Dole’s Schwartz Sheds More Light On Recent Recall

September 21, 2007

We ran Dole Hit With Another Recall as soon as it was learned that the Canadian Food Inspection Agency had identified E. coli 0157:H7 on “Dole Hearts Delight” salad mix in a warehouse operated by Loblaws.

With a few days having passed, we asked Pundit Investigator and Special Projects Editor Mira Slott to again speak with Eric Schwartz, President, Dole Fresh Vegetables, Inc., in order to gain more perspective on the situation:

Q: After learning of the presumptive positive in Canada, how did you determine the scope of the recall? Why did you limit the U.S. recall to those 755 cases?

A: On Friday night, all we knew was one sample tested presumptive positive, the lot number and the code date, and we got all this information from Loblaws. CFIA alerted Loblaws Friday night and in turn our customer called us. CFIA did not release their alert until Sunday night when their test was confirmed. We also got this updated news from Loblaws.

When we got the call Friday night from Loblaws on the presumptive positive, we made our contacts with FDA. FDA leaves it up to the company whether to make this information known to customers. Since we didn’t have all the facts from CFIA, we determined how to proceed by conducting a risk-based analysis.

Our Canada production on that product was isolated to one individual lot, a total of 88 cases in that particular lot, and all product went to one customer, so it was easy to contain.

When we got the positive confirmation Sunday night, we decided to bracket around that 88 case lot and build a safe zone around it. The entire batch of related product covered roughly 14 big steel baskets. We had records of those 14 baskets. We said let’s grab that whole batch.

The actual product in question involved the initial 88 cases in Canada. We were casting a very wide net by recalling the U.S. portion. Based on what information CFIA is sharing with us, we had enough samples to suggest this was not a widespread problem.

Q: Could you clarify how you narrowed the U.S. recall to those 14 baskets? For example, did you recall all product that used the same blend; that went through the same production line during a confined period of time before the plant was sanitized? How are these decisions made?

A: In determining what action to take, we have to look at how far we should cast the net to protect the public. This is a very subjective call of what would make us comfortable, such as the number of days from shipment when no one is sick. When we think of how the product was produced, that the problem involved 88 cases, or 500 bags and CFIA found one problem bag, the recall decision was actually casting a pretty wide net.

When we talked about doing a recall in the U.S., it was a very subjective call; nothing scientific behind it. Those baskets were all produced for that particular blend, the same blend as those 88 cases in Canada. The scope of the recall in the U.S. covered all that same blended product run at the same time as those 88 cases in Canada.

Q: What if that blended product contaminated other product being processed in the plant? Wouldn’t you need to set parameters based on when the production lines that ran that blended product, or components of it, were last sanitized, etc.?

A: If we make the assumption that the product could have contaminated machinery that other product went over, then the recall could involve significantly more product. CFIA regulators don’t share all their information until their investigation is completed. We assessed the scope of the recall based on our analysis of the information as we became aware of it.

CFIA met with our outside consultant on Tuesday to discuss the case. We learned at that time, CFIA had tested at least 40 bags of that 500 bag lot. Since we knew CFIA had tested multiple bags time-stamped in different time frames, and only one was a problem, we knew we were dealing in a limited scope.This suggested an isolated case.

When there were no illnesses called in from either Canada or the U.S., we saw no wider scope than that. If we had reason to believe there was danger, or that illnesses were reported, we would have done a wider recall.

When our customer in Canada called, we knew the particular batch. We hadn’t tested our own product yet, but we had companion samples to do our own testing.Companion samples are random bags we pull of the packaging line when product is being produced so it is representative of the particular batch that was mixed and packed.

Q: Why the gap between the knowledge of the presumptive positive on Friday night and the decision to expand the recall in the U.S. on Sunday night? Why wait until the positive test came through?

A: It looks like we delayed expanding the recall, but in fact we had no reason to believe the problem went beyond that 88 cases in Canada. We just wanted to cast a wider net on our part to be extra conservative.

