Re: {MEDSTATS} Powering new studies based on old event rates...

[John Whittington]

At 05:02 24/05/2012 -0700, irene stratton wrote:
>Some time ago I was sent a paper to review which was about powering
>studies for an outcome in diabetes. Due to landmark trials in Type 1
>and Type 2 diabetes published in 1990s control of hypertension and
>hyperglycaemia is now better than it was. Hence trials powered on
>event rates in previous studies end up being underpowered because
>event rates are now lower. I can't find this paper and wonder if
>anyone else knows of something like this.

A couple of thoughts. Firstly, if one actually _knows_ (i.e. has adequate
evidence) that event rates 'are now lower', then one should presumably use
those more recent 'known' event rates for determining power.

Secondly, is it not probably the case that, in practice, a study 'powered'
on the basis of too high event rates will only be 'underpowered' in
relation to the detection of the same absolute amount of difference/effect
- in which case one might consider whether that was a reasonable
expectation if event rates are generally lower?

Kind Regards,


John

----------------------------------------------------------------
Dr John Whittington, Voice: +44 (0) 1296 730225
Mediscience Services Fax: +44 (0) 1296 738893
Twyford Manor, Twyford, E-mail: Joh...@mediscience.co.uk
Buckingham MK18 4EL, UK
----------------------------------------------------------------