Random Effects and Recurrent Events in Survival Analysis

Music City Center 

Room: CC-110B 

A key feature of time-to-event data is the presence of censored/truncated observations and standard methods of analyzing such data are covered in introductory statistics courses. However, methods for analyzing recurrent events some of which may be censored, are often not covered in introductory classes. Likewise, random effects are well-studied when modeling continuous or dichotomous outcomes with many books/resources available. But when analyzing time-to-event outcomes, incorporating random effects is challenging beyond the use of few basic models. Theory is well developed, however there is a need to adapt the myriad methods available in literature to real life applications. As the field of Biostatistics & Data Science advances, recurrent events and clustered survival data are increasingly seen in practice. As understanding theory is onerous, many practicing statisticians are tempted to fit only the very basic models to such data often ignoring the rich possibilities of more advanced models. This course provides an opportunity to learn about advanced modeling for such data keeping in mind the underlying assumptions of the models. Real life examples will be covered using the R/SAS software and more advanced models (including Bayesian) will be discussed. The course will be taught at an Applied level with minimal theory.