Skip lot sampling Plan for Micro testing

Hello all,

I have been trying to educate myself about the logistics of creating a Skip Lot sampling plan. This is in reference to a company that makes some personal care products and some dry mix formulations for energy drinks.

We get tons of raw material in, some formulation are complicated.

We are currently sending out samples for micro testing on every single lot that comes in. The first test for each lot includes ID testing. I have found quite a bit of information, but It is an entire discipline in itself. I am just easing in to details on AQL tables when I felt my brain start to melt down a bit....

I have a flow chart (I borrowed from an FDA site) and other attributes dictating the path of approval to initiate skip lot testing.. but as I start to read over matters relative to AQL I'm not sure I am going the right direction any more.

Do any of you institute Skip lot testing specifically for Micro?

Could any of you share or point me toward a generic SOP for this to highlight a fully approved (FDA, SQF level 2) Skip Lot Program?

If so It would be greatly appreciated and if you are ever in the Utah area I will personally buy you a Chicken-pot-pie and a cold beer of your choice.....

Hi PD,

You are venturing into highly murky statistical waters. "Dodge" by name and perhaps also the preferred action.

From a quick look in my Beginners book of SQC (Grant-Leavenworth) i quote -

"skip lot sampling is heavily dependent upon the assumption of homogeneity among lots and a good quality history."

IMEX the scatter intrinsic in process micro. data is rarely compliant with such a caveat. But perhaps you have found the Golden exception.

I appreciate you expert opinion Charles.

I am biasedly arguing for what I wish to be possible (perhaps the opposite of devil's advocate?)

What if the raw material were say a twenty pound sack of salt.

After ten lots from the same supplier and ten Micro tests returned with nothing...

And approval from the customer..

Could I?

Say yes and that Chicken Pot Pie is all yours :-)

PD

For salt Plastic Ducky, absolutely perform skip lot....you can get randomizer software that can generate a random lot to sample = completely unbiased sampling. 

We use tons (literally) of salt a season---the mine happens to be 2 hours down the road, inherently safe product with a 200 year old history same location

If you have COA or COC from your suppliers and all of your due diligence has been performed, I say yes, go for it.

Hi Ducky,

What you just described in post 3 feels less like a skip lot sampling plan, which would be statistics based looking at the likelihood that you would discover a problem and what the scope would then be, and instead more like a validation/verification schedule, which is more risk based and much more common.

Per your example, if I validate a supplier by testing their first ten deliveries to ensure they're reputable, then monitor/verify annually to see if any issues pop up, that feels a lot nicer since you're assuming a level of risk in your program and the scope of issue would expand accordingly.

When it comes to skip-lot for micro, comes back to purpose of testing. If you're pursuing skip lot sampling because you never expect to see a positive and thus not deal with the increased scope of one, then I always ask why test at all? Testing implies you consider it enough of a threat to monitor, or otherwise you may be testing simply to satisfy customers who expect it.

There's certainly precedent for skip lot micro in water and milk testing, most states require only 1/week/product line testing, and presumably the scope covers the entire week since last good test because the products are so homogenized. My biggest issue with these is again, positive results don't result in increased testing moving forward, when in fact they've probably revealed low level contamination that simply wasn't represented in previous samples.

IMEX: Either test every lot because conditions and materials change day to day, or validate your processes/suppliers to eliminate micro-testing all together. Either you think it's worth monitoring or you don't, skip lot tends to be the latter but with hesitation.

www.itl.nist.gov/div898/handbook/pmc/section2/pmc27.htm

purpose and scope of a skip lot sampling plan

I still say go for it.....if it's good enough for meat between CDN and USA, it's good enough for you

IMO, I thought this was the reason for the introduction of HACCP.

It's simply not possible to test every bit of food produced for astronauts.
Why test each and every lot of material?

Unless you are dealing with historically high risk materials, why not do a hazard analysis, risk assess the hazard and come up with an appropriate control measure? (testing schedule).

If the company in question is "required" to test each lot, or does not mind the cost of doing so, I guess it's a moot point.

Marshall

For salt Plastic Ducky, absolutely perform skip lot....you can get randomizer software that can generate a random lot to sample = completely unbiased sampling. 

 

We use tons (literally) of salt a season---the mine happens to be 2 hours down the road, inherently safe product with a 200 year old history same location

 

If you have COA or COC from your suppliers and all of your due diligence has been performed, I say yes, go for it.

Just to inject some humor here. In a bakery we use a lot of salt as well. 
Each pallet weighs one ton. There are generally 40 pallets on a shipment, We probably get 10 shipments a year.

So that's 400 tons of salt a year.

Marshall

Hi PD,

Did a mini-search yr interest via Google.

