Back to the Top
The following message was posted to: PharmPK
Sir,
I want to know about the sample size for a three way cross over study
in human. Three treatments are to be given at an interval of 20 days.
Isn't 5 volunteers be sufficient?
Swati Kotwal,
Department of Biochemistry
RTM Nagpur University
Nagpur
India
Back to the Top
The following message was posted to: PharmPK
s.o.o
In statistical terms it is not sufficient however it will give you
the data you want. Whether you use 20 volunteers or 5 volunteers the
data you get will be basically similar.
One of the challenges in Biovailability/Bioequivalent studies is
getting volunteers so where you have failed to get more than five
volunteers you can go a head and do the study but handle the issue of
no of volunteers in your discussion.
Back to the Top
The following message was posted to: PharmPK
Swati
A fundamental relationship equates the confidence of a given trial to
the ratio of the magnitude of the signal to the magnitude of the
noise, multiplied by the square root of the sample size. Deciding
ahead about a desired confidence level, particularly with biological
systems, all boils down to estimating as good as possible the signal
versus the potential noise. So although your study design (just
briefly described) seems to me perhaps not the best one, as far as
sample size you must look at the intrinsic sources for variability,
about your drug first, and about the physiology second.
Luis
--
Luis M. Pereira, Ph.D.
Assistant Professor, Biopharmaceutics and Pharmacokinetics
Massachusetts College of Pharmacy and Health Sciences
179 Longwood Ave, Boston, MA 02115
Phone: (617) 732-2905
Fax: (617) 732-2228
Luis.Pereira.aaa.bos.mcphs.edu
Back to the Top
The following message was posted to: PharmPK
Folks:
Re: Sample size in BE studies
Apart from the statistics there is the issue of the effect of an
outlier and
indeed whether or not you want one in the study. One argument is that
for
predicting safety the study should ideally detect at least one so
that the
PK profile can be measured accurately and compared reliably with the
average
metaboliser. The counter argument to that is that they produce wide
group
variances and so prevent a statistically valid proof of
equivalence. In my
view the counter to that is to take the outlier out of a calculation
with
the explanation for doing it and moreover to include it in the protocol
statistical plan.
Traditionally in the UK we have used 12 subjects per group but in the
US it
has been 24, with the aim I think, of increasing the odds of
including an
outlier. I am not sure that doubling the size makes a big enough
difference
to justify the cost and exposures of volunteers, though increasingly
we do
seem to be using 24 anyway.
If anyone feels like preparing one, I would be interested to see a
table or
graph showing the group size required to detect one outlier with an
probability of 95% and possibly a second graph doing the same for
detecting
two outliers.
Andrew Sutton
Guildford Clinical Pharmacology Ltd.
The Technology Centre, Occam Road
Guildford, Surrey, UK. GU2 7YG
Tel: +44 (0)1483 455375. Direct: 688303
URL: www.gcpl.co.uk
Back to the Top
Res. Sir/Madam,
is it possible to determine a sample size for BE study when mean,sd,%
cv are unknown? also pilot study and any previous data is not
available.so is there any procedure or formula to calculate sample size.
help me
thank you,
sushil A. Kale
Back to the Top
The following message was posted to: PharmPK
Sushil,
In this case, you need either run a pilot study or follow FDA's
guidance to have a minimum 12 subjects assuming you are using a cross
over design.
CQ Deng, PhD
Talecris Biotherapeutics, Inc.
Research Triangle Park, NC, USA
Back to the Top
The following message was posted to: PharmPK
Sushil,
In this case, you need either run a pilot study or follow FDA's
guidance to have a minimum 12 subjects assuming you are using a cross
over design.
CQ Deng, PhD
Talecris Biotherapeutics, Inc.
Research Triangle Park, NC, USA
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)