B1836
Title: Predicting survival risks in the presence of multiple longitudinal markers: A simulation study
Authors: Ayca Pamukcu - Acibadem University (Turkey) [presenting]
Ozgur Asar - Acibadem University (Turkey)
Abstract: In prospective medical studies, typically baseline covariates, multiple longitudinal biomarkers and time to a survival event are collected. Available literature has emphasized the importance of using repeated biomarker measurements for predicting survival outcomes. Nonetheless, majority has considered the use of a single longitudinal biomarker. We aim to investigate the predictive ability of more than one biomarker. We conduct extensive simulation studies to inspect this contribution under several scenarios, e.g. under different correlation structure between multivariate markers, and different associations between markers and risk of survival event.