Software : Extensions


The links below provide access to a number of extensions to RCTdesign that have not yet been incorporated into the standard release. In each case, you may download a zip file containing R script files and rudimentary documentation. These routines are currently being enhanced and validated for inclusion in a future release, but in the meantime we welcome your exploration of these functions, as well as any feedback you might provide for improvements.

  1. Bayesian evaluation including functions used to produce the contour plots displayed in our Statistics in Medicine tutorial on “Bayesian evaluation of group sequential clinical trial designs”.   

  2. Adaptive clinical trials including specification of designs that implement the Cui, Hung, and Wang (1999) approach to adaptive repowering of a clinical trial, along with inference appropriate that is possible based on pre-specified adaptation rules. These methods represent extensions of the Whitehead (1987) bias adjusted estimates and confidence intervals and p values based on the MLE, stagewise, and likelihood ratio orderings of the outcome space. Implementation of inference based on the Brannath, Mehta, and Posch (2009) ordering of the outcome space for fully adaptive (not pre-specified) designs will also be available in the next release of RCTdesign. (See G. Levin’s 2012 unpublished PhD dissertation on the RCTdesign Methodology webpages for details.)

  3. Sensitivity analyses for informative censoring including routines that compute and display inference under assumptions of missing not at random (MNAR) models based on a proportional hazards model that could be used in clinical trial data. Our preliminary research suggests that this model is very robust to departures from proportional hazards. (See E. Meier’s 2012 unpublished MS thesis on the RCTdesign Methodology webpages for details.)

  4. General utility functions in R. Okay, so these are not at all specific to clinical trials, but they do provide some general routines for descriptive statistics and plotting that we sometimes include in our other script files.