Longitudinal studies follow individuals over time and repeatedly measure health status. Analyses of longitudinal data are often complicated by several factors that can threaten the validity of standard analysis methods. First, missing data in longitudinal outcomes can arise when individuals are lost to follow-up, either due to drop-out (e.g., in randomized trails) or death (e.g., in long-term observational studies). Second, when modeling intermittently measured time-dependent covariates in a survival analysis, biological variation can lead to measurement error. Joint modeling of longitudinal and survival outcomes has emerged as a novel approach to handle these issues.

We will detail the use of mixed-effects models for the analysis of repeated longitudinal measures, Cox regression models for the analysis of event-time outcomes with longitudinal measures as time-dependent covariates, and their combination in a joint modeling framework. An in-depth data analysis (conducted in R) will be used to discuss analysis strategies, the application of appropriate analysis methods, and the interpretation of results.

This course is targeted toward individuals with some prior experience with statistical methods for longitudinal data analysis and survival analysis. Individuals with no prior experience with longitudinal data analysis should consider Module 9: Generalized Estimating Equations and Mixed-Effects Models for Longitudinal Data Analysis. Individuals with no prior experience with survival analysis should consider Module 12: Survival Analysis with Emphasis on Applications to Clinical Trials.