Well, I would argue that validation of SaMD, as with any software, is always required to some degree or other. After all, validation is basically "demonstrated proof the software/device meets the needs of the user" and that seems fundamental, particularly when the entire device IS the software. Why would you not want to do that? Obviously, automated test can do most (all?) of your verification activity, but validation requires actual users or their surrogates.
Now, that said, specifics matter when determining what comprises sufficient validation. If your release is an initial product, generally a number of actual users should test the complete function of the device and assess that it met the "user needs" or "voice of the customer" (if you do systems engineering approaches). However, if the release is a small patch, you may be able to leverage the previous validation. And if it is a single new feature, the validation could be customized to that feature alone.
I encourage you to focus on the intent of the standards/regulations, and figure out what makes sense to assess - then go to the standard and see how it fits and aligns, rather than trying to use it as a "cookbook" or the "maximum" effort that is needed.
Ginger
------------------------------
Ginger Glaser RAC
Chief Technology Officer
MN
------------------------------
Original Message:
Sent: 25-Feb-2024 04:53
From: Anonymous Member
Subject: SaMD-validation
This message was posted by a user wishing to remain anonymous
Hello RAPS,
Could you please share your thoughts when the validation of SaMD is necessary and how does the clause 7.5.6 of the ISO 13485 apply (if)?