sample size for BA/BE studies [Regulatives / Guidelines]

posted by Helmut Homepage – Vienna, Austria, 2008-11-17 16:47 (6057 d 22:10 ago) – Posting: # 2695
Views: 8,502

Dear Muneesh!

❝ […] we have done a pilot study with 11 subjects and results from the pilot study are as follows:

❝ intra subject CV: 0.166203 %T/R Ratio for Cmax: 115.89

❝ Now my query is that should we consider only the Intra-subject CV or both the Intra-subject CV and T/R ratio observed from the pilot study or the expected T/R ratio (0.95-1.05) to calculate the sample size.


If you want to talk about an expected deviation of 5%, it's always better to go with 1-0.05, rather than with 1+0.05. In other words you get the same power for T/R-ratios of 1-0.05=0.95 and (1-0.05)-1=1.053.
Why do you expect a deviation of 5% after a pilot study where you have already observed almost 16%?
Not only the point estimate, but also the CVintra obtained in the pilot study carry some degree of uncertainty.

❝ Because if we consider both the Intra-subject CV and observed T/R ratio, sample size comes much more than what is coming from considering only the Intra-subject CV.


So if we take both numbers as carved in stone - which they aren't - we get n=62, right?
If you believe in a T/R-ratio of 0.95 you'll get n=14, but if you calculate an upper 80% confidence limit of CVintra you'll get 0.21596 and therefore n=24.
How will you convince an Ethics Committe to run a pivotal study based on a sample size estimation based on believes rather than your pilot study's results? With the actual PE from your pilot study you'll get n=104.
So that's a nice demonstration about the steepness of power curves if moving away from unity and the uncertainty of CVintra from small pilot studies.
I would reformulate.

❝ Please guide us what is the globally accepted method of calculating sample size.


Sorry, there's none. See the Guidance page for examples. As an example in the European Draft Guideline in the section 'Number of subjects' detailed specifications in the current NfG (variance, α, expected T/R ratio, power) are replaced by a laconic 'The number of subjects to be included in the study should be based on an appropriate sample size calculation.'
I see!
—
Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

Complete thread:

UA Flag
Activity
 Admin contact
23,424 posts in 4,927 threads, 1,676 registered users;
39 visitors (0 registered, 39 guests [including 5 identified bots]).
Forum time: 15:58 CEST (Europe/Vienna)

Philosophy, like medicine, has plenty of drugs, few good remedies,
and hardly any specific cures.    Sebastien-Roch Nicolas de Chamfort

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5