PharmPK Discussion - Outlier in bioequivalence

PharmPK Discussion List Archive Index page

On 4 Aug 2004 at 10:13:14, "Kanimozhi" (kanimozhi.-at-.gvkbio.com) sent the message

Dear all,
I have a doubt, when there is an outlier in bioequivalence, what we
have to do, do we have to remove that one from statistical analysis? .
It will be nice if anyone could suggest me.
Thanking you,
Kanimozhi.A

Back to the Top

On 4 Aug 2004 at 13:13:20, "Paula Macedo Cerqueira" (Paula.Cerqueira.aaa.anvisa.gov.br) sent the message


According to Brazilian Guidances, you have to present two statistical
analysis: one including and the other one removing the outlier.
Paula

Back to the Top

On 4 Aug 2004 at 12:05:32, "Bonate, Peter" (pbonate.-a-.ilexonc.com) sent the message


This topic is covered in the FDAs Guidance on Statistical Analysis of BE
studies (http://www.fda.gov/cder/guidance/) and in the EMEA guideline on
BE (http://www.emea.eu.int/pdfs/human/ewp/140198en.pdf).
Removing an outlier in a BE study is a tricky question.  Unless you have
a good reason, such as the subject vomited shortly after ingestion, then
I would make the argument that you cannot remove the subject.  The FDA
guidance discourages outlier removal for nonreplicate designs.  For
crossovers you need to look at all periods.  The EMEA guidance
discourages post-hoc removal of outliers.  You need to define a priori
how outliers will be defined and how they will be handled.
If you do, however, remove the outlier you need to present the analysis
both with and without the outlier to make the analysis ICH compliant.
Hope this helps,
Pete bonate

Back to the Top

On 4 Aug 2004 at 10:56:53, "Osama Mohamed" (mohamedo.aaa.ohsu.edu) sent the message


Dear Dr. Kanimozhi:
For outliers strategy in statistical analysis, it is recommended that
you do your analysis with and without the outlier cases and see if they
make a significant difference in your conclusion or inference. if they
make a difference, you can remove the outliers from your analysis but
you should report them with a logical suggested explanation about why
outliers had these different results (e.g they are of different race or
if the results were analyzed by different apparatus etc.).
Good luck,
Osama
Osama H. Mohamed
Ph.D. Candidate in Clinical Pharmacokinetics
College of Pharmacy, Oregon State University

Back to the Top

Want to post a follow-up message on this topic? If this link does not work with your browser send a follow-up message to PharmPK@boomer.org with "Outlier in bioequivalence" as the subject

PharmPK Discussion List Archive Index page