Bioequivalence and Bioavailability Forum

Dear Ohlbe!

❝ I think the idea would be rather to report the repeated result the way Canada used to request it: the first value obtained should be used for PK calculations, and there should be a table in the report showing the initial and repeat value for random repeats.

❝ Now the question is, what do you do with such a table ?

Oh, things get clearer now!
This is the Canadian approach – the way we have interpreted and applied it for many years – without any actual use of the data afterwards.
We split the concentration range of the unknown samples in a such way, that the medians of the found concentrations were just about the three QCs-levels. Then we calculated the CV% of replicates within these three ranges.
We always simply looked at this table, waiting for some ghost to be released out from the lab-notebook explaining us what the difference means. Actually in thousands of samples there was no obvious difference at all (sometimes the CV was higher, sometimes lower; did not got any clue out from it).

❝ ❝ I always suggest to include two adjacent values, or - if the suspect values is the first/last one in a profile - two succeeding/preceding ones.

❝

❝ I agree with you. We used to do this in a lab where I worked some years ago, but I can't remember seeing it done at any of the labs I have visited since.

I’m observing some decline in scientific standards.
I know only two labs any more following some basic plausibility assessments – of valid batches – before releasing their data.
IMHO the ‘lights on, nobody at home’-attitude becomes increasingly fashionable. Many analysts in recent years are so scared, that they forget (or never were taught about) basic scientific principles.
No, following SOPs is no warranty for anything (except getting a warm handshake from an inspector).
Sometimes I’m getting the impression, that all our validation efforts during the past 20 years – at least to some extent – are nothing more than a ‘ship painting activity’, both for regulators and the industry.
If we would be really and honestly interested in confirming our results, we should not spend a single cent in validation, but analyze all samples with a second method, which is entirely independent from the first one (different sample extraction, separation, detection principle).
I’m talking about a true experimentum crucis (let’s say a RIA instead of LC/MS), not this fiddling with robustness (RP8 instead of RP18, slight changes in mobile phase composition, blahblah).
I bet on seeing some fascinating results leaving SOP-driven validation maniacs just speechless…