A Bayesian Model for Joint Analysis of Multivariate Repeated Measures and Time to Event Date in Crossover Trials

Joint modeling of longitudinal and survival data has become a popular technique in analyzing clinical longitudinal trials. In this discussion, the potentials of joint modeling is explored for analyzing time to event (TTE) and multivariate repeated measures in crossover studies. The work is motivated by a real-life crossover study with three visual analogue scale (VAS) responses and a TTE response. To recover the information lost due to censoring of the TTE variable, we propose a Bayesian joint model (BJM) to analyze the VAS and TTE responses jointly, leveraging the moderate associations among the responses. The joint model links the TTE variable to the VAS repeated measures via multi-layered subject-specific random effects. We show the BJM produces more efficient inferences with satisfactory goodness of fit in general with comparison to modeling of the VAS and TTE responses separately. A simulation study is performed to demonstrate the inferential advantages of BJM over separate modeling and maximum likelihood approaches via nonlinear mixed modeling in the crossover setting. This work also demonstrates the flexibility and usefulness of zero-one inflated beta regression in modeling non-Gaussian fixed-boundaries-inflated outcomes in general.