Censored survival time and other failure time data are commonly encountered in clinical and epidemiological studies. The Cox proportional hazards model is the method of choice to analyzing such data, but has important limitations. The goal of this SBIR is to incorporate cutting-edge research to extend the Cox model to transformation models allowing non-proportional hazards, to random-effects models for multivariate failure time data, and to joint modeling of repeated measures and event times and to produce a robust, well-tested and user-friendly software.
