Back to the Top
I have two questions regarding deviations in time for collected
samples for bioequivalence studies:
1. Is there a specific statistical method or the sort for determining
significance in deviations in sampling times
2. How can such deviation if determined of significance be handled on
PK softwares like kinetica and winnonlin
Thanks
Raja' Sammour
QA Supervisor
Triumpharma
[Interesting question. If there are significant time differences
between subjects (and I think we all want the time to be recorded
accurately) how does a non modeling approach to BE statistics work
(AUC aside)? Modeling with Kinetica, WinNonlin et al. shouldn't be
affected as the actual times can be used. - db]
Back to the Top
[There will be some delay in messages being sent out for the next
couple of days, no need to resend your message - db]
The following message was posted to: PharmPK
We always use the relative sampling times (time between actual sampling
and actual dosing) and never work with nominal times. I would be very
interested to know how others feel about this. At the moment we are
migrating to Watson LIMS, and we noticed that the interface to Kinetica
only exports nominal times. I was very astonished about this: does it
mean that most colleagues use nominal times???
Regards,
Bert Meijer
Dopharma Research
Back to the Top
Further to the notes I=92ve received:
Since we use nominal times in our analysis would the following =20
approach be acceptable from both scientific and regulatory point of you:
For example,
With a sampling time : 0.25, 0.5, 0.75, 1.00, 1.5
A shift in time so that the 0.25 was withdrawn .at. 0.5 & 0.5 .-at-. 0.75, =20
0.75 .at. 0.75 and the rest was on time
Can we consider the 0.25 as missing sample and carry on using the =20
nominal times or regardless we should use the actual times no matter?
How do colleagues present their mean profiles using actual times?
When you do actual times do you actually enter 8:15 as 8:15 and not =20
0.25?
Thanks 4 all ur help & Best regards
Raja' Sammour
QA Supervisor
Triumpharma=
Back to the Top
Dear Bert,
I also agreed upon the same. Since we are plotting the graph and
capturing the log tranaformed data then it does not make any difference.
All the thing is we just have to have a PK data in a graph derived
manner. The importance is the time differenne between dosing and
collection of blood.
Hiren
Cadila
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)