Learnaboutgmp Community

Need Validation Advice

I have just been appointed validation manager of our site for Facilities, Utilities and Equipment.

We are under close scrutiny since we are a medical device manufacturer but I want to make sure we are still doing the right things and not just what we hired our paid Ex FDA consultant to tell us to do. From my understanding most of the FDAs understanding of validation is taking it from the Bulk Manufacture side. Meaning pharma. We build electromaechanical devices and test them before they leave. Does this truly warrant a (IQ, OQ,PQ) validation if you are using software designed for the equipment or for the testing to perform functional test?

To me and my engineering background we need to establish that we have installed the equipment properly that is used to test the devices and that our software is adequate to perform its desired function. I don’t really understand how a PQ comes into play if you have an operator walking through instructions repeatidly on an electromaechanical device? How do you have variation?


Yes of course, when you say you test what do you mean, how do you test?

Is your process validated? Are your PLC’s validated? Is any of the software validated?

Thats the IQ

That is used to test the devices and that our software is adequate to perform its desired function. I don’t really understand how a PQ comes into play if you have an operator walking through instructions repeatidly on an electromaechanical device? How do you have variation?[/quote]

How many devices are you making a day?

on 80% of our product we test, we actually use a dedicated device as the subtester to test individual components. The operator has instructions on what to select. For final test of our devices that go out to our customers, we actually test with the system software (validated and verified software) and for any softwre that isnt going to a customer we use our non product software process to verify SRS, URS and then teest to the URS and trace that and then approve the val package.

In my opinion we fall under a qualification of did we IQ the equipment and software and do we have our validation and verification that all software performs as intended. However I keep getting told by the specialist that we should be performing a PQ, to me that has no bearing on this situation. I am basically operating with an electrical Go/No Go gauge, there is n variability, the part either turns on or turns off? I am simplyfiying the part since it is an instrument but the software basically tells the equipment what to do. If the test pass then the part moves on for release?

I think I am running into a situation of applying pharma rules to device manufacture?

we have several product lines but each one builds and test at most 5 units per day.

I would ask your validation expert, the exact reason why a PQ has to be performed, in my view if you are only producing 5 products per line that is not alot in comparison to to the manufacturing lines that produce thousands of units per day.

Would be interested to know.


I don’t think they’re necessarily applying pharma rules to medical device manufacturing. It’s common on the device side to do IQ/OQ/PQ on manufacturing equipment.

The idea would be to run a PQ to ensure no variability over time. Variability could be introduced in a number of ways; however, since you have a functional test on each unit and based on your description, I would agree that it’s most likely that any variability introduced would be caught.

That said, there is a comfort level with inspectors to see the full IQ/OQ/PQ suite. It might be easier in the long run to do, say, 3 full-day shifts (15 units). It’s my understanding that you could still sell these units if they were done per procedure and were quarrantined until after your validation assessment.

If you want to stand firm against doing a PQ, you should probably perform a risk analysis focusing on effects of full shift runs. Analyze (and document!) what might go wrong and how all are mitigated (presumably) by the go / no-go test (or whatever other controls you have). Include in the analysis anything that might introduce variability (machine heating up, dust accumulation, etc.). You will then have evidence that you considered the risk and made a sound judgement to scope your validation.

In the past we would run three boxes and call it a PQ but when we spoke to other validation specialist they said we arent validating anything since all we are doing is having them follow instructions on how to use software. The software we use is validated product software and any non product software is also validated. I guess I just dont see the value, to me the value is in how did you install the software and/or equipment and this should be a qualification. But I keep getting beat over the head with a validation has all three.

After re-reading the posts, I think I’m confused now about what we’re discussing. There are several ‘objects’ in the discussion:

  1. the device itself

  2. the manufacturing equipment / processes

  3. the test equipment / software / processes

Each should have a validation package. I would expect that only the manufacturing equipment / processes would ‘need’ a PQ (if at all). I would not expect the test equipment / software to have a PQ.

When you did in the past what you called your PQ (3 boxes), you said that it was considered insufficient since they just followed the instructions on how to use the software. This sounds more like you were validating the test software.

What seems to be the norm for manufacturing equipment validation is to do 3 full-shift (or equivalent) runs with a complete shutdown / teardown after each run. I am just presuming you can manufacture more than 1 device in a production run.

Do you have a manufacturing process map showing the start-to-finish activities? Those things usually help me get my arms around the problem better. I’m not saying you need to share it but maybe work with your consultant to scope the problem?


I am quite confused as to your reasoning. 820 the regulation for Medical Devices is just as comprehensive as 210/211 the pharma regulations.

820 requires all software used in a Medical Device or in production of a device to be validated.

I suggest you read through 21 CFR Part 820 Quality System Regulation - in full.

Alex Kkennedy