Back to the Top
Dear All
Does anybody know of a guideline which gives the method for handling
outlier QC's during subject analysis
At the Method Validation stage the outlier QC's are checked using
Dixon's outlier test before excluding from precision and accuracy
calculations
Should the same procedure be followed for study samples also ?
Regards
Nilima
Back to the Top
Dear Nilima,
As outline in the FDA Bioanalytical method validation, page 10 second
to last paragraph, all QC (including outliers) should be reported.
What we have done lately is to prepare two tables, one with all QC's
included and below the same table without the outliers.
Hope this may help you,
Kind regards,
Sylvain Mandeville, Ph.D.
LAB Research Inc.
sylvain.mandeville.at.labinc.hu
www.labresearch.com
Back to the Top
Dear Nilima,
You can find a reference for outliers test in ANVISA guidelines
(i.e. Grubbs-1 test). one more thing i would like to mention is that
this Grubbs-1 test works only for single outlier in a data set. If
you have more then one outliers, that may be either on higher and
lower ends or both on same side, you can apply Grubbs-2 and Grubbs-3
test. For details you can refer " Missing values,outliers, robust
statics & non-parametric methods" by Shaun Burke, LC-GC Europe.
I agree with Sylvain Mandeville that FDA guidelines state that
outliers can be excluded from a data set ( As per a predefined SOP
criterion) and the results of both with and without outliers should
be reported in final report.
with due regards,
Kuldeep Sharma
Back to the Top
The following message was posted to: PharmPK
Dear Sylvian,
What we have observed in recent audits that auditors recommend that
Precision and Accuracy of QC should pass even after including outlier!.
If we have to exclude outlier QC value then a strong analytical
justification is required.Please comment
Dr Ashish
Back to the Top
The following message was posted to: PharmPK
Dear Dr. Ashish,
you wrote:
>What we have observed in recent audits that auditors recommend that
>Precision and Accuracy of QC should pass even after including
outlier!.
>If we have to exclude outlier QC value then a strong analytical
>justification is required.
>
Would you please specify 'auditor', e.g., coming from (which) authority,
a sponsor, or an external 'quality management'-company on behalf of the
sponsor.
If ALL QCs should have to pass, we also can throw away the guidelines on
bioanalytical validation from 1991 onwards (yes, I've been at the
Crystal City Conference in Arlington 1989, it was a nice experience -
screaming included).
It's good scientific practice to expect the unexpected. There always
will be situations, we cannot explain ('strong analytical
justification').
We are not living in a deterministic universe.
IMHO throughout the last decade an increasing number of 'auditors' made
a stage appearance with only little scientific background, but being
really strong in sticking on excessive formalism ;-)
--
Helmut Schuetz
BEBAC
Consultancy Services for Bioequivalence and Bioavailability Studies
Neubaugasse 36/11
1070 Vienna/Austria
tel/fax +43 1 2311746
Web http://BEBAC.at
BE/BA Forum http://forum.bebac.at
http://www.goldmark.org/netrants/no-word/attach.html
Back to the Top
Dear Dr. Arnish,
When dealing with outlier, there is normally an issue with that
particular sample such as an extraction issues, injection issues,
etc. This is why when faced with those situation we look at the
possible cause of errors.
As I said previously, we always include all values (which are outside
the +/- 15% acceptance criteria) in the statistical evaluation and if
the CV % is within 15% , then all values are reported and the outlier
test is not performed. It is only when you observed a value such as
1.5 or 2 times higher or lower) than normal that this QC value become
suspicious (would become an outlier after the Grubbs test) and can be
concluded that there is something peculiar with this QC sample.
I personally never been involved in such a situation (i.e. sponsors
requiring that a validation much pass even with outlier) but maybe
other members have and they can share their experience in dealing
with such recommendation from the sponsors.
I hope this may help,
Regards,
Sylvain Mandeville, Ph.D.
LAB Research Inc
sylvain.mandeville.aaa.labinc.hu
www.labresearch.com
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)