Many medical decisions involve using accumulated information on patients under surveillance to predict transitions in future health status, such as progression of disease or advancement to death. Longitudinal studies allow investigators to correlate changes in time-dependent exposures or biomarkers with subsequent health outcomes. At any given time, an individual’s longitudinal measures up to that time may be used to update the predicted risk of future adverse outcomes and guide medical decisions regarding monitoring and treatment. For example, high-risk individuals may be targeted for preventive strategies or aggressive treatments, whereas less frequent follow-up may be recommended for low-risk individuals.
In this course, participants will learn:
(1) Flexible approaches, such as joint models and partly conditional models, for modeling dynamic prediction rules for risk of a future adverse outcome using longitudinal trajectories up to the time of prediction, and
(2) Methods for evaluating predictive performance using summary measures that are appropriate for censored survival outcomes, with a focus on predictive accuracy using time-dependent sensitivity and specificity for prognosis of a subsequent event time.
Methods will be illustrated using examples from HIV, end stage renal disease, and organ transplantation settings. The course will include hands-on training and demonstration of relevant R packages for answering research questions. Real-data examples for analysis will be provided and the instructors will discuss implementation and interpretation.
This course is designed for those with basic understanding of methods for correlated data and survival analysis. Some experience with programming in R is preferred, but not required.