I received a question the other day by email regarding the parameterization of the Weibull survival model which is included in both SurvivalStan (a Python package) and biostan, an R mini-package I put together for a BioConductor workshop held in 2016.
The question read:
Hi Jacki,
I wonder if you might be able to help me with a question? I have been looking at stan recently for survival models and I noticed the biostan repository, which is quite a helpful resource.
Survival analysis is an important and useful tool in biostatistics. It is commonly used in the analysis of clinical trial data, where the time to a clinical event is a primary endpoint. This endpoint may or may not be observed for all patients during the study’s follow-up period.
At the core of survival analysis is the observation that the observed time to event $t$ is the result of an accumulation of event-related risks at each moment up to that time $t$.