There are lot's of assumptions in your question and therefore, not a simple yes/no answer. You need to go back to the design input, output and verification before you can answer validation. Design Verification is focused on the design output meets design input (ie. “Did I make the product right?”) While Design Validation is verifying design output meets user needs and intended use(s) (ie. “Did I make the right product?”). Refer to the FDA website for specific guidance on Design Control Guidance. They specifically mention dimensions in design input and output.

You must verify that it will not affect the fit, form, or function, or increase any associated risk, then based on the findings determine if additional validation is required.
Hope this helps,
Jãna Fernelius

Ee Bin Liew did this tolerance change have any change to the intended use, indications for use, or any impact to the contraindications on the product?

Somashekar BV I will interested to know why .... The design engineers decide to widen the tolerance of a particular dimension, x.
1. Blunder ... Wrong at the first place
2. Costly to make / buy
3. Unable to meet the tolerance consistently because of process capability
4. Is that dimension critical to safety and performance
5. Looking for cost reduction
6. Any other specific reason not listed ?

May want to verify and validate again since previous verification and validation may have been inadequate (missed something).

Stephane Morvan I agree with Somashekar...

The root cause of that change plays an important role in that cascading analysis, and it would probably be easier to start from this root cause and drill down on its impact on product performance based on a particular regulatory framework.

Matthew Romey The other question would be: how and under what circumstances do you notify the FDA?

Hema Lakkaraju May want to update tolerance related risk impact which become the platform for validation strategy

Karli "KC" C. I agree with Somashekar and Stephane - the why matters a lot. Generally speaking, if the drawing is being updated to match the "as-built" condition, it is my understanding that validation may not be required.

Karthik Venkataraman just go by your do as you say say as you do philosophy - for such a design change how do you evaluate its impact on end users (patients) & (operators) - as per your qms if you have one. if you might not ...point 4 in Somshekars list - the risk perspective from a regulatory side...you may then also need to address the process part to accommodate such engg changes in the days to come..

Annika Fahlén you need to address the question and come up with a very good reason with supporting evidence data if you don't validate (use risk management)

Shrikant Kalegaonkar Hello All -- Thank you for your comments and questions. I find them all valuable as they touch on different aspects of the change. You've given me much to think about. (I am also fascinated by the diversity of response.) -- Best regards, Shrikant Kalegaonkar

Tolerances should be established during the design & development process--which requires design verification and design validation.

My advice would be to go the other way. Start with broader tolerances that are acceptable during D&D and then "fine tune" for manufacturing process controls.

Another question to consider is this: Is the dimension a critical dimension?

George Hummel, ACQI A design change would require validation that input requirements were s.till

George Hummel, ACQI Still met

Matthew Romey I agree with Jon and George. I really like Jon's point about starting broad. Sometimes people set their tolerances based on what they can do today, not on what the design requires. This can lead to overly tight design specifications. Why make tight tolerances when you don't need to? Remember that things can change: processes, vendors, operators, all of these can change your future output such that you no longer meet the tight tolerances you set originally. And it's much easier to justify tightening a specification than it is widening it.

Absolutely. Consider also all the interface requirements as well

Antonin Cuc I am a State investigator oriented on the causalities of technical forensic evidences for the mortality injuries in the illegal medical usage Medical Devices. I have analysed firm research of technical stability of implants by patient usage...By the firm results there are evidences - the coaxiality between femoral axis and stem is for optimal long being stability recommended +/- 0.8 grads; with guarancy the functional ability about 16 years - but in defined acceptual tolerancy +/- 1 grads there are then the long being usage of probable functional guarancy only about median average 8 year only! Thatsfor the firm is recommending for installation in patient body the computerized assistasnce for assembling with continually positional controll of implant processing in hip bones on surgeon hall! There are 45 years alloplastic history and this dimension should be inforced in the orthopaedic praxis as mandatory duties to declare in the EHRs! The orthopaeds are ignorants!

Agree with Jon, Matthew, and Andys comments. Design the part with the broadest possible tolerances to start off with. If too tightly specified it causes real issues in manufacturing.

Thomas Marmor I agree with Jon and the others. I would start to define the tolerances in a product spec as broad as pssible. Based on the product spec I would create a production/manufacturing spec used internally with smaller tolerances. This would also help to Monitor and Control the production.

Charlie Lo Hi guys, I coundn`t agree more with your suggestions to define the tolerances in a product spec as broad as possible, however in the market where I work for, customers are always asking for a more portable and small device, which usually are taking out the tolerance we keep. What would you say about this?

Matthew Romey Charlie: If the customer need is for it to be small, then you should do your best to design it to be small. While this might create more challenges by tightening your tolerances, that is something you will have to address in the design and manufacturing of the device. My point is to let the design requirements (e.g. it needs to be small) be the primary driver of tolerances. Of course there may be exceptions that I haven't thought of, but take it as a rule of thumb.

Charlie Lo Hi Paul, Hi Matthew, appreciate your precious feedback. I guess it`s ourselves who made the life too tough, not the customers.