Back to the Top
Dear All,
I would like to know how to calculate outlier in bioequivalence study?
Before few month the same question was there in this discussion but
"Pharmaceutical Statistics by Bolton" for outlier is not that
muchsatisfactory. I have seen the outlier calculation in other book by
Shein-Chung Chow but still I need few more research paper and articles
on outlier for perticular bioequivalence and method validation. It will
really be appreciated ifanyone can help me.
Thanks
Priti, Biostatistician
Aurobindo Pharma Ltd.,Hyderabad, India
Back to the Top
The following message was posted to: PharmPK
Priti,
Go the FDA guidance index page (see the URL below).
Search for the guidance titled: "Statistical Approaches to Establishing
Bioequivalence" dated January 2001.
Go to page 14 of the guidance and read paragraph C titled: "Outlier
Considerations".
Since the existence of a subjects outlier could be due either to
product failure or to subject-by formulation interaction, if you do not
have a replicate design and no protocol deviations, you will have to
agree with the Agency on how to handle the outlier.
As a general rule, the exclusion of outliers, particularly in
nonreplicated designs, is discouraged.
I hope hat this helps.
http://www.fda.gov/cder/guidance/index.htm
Stefano Persiani, PhD
Director, Drug Metabolism and Pharmacokinetics
Rotta Research Laboratorium SpA
Via Valosa di Sopra, 7-9
20052 Monza (MI)
ITALY
Tel. ++39-039-7390-318
Fax. ++39-039-7390-371
e-mail: Stefano.Persiani.-at-.rotta.com
Back to the Top
The following message was posted to: PharmPK
Dear Priti
Are you speaking about subject outlier or any sample outlier?
Thanks
Bapuji AT
Group Leader- Bionalytical Research
Ranbaxy Labboratories
Back to the Top
Dear Mr. Stefano Persiani & Mr. AT Bapuji,
Thanks for your reply. I have gone through 2001 FDA statistical
guidelines, in which they have not mentioned how to determine outlier
from data. When to apply outliers test that I know. But method for
outlier for bioequivalence I need. Even in the “Pharmaceutical
Bioequivalence by Peli G. Weling, Francis L.S. Tse, Shrikant V. Dighe”
Page 371-372 also indicate when to apply outlier test but my question
is “what is the method to detect outlier?” I want to know within
subject outlier (for pharmacokinetic data) as well as QC sample value
outlier in method validation.
Regards,
Priti
Back to the Top
The following message was posted to: PharmPK
Dear All,
i am interested in knowing the source of published
information on Biopharmaceutical Classification of
drug substances, other than the FDA Guidane documents
Thanks,
Back to the Top
The following message was posted to: PharmPK
Dear sir,
we have small statistical problem in one BE study, and this
study
having 46 subjects data with this data BE acceptance criteria is
not
meeting due to six subjects showing test formulation
concentrations
were very low when compared with the reference formulation
concentrations.
If we are excluding those six subjects data this will meeting
acceptance criteria. So, how can we consider those six subjects
data
as outliers? what is procedure for cosidering as outliers?
Please give
us some suggesions immediately, we are eagurly waiting for your
reply
on this issue.
with regards
M. ram babu
Scientist
INDIA
e.mail: mrbabu.aaa.rediffmail.com
Back to the Top
Dear Mr. Ram Babu,
I am afraid you cannot just regard 6 out of 46 (= 13%) as outliers.
Regulatory agencies (at least European/FDA) will start questions when
you do
that. So you must have a very good reason to ignore them. Maybe because
of
differences in your formulation, the drug is not metabolised or absorbed
very well by a certain group of patients.
Kind regards,
Berber Snoeijer
Biometric Support
The Netherlands
bsnoeijer.at.biometricsupport.nl
Back to the Top
The following message was posted to: PharmPK
Dear M. Ram Babu,
exclusion of outliers in BE studies generally is discouraged by
regulatory bodies. However, if you want to exclude outliers, according
to Good Statistical Practices you have to lay down your decision rules
in the statistical protocol (i.e., before start of the study).
We often apply something like the following:
o) detection of outlier/s (based on the studentized
intra-subject.residual errors from ANOVA on ln-transformed data)
o) if outlier/s are identified, check your source data in a backward
procedure: start looking at potential errors in data
transmission/transcription, continue with the analytical conduct of the
study (e.g., sample storage, problems with interferences in
chromatography, extraction problems...), and finally check the
clinical part (emesis after an oral application, diarrhoe after an MR
product...).
o) if you find anything explaining the discrepancy in responses between
treatments AND have stated that beforehand in the protocol, you may
exclude the subjects from evaluations and pray (there is absolutelly no
guarantee that authorities will follow your arguments).
o) if you do not find an explanation OR you did not state the procedure
in the protocol, I am afraid, nobody will accept exclusion of outliers
based on statistical grounds only. If you perform ANY outlier test to
alpha=0.05, it means in 1 out of 20 studies you will get a (false)
significant result...
btw, I think it will be rather difficult finding regulators willing to
accept the exclusion of 13% of subjects from the evaluation, even if
you could justify it...
Good luck
Helmut
Helmut Schütz Biokinet GmbH / Dept Biostatistics
Neubaugasse 36/11 Nattergasse 4
A-1070 Vienna/Austria A-1170 Vienna/Austria
tel/fax +43 1 9713935 tel +43 1 4856969 62
no cell phone ;-) fax +43 1 4856969 90
http://www.goldmark.org/netrants/no-word/attach.html
Back to the Top
The following message was posted to: PharmPK
Exclusion criteria have to be set as an SOP prior to data analysis. If
you don't have one in hand, then you are out of luck because you are
now trying to salvage something that should be redone.
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)