Serà el dimecres 23 de febrer a les 12.30 i en la Sala de Graus "Manuel Valdivia" de la Facultat quan Valeria Jocelyn Leiva Yamaguchi ens parle de "A novel two-stage estimation for joint models of longitudinal and survival outcomes: An accurate approach".

"In survival biomedical studies it is usual to collect measurements of diferent biomarkers of the same individual over time. These repeated measurements can be dependent on the survival process. In this case, a popular way of modeling the longitudinal and survival data is through a joint model approach. Typically, this model consists of two submodels: one for the longitudinal outcome and the other for the survival outcome. The longitudinal submodel is usually a linear model with random and fixed effects, while for the survival submodel, Cox proportional hazard is commonly preferred. In the estimation of a joint model, the shared information by the two submodels can make the inferential process expensive and computationally unstable. So, we propose a two-stage approach in order to make the inference less time-consuming and more accurate. To validate the robustness of our proposal, the computational time and bias produced by our methodology are compared with traditional joint model approaches through simulation studies."

•