Group sequential design for safety monitoring - Whitehead's Triangular test

[Jen Bell]

Hi,

I need to use Whitehead's Triangular Test (http://www.ncbi.nlm.nih.gov/pubmed/11042622) for a group sequential design monitoring safety in a trial (so a one-sided test). I cannot find anything that would assist me with doing this in Stata, and wondered if anyone out there had experience of this who could share any advice please?

It was originally implemented in PEST (no longer available) and then a paper was published with guidance for SAS (http://www.ncbi.nlm.nih.gov/pubmed/20876163) but unfortunately I do not have access to SAS. I have tried reproducing this in Stata using Mata, but it's the 'call seq' part that there doesn't seem to be a Stata equivalent for.

There is landemets in Stata for other stopping rules, but that doesn't have an option for this particular test sadly.

Any help or insight anyone could offer would be very much appreciated!

Thanks,

Jen

[John Moran]

Hi Jen
Not exactly what you may want, but Michael Grayling (http://www.stata.com/meeting/uk15/ab...5_grayling.pdf) has described this in his presentation "Efficient multivariate normal distribution calculations in Stata" at the 2015 UK SUGM as a part of illustrating the mvtnorm module.

[Jen Bell]

It's possibly something to work from - thank you John!

Any other suggestions still welcomed!

Jen