13 replies to this topic

[Gelato Quality Lead]

Hi,

 

Looking for ways to reduce costs without compromising our results.

 

We manufacture multiple flavors and types of frozen desserts per day. Currently, we are pulling two samples from each batch and testing both for coliform, aerobic plate count, and yeast and mold. We test one product type per production line per day instead of testing from every single batch.

 

We would like to be able to start testing composite samples. Would we be able to combine different flavors and product types for a composite sample? Or would it be best to just combine the two pulled samples from the same batch? 

[olenazh]

If it's in-house testing, you may combine different flavors and test for micro to ensure the batches are clean. However, I've personally never heard of testing of a several batches mix: we test (in-house lab, testing for TC & E. Coli) our batches of ice cream individually.

Keep in mind though that your in-house testing results would not be considered official - unless your internal lab has all mandatory accreditations. If not, you'll need to send product samples to the accredited lab for micro testing on a predetermined schedule (for instance, we send 3 samples a month).

[Scampi]

You cannot meld multiple products into a composite as you'd never be able to tell where the issue came from 

 

 

As for individual product types, I would be doing a composite sample for each

 

20 grams every X minutes/volume until changeover.  Send that whole thing to the lab and mark your COC as a composite- they will then blend it appropriately and sample from what is now a homogeneous composite sample

 

Repeat for all products run

 

Now---if saving money without compromising is what you're really after, then I would suggest having a really good look at production scheduling against historical sales.

 

Make larger production runs whenever possible---less changeovers is automatically more productive and therefor less costly

[Charles.C]

Hi,

 

Looking for ways to reduce costs without compromising our results.

 

We manufacture multiple flavors and types of frozen desserts per day. Currently, we are pulling two samples from each batch and testing both for coliform, aerobic plate count, and yeast and mold. We test one product type per production line per day instead of testing from every single batch.

 

We would like to be able to start testing composite samples. Would we be able to combine different flavors and product types for a composite sample? Or would it be best to just combine the two pulled samples from the same batch? 

Hi GQS,

 

Microbiological sampling is a complicated topic.

Some disadvantages regarding "combining" have already been noted. Any benefit may depend on your specific objective (eg intra-batch variations, batch mean parameter values, acceptance/rejection criteria ).

Combining as per last sentence in the OP was (briefly) mentioned in this earlier thread -

https://www.ifsqn.co...ch/#entry106275

 

Some diverse Literature viewpoints are presented below -
 

Pooling of samples can be performed when more than one 25g test portion from
a specified product batch is to be examined and when evidence is available that
combining test portions does not affect the "result" for that particular food. This can
either be done by combining 25g of each sub-sample into a larger volume, and
then preparing an initial suspension using the entire combined volume of sample,
or (for enrichment tests) by preparing separate enrichment broths for each sub-
sample and then combining a portion of each broth for subsequent testing

(Public Health,England,2019)

 Preparation_of_Samples, Public Health,England.pdf   546.15KB   23 downloads

and

 

The concept of compositing of sample units is not new. It is widely used in the preparation of analytical samples for chemical and physical analyses because by efficient blending of a number of sample units, an analytical sample can be taken that is more truly ‘representative’ of a‘lot’ (Patil 2002; Anon. 2003). The effect of compositing is to generate a sample that will be more ‘typical’ of a batch than will individual samples tested separately as the‘between sample’ variance is reduced and the analytical result reflects more closely the true composition of the lot. Such procedures have been used widely in chemical analyses including those for trace contaminants such as mycotoxins in foods, for generic assessment of microbial populations in soil and water, for environmental studies of avian faecal microflora, etc. Note, however, that these analyses are quantitative in nature.

