Safe Foods for Canadians

This is the second in our summary of the new Canadian Food legislation.

Here we will address one of the stated reasons for the passing of the Act.  From the CFIA’s (Canadian Food Inspection Agency) summary of the Act:

“Strengthened food traceability:  Current legislation does not require food manufacturers and other regulated parties to have traceability systems.  The legislation provides the CFIA with strengthened authorities to develop regulations related to tracing and recalling food, and the appropriate tools to take action on potentially unsafe food commodities.  This includes a prohibition against selling food commodities that have been recalled.”

Boy have we come a long way.  Not so long ago, MS Word would tell me that “traceability” was not a word (back when we were talking about it as a neat feature of our SPC systems.)  No longer!  Now we offer complete systems for traceability, not just an extra feature of another solution.

The provisions of the USA Bio-terrorism Act of 2002, USA FSMA, and now Safe Foods for Canadians, all consider traceability to be a major concern.  Not to mention that the big box companies have even more stringent requirements.  Many food companies think they have it nailed, but there are serious holes and issues in most paper based systems.  The first and foremost is of course the possibility of error, transcription, or illegibility.

If you manufacture your own products, under BTA you are allowed a maximum of 24 hours to produce a complete recall list with SKU # and location.   If you co-pack, or ship to the Walmarts and Costcos of this world, you may have less than 4 hours.

In a paper world, production batch sheets are usually sorted by shift and placed in envelopes with supporting documentation.  Although I have met people who swear they can do a mock recall in less than 4 hours purely on paper, I can’t conceive of how they can do so accurately and completely:  It is easy to recall all of production lot 2012076, (it may all have shipped to the same big box warehouse!) but to recall all SKUs that contain or may contain liquid sugar 211093882 that entered the tank on May 15 2013, and the tank hasn’t been cleaned since then – that would be a neat trick.

Even worse: What do you do with this if the recall is for any product delivered by driver John Smith?  Or all products of JS Company delivered between Jan 14 and Jan 19?  And what if somebody dropped the batch sheets into a vat?

Canada isn’t there yet, but between the Safe Foods for Canadians Act, and the CFIA stepping up to provide regulations, I don’t think Canadian food processors have a very long window… and besides, it is just smart business, good manufacturing practice, good corporate citizenship, and limits exposure if you do need to do a recall.  Lack of clear traceability by definition means that a recall can be orders of magnitude larger than it would be with contained traceability.

Dealing with Bulk Storage and Continuous Process for BTA 2002

This is the most difficult part of any traceability system.

If complying with BTA 2002 is the objective, neither the Act nor FDA guidelines stipulate definitive threshold levels for any ingredient, at least not that we have found.  But the reality of a continuous process is that there is always a remnant of any previous ingredients in the pipes, silos, or bulk tanks until that vessel is cleaned.  So for strict compliance with the Act (which does carry criminal penalties for non-compliance), every lot that could potentially have comingled with subsequent batches needs to be identified.  The only acceptable cut-off is a cleaning of the vessel, destruction of its contents, and re-filling with fresh lots.

If “reasonable” recall is the objective; where product may be substandard, slightly contaminated, but certainly non-lethal, a certain amount of common sense may be applied to recall traceability.  This becomes a risk management tradeoff of expected impact against costs.   If the liquid sugar used a month ago had a bad odor, perhaps recalling any product made yesterday from the same tank the tainted sugar occupied 400,000 gallons ago may be overkill.   But if that bad odor was a fast growing lethal biological, it could certainly still be in the tank – maybe even in greater concentration…  To comply with BTA 2002, a company must be able to track any contamination, fully within 24 hours.

Intuitively, for solids, one would expect that if we had a 300,000 pound silo, and lot 1 was 100,000 pounds, followed by lot 2 at 50,000 pounds, and lot 10 at 70,000 pounds, one could derive what went into subsequent lots on some kind of a FIFO (First In First Out) basis. (Perhaps with some small overlap for comingling at the imaginary line between lots.)  So once we had consumed 100,000 pounds, the assumed current lot would be lot 2; once we had consumed another 50,000 pounds, lot 10 would be in play.

