The Nunes lettuce recall, which we dealt with here, points out the weakness of testing inputs and then seeing results after product has been shipped. First, it doesn’t guarantee safety because the product could have already been consumed. Second, it doesn’t stop the bad publicity related to recalls.
My read of the regulatory agencies is that they really want product testing. As part of its own efforts to appease regulators and rebuild the confidence of consumers, Natural Selection Foods has created a product testing “firewall” for its facility:
“Most important is what we are calling the “firewall.” We will be testing all of the freshly harvested greens — spinach and everything else — that are brought to our facility before they enter our production stream. If pathogens are detected, the lot will be discarded. This program is modeled on the program successfully implemented by the beef industry and approved by the International Commission on Microbiological Specifications for Food. This “firewall” will prevent anything like thisE. coli-contaminated produce from ever entering our facilities.”
This is one way of doing it, and if it keeps pathogens out of the plant, it may even be the best way. My sense, however, is that in the end the FDA really wants product testing after the product is processed. The big cost to this is not the testing itself; it is that you really want a hold-and-ship system implemented, in which product is held until the test results come back “all clear”. (The test results take about 48 hours, so you are talking about the loss of two days of shelf life.)
This is a big problem for certain prepared foods that only have seven or eight days of shelf life — but most fresh-cuts have 14 to 21 days, so a day or two delay shouldn’t be impossible to deal with. If the system was done properly, we could possibly even allow shipping to cross country destinations as long as the product isn’t released from the truck until the “all clear” was given. This would mean no reduction in shelf life to the East Coast, for example. Recalled product should be so infrequent that the cost of the truck on the occasional recall should be inconsequential.
Still, if we could speed up test results, it would be better, so I turned to my friend Lou Cooperhouse, Director of the Rutgers Food Innovation Center, and asked him about the prospects for faster testing:
As you know, most microbiological testing for food products done at food companies today occurs with Petri dishes or Petri film that typically takes 24-48 hours, or longer. However, for the perishable food industry, this is far too long a time period, as simply waiting for microbiological results can easily consume 25% to 100% of the total shelf life of a perishable food product. This is obviously not a realistic situation.
As a result, many companies in the perishable food industry commonly ship products before finished product microbiological testing has been completed. They rely on “statistical process control” that occurs post-shipment, instead of microbiological testing that should occur (and be completed) pre-shipment for 100% of the lots that have been produced. Many companies don’t even test-finished products for pathogens like E. coli 0157:H7 and rely instead on environmental testing on a random and statistical basis.
An interesting model in the food industry is the sprouts industry, as sprouts have a very short shelf life and have been implicated in a number of food safety recalls in the past as well. There is an interesting (but technical) PowerPoint on the FDA website — link to http://www.cfsan.fda.gov/~dms/sprfuto/sprfuto.ppt — about rapid methods for testing spent irrigation water as an indicator for the microbiological condition of sprouts. Rapid methods for testing sprouts for E. coli are currently being practiced in the industry, and FDA has supported and promoted these efforts. However, such methods are not utilized in other perishable foods industries.
Another interesting rapid-testing model in our society is drug testing — which many of our companies utilize for evaluating prospective new employees. Drug screen results are given within minutes, and if an individual “fails” in the instant “presumptive” test, samples are sent to labs for confirmation. So such technology is used commonly in industry today
In my opinion, we should learn from alfalfa sprout growers, and even the drug testing industry, and adopt a quick-testing system across the perishable food industry, and insist that manufactures adopt procedures for microbiological testing of all lots of finished products before shipping.
Such test kits are now available, and presumptive testing can tell a food producer within minutes if harmful levels of E. coli or E. coli 0157:H7 are present. I believe that such kits are available for less than $5 per test. An organization called the AOAC is a not-for-profit scientific association that is the recognized clearinghouse for approved scientific methods used in analytical testing. They list such testing methods on their website at http://www.aoac.org/testkits/testedmethods.html #Microbiological
This site also has links directly to the manufacturers of these products as well.
So this is what we are looking for:
“…a quick testing system across the perishable food industry, and insist that manufactures adopt procedures for microbiological testing of all lots of finished products before shipping.”
Lou mentions the sprout industry and when I think sprouts, I think of my frequent correspondence with Bob Sanderson at Jonathan’s Sprouts in New England. I asked him if he thought that the experience of sprout growers might be helpful to spinach and lettuce growers. He was a little skeptical:
FDA recommends that spent-irrigation water from all sprout-production batches be tested for Salmonella and E. Coli 0157:H7, and that results be back in house prior to shipment of the sprouts being tested.
In some ways this seems analogous to the packaged salad situation, but it is different in some key respects. We can put collection cups at every drainage point on our growing systems, and so, in effect, get a water sample that has come into contact with every sprout in the batch. The sprouts sit there for another 48 hours after we take this sample, and so we can know exactly which sprouts are involved if we get a positive test result.
The “rapid” tests, which are widely used, still take the better part of 48 hours. There are 24 hour-methods, but I don’t know of labs that offer them yet. I read about even faster methods; don’t know about their “limit of detection”, pre-enrichment, etc.
I think the salad mixes present much greater challenges in terms of high-confidence sampling and testing. But people are clever…
Bob is clever too and serious about sprouts and food safety. Take a look at this page from his website that details, with photos, the food safety regimen for sprouts.
What is clear is this:
- We need product testing on vulnerable things such as fresh-cuts and carrot juice.
- It must be reasonably priced.
- It might take away some shelf-life, but a livable amount.
- It would be best if we can find quick-read tests
I think we need to stop chattering and just make it happen. Wal-Mart, Costco, Safeway, Kroger, Supervalu — call in your suppliers and ask them for a schedule of how quickly it can be implemented.
Resistance is futile — and foolish.