By Anonymous on Monday, March 29, 2004 - 07:55 am:
Group,
I have a question about method validation. The ICH guidelines mention that six determinations of 100% concentration is required for precision.If my method requires duplicate sample preparations, does that mean I have to do 12 sample preparations to meet the six determinations requirement?
Thank you.
By labcat on Monday, March 29, 2004 - 11:37 pm:
In my opinion, yes.
The reason is simple: this experiments is aimed to evaluate the scattering of your results if you perform several times your "final" measurement. The duplicates could be considered as an "intermediate" result.
So, if the measurement is made with two replicates, you have to calculate the mean value of six measurement, each made of two replicates.
By DR on Tuesday, March 30, 2004 - 09:54 am:
I agree - but I have a question for your method author (and other fans of duplicate preps) - WHY?!?
Duplicate preps are of no use.
Either go with standard RSDs & rē values and trust your analysts & systems or use triplicate preps so you can calculate %RSD on the results.
Duplicate preps should only be done to confirm an original result that looks odd (imho).
By Anonymous on Tuesday, March 30, 2004 - 04:48 pm:
I think I would disagree with that last point to a certain extent. When laboratories make duplicate injections they generally have an SOP which stipulates what is the acceptable percent difference (at least they should). Then in cases where the two injections do not agree within this specified tolerance, additional injections would be made. And one of the values may be invalidated.
I agree that generating two results that don't agree with one another is a dead end. But if additional injections are made it's a different story.
By Tom Mizukami on Wednesday, March 31, 2004 - 10:37 am:
Anon 3/29,
Method precision based on six replicates at the 100% level and duplicate sample preps are separate issues.
The reviewer guidance Validation of Chromatographic Methods recommends at least 10 injections for valiadation. http://www.fda.gov/cder/guidance/cmc3.pdf
Single sample precision is measuring the performance of the method and instrument and is explicitly not taking into account sample preparation. I would just make 10 replicate injections of one of your samples and call this your injection repeatability study. If you are going to use more than one sample I guess you could look at the precision of the response factor but I think this is confounding injection repeatability with an element of analysis repeatability. Good luck.
By Anonymous on Friday, April 23, 2004 - 02:06 am:
What is cramer's rule ?
How height is calculated in first order derivative spectroscopy?
What is Q-quntitative method ?
How Sandell's sensitivity is calculated
By Anonymous on Friday, April 23, 2004 - 11:50 am: