I got a question from a customer and I ask him to post to community too. However may you have some insights or literature to recommend too (on bold topics):
"I am currently in DSP development team and working on the scale down model qualification. This is the first time we will do this exercise.
To qualify this scale down model, our CMO propose to perform at small scale 3 runs in duplicate using material from 3 different batches at large scale.
We would like to know, based on your experience, if this is an industry common practice to use material from 3 different batches (“must-have”) ? or if to demonstrate statistically the scale equivalence, we can consider that point as a “nice to have” ? As you understand at this moment, only material from 1 batch at large scale is available to perform run at small scale.
We are trying to find some literature regarding this point in order to use as support. Please, do not hesitate if you have some literature or feedback regarding this specific point or about statistical approaches for scale down model qualification."
I guess the answer is it depends ... (e.g. on general variability of the material from batch to batch), however you might have a good reference to direct him to.
Thanks for any suggestions.
