FDA Hints at Peanut Field Problems
In a post a couple weeks ago describing my company's reaction to the initial PCA disaster, I reported my thoughts at the time:
But now we had a second recall in two years. Had something changed in the peanut supply chain -- practices in the peanut fields, in the shelling process, in the heat resistance of salmonella? -- that meant that industry practices which had previously prevented illness were no longer effective? Was general nastiness in the PCA plant enough to explain to their problems? (Maybe, but I have my doubts.) If it is general nastiness, why was the FDA considering moving peanut butter to the "high risk"category? Political cover? Or do they share some of my questions? (Emphasis added.)
The Atlanta Journal-Constitution reported yesterday that officials at the FDA shared some of my questions:
The FDA is examining peanut shelling operations that supplied raw peanuts to the Blakely plant, an official with the agency said. Depending on what they find, they’ll look at farms that supplied peanuts to the shellers.
According to the paper, the FDA is looking beyond conditions in the PCA plants to practices at the shellers, and ultimately in the fields, because the strain of salmonella found in one unopened jar of PCA-produced peanut butter is a genetic match to the same salmonella that was implicated in the Peter Pan recall two years ago. This suggests a source upstream from the two plants.
And, if the FDA finds problems upstream, we have to ask what this means for the scope of the current recall?
Plainview Recall Moves Forward; Is FDA Overwhelmed?
The FDA has updated its narrative account of the PCA crisis to focus on the recall from the Plainview, TX plant, including the following:
The Texas DSHS is currently notifying all first-level customers of PCA Texas products from January 1, 2007 forward that all products are subject to recall. FDA will audit 100 percent of those PCA customers to facilitate the removal of product from the marketplace.
There are several recall notices already posted on the FDA website, and many more press releases out on the web that haven't been posted to the FDA site.
Few have noted that the FDA field staff which is carrying out these recall audits is exactly the same staff responsible for routine inspections and other investigations. State inspectors are also contributing to the follow-up work, based on a couple of conversations I've had with field inspectors.
This is a vicious cycle -- the underfunded and overworked field staff is unable to conduct routine work in a way that might prevent a recall of this size, so the recall happens, pulling that staff for weeks on end away from routine work, so....
Meanwhile, the political classes and editorial writers yammer on and on about federal level reorganizations and a laundry list of other fairly useless "reforms."
Time for an FDA field staff surge. We can worry about rest later.
Just When I Thought the PCA Story Couldn’t Get Any Weirder
This morning the press is reporting that the same salmonella strain that caused the initial outbreak has been isolated from peanut butter made from peanuts that were shipped from the Texas plant. The same strain from Georgia and from Texas?
From the AP story:
An opened container of Vitamin Cottage peanut butter tested positive for the outbreak strain, which came from a Colorado resident who got sick, company vice president Heather Isely has said. Earlier, the same strain of salmonella bacteria was detected in containers of peanut butter that had been produced at a Peanut Corp. plant in Blakely, Ga.
U.S. Food and Drug Administration spokeswoman Stephanie Kwisnek said two samples of Vitamin Cottage peanut butter from two different consumers tested positive for the outbreak strain, but it was not clear how many containers were involved.
It's possible the Vitamin Cottage peanut butter was contaminated after it was opened, health officials noted. But the latest test results raise questions about how many of the outbreak illnesses — which have been attributed to the Blakely plant — came from other production facilities.
"Because of the public health risk posed by positive findings of salmonella associated with the outbreak strain at PCA's plant in Blakely, Ga., the FDA expanded its scope of inspections to include other PCA plants," said FDA spokeswoman Stephanie Kwisnek.
And from The Dallas Morning News:
Doug McBride, a spokesman for the Texas Department of State Health Services, said peanut meal at the plant in Plainview, about 50 miles north of Lubbock, was sampled Feb. 12 and tested positive last weekend.
"Our lab determined that it is the outbreak strain of typhimurium," McBride said. "We don't know what that means yet."
However, the fact the identical salmonella strain was found at both plants, more than 1,100 miles apart, suggests a cross-contamination of their product lines, health officials said.
Health officials are looking for some kind of link between the two plants.
