DISES | Home | Seminari e Convegni
15 Maggio 2018, h. 15.00 - Sala dei Consigli DISES
Marcella Mazzoleni - Università di Milano Bicocca
"Joint models for time-to-event and multivariate longitudinal data: a likelihood approach"
The joint models analyse the effect of longitudinal covariates onto the risk of one or more events. They are composed of two sub-models, the longitudinal and the survival sub-model. For the longitudinal sub-model a multivariate mixed model is used, considering fixed and random effects. Whereas for the survival sub-model, a Cox proportional hazards model is proposed, considering jointly the influence of more than one longitudinal covariate onto the risk of the event. The purpose of the presentation is to show the extension of a univariate estimation method based on a joint likelihood formulation to the case in which the longitudinal sub-model is multivariate. The estimation method is based on the maximisation of the likelihood function achieved through the implementation of an Expectation-Maximisation (EM) algorithm. In the M-step a one-step Newton-Raphson update is used, as for some parameters estimators, it is not possible to obtain closed-form expression. In addition, a Gauss-Hermite approximation is applied for some of the integrals involved. Some simulation results and an application to a well-known dataset are shown.
This is a joint work with: Mariangela Zenga.