Back to the Top
Hi
Kindly guide me about T/R Ratio.
When we anaysis data for pilot study, we need to estimate sample size
based on this pilot analysis. I use SAS program to estimmate sample
size. When I use this prog I need to specify alpha, beta, intra
subject cv, T/R ratio and confidence interval.
For estimating sample size I take T/R ratio as 0.90,.95 or 1.00.
Can somebody thro some light, which T/R ratio to be used while
estimating sample size, the one which we get from analysis or those
which I have specified above?
It will be greatful if somebody gives the reference for this.
Regards,
Shalaka Kale
Statistician
Back to the Top
Hi Shalaka,
For sample size estimation of pivotal studies, one would normally use
the CV% from the pilot results but better to use the T/R ratio of
0.95. Though, you should take into consideration your actual ratio
and see how this affects your sample size estimates. By using 0.95 as
the ratio, you will be assuming that the two formulations will be
within 5% of each other. If the ratio from your pilot results is
quite different from one then this suggests a problem with the
formulation. Also, note that as you deviate from ratio of 1 (higher
or lower), the sample size will increase dramatically.
Hope this helps!
Nav
--
Nav Coelho, M.Sc.
Biostatistician, Pharmacokinetics and Statistics
Biovail Contract Research
Tel: (416) 752-3636 Ext. 369
Email: navdeep.coelho.aaa.biovail.com
URL: www.biovail-cro.com
Back to the Top
The following message was posted to: PharmPK
Hi Shalaka!
>For estimating sample size I take T/R ratio as 0.90,.95 or 1.00.
>Can somebody thro some light, which T/R ratio to be used while
>estimating sample size, the one which we get from analysis or those
>which I have specified above?
>
It depends ;-)
Even from a reasonable sized pilot study, estimates of CV are less
reliable than the T/R ratio.
If your main study will use the same batches, I will suggest using the
ratio from the pilot study.
But: since the CV of such importance, I would use not the calculated
value itself, but some upper confidence limit.
If your main study will not use the same batches, it's up to you.
Since batches of both test and reference may be released with an actual
content of +/- 5% of nominal, you may choose whatever you like, but 0.90
is a conservative - but not unreasonable - assumption based on differing
content. In my experience most people go for 0.95.
Some guidelines (Canada, WHO) suggest using a ratio of 1 in sample size
estimation and allow for a correction of the point estimate for the
measured content.
Good luck!
Helmut
--
Helmut Schuetz
BEBAC
Consultancy Services for Bioequivalence and Bioavailability Studies
Neubaugasse 36/11
1070 Vienna/Austria
tel/fax +43 1 2311746
http://BEBAC.at
Bioequivalence/Bioavailability Forum at http://forum.bebac.at
http://www.goldmark.org/netrants/no-word/attach.html
Back to the Top
The following message was posted to: PharmPK
One which you have got from analysis of pilot data.
Mitesh Gandhi
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)