Acceptance criteria of comparative method transfers
There are no requirements regarding the sample size (= number of determinations as well as number of samples). A single batch is sufficient according to USP <1224> and there are recommendations in the WHO Guideline 961, Annex 7. It’s important to select a sample size where reliable results can be produced.
In case of a comparative transfer, it needs to be demonstrated that both laboratories have equivalent results. There are different strategies and statistical tools for the definition of the required acceptance criteria. For example, absolute values can be compared, a percentaged deviation from the average value or a relative standard deviation (RSD) can be determined, significance tests (t- & F-test) consulted or equivalence tests according to USP <1010> be implemented.
The decision how the acceptance criteria are defined is based on the complexity of the method. In this process it’s helpful to implement a risk analysis at the beginning of the planning phase. In case of less complex methods in which scatter is assessable, it can be sufficient to compare the average values of the results with the limit values of the specification (= absolute limit). The previous validation of the donor site (DS) as well as the results from the familiarization trials or even experiences with similar methods can be used for orientation to assess the statistical spread.
In case of complex methods - if the scatter during implementation is unknown or can’t be assessed accurately enough – it’s recommended in a position paper of different pharmaceutical companies to use equivalence tests.