Robust Joint Modelling Framework Joint modelling of longitudinal time-to-event outcomes typically combines a linear mixed-effects model for repeated measures and a Cox model with time-varying frailty for time-to-event outcome (Asar et al., 2015). Typical distributional assumption is that random-effects and measurement error terms in mixed-effects model are Gaussian. However, this assumption might be restricive for real-life problems, where it is quite likely to have
subjects who do not conform the population averaged trends (they are examples of outliers in the random-effects), and