C. Armero i Cervera, T. Kneib, J. García-Seara
Joint models for longitudinal and survival data provide a flexible framework for the joint analysis of repeated measurements and time-to-event outcomes, allowing inference when both processes are associated. We propose a Bayesian multivariate competing risks joint model with unit-bounded and ordinal longitudinal responses. The longitudinal component for the unit-interval response is modeled through Beta, Kumaraswamy, and Unit-Weibull distributions, enabling the assessment of different distributional assumptions. The ordinal variable is modelled via a cumulative mixed logit regression model. The survival sub-model is a cause-specific hazards model for competing risks linked to the longitudinal processes via random effects. Posterior inferences are performed using Markov chain Monte Carlo methods with JAGS. Model assessment is performed using WAIC and LPML criteria. The methodology is applied to a study of heart failure patients undergoing cardiac resynchronization therapy.
Keywords: joint modeling, cardiac resynchronization therapy, competing risks
Scheduled
GT Bioestadística SEIO - BIOSTATNET: Análisis de Supervivencia
September 2, 2026 3:30 PM
Aula B
Other papers in the same session
D. Antonio, F. Blanco, M. Saez Zafra, M. A. Barceló
S. Saavedra Martínez, A. López Cheda, M. A. Jácome Pumar
G. Gómez Melis, I. Arostegui Madariaga, K. Langohr
A. Alcaraz Peñalba, A. M. Franco Pereira, M. C. Pardo Llorente