Once again, putting on my wild speculation hat, I don't think they will find one. I can think of no reason to ship product from one plant to the other. If press reports are correct, both had the capacity to roast, package, and granulate. And freight, especially in less-than-truckload amounts, would be prohibitively expensive, I'd think.
I'm back to one of my original how did this happen? thoughts -- that these peanuts were purchased already roasted from somewhere and shipped directly to the two plants.
In the meantime, I'll be watching the CDC site to see if they post an update on the strains.
UPDATE: Another thought. Although the Virginia plant has not been implicated, it is at least possible that roasted nuts were shipped from Tidewater Blanching to the other two plants. Financially, I can't see that making sense, but at this point, who knows?
The Paradox of More Regulation
Fascinating talk, and something to consider while working out how to improve food safety after the PCA scandal.
Can We Test Our Way to Food Safety?
The Georgia Senate has passed a bill to require food testing and submission of positive test results to the state. If properly funded -- and this is a huge if -- I have no doubt such a plan would help improve food safety. But it is not a silver bullet.
The testing of food destroys the food. So you can never test 100% of the food and say that it is 100% safe. You have to take samples of the food (called your sampling plan), and based on the results of the test, you can infer that the food is safe. But different foods and different processes require different sampling plans. It also matters exactly what you are testing for.
Determining effective sampling plans for probably hundreds of different foods in hundreds of different settings would be massive, and I would add, highly technical task. (Unless Georgia goes to some crude one size fits all regime, or maybe S, M, L.) It is worth pointing out that a deep knowledge of statistical inference and sampling strategy design is not going to part of the skill set of the typical environmental health specialist.
The bill provides that if a company submits a testing plan as part of a food safety plan, and if the state approves it, then they can use that plan. The approvals would be a slightly less massive job. The prime benefit would be the advantage early on in getting a look at everybody's testing strategies. Some would be highly sophisticated, and could be used as models.
Before everyone jumps on the more testing bandwagon we need to be clear that monitoring a test regime of this magnitude will require a lot of staff and a lot of computer hardware and office space.
I'll also point out that the easiest thing in the world for an unscrupulous food manufacturer to do is to evade a regime like this. It is just as easy to pull your samples from a bulk package of previously tested product as it is off your line.
I'd rather see the money and effort go into boots on ground -- more, better trained inspectors and more frequent inspections.
Is Third Party Certification the Answer? An Answer?
There has been some coverage of the fact that PCA's Georgia plant had obtained a "superior" rating on an inspection by AIB, apparently required by Kellogg's. Read this fabulous post by Jim Prevor at Perishable Pundit on this topic.
When my company used to be an ingredient supplier we were required by several customers to be inspected by AIB International (then the American Institue of Baking). This was before my time, so I don't really know a lot about the effectiveness of private third-party food plant inspectors. Nevertheless, all kinds of pundits and bloggers, who know even less than I do, seem to be advocating for more extensive certification as part of the solution to improving food safety. Like many of the supposed solutions, this is no magic bullet.
My Post From the Daily Kos
My wife suggested I put some of my thoughts about the peanut recall on Daily Kos. The text of my post follows. There were a number of interesting and helpful comments, several of which I responded to. I hope to copy some of the comments and my responses over to this blog later today.
One long month in the peanut butter business from an insider:
Last month I was trying to sell peanut butter at the San Francisco Fancy Food Show when people started showing up at my booth and asking me if my peanut butter contained salmonella. This was not a good way to find out about a recall. Since we do all of our own roasting from raw peanuts, I knew our peanut butter wasn't affected by the recall. So I wasn't terribly concerned.
I then got a couple worried phone calls from our staff back at the office in Grand Rapids. Customers are calling. They want to know if our peanut butter is "safe." Safe is a hard word. Yes, our peanut butter is safe as in "he is a safe driver -- he's never had an accident, doesn't drink, drives a Volvo, doesn't speed." And we haven't had a problem in our 80 years of making peanut butter. But safe as in "100% guaranteed safe absolutely no risk under any circumstances" -- as the mother of a child with a chemotherapy-impaired immune system asked us -- no. There is no responsible food manufacturer who can say that.