Yes, skip-lot testing is being applied in practice for (a) acceptance sampling of various quality factors including micro.data of received product lots, (b)  internal Production Control.

Unfortunately the accessible articles I found so far gave no  practical  examples of  designing/implementing skip-lot procedure for (a).

(Scampi's link has useful-looking formula/performance graph but afai could see lacks an example of typical usage).

Some representative extracts -

2.6.3   Tightened inspection/skip lot sampling
When we have an ongoing relationship with a supplier, or other ongoing sampling
situation, we can, over time, develop a level of confidence in the performance of
that supplier. That may lead us to relax the rate of sampling, including skipping
sampling of some lots altogether, using the freed sampling resources where they
can  be  more  beneficial.  If,  however,  the  reduced  rate  of  sampling  finds  a
`defective' lot we will revert to the initial sampling rate, or even to a more stringent
level of sampling, until we develop confidence that the supplier's performance has
returned to an acceptable level. Formal methods for establishing such plans are
described by Schilling (1982) and Montgomery (1996).

 

It  is  also  worthwhile  to  note  that  sampling  in  a  microbiological  context  is  no
different  to  sampling  for  other  quality  or  safety  characteristics  of  food,  or  of
non-food  products.  Codex  has  developed  guidelines  on  sampling  (CAC,  2004),
which includes references to various related International Organization for Stan-
dardization (ISO) standards. These international standards include information on
various “sampling systems” that allow the lot-by-lot sampling requirement to be
relaxed when a history of good compliance has been demonstrated, e.g. skip-lot
sampling. However, such systems are outside the scope of this document and in-
terested readers are referred to Montgomery (2012), for example. Nevertheless, the
more generic term acceptance sampling (and testing) encompasses these alterna-
tives systems, which are also based on same underlying statistical methodology.

 

 

Measure ALPHA 4.  Increase security spot checks of vehicles and persons entering installations under the jurisdiction of the United States.
 
Veterinary Support Measures:   
 
1.  Weekly activities by Food inspection personnel include skip lot style sampling plans
for examining commercial vehicles delivering food and water to the installation at the
commercial vehicle entry control point.  Skip lot sample size shall mimic the sampling
patterns used by local security forces; or a normal severity sample plan shall be
developed using ANSI/ASQC Z1.4, Sampling Procedures and Tables for Inspection by
Attributes.

 

Much of the literature regarding food safety sampling plans implicitly assumes that all lots entering commerce are tested. In practice, however, only a fraction of lots may be tested due to a budget constraint. In such a case, there is a tradeoff between the number of lots tested and the number of samples per lot. To illustrate this tradeoff, a simple model is presented in which the optimal number of samples per lot depends on the prevalence of sample units that do not conform to microbiological specifications and the relative costs of sampling a lot and of drawing and testing a sample unit from a lot. The assumed objective is to maximize the number of nonconforming lots that are rejected subject to a food safety sampling budget constraint. If the ratio of the cost per lot to the cost per sample unit is substantial, the optimal number of samples per lot increases as prevalence decreases. However, if the ratio of the cost per lot to the cost per sample unit is sufficiently small, the optimal number of samples per lot reduces to one (i.e., simple random sampling), regardless of prevalence. In practice, the cost per sample unit may be large relative to the cost per lot due to the expense of laboratory testing and other factors. Designing effective compliance assurance measures depends on economic, legal, and other factors in addition to microbiology and statistics.

 

The refs mentioned above are mainly to standard text Books. Will revert if any accessible/useful.

PS - an example of use in internal production control is attached -

 skip-lot2.pdf   170.09KB   72 downloads

addendum

Hi PD,

I don't know what particular micro.testing you are considering so the following add-on may/may not apply.

Despite the boom in statistical acceptance plans driven by the process crowd  I noticed this (2016) caution regarding micro. potential in a well-respected text on statistical aspects of microbiological food analysis -

 micro limitations.pdf   162.08KB   51 downloads

It's undeniable that for any substantial Lot of material, AQL-type plans require extremely massive sample sizes from a micro. POV. Even the  skip-lot plans I saw usually involved n>10.

To illustrate afaik the current  sampling scene I have attached a few compilations of food sampling schemes in "common" use for sensory and  micro.aspects.

 sm1 - Codex Guidelines on Sampling,2004.pdf   507.79KB   57 downloads

 sm2 - Practical Examples of Sampling Plans.pdf   529.28KB   70 downloads

 sm3 - NKML Procedure Guide on Sampling Food for Analysis ,2014.pdf   650.32KB   46 downloads

PS - a quite detailed exposition of the ISO Procedure (2859-3) for implementing  skip-lot sampling is at the link below. Some paracetamol is recommended to be available close-at-hand.

http://ifsscan-odemo...nceSampling.htm