What if I find a defect after AQL (Inspection)? Pharmaceuticals

[stampin]

Apologies if this has been asked before - if so I couldn't find it.

In our pharma operation we use AQL inspection at the primary packaging stage. We have a Major defect limit of 0.1%. We pack into trays of 2200 units. Then, we use another line to pack the units into secondary packaging.
The question we're struggling with is this: What should we do if we find defects during the secondary packaging process? The product has already passed AQL inspection, but is there a defect level that would then lead us to either believe that we have a previously unidentified 'cluster' of defects, or that our AQL process has been in some way inadequate? In other words, what level of defects can we ignore?

I'd really appreciate opinions on this!

 

[Ronen E]

Hi & welcome to the cove

Sampling plans are statistical tools. They are designed to provide confidence that at a certain probability level the AQL you are looking for will not be exceeded in the overall population. Probability, not certainty. they also involve some assumptions, including, in some cases, regarding the distribution of certain properties across the population. As always, assumptions can prove wrong.

If you feel there is a real issue then you may want to conduct some bigger scale investigation at the secondary packaging level. Consider checking some bigger samples, trying to analyze the results and re-looking at any assumptions.

Cheers,
Ronen.

 

[Burcu Sevim]

Apologies if this has been asked before - if so I couldn't find it.

In our pharma operation we use AQL inspection at the primary packaging stage. We have a Major defect limit of 0.1%. We pack into trays of 2200 units. Then, we use another line to pack the units into secondary packaging.
The question we're struggling with is this: What should we do if we find defects during the secondary packaging process? The product has already passed AQL inspection, but is there a defect level that would then lead us to either believe that we have a previously unidentified 'cluster' of defects, or that our AQL process has been in some way inadequate? In other words, what level of defects can we ignore?

I'd really appreciate opinions on this!

Hi Stampin, I am struggling with the same question and wonder if you have any answer for this. I appreciate if you can share the answer.

 

[Jim Wynne]

Hi Stampin, I am struggling with the same question and wonder if you have any answer for this. I appreciate if you can share the answer.

Welcome to the Cove. The post you're replying to is over 11 years old, and the user is no longer registered. There is almost no chance that they will respond. Nonetheless, perhaps someone else will have some advice.

 

[Alan_DB]

Hi Stampin, I am struggling with the same question and wonder if you have any answer for this. I appreciate if you can share the answer.

If you can give some details I may able to help - what sampling plan are you using, what do your existing documented procedures point to, etc

 

[Burcu Sevim]

If you can give some details I may able to help - what sampling plan are you using, what do your existing documented procedures point to, etc

Hi Alan,
thanks for your support! We use third party manufacturers for drug product manufacturing. They perform 100% visual inspection of the finished units and the AQL inspection. Most of them sample appr. 315 units from B/M/E and use 0.65% acceptance limit for major defects. Wenn we receive the final product after a successful AQL inspection, we perform a manual packaging of the product. In some cases, we observe defects like particles between stopper and sysringe wall and initiate a complaint for a single particle found in a single unit. My question is if it might be possible to establish predefined acceptable risk to which we can refer.

 

[Alan_DB]

Is the rate you are seeing defects within the acceptable quality limit that has already been established? What would trigger your contracted manufacturer to use a tightened inspection? How have they demonstrated competency?

The effort should be put to solving the problem - both where are the defects originating from and why the visual inspection is not detecting it. You could decide to accept it based on risk but its always hanging over your head. As the product owner it is at your discretion but it will be poked at and questioned a lot, especially when viewed what is categorized as a critical / major / minor reject.

 

[Burcu Sevim]

yes, the defects we are seeing, are within the AQL. The particles are intrinsic, either they originate from the process steps or from the packaging materials. And we know either the 100% visual inspection or the AQL inspection cannot ensure a 100% defect free batch. I agree that we put our efforts to solve the problem with a goal to eliminate all defects. But the complaint process is a time consuming process. the batch is put on hold until the complaint is completed. is it really necessary to initiate a complaint at TPM for a defect which is within the AQL?

 

[Alan_DB]

Ah OK, in that case I would just reject these from your packing process and calculate the batch yield - if acceptable then I wouldn't go any further.

Other people may have differing opinions

 

[VSchultzm71]

Ah OK, in that case I would just reject these from your packing process and calculate the batch yield - if acceptable then I wouldn't go any further.

Other people may have differing opinions

Hello, Alan,
I am familiar with Trigger Alarm Systems, rather than AQL.
My question is: If my sampling lot size is 200, with acc as 7, rej as 8, found 8 defective units when I am at the 50 count, do I stop the inspection and reject the whole batch? And also if after taking a fresh sampling batch, another 200, should I use the same method?
Thank you,