During IQ/OQ/PQ training, it was mentioned to use 3 production sized runs as sample size for PQ. I have tried to find a document that specifies this. Is anyone else aware of how/where this requirement originated? The GHTF guidance document refers to ensuring production conditions and challenging the process enough times to assure the results are meaningful and consistent. I think that running 3 trials in production quantities is overkill. I see more value in running exhaustive testing which would invoke more variation than would normally be seen during a production order. We use formal DOE during OQ which determines our approved operating window. If we were to test x hours of production at y settings including cold start, run, shut down, at 3 replications, I would be comfortable that the window will be able to produce good parts in the production environment.

Any help is appreciated.

Not sure where or if it is documented formally, but I know I have seen it in writing somewhere. The FDA requirement is 3 successive lots must be acceptable and European requirements are as many as 5 successive lots.

I agree that increasing sampling within a lot will give increased assurance of product quality, but it is lot to lot variability that needs to be consistent.

Though your DOE did identify the critical parameters that need to be controlled, the PQ validation confirms at the large production scale that the process is robust to the window.