< 1 min reading time

As originally asked by Marc Henn.

All: Have a class 1, GMP exempt medical device. Also a device per MDD Class 1. This product, regardless of the failure, will not cause harm/safety issues, only complaints. We are correcting a manufacturing nuissance issue, the correction will be invisible to the end user. Indications, specifications etc. stay the same. Risk analysis does not yield anything significant. With this said, of course we must/need to validate this change to assure quality/safety of the product. My question centers around how extensive a validation do we need to perform? i.e. multiple batches to measure within lot and between lot variations. Do not want to “over”/”under” validate, however, Just wanted to “benchmark” thoughts and how others are handling low risk device validations.

---

Steve Doppelt
Schuremed Domestic and International Marketing Consultant
Please send me more information about the product Doppmed@aol.com

Mark A. Chipperfield
Principal Consultant and Company Director at Corvus Device Limited
Also consider that as a Class 1 Design Controls exempt device, Design Validation per 21CFR820.30 is in principle not applicable. If you have established DVal for prior US submission then you have set a rod for your own back and need to reflect this change against that package.
Of course, under ISO13485 the requirement remains for EU market.

Mark A. Chipperfield
Principal Consultant and Company Director at Corvus Device Limited
Separate Process Validation from Design Validation – these are different concepts.
Agree that PV needs to be reviewed for any gaps based upon process changes – likelihood is that you at least re-validate the relevant module, process step, specification elements, etc.
Perform a risk-based assessment of the product post-change, relative to prior Design Validation (and any criteria). It sounds as if you have all of the inputs you need (RA, User needs, past DVal, Clinical severity, etc). This is a ‘Validation Assessment’ and should conclude (based upon your comments) that there is no DVal gap and therefore no requirement to perform additional studies. But ensure that this is the case and you can stand behind your conclusion.
All of the above should be part of your Change management documentation and should be captured within your DHF/Conformity Dossier, Risk Management File and Usability Engineering File.

Pedro Mendes
Technical Steward Devices at Novartis
Since you are changing the process it should be validated. With this validation you will demonstrate that you moulded parts are in specification. Moreover you can demonstrate that nothing will be different in the moulded part.
You must have a change control to perform the proposed modification. This change control should assess the impact on design verification and design validation. From what you say I should consider that design validation is not need (only the assessment) and you can use the results of process validation to support it.

Tom Haggerty
Engineering Manager at Terumo Cardiovascular Systems
Good point Pedro!

Pedro Mendes
Technical Steward Devices at Novartis
Hello Marc,
Do you mean design validation or moulding process validation?

Tom Haggerty
Engineering Manager at Terumo Cardiovascular Systems
If the changes don’t effect the end user and don’t effect the user experience, no validation appears necessary. What appears to be needed is to determine what design requirements were affected by the change. Those requirements need to be re-verified to ensure the device still meets it’s design requirements following the change.

Dean Miller
Sr Principle Quality Engineer at R & Q Solutions
I typically try to be consistent in validation approach, for example an OQ will always reflect worse case range of processing conditions, and PQ will always represent the breadth of normal processing. Then, my sampling for each run will reflect the defect detection level desired based on risk to patient or user. I use LTPD sampling for this purpose. Of course, it is necessary to have LTPD levels established based on the risk of a defect escaping detection. Common numbers are 1, 2.5, 5, and 10. It sounds as though you may be in the 5 range, where 10 won’t even be noticed during use.

Jean Bigoney, RAC, CQE
Regulatory Affairs Specialist.
Marc, I think the answer is contained in your description and question.

You state that you wish to make a correction to a manufacturing “nuisance” issue which thee end user will not notice.

As such, your validation has to be extensive enough to determine whether the change removes the nuisance. You might want to take a page from Six Sigma methodology and apply “Define, Measure, Analyze, Improve, Control.” For example, you ask whether multiple batches need to be evaluated. I’d turn the question around and ask if your manufacturing nuisance is consistent between batches, absent in some batches, varies between batches etc. Only if you know that will you have a handle on the extent to which you need to validate.

If you have the assurance that the change has no regulatory impact, this sound like it is purely a QA issue.

Sabina Tall
Validation and Quality Engineer på Knightec
If the risk analysis does not yield anything, that sounds as no critical risks for the patient and only some qualification can be done securing the process and product.
Difficult to say with not having all information 😉

Sabina Tall
Validation and Quality Engineer på Knightec
I would have started with an risk assessment regarding the change!
How does the change effect the patient in the end?

That would give me a hint of the extent of the re-validation.
Also take in consideration of the validation before the change.

Gentian Muca
Trainer and Lead Auditor (Quality, Environment, Health-Safety & FS)
You must use the same validation criteria as before change of MDD.