Back home from the food show, I watched this recall unfold like a slow-motion train wreck. While our peanut butter wasn't affected, we needed to check our purchasing records for our retail stores. A couple of our suppliers issued recalls on some products, but nothing we had purchased. One bullet dodged. But we are still watching the FDA Recall List like a hawk.
We put a little notice up on our web site indicating that we weren't affected by the recall, and put ourselves on the primary list of unaffected products. We assured our wholesale customers that we were unaffected.
I next scheduled a series of small group meetings with all of our employees to talk about the recall, our responses, and to re-emphasize how important our sanitation and personal hygiene rules are. I am fond of saying about business problems "lighten up a bit, nobody is going to die." In the case of sanitation, that's not so.
I also told our employees that for a couple weeks every inspector in the nation was going to be tied up chasing down all the recalled products. Once that was over, we could expect the FDA to show up.
For the next two weeks I spent most of my time on line following the reporting as carefully as I could. This wasn't rubber-necking, but an attempt to figure out exactly what had happened at the PCA facility. In the case of the Peter Pan recall in 2007 the explanation given for the salmonella contamination was a roof leak. Fine. Keep the roof from leaking -- we're on it.
But now we had a second recall in two years. Had something changed in the peanut supply chain -- practices in the peanut fields, in the shelling process, in the heat resistance of salmonella? -- that meant that industry practices which had previously prevented illness were no longer effective? Was general nastiness in the PCA plant enough to explain to their problems? (Maybe, but I have my doubts.) If it is general nastiness, why was the FDA considering moving peanut butter to the "high risk"category? Political cover? Or do they share some of my questions?
On February 5th we shut down our nut roasting, confectionery, and peanut butter lines for our annual preventative maintenance period in which we really tear apart the equipment. Since we would be shut down for a couple weeks at least, I decided this was time to review all of our practices to see if we had holes in the safety net. In addition to a complete internal review, I have hired a sanitation expert from the largest food safety consulting company to come and review our practices as well.
Last Thursday (2/12) I was just finishing up a conversation along these lines and informing my senior managers of our plans when there was a knock at the conference room door. "Jeff, the FDA is here." Even though I was expecting this, I felt like I'd been called to the principal's office.
I met the inspector at the door with my production manager. The inspector showed his ID and badge (yes, they have badges, just like the police), and we went to a conference room. He explained that we were due for our annual inspection and that today was the day because his supervisor had picked us out and moved us up the list. The initial concern was that we had covered all the bases in checking to be sure we had not purchased or sold any recalled items. But in asking questions our inspector realized that we actually made peanut butter ourselves. After an hour or so of additional background questions on the company, products, distribution, and such matters, it was time to do the walk through the facility looking for Good Manufacturing Practice violations. We gave him everything he asked for, and answered his questions forthrightly.
That took another hour (it would have been longer had we been operating). The inspector left for lunch and checked with his office for additional guidance. The main issue was whether to return and take environmental swabs for testing, and whether to pull product samples for microbiological testing. Since we were shut down and hadn't run for a week, the decision was no. We will, as a courtesy, inform them of our next few peanut butter runs should they desire to inspect. We haven't received the written report of the visit yet, but I expect it will be clean, just as our last nine annual inspections have been.
A long story. What's the point? Much political energy is being put into reform proposals. But the rubber hits the road when an individual inspector walks into a particular plant. In our case, that is for 3 or 4 hours, once a year. The other 364 days it is up to me and my employees, working on food safety issues as I've tried to work through this one.
Is that a problem? In PCA's case, yes. But would an additional annual inspection have prevented it? Two more inspections? Three? Would combining USDA's food regulation authority with the FDA's have prevented this? How, exactly? Would having federal inspectors replace state inspectors help? The inspector who called on us had worked for the State of Michigan for years before moving to the FDA. Would some enhanced ability to trace food help? Only after the fact, when it is too late. HACCP plans for all food manufacturers? PCA had one. How about reporting all salmonella and other microbial tests to the FDA? How many people would it take to police and follow up on millions of tests conducted by hundreds of labs, and by companies in house?
I'm OK with all of the above, and with having my taxes raised to pay for it all. But the safety of our food is still going to depend on ordinary people struggling to do the right thing every day.
