Hi ! I'd appreciate your thoughts on this.
Say you were involved with setting up a small trial that looked at the effect of a drug on loss of weight of the participants. This would be an RCT with two arms: one treatment arm and one placebo arm. Say the participants were losing weight, generally, as a result of illness, and that there were two stages of the illness. Participants would be expected to lose a lot more weight if they were Stage II rather than Stage I, and so the decision is taken to stratify by Stage. So participants are randomised in a stratified way, with one randomisation list for Stage I and a second randomisation list for Stage II.
Putting modelling to one side for now, an analysis of the above design might be to perform two t-tests of the change in weight across the two arms, one for Stage I and the other for Stage II, separately. Analysing the two Stages separately would, however, dramatically decrease the power of the study. Is it permissible to analyse with a t-test that compares the change in weight on the treatment arm with that on the placebo arm, ignoring the stratification ? (I've come across a maxim (in the field of cluster randomisation) that says, "Analyse as you randomise", which would argue against doing this.)
This is a hypothetical example, chosen to have a continuous outcome, so that I don't think that we are in the province of the Mantel-Haenzel test. Could you suggest how one might proceed to analyse with the above scenario ? References ?
TIA,
Martin Holt
Medical Statistician
Southern Derbyshire Acute Hospitals NHS Trust
[log in to unmask]
********************************DISCLAIMER********************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. However, any views or options presented are
solely those of the author and do not necessarily represent those
of the Southern Derbyshire Acute Hospitals NHS Trust. If you are
not the intended recipient or the person responsible for delivering
the email to the intended recipient, be advised that you have
received the email in error and that any use, dissemination,
forwarding, printing or copying of this email is strictly prohibited.
SDAHTwill not be liable for direct, special, indirect or consequential
damages arising from alterations of the contents of this message
by a third party or as a result of any virus being passed on.
If you have received this email in error please contact the SDAHT
administrator. Email or telephone on 44 (0)1332 340131 ext 5777
*******************************************************************************
[log in to unmask]
|