Method validation accuracy

[foodchem]

It is acceptable to determine method accuracy by comparing the results to the historical mean obtained by the method under validation rather than against a known reference standard?

[tom jupille]

That question is a bit too broad to allow a simple answer.

Is this an existing method that you run regularly or are you re-validating an old method or are you validating a new method?

Is this a "regulatory" method (i.e., one that will have to meet FDA (or equivalent) scrutiny?

What are you using as calibrators (standards) on a day-to-day basis?

[foodchem]

This is a retrospective validation for a protein (P) used for regulatory compliance. Accuracy is determined by using known conc. of P and calculating the conc. against a standard curve of BSA. However, the known conc. of P is determined based on the historical assay results. There has been no independent verification of P’s value. Can accuracy be properly assessed is the known value of P has not been independently verified?

[tom jupille]

Well that helps provide the basis for a short answer:

"No".

As I read ICH / FDA, you either a known standard or comparison with an independent (known to be accurate) technique.