Protocol Deviation Handling – A Premier Validation Template


So how do you handle all of those nasty deviations when you start to execute your Validation Protocols? Have you got individual deviation handling forms to account for deviations or do you just handle the deviations at the back of the protocol. Most companies have their own way of handling deviations, some prefer to have separate deviation handling forms while other like to have deviations as part of the main protocol.

I’ve seen this ‘deviation’ process handled both ways. To start, I hate the term deviation in a non-validated system. To me those nasty deviations should be called incidents, observations, descrepancies…anything but deviations. If you call the errors ‘deviations’ and enter them into your deviation process, it provides a direct link back into your validation activities. I STRONGLY suggest containing the validation process by using incident reports until the system is released then using the deviation process.
My nightmare experience was with a state run lab when I was performing software validation. They insisted on entering every minor incident as a deviation. 108 deviations later, the validation was complete. I wouldn’t want to explain to an auditor why there were so many issues with the validation process and state of control around the validation (some incidents were IT changing HW/SW during the OQ/PQ).

good point Gregg, sounds very similar to a project I was working on. I like the term incident instead of deviation it doesn’t sound as bad.

Good suggestion for the next project!

We run a system where we have a local deviation system contained entirely within our validation system - and not linked in any way to our site deviation system.

So we could get a high number of deviations against a single valdiation exercise potentially - ie differences between the approved protocol and the actual work / results - but they would reside only within that validation exercise.

As an example, if the validation protocol says “press the blue button to do x” and actually what’s present on the machine is a green on-off switch which does x we would record that as a deviation from the protocol. There’s no impact, other than to consider updating the procedure for operators so they know that to do x they use the green on-off switch (not a blue button).

Because we record these minor differences as deviations, we can get several per validation - but we think that’s OK, and defendable in any audit.

I don’t think it’s an “ideal” - but it does work for us as a company. (If anything, we over-record rather than under-record - bit of a waste possibly, but minimal regulatory risk).


The link to this article is no longer active. Can you please reactivate?