Peanut Corporation of America and the Ruin of Being the Low Cost Provider
This is not a defense of PCA or Mr. Parnell. But from the public coverage I can see exactly how he put himself in a box where shipping bad product might have seemed like a risk worth taking.
The peanut business has been brutally competitive on a price basis for 30 years. A penny a pound in price can make the difference between keeping business and losing it. So PCA bought peanuts cheap and produced cheap. Part of the produce cheap was, it appears, skimping on training, wages, sanitation, and plant maintenance. The other part was getting raw peanuts in the door and finished product out as fast as possible so as to keep inventory low and cash flow strong.
Everything goes fine for quite awhile. PCA makes money and grows rapidly. Then they get a positive salmonella test in the summer of 2006. At this moment, PCA was effectively out of business. Here's why: to deal correctly with that test, given the conditions cited by the FDA at their plant, the plant had to be shut down, essentially be renovated, and then all the equipment cleaned from top to bottom, inside and out.
This is a huge job. It would take a month or two, minimum. During that time, all of PCA's customers, used to fast turn around on orders and cheap prices, are long gone. Could they shift to the Texas plant? Probably not, given conditions, but especially given freight costs.
PCA was running with no room for error -- and errors in business are inevitable.
PCA was finished when that 2006 test came back -- or at least in deep trouble. But that is a hard thing to admit. Maybe a retest -- we've never had a problem before, and labs aren't perfect. The rationalizing begins, and is repeated, becoming habit. PCA keeps digging the hole deeper, like a hooked gambler, until they are busted. Busted financially, busted legally, and busted ethically.
What Actually Happened at PCA?
A long and completely speculative post concerning a situation about which we are never likely to learn the full story:
I remain a bit skeptical that just general dirty conditions could account for the amount of salmonella-tainted product found coming out of the Georgia plant. This is because in most cases nut production lines are largely closed to environmental contamination and because the first step in the process -- roasting at high temperature -- basically sterilizes the nuts. So the plant can be dirty, but "inside" the process, you have a continuous flow of "clean" material.
Never having been in the PCA facility, and not having seen their product flow, I'm talking through my hat, but I think that some of their internal practices spread contamination around and made the problem much worse.
The most obvious examples are the tests from 1/24/08 and 1/25/08, and the test from 9/24/08 and 9/26/08. In those cases, there was a positive test on granulated nuts, followed by a positive test on peanut paste. When you granulate nuts the nuts are passed through screens to separate the pieces based on size. You are left with too big, too small, and just right. One wouldn't want to waste the too big or too small -- they would go right into the peanut butter or peanut paste lines. One could also appropriately and routinely move product from the peanut paste line to the peanut butter line. So, once PCA had a positive result on granulated nuts, the contamination could easily be spread to other lines in the ordinary course of business.
So, if the granulated nuts were the source of at least some of the positive tests for peanut paste, and perhaps for peanut butter, how did they get contaminated?
Poor storage of roasted nuts in a generally nasty environment could do it, but I can think of a couple of other ways.
One would be "totes." Totes for raw peanuts are big sacks that hold 2000 pounds of nuts. When we move nuts from our oil roasting to our packaging equipment, we use very expensive covered stainless steel totes. But if one really wanted to save money, one could perhaps re-use fabric totes that had previously been used to hold raw product.
Another way would be processing contaminated roasted peanuts purchased by PCA from another source. According to some press accounts, PCA was always looking for a deal on raw peanuts. But it is also possible to buy roasted peanuts. And to find them on a deal. So, were I the FDA, I'd want to trace back upstream from PCA to see what I could find.
Finally, I thought of floor scrap. It is almost impossible, when granulating, to keep nuts from flying all over the place. Once they hit the floor, they should go straight into the trash. But an inspector would have to get awfully lucky to catch someone sweeping them up and dumping them back into the process.
None of these possibilities, of course, are mutually exclusive.
It’s A Wonderful Life
Watching the former employees come forward in the PCA disaster, I can imagine one of them saying to Mr. Parnell:
Do you realize what this means? It means bankruptcy and scandal and prison. That's what it means. One of us is going to jail - well, it's not gonna be me.