Is a DOE’s confidence interval statisticlly sufficient for an OQ or is the DOE only used to find the operating window.
In my opinion, it could be either!
As that is not helpful … it depends on what confidence you want to have, what confidence the DOE delivers (and whether these match up), and what your site policies and procedures say about the use of results / statistics.
We would use a good DOE to determine those experiments to be done at the OQ / optimisation stage. We have a procedure that states that provided the statistical quality meets our requirement, then the results can be used directly in the PQ stage (allowing for calibration).
Cat