Dear all,
Dave Doggett asked whether there exists any
information regarding estimating sample size for
a meta-analysis.
While little has been written about this issue, to my
knowledge, Pogue and Yusuf have addressed the issue
most explicitly in the following article:
Pogue J, Yusuf S. Overcoming the limitations of
current meta-analysis of randomized controlled
trials. Lancet 1998;351:47-52.
This is a must-read for anyone interested in meta-
analyses.
To quote from their article, regarding sample size for
meta-analysis:
"...a reasonable assumption is that the sample
size required should be at least as large as that
of a single well-desigend and optimally powered
RCT....In fact, given the greater potential for
several additional biases and the heterogeneity...
in a meta-analysis compared with a single RCT,
more information may be needed. We propose
the concept of optimum information size (OIS) as
the minimum amount of information required in the
collective literature for reliable conclusions about
an intervention to be reached. OIS calculation can
be similar to sample-size calculations....Most
clinically important interventions are likely to reduce
the relative risk of major outcomes, such as
myocardial infarction, stroke, or death, by about
10-20%. On the assumption of reasonable event
rates for patients after an acute myocardial infarction
(such as a 5-10% annual mortality rate) and a type I
error rate of less than 0.05, these can be fitted into
a traditional sample-size formula for a z test
difference between two proportions. This formula,
commonly used to calculate sample size for RCTs,
an be adapted with the above assumptions to
provide an OIS for the meta-analysis. Although the
use of a single istribution formula may be questionable
for a meta-analysis that is probably the product of
several different statistical distributions, the OIS
provides a first approximation of the minimum sample
size required."
Hope this is helpful.
I would be interested to hear your comments on
the above recommendation.
Regards,
Janet E. Martin, BScPharm, PharmD
Project Leader
Evidence-Based Prescribing Initiative
London Health Sciences Centre
London, Ontario, CANADA N6A 4G5
tel: (519) 685-8500 ext 77814
fax: (519) 667-6811
[log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|