Factorial trials are routinely analysed to investigate the main effects of
interventions and interactions between them, the latter albeit often with
reduced power. We are presently working for a DMC on a 2*2*2 factorial
trial and have been discussing whether this is a suitable approach for the
DMC. Our concern is that we need to look at each of the 8 'individual'
groups for safety reasons, so that we can quickly ascertain whether a
particular combination of treatments is harmful. Again, fully accepting
that statistical power will be compromised - our first look will be after
outcome is available for 1000 patients.
We wonder if other people have experience of this problem and can offer
advice or point us to some key references.
Many thanks in advance,
Ed Juszczak.
**********************************************
Ed Juszczak
Medical Statistician
ICRF/NHS Centre for Statistics in Medicine
Institute of Health Sciences
Old Road, Headington
OXFORD OX3 7LF
UK
e-mail: [log in to unmask]
Tel: 01865 226794
Fax: 01865 226962
www: http://www.ihs.ox.ac.uk/csm/
*********************************************
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|