Bayesian joint modelling of multiple longitudinal outcomes and dependent competing risks using bivariate Marshall–Olkin weibull distribution

Abstract

Joint modelling of multiple longitudinal variables and time to an event under two dependent competing risks using bivariate Marshall–Olkin Weibull distribution is investigated. This joint modelling, enables researchers to consider the possible simultaneous time of failure under both risks. Considering dependent competing risks which is a more realistic assumption is another advantage. The shared parameter approach is applied, which links the multivariate mixed-effects model, considered as the sub-model for multiple longitudinal variables on the bivariate Marshall–Olkin Weibull distribution as the survival sub-model under two dependent competing risks. Bayesian inference is an asset to overcome computational challenges. A simulation study and a real application are presented to show the performances of the presented joint model.

Keywords:

• Joint modelling
• multiple longitudinal outcomes
• depndent competing risk
• simultaneous failure time
• Bayesian inference

Disclosure statement

No potential conflict of interest was reported by the author(s).