Thanks for this very accurate post.
At Matrix Requirements, we offer our users a test plan so that *they* can validate our application.
We also deliver a validation plan when we upgrade their software (it's a cloud application) so that they can re-validate it.
This fulfills their obligation, and since the test results are documented inside our application they can show the inspectors how/when it has been done.
Yves
https://matrixreq.com

Richard Young Very good

Edwin Bills, ASQ Fellow, RAC As Yves indicated many software firms offer a test plan. But, it has to be customized for the "intended use", including the configuration that is actually installed at the customer site. Many software packages offer many fine features, but customers do not always use all of the features, so the parts not used do not need validation. This is but one example of the customization of the test plan. The actual test cases in the formats of the documents used by the customer are but one part of test cases that need to be developed by the software customer.

Dan Brown I agree with the entire article, especially the last paragraph that failure to perform effective validations of software is a very common 483 citation. The FDA is negatively impacting the industry by failing to provide us and its own investigators with clear, consistent expectations for what constitutes effective software validations. I have strong negative feelings towards the auto and aerospace industries, yet they do provide guide books for APQP, PPAPs, and FMEAs. Do these and follow the rules and your processes are validated and accepted for keeping cars running and planes flying. The FDA must provide similar clear guidelines for validation. I have read the “General Principles” document and it fails to explain what must be done to validate. Every company does it different and every investigator has different expectations.