ISO 9001:2015 8.3.4 & validation of new equipment

[ScottK]

I think they are waaaay off base too - you can physically check a die cut part for quality. Doing IQ/OQ/PQ makes zero sense. It's really for special processes.

I dunno Andy... I've been in this poster's shoes....
having dealt with a lot of big device customers and lots of audits by them they tend to expect IQ and OQ on just about any equipment used to make product. Even qualifying any new dies we had for this same process.

PQ only for special processes.

 

[Deli88]

It was the customer that drove the validation of the machine. At the time they signed off on the validation but now, after the audit, they are saying that the validation package isn't strong enough. I don't think they want us to go through revalidation as the machine has proven itself over the last 3 years that it is capable of doing the job.

 

[ScottK]

It was the customer that drove the validation of the machine. At the time they signed off on the validation but now, after the audit, they are saying that the validation package isn't strong enough. I don't think they want us to go through revalidation as the machine has proven itself over the last 3 years that it is capable of doing the job.

I would push back saying that if the initial validation was good enough it's good enough today unless there have been rejections showing that this machine is at the root of the problem.

 

[Deli88]

We have had zero issues with the machine since it has started up in 2015. I think they were expecting to see a 30-40 page report outlining the reasons for the exercise, reason for sample size etc..

 

[ScottK]

We have had zero issues with the machine since it has started up in 2015. I think they were expecting to see a 30-40 page report outlining the reasons for the exercise, reason for sample size etc..

Then the auditor's expectations need to be managed.
Based on 3 years of history this is not a value added activity.

 

[Candi1024]

We have had zero issues with the machine since it has started up in 2015. I think they were expecting to see a 30-40 page report outlining the reasons for the exercise, reason for sample size etc..

Yep, statistical reasoning behind sample size would be required. An IQOQ wouldn't require sample size, so you must be referring to a PQ.

So the finding wasn't that PQ didn't exist, it was that it wasn't completed well. Validations have grown in the 20 years I've been doing them. Medical device company's are now expecting a lot more, including a risk analysis and links to each test that would reduce that risk.

Did you write the protocol first and have it approved before completing the testing?

Very few suppliers have the experience to write thorough validations. Yet the purchasers demand them. Typically this results in useless documents that only exist to fulfill a customers checklist.

 

[Deli88]

Yep, statistical reasoning behind sample size would be required. An IQOQ wouldn't require sample size, so you must be referring to a PQ.

So the finding wasn't that PQ didn't exist, it was that it wasn't completed well. Validations have grown in the 20 years I've been doing them. Medical device company's are now expecting a lot more, including a risk analysis and links to each test that would reduce that risk.

Did you write the protocol first and have it approved before completing the testing?

Very few suppliers have the experience to write thorough validations. Yet the purchasers demand them. Typically this results in useless documents that only exist to fulfill a customers checklist.

Hi Candi,

No, there doesn't appear to have been one done at the time, I've spoken with the customer and they have pretty much laid out what you have said. They're not going to request that we go back and redo everything but that we show we have considered the NC and will act on it now for any future events of validation

 

[John Broomfield]

PQ only for special processes.

...and for organizations investing in PQ to reduce their cost of testing and inspection.

Customers like to share these cost savings too.

John

 

[Candi1024]

I disagree. I have completed retrospective validations that caught issues that have been hidden for years.

 

[ScottK]

I disagree. I have completed retrospective validations that caught issues that have been hidden for years.

on extremely straight forward operations like die stamping?
you don't get much easier as far as verification.