Print

Print


Many thanks to all those who replied to my query on interim
analysis.

Lada Smoljanovic wrote:

The following reference might help you:
S.C.Choi (1997). Interim Analysis and Early Termination of
Clinical
Trials. Journal of Biopharmaceutical Statistics, 7, 533-543.

--------------------------------------------------------
Stuart Barber replied:

The density function of the test statistics is multivariate
(and often assumed to be multivariate normal).  Evaluating
it usually
requires repeated numerical integration; an efficient
algorithm is given in

Jennison, C. (1994).  Numerical computations for group
sequential
tests.  In Computing Science and Statistics vol. 25,
(Tartar, M., & Lock, M.D., eds), pp263-272.

This information also appears in chapter 19 of Jennison, C.
& Turnbull, B.W. (2000).  Group sequential methods with
applications to clinical trials.  Chapman & Hall / CRC:
London.

I believe Chris Jennison has made some FORTRAN software
available at his web page:

http://www.bath.ac.uk/~mascj

--------------------------------------------------------
Jon Cooke replied:

I looked into the subject of Interim Analysis a few months
ago and found the following texts very useful.

Group Sequential Methods with Applications to Clinical
Trials by Chris Jennison and Bruce Turnbull
The Design and Analysis of Sequential Trials by John
Whitehead

Both are excellent texts, but do require a high degree of
understanding of mathematics if you want to understand the
calculations.
Most people use tables to find alpha values as the
calculation to find them involves recursive numerical
integration. There are computer programs to evaluate them
(EAST and PEST), but they tend to be very expensive.

To answer your original question. I believe that a normal
distribution is assumed for the data hence Z values and
would therefore have the standard density function for the
normal distribution (I think that is the case anyway). You
may want to check in the book references I gave.


~~~~~~~~~~~~~~~~~~~~~~~
Jochen Weber, PhD
Dpt. of Biometrics
IMFORM GmbH,
Birkenweg 14, D-64295 Darmstadt, Germany
++49-6151-8280-23 (tel)
++49-6151-8280-10 (fax)
~~~~~~~~~~~~~~~~~~~~~~~




%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%