Difference between a "check" and a "test"

[Mark Meer]

Our assembly process instructions includes lots of "checks" throughout, to ensure that any mistakes can be caught early in the process.

In a recent inspection from the US FDA, the inspector was of the opinion that these "checks" were in fact in-process acceptance tests, and should therefore be documented. I see little value.

Consider the following illustrative (but extremely simplified) example:

1. Visually inspect all parts for defects or blemishes.
2. Check the LED component functions prior to installing by applying xV power and confirming that it lights.
3. Install the LED into the sub-component.
4. Check the sub-component. Plug in and confirm the LED lights.
5. Install the sub-component into the final assembly.
6. Test the final assembly. Plug in and confirm the LED lights.

So, I'd like to hear some Covers weigh in:
In the above example, what would you consider a "test" that would have to be documented?

 

[somashekar]

In the above example, what would you consider a "test" that would have to be documented?

You determine ... You Document ... You Record same ...

Consider the following illustrative (but extremely simplified) example <modified for you>:

Visually inspect all parts for defects or blemishes.
Check the LED component functions prior to installing by applying xV power and confirming that it lights.
Install the LED into the sub-component.
Check the sub-component. Plug in and confirm the LED lights. Record same
Install the sub-component into the final assembly.
Test the final assembly. Plug in and confirm the LED lights. Record same

 

[Mark Meer]

Thanks for your input Somashekar. Your opinion seems in-line with the FDA inspector.

But I still feel there's a somewhat arbitrary interpretation being applied to distinguish in-process acceptance testing from just a simple check.

My position is still that only the final test has any value to document, as it's the only one that ultimately affects final product quality.

Anyone else want to weigh in?

 

[normzone]

If you don't document it, you can't prove that it's happening. If it's important enough to do, it's important enough to document.

If the final inspection is the only one that's important enough to document, then only inspect it at that stage.

 

[AndyN]

Thanks for your input Somashekar. Your opinion seems in-line with the FDA inspector.

But I still feel there's a somewhat arbitrary interpretation being applied to distinguish in-process acceptance testing from just a simple check.

My position is still that only the final test has any value to document, as it's the only one that ultimately affects final product quality.

Anyone else want to weigh in?

Some products can't wait until final inspection to determine quality! Plus doing ISO has always been about preventing problems, so you'd better perform in process (or earlier) checks!

 

[Mark Meer]

If you don't document it, you can't prove that it's happening. If it's important enough to do, it's important enough to document.
If the final inspection is the only one that's important enough to document, then only inspect it at that stage.

I disagree. Every step in the assembly instructions is "important enough to do", but not all necessarily warrant records to "prove that [they're] happening". ...Otherwise, you might as well have a checklist for every screw fastened!

And this is the point of discussion: there needs to be a distinction between a "check" (something to catch issues early on), and a "test" (which is actually confirming quality of devices).

 

[somashekar]

And this is the point of discussion: there needs to be a distinction between a "check" (something to catch issues early on), and a "test" (which is actually confirming quality of devices).

Going with your own definition, which I like too., this distinction needs to be in place and purely based on your own technical and process control judgement.
Every operator in every operation does a check involuntarily. When you decide a stage for a set of specific checks (together called as test) the evidence for the same stage completion (test record) will constitute your inprocess inspection records. This could be as simple as a check mark against a checklist.
I hope we both agree that we need inprocess inspection to assure the expected final yield (Post Number #5) in the final inspection.
Such stage / stages are purely what you decide for your product.

 

[Mark Meer]

I hope we both agree that we need inprocess inspection to assure the expected final yield

I don't know if we can agree. In my opinion, (unless a specific quality system objective), production "yield" is quite separate from product quality.

It's not hard to conceive of a situation - say, for example, extremely high final acceptance criteria - where production yields are low, but product quality is very high.

As such, I'm of the opinion that any in-process acceptance checks/tests/whatever that have to do with yield, but DO NOT impact final product quality, need not be in the scope of 21 FDA CFR 820.80.