Hi - I wonder whether there is any implementation of the Whitehead double
triangular boundary in S that is available to be downloaded (too bad that
the beta version of SeqTrial is no longer available). I'd appreciate your
suggestions and hints. The details of what I want to accomplished is as
follows:
I would like to analyse a clinical trial by the Lan-DeMets "alpha-spending"
method (actually, I already have done this analysis) and Whitehead's
Christmas Tree method, to compare the conclusions. There are two
peculiarities that make the calculation difficult:
1) Assymetrical tails: The Lan-DeMets calculation is performed with a test
size of 0.025 in the upper tail and 0.10 in the lower tail (to ensure rapid
stopping for harm). Thus, I need an unusual type of Whitehead double
triangular design for the comparison.
2) Premature stopping: The trial was halted by the DSMC because of low
recruitment. In the Lan-DeMets procedure, we were able to handle this by
spending all remaining alpha on the final look (we have some simulation
results and power calculations by which we can judge how much damage this
did). Thus, in the Whitehead approach, we have to terminate the analysis
with too little information to guarantee closing out the indifference
region in the upper tail.
Thanks.
Martin Ho
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
|