|Title||CureBayes: Bayesian phase 3 design using the cure rate model. (SAS and R)|
|Year of Publication||2017|
|Authors||Ibrahim, Joseph G., and Matthew Psioda|
This software package can be used to calculate the required sample size for a future trial using a time-to-event endpoint where the analysis is to be performed using a Cure Rate model and where an existing trial is available that can be used to inform the design and analysis of the new trial. The software allows one to construct a fixed, maximally informative power prior a priori while simultaneously identifying the minimum sample size required for the new trial so that the design has high Bayesian power and Bayesian type I error control. These Bayesian operating characteristics are weighted averages of the type I error rate and power for fixed parameter values with weights determined by the posterior distribution of the parameters given the historical data after conditioning on the relevant hypothesis being true.
This software was written in SAS and R for used on a high-performance computing cluster (HPCC). All of the SAS and R programs were tested in both Windows and Linux environments. The programs were developed using SAS version 9.4 (TS1M1) and R version 3.3.1. The R programming requires Rcpp (version 0.12.10) and RcppArmadillo (version 0.7.800.2.0). All of the programming was developed first on a Windows machine and then ported to Linux for large-scale simulations on a Linux-based HPCC running a SLURM scheduler. SLURM array scripts are provided with the programs to help users with a similar implementation. After minor modifications for directory paths, the programs should run in both Linux and Windows environments. All programs make use of random number seeds so that our results should be 100% reproducible. Programs need to run in sequence (see readme file contained within) as the output from one step generally defines the input for the next step.
CureBayes: Bayesian phase 3 design using the cure rate model. (SAS and R)