What does "fully verify" mean exactly in process validation?

[terrytw]

Hi everyone, I know this has been discussed before, and I have a general understanding of it, but there is one thing that I cannot wrap my head around.

FDA says "Where the results of a process cannot be fully verified by subsequent inspection and test, the process shall be validated with a high degree of assurance and approved according to established procedures. "

For example sterilization process needs to be validated, and manual cutting does not need to be validated, because the testing of sterility will destroy the product, whereas the manual cutting process can be tested for every piece by an IPQC of dimension testing (which corresponds to "fully verified"). So far so good.

But, what if I do not conduct an IPQC for dimension in manual cutting? Am I forced to conduct process validation for manual cutting? In other words, do I have to test every piece if the process is not validated? If not, how can I be sure that it is "fully verified"?

Thanks for anyone chiming in!

 

[chris1price]

Essentially, yes. Fully verified implies 100% inspection or test. And don't forget that inspection and test are also processes that will need validation.

 

[terrytw]

Essentially, yes. Fully verified implies 100% inspection or test. And don't forget that inspection and test are also processes that will need validation.

Thanks a lot!

That is what I thought. There is a senior quality manager at my company who says otherwise, which is why I am asking. She says you don't have to test each piece if the process is not validated, and proceed to name an example of appearance, which does not correspond to any specific process. Her logic does not really make sense to me, so I stopped asking her.

 

[d_addams]

Thanks a lot!

That is what I thought. There is a senior quality manager at my company who says otherwise, which is why I am asking. She says you don't have to test each piece if the process is not validated, and proceed to name an example of appearance, which does not correspond to any specific process. Her logic does not really make sense to me, so I stopped asking her.

Unless your operators are blindfolded, each one is 100% inspected for appearance. Which then brings up the question, what are the appearance criteria? Subjective visual criteria can be a pain to do test method validation on, but if you have visual acceptance criteria on your specification/process (like 'check for appearance') you'll need defined acceptance criteria.

 

[terrytw]

Unless your operators are blindfolded, each one is 100% inspected for appearance. Which then brings up the question, what are the appearance criteria? Subjective visual criteria can be a pain to do test method validation on, but if you have visual acceptance criteria on your specification/process (like 'check for appearance') you'll need defined acceptance criteria.

The acceptance criteria is something like "no hook, no black spot", which is not easy to spot (we have an accept/fail sample grid, even if it is a small black spot, it's still considered fail) unless the operator is closely examining the product. It is usually tested at IQC and FQC with a sampling plan. There is no single process that affects the appearance because it is rather just the catheter shaft that we receive as raw material might be defective.

 

[Tidge]

But, what if I do not conduct an IPQC for dimension in manual cutting? Am I forced to conduct process validation for manual cutting? In other words, do I have to test every piece if the process is not validated? If not, how can I be sure that it is "fully verified"?

To add to (and emphasize) the already good advice that was given: In order to be "Fully verified", the characteristic (being verified):

1. Requires a specification
2. Requires a record of the result of the verification

As was mentioned by @chris1price, the method used to find the result will have to be valid(ated), and must not be allowed to become unvalidated. Practically this means that equipment is maintained and calibrated, and operators remain continuously trained and assessed.

I have occasionally encountered people who say that a "first and middle and last piece" inspection is 100% verification... it is not. At best this practice is a sort of sampling mechanism for a process that is (hopefully) validated (or could be).

 

[Medteresting]

Hi all,
Im a bit new to this, and have a follow-up question here. What aspects of the product has to be verified for 100% of the devices, to avoid doing validation?

Saiing your end product is a Control unit (screen with hardware and software, made to do spesific measurements). You have a measurement equipment plugged into the screen. Could you defend doing verification of 100% of the devices, on just the performance of the equipment? meaning just checking the device turns on, and measurement from the device is correct. Or do you have to check all other aspect/requirement set for the product?

 

[yodon]

I think you may be conflating product V&V with (manufacturing) process validation.

In your example, what is the (manufacturing) process involved?

The testing you describe sounds more like acceptance testing.

 

[propdto]

To add to (and emphasize) the already good advice that was given: In order to be "Fully verified", the characteristic (being verified):

1. Requires a specification
2. Requires a record of the result of the verification

As was mentioned by @chris1price, the method used to find the result will have to be valid(ated), and must not be allowed to become unvalidated. Practically this means that equipment is maintained and calibrated, and operators remain continuously trained and assessed.

I have occasionally encountered people who say that a "first and middle and last piece" inspection is 100% verification... it is not. At best this practice is a sort of sampling mechanism for a process that is (hopefully) validated (or could be).

Great information. Not much to add other than that FDA has written several warning letters addressing this topic. From the warning letters it is pretty clear that inspecting parts at the beginning, middle, and end of the manufacturing run is not full verification. One of the warning letters I read stated something like...you have to measure each part or you have to validate the process. Sampling or SPC is not allowed either to make the final quality decision unless the process is validated.

 

[propdto]

Hi all,
Im a bit new to this, and have a follow-up question here. What aspects of the product has to be verified for 100% of the devices, to avoid doing validation?

Saiing your end product is a Control unit (screen with hardware and software, made to do spesific measurements). You have a measurement equipment plugged into the screen. Could you defend doing verification of 100% of the devices, on just the performance of the equipment? meaning just checking the device turns on, and measurement from the device is correct. Or do you have to check all other aspect/requirement set for the product?

Are you performing a functional test at the end of the production line to verify that the device as built meets specifications?