(Jarvis,2007)

 Compositing of Samples,Jarvis,2007.pdf   1.81MB   23 downloads

 

And to illustrate the possibility of false illusion -

 Compositing (Pooling) Micro-Samples.pdf   220.48KB   43 downloads

 

As a further discussion of the potential loss of detail via compositing can see the early (2002) but nicely readable intro attached below. Possibilities for regaining significant lost information are briefly mentioned and have been  analysed in considerable depth (heavy maths !)  in some later papers however the necessary extra efforts involved looked IMO to possibly be an example of the "Law of Diminishing Returns".

 composite-sampling,Patil,2002.pdf   30.8KB   30 downloads

Edited by Charles.C, 21 July 2022 - 06:11 AM.
added pdfs

[sqflady]

I have seen "BME" testing which stands for "Beginning, Middle, End".  When making multiple batches of the same item, say 50 batches, we would sample batch 1, batch 25, and batch 50.

[Charles.C]

I have seen "BME" testing which stands for "Beginning, Middle, End".  When making multiple batches of the same item, say 50 batches, we would sample batch 1, batch 25, and batch 50.

Hi SQF Lady,

= a condensed version of systematic sampling if a continuous flow. Folllowed by compositing ? (This is how minerals are often handled)

 

Actually I had assumed the OP was faced with a completed Lot which simplifies the statistics a bit.

 

PS - JFI, The BME system is also briefly discussed here -

https://www.ifsqn.co...ch/#entry106148

(also see post 2 of same thread)

Edited by Charles.C, 21 July 2022 - 06:14 AM.
added PS

[Tony-C]

Hi Gelato Quality Specialist,

 

Before discussing composite samples, my question would be how often do you get out of specification results?

And are the two results from the same batch consistent?

 

Kind regards,

Tony

[Gelato Quality Lead]

Thank you all for the responses and resources. This will be very helpful in my determining the direction we will go in.

 

The idea was that we would plate the composite sample with the multiple flavors in it and if the composite sample plate were to be out-of-spec, then we would have to plate each lot individually to find the culprit. 

 

We rarely get out-of-spec results and the two results (one sample from beginning of batch and one sample from end of batch) are consistent with each other. 

 

For these reasons we were interested in cutting cost and time dedicated to micro plating.

[Charles.C]

Thank you all for the responses and resources. This will be very helpful in my determining the direction we will go in.

 

The idea was that we would plate the composite sample with the multiple flavors in it and if the composite sample plate were to be out-of-spec, then we would have to plate each lot individually to find the culprit. 

 

We rarely get out-of-spec results and the two results (one sample from beginning of batch and one sample from end of batch) are consistent with each other. 

 

For these reasons we were interested in cutting cost and time dedicated to micro plating.

Hi GQS,

 

Quant. micro. data can be notoriously variable IMEX.

 

Unless yr products all have v. low counts (eg heat treated) and/or yr spec. is unusually flexible, I am surprised you could achieve "consistency" with yr sampling procedure.

It all depends on the numbers of course.

[AJL]

Revisiting here.
Pooling samples for listeria and salmonella (we expect that they will both be negative of course).
Legally is this a challenge?
Of course we won't know where the issue comes from if there is a positive, but we are not expecting a positive.
But it is not quantitative, test, so surely combing say 3 *25g is actually fine?

[jfrey123]

None of us "expect" a listeria or salmonella positive (or at least I hope we don't lol).  I like composite samples from one batch for better accuracy across the production run, but I wouldn't ever consider pooling multiple batches.  A single positive there affects all batches from the composite, and the products I deal with now don't have sufficient shelf life for us to hold and then resample each batch individually.  Subsequently, say you composite all batches from a week's worth of production and it comes back positive:  you won't know which day it occurred, so it prevents you from narrowing down which day a sanitation deficiency or a temperature issue or other factor led to the positive.

[AJL]

Yes but legally no issue? We produce over 3 days (new batch each time) but I was hoping to be able to composite these. We don't have a high risk for listeria, it's a customer requirement

[Gelato Quality Lead]

Hi AJL,

 

As mentioned by jfrey123, if you are compositing multiple batches and you end up with positive results, then you would need to place all three batches on hold for retesting to determine which one is the culprit. We don't have enough space or bandwidth to be able to do this, if it were to happen, so we composite samples only from the same batch. 

[MOHAMMED ZAMEERUDDIN]

Composite sample of different batches for the same product is preferable if at least one months result trend shows no deviation from the regulation standard.