However, this is not necessarily true.  For solid storage systems that release contents from the bottom of the vessel, product tends to flow in a conical funnel:  newer lots flow down the middle while previous lots hug the sides of the silo.  Therefore, we might find that production after 75,000 pounds contains elements of both Lots 1 AND 2 (long before the 100,000 pound line).  Also, production quantities after 160,000 pounds may still have some quantity of Lot 1 AND 2 as well.  It is NOT true that only lot 10 is flowing after 150,000 pounds…all three  may be  in use.

Still, depending upon the ingredient and environmental conditions in the silo, one could come up with a mathematical model that could provide a realistic estimate, but it would only be accurate under controlled conditions.

 

This is compounded in the case of liquid or gaseous ingredients.  They comingle easily and completely.  Any math assumed would have to know the exact quantity in the bulk tank when any additions were made and could only estimate dilution levels.

So in reality, any product consumed after any receipt will contain traces of any lot from the same vessel up to that receipt, back to the last cleaning, perhaps 3 years ago!

 

But it is not all bad news!  An alert on an ingredient does not necessarily require a recall of everything made in the last 3 years.  We can state categorically that a lot made on Jan 14, 2012 cannot possibly contain even a miniscule trace of any lot received on Jan 15, 2012.  That is simply not possible.

It may be possible to interpret some regulations pertaining to continuous processes  to exclude more previous lots, but any conclusion must be carefully considered, and defensible; and presumably specific to the process in question.  A silo that contains 40 different lots, that is filled daily, but “emptied” over a month, will have a different mathematical algorithm than a silo that contains 3 lots and effectively empties 4 times a day.

It is more likely that a terrorist chemical or biological contamination will be in a minor ingredient that is added discretely in smaller quantity to the process (or from the water supply which virtually nobody traces in lot traceability).  Smaller, concentrated quantities are easier to acquire, move, and hide.  Knowing what bulk ingredients are definitely NOT in a lot is substantially better than just guessing from paperwork… especially when there are only 24 hours with which to respond with complete information (as few as 4 hours if you are a co-packer).

 

Traceability and the small food company

You are the person responsible for food safety and regulatory compliance at your small family owned food company, and your firm makes cookies. Really good cookies. Consumers love your product, the food stores that distribute your cookies like your sales reps, the company owner is happy with your work, and you are looking forward to a little vacation with your family. It’s Friday afternoon before a long weekend, and life is good.

Then you get the phone call. A company that supplies you flour was just found to be in violation of cleanliness regulations, and their product is being recalled. Fortunately, you buy flour from two different sources so you should still be able to keep your loyal customer base supplied with cookies.

All you have to do is to trace which cookies the flour from the company in question went into.

No problem; you have all of the paperwork for the shipments, and all your batch tickets filed in your office. The owner’s son is home from university for the weekend, and if you draft him (and his girlfriend), and get the office staff and sales reps from the tri-state area to come in on Saturday, you should be able to produce the lot list for recall by Monday morning.

Good thing the recall was not initiated by the FDA under the Bioterrorism Act, or you would have had to be ready by 2 PM Saturday!

As long as none of the paperwork is lost, you’ll be fine.  Too bad about the vacation though.

There is a better way. A way that involves a few keystrokes, and about 10 minutes.

For the small food company that doesn’t yet run any other manufacturing quality software (Net Contents Control, Statistical Process Control, Equipment Downtime, Overall Equipment Effectiveness), we present QIC Trace Lite.

Using QIC Trace™ provides complete traceability on product from where the raw ingredients enter your building, all the way through to the lots that you ship out to your distribution channel.

Product traceability in the food industry is no picnic. Grains of salt do not have serial numbers, and many similar ingredients end up in a wide variety of products. QIC Trace™ was designed and written for the food industry.

Look at QIC TraceLite.   The company owner will appreciate you more.