As per the draft guidance document on process validation issued by FDA in Nov '08, one of the recommendation is to continue ‘monitoring’ & ‘sampling’ at same levels used during initial process qualification stage until sufficient data is available to generate significant variability estimates.
My questions - i) How many batches should generally be monitored & sampled after the initial qualification?
ii) What is the simple meaning of ‘to generate significant variability estimates’?
iii) How many batches should typically be considered for the initial process qualification? Manufacturers were typically studying 3 batches till now, but I am not sure if that would that be sufficient to demonstrate control of variability…
iv) Also, the draft requires ‘application of statistical metrics for defining inter-batch variability’ - what statistical metrics should be used? RSD for key analytical parameters across the validation batches would be sufficient?