A1171
Title: Joint models in longitudinal studies for efficient and robust inferences
Authors: Lang Wu - University of British Columbia (Canada) [presenting]
Abstract: In many longitudinal studies, within-individual repeated measurements often exhibit large variations that change over time. Understanding the nature of within-individual systematic and random variations enables more efficient statistical inference. A nonlinear mixed effects (NLME) model is considered for modeling longitudinal means, together with a model for within-individual variances that addresses outliers in repeated measurements. Additionally, joint models for longitudinal data collected at different phases of studies are considered. Statistical inference is based on the joint likelihood for all models. The proposed methods are applied to a recent AIDS dataset, with interesting new findings.
