Survival modeling (https://en.wikipedia.org/wiki/Survival_analysis) is a standard component of clinical data analysis, but often proves challenging in practice. Not only is there a variety of approaches to modeling the baseline hazard function, but model checking and evaluation is complicated since we are often not modeling the censoring process explicitly.
This session will work through several examples of doing survival analysis using Stan, including both parametric and nonparametric approaches to modeling the baseline hazard. We will describe practical approaches to posterior predictive checking and model evaluation, as well as several extensions to the standard proportional hazards model such as incorporating nonlinear covariate effects, time-varying covariate effects, and stratified analysis with/without varying-coefficients.
The presenter Jacki Buros (https://www.researchgate.net/profile/Jacki_Buros2)is a biostatistician at the Hammer Lab (http://www.hammerlab.org/), which is affiliated with the Icahn School of Medicine at Mt. Sinai. These examples are motivated by her work evaluating biomarkers for response to immunotherapy.
We will have pizza at 5:30 and will start the talk at 6:00 PM. Please, come on time as we will be unable to accommodate late arrivals. Thank you!
P.S. We are instituting a small registration fee to help pay for the food and get a more accurate attendee count.