[R] Triangular Test

Marc Schwartz marc_schwartz at me.com
Tue Feb 21 17:19:23 CET 2012


On Feb 20, 2012, at 7:45 AM, kende jan wrote:

> Hello,
> 
> I would like to perform triangular test for clinical trial with R.
> can you help me please ?
> 
> Jan


Hi,

I am not aware of any specific implementations of Whitehead's Triangular Test[1] in R, though somebody else may be aware of something locally. There are implementations in SAS of course, either using Whitehead's own PEST (http://www.maths.lancs.ac.uk/department/research/statistics/mps/pest) or the PROC SEQDESIGN implementation, neither of which are free.

There appears to be an online implementation, at least of potentially limited scope here:

  http://www.stattools.net/SeqUnpaired_Exp.php

but it would be reasonable to validate the implementation before proceeding with use.

For R, there are other group sequential boundary and alpha spending based methods implemented and there is a CRAN Task view that you might find helpful, which links to R packages that implement some of these procedures:

  http://cran.r-project.org/web/views/ClinicalTrials.html

I have used Keaven Anderson's gsDesign package for example, which includes a nice document with worked examples of designs and subsequent monitoring.

HTH,

Marc Schwartz

[1]  Whitehead J. The Design and Analysis of Sequential Clinical Trials. 
     Revised 2nd Edition. Chichester: Wiley, 1997



More information about the R-help mailing list