Determining Sample Size in Design V&V activities

Aphel

Involved In Discussions
Hello,

I have a big problem...

After an FDA inspection my company has to define a procedure for defining sample sizes regarding design verification.

My question is...how did you find a solution?

What are reasonable statictics standards i can use or refer to for defining sample sizes vor design verification tests? For example: electrical safety tests, software verification, mechanical durablility and loadability tests,...?

I am very thankful for every hint you can give me...

Best regards,
Aphel
 

Aphel

Involved In Discussions
Hello...

No, I am talking about design verification...

During our last FDA inspection three weeks ago, the inspector expected to see a design verfication plan, which contains test cases with sample sizes that are based on stastic justifications...

During the inspection we showed a verification report. And the inspector asked for our justification (based on statistics) regarding our sample size in the test cases...
 

Ronen E

Problem Solver
Moderator
Hello...

No, I am talking about design verification...

During our last FDA inspection three weeks ago, the inspector expected to see a design verfication plan, which contains test cases with sample sizes that are based on stastic justifications...

During the inspection we showed a verification report. And the inspector asked for our justification (based on statistics) regarding our sample size in the test cases...


Hi,

I'd very much like to help you, but...

You could phrase some general guidelines (statistical and other) in a SOP to determine Design Verification sample sizes, but the truth is that it's very much specifics-dependent. And any decent FDA inspector would know that. E.g., some procodes have special controls that prescribe it for certain aspects; these would of course not apply for other procodes. Etc. etc.

To really have a solid justification, one must look at every plan and every test, individually. Of course if some tests are replicated between devices or product lines, justifications may be replicated (not always).

If you were a client of mine I'd first ask to see exactly what that inspector wrote (and further understand the context, e.g. verbal discussions held). I'd also need to know the specifics of the device(s) at stake. Some general recommendations can be made but that requires, as a minimum, understanding what sort of devices you're making. Software is different to mechanics, etc. And again, these would only be a starting point that may take you, I believe, about 30-40% into the process of determining and justifying all sample sizes in any given plan.

Some might view it differently, though...

Cheers,
Ronen.
 
Last edited:
S

shala

Hello. Would you please help me in sample size selection based on statistical analysis for design validation and verification.
Thank you.
 

Bev D

Heretical Statistician
Leader
Super Moderator
Hello. Would you please help me in sample size selection based on statistical analysis for design validation and verification.
Thank you.

please read the thread and attachments that Harry referenced above and then if you could give some more specifics - as Ronen requested - we can provide you with useful advice...
 
S

shala

Thank you. we want to validate the package and also we are are testing Severe condition for failure of electronic part .( reliability study)
We need to select sample size for both cases and we are in design phase.
 
Last edited by a moderator:

Bev D

Heretical Statistician
Leader
Super Moderator
Have you read the attachment that Harry referenced on page 1?

also we really need very specific details of what you are trying to validate. Sample size determination is not a straightforward process.
 
Top Bottom