One aspect of the GHTF guidance (which I otherwise find very valuable and informative) that always sat wrong with me was that while it was excellent at explaining how to do a validation, it (IMO) confused the concepts of Design Requirements with Production Requirements, especially with this line (emphasis mine):
For some types of devices, there can be many steps that get to "individual device is made per the DMR as is working per the validated design of the DHF"... and along the manufacturing path there can be all sorts of intermediate process checks for things like "gross assembly defects" that I feel don't always need to be "validated", because:
The guidance is circa 1999 (early in "design controls" for medical devices), so I can somewhat forgive the point... but a strict reading of the guidance could easily lead people to argue that (my simplification) any process in manufacturing that results in an output that is supposed to conform to a specific individual design requirement but doesn't have that a process output from that requirement explicitly measured and recorded means that the process has to be validated.
For some types of devices, there can be many steps that get to "individual device is made per the DMR as is working per the validated design of the DHF"... and along the manufacturing path there can be all sorts of intermediate process checks for things like "gross assembly defects" that I feel don't always need to be "validated", because:
- Generally, good design requirements aren't expressed as "negative requirements"
- The process checks are often for things explicitly NOT in the DMR (i.e. there isn't a DMR for non-conforming product)
- The process checks are often to reduce the burden of rework and retesting because of sequential tests (so business requirements more than product safety requirements)