5 replies to this topic

[Totes716]

I work in an facility where we are being required by an auditor to start an annual micro testing program for incoming drink mix materials.  There isn't really much guidance for how much testing I need to be doing, but does anyone have any experience with creating annual testing programs for incoming materials? We're generally a low risk industry and the water activity of the materials is typically around .9, so it micro issues don't crop up much.

[kingstudruler1]

Was there a particular reason he/she thinks you need an annual ingredient testing program?   I'm not sure that makes sense other than he thinks you should be verifying ingredients or suppliers by micro testing.  You may need to provide more info on the auditors finding.   It may also depend on the product produced and ingredients used.   

 

In general:

First you need to come up with a sampling plan.   Without knowing the auditors' expectations that is hard to determine.   Some people might say one sample per lot is ok.  Some may state that you have to sample the square root of the number of containers in a lot.  There are quite a few items on this forum regarding sampling.  

 

Again, without knowing the auditor's logic or criteria what you test for depends on the ingredient.  I would test the product for the organisms listed on the COA or spec sheet.    Or maybe those listed in appendix 1.  If there are none present, APC, coliform, Y/M and possibly pathogens of concern would probably be acceptable.   If you test for pathogens, I would do it when the lot comes in and place it on hold until the results come back negative.   

 

Also, 0.9 aW seems quite high for dry products.   There are a lot of microorganisms of importance that will grow at that level.     

[Totes716]

Was there a particular reason he/she thinks you need an annual ingredient testing program?   I'm not sure that makes sense other than he thinks you should be verifying ingredients or suppliers by micro testing.  You may need to provide more info on the auditors finding.   It may also depend on the product produced and ingredients used.  

 

That's an excellent question.  We tend to panic when an auditor suggests something and our data to push back on audits is extremely weak at this stage - it's growing.  We currently pull one sample per lot, perform some physical testing, and then release.  The finished good obviously gets micro tested, at a third party lab, before it goes out the door.  We haven't had any real issues with the finished good.  The product is very shelf stable.

 

The auditor requested though that we do annual micro testing on our vendors to support the release of raw materials.  What I'm interested in is what kind of schedule this might entail - how many lots per year per vendor etc. I think there are risks for salmonella and listeria - though our vendors do test for it.

 

 

[Charles.C]

That's an excellent question.  We tend to panic when an auditor suggests something and our data to push back on audits is extremely weak at this stage - it's growing.  We currently pull one sample per lot, perform some physical testing, and then release.  The finished good obviously gets micro tested, at a third party lab, before it goes out the door.  We haven't had any real issues with the finished good.  The product is very shelf stable.

 

The auditor requested though that we do annual micro testing on our vendors to support the release of raw materials.  What I'm interested in is what kind of schedule this might entail - how many lots per year per vendor etc. I think there are risks for salmonella and listeria - though our vendors do test for it.

 

 

Hi Totes,

 

For an unknown, isolated Lot, The most commonly published (semi-statistical) acceptance sampling scheme per individual micro-characteristic is that based on nmMc  (I assume no need to explain the detailed meaning of the letters). The typical value of n (number of samples) = 5  but the rest will depend on the material specifics.

 

The above is obviously a lot of work/expensive so IMEX is seldom used for routine monitoring purposes where  typical, relevant, micro data has already been established and compliance to a specification is applicable. Initially I usually take a minimum of 2-3 samples to enable an idea of composition variability and then (assuming results OK) reduce/compromise as "appropriate". (Obviously if one could assume perfect homogeneity/analytical techniques, n=1 would always suffice).

 

If you only had one vendor/one raw material/one "multi-analysis" per year I would initially suggest as per 1st paragraph above. But it immediately depends (eg financially) on how many analyses involved ? I anticipate many processors simply take one sample per vendor/raw material and (assuming OK results) that auditors go along with this. If multiple raw materials are involved per vendor.of a "similar" micro.characteristic use of a rotational sampling may be appropriate.

 

It likely all depends on the product details/history.

[kingstudruler1]

That's an excellent question.  We tend to panic when an auditor suggests something and our data to push back on audits is extremely weak at this stage - it's growing.  We currently pull one sample per lot, perform some physical testing, and then release.  The finished good obviously gets micro tested, at a third party lab, before it goes out the door.  We haven't had any real issues with the finished good.  The product is very shelf stable.

 

The auditor requested though that we do annual micro testing on our vendors to support the release of raw materials.  What I'm interested in is what kind of schedule this might entail - how many lots per year per vendor etc. I think there are risks for salmonella and listeria - though our vendors do test for it.

 

I think what he/she is suggesting is once sample of each raw material per year is tested.   An annual one sample per lot testing could serve as some validation of a raw material supplier.  However, it is statistically stupid.  But, it's better than nothing.  

 

With this logic - Nothing at all < once per year < once per quarter < once per month < every lot, etc.   

 

Because you don't have a lethality step (correct?), the finished product testing can also serve as a validation.  

 

If you have the resources, just test every lot received and be done with it.   

[Totes716]

We test every lot for certain physical characteristics at this stage due to the nature of some of the raw materials.  The auditor wasn't very clear though.  They requested an annual testing program to exist to validate that our vendors were selling us clean product - some near misses have involved the vendors not potentially being too honest on their certs.  My question really was what is 'annual vendor testing' to you guys on raw materials ?

 

In answer to you, we do not have a kill step, but microbiology is done on the finished product. 

 

I think @Charles.C gave pretty good answer to some degree.  Maybe the terms I'm using are a bit a expansive.  

Edited by Totes716, 22 April 2023 - 07:30 PM.