A reporter asked me, since there might be a problem with romaine, why don’t you recall all your romaine?It’s in 80 percent of products we make. There has to be a line drawn so that you’re not recalling for recalling sake.

The fact is that in this particular case in Canada, there was one positive sample, not multiple samples. This suggests it wasn’t something in the plant streaking other product.

If regulatory folks would be more open about what information they have upfront, we’d learn much faster the extent of the problem in order to take the best action.

Q: What information from CFIA would have helped you in determining what actions to take in this instance?

A: Their testing methods, how many samples, where they collected product samples from. That recall net we cast on the U.S. wasn’t based on a collaborative discussion with CFIA. All they said was they got a presumptive positive. We had no idea how big the sample. We heard from Loblaws first. CFIA didn’t call us. It was our customer that informed us of the problem.

The industry could be helped in making the decision if regulatory folks in Canada and the U.S. share what they know immediately. FDA does the same thing. They release information on a what-they-think-you-should-know basis as opposed to sharing all information.

Q: From what you describe, recalls and their scope often come about in subjective, arbitrary ways, and the FDA plays a peripheral role unless illnesses are involved. Don’t you think this kind of system could make a consumer uneasy?Even the most ethical and well-informed companies will have conflicting interests that could bias their recall decisions, or at the very least the perception would be there.

A:The FDA will make an announcement if illnesses are involved, as they did during the spinach outbreak last year. I don’t know if there’s an official policy here. Recalls are voluntary but if the company doesn’t act, they do.

It comes down to a risk-based assessment. When processing product, we use chlorine to kill pathogens, as the whole industry does.The big mystery is obviously that the wash system is not full-proof. A big symposium in Washington D.C. [Thursday] with regulators, scientists and industry folks is examining how we accelerate food safety research.

Think about the meat industry. There’s a U.S. inspector in each plant and that hasn’t stopped recalls. And consumers generally cook meat products.

The produce industry doesn’t have a kill step.We can minimize factors, but more research needs to be done. Irradiation doesn’t work on produce in the same way it does on meat. Produce has natural bacteria, and irradiation can’t properly distinguish between the good and bad bacteria.

We’re doing research on this like other companies to discover some form of irradiation that is proven to be a kill step. Even then, there will be resistance because consumers associate irradiation with radiation and cancer.

This outbreak is serious in the sense that a confluence of circumstances — the proximity to the one-year anniversary of the 2006 spinach outbreak, the prominence of the Dole name and the fact that Dole was implicated in the 2006 spinach outbreak, the fact that many politicians and interest groups are not satisfied with the current produce regulatory system and so are looking for any excuse to push their agenda — has added up to much more publicity than an incident of this size normally would muster… 1,061 news articles in the past few days according to Google.

This can’t be good for sales. Consumer confusion would be expected to impact both other Dole fresh-cuts and other brands, as consumers may not memorize the exact brand and exact mix that is implicated. Even if they aren’t confused, it will surely make many consumers hesitant to shop the category.

Bruce Peterson argued here that the produce industry could limit the impact of recalls with excellent traceability systems. To an extent, that is inarguably true.

Yet one of the lessons that is clear from this situation is that it is not completely true. When one brand of soup has botulism, the assumption by regulators, the media and consumers is that the problem is confined to one company. Something in their factory is wrong.

In this case, though, the focus by the media was immediately on the California Leafy Greens Marketing Agreement and its adequacy to the task.

Partly this is because opponents of the Marketing Agreement brought the subject up, and partly it is because the whole industry has been bound together under one safety protocol.

Although, technically, it is true that the CLGMA is just a base, and anyone can add to its standards, to publicly promote that one has higher standards than required by the CLGMA would be to imply the inadequacy of overall industry standards.

So our fates are bound to one another in the industry — even if we have perfect traceability.

Some specific points:

There seems to be a problem, and we heard this with the Metz Fresh recall as well, with getting information out of CFIA and FDA. Companies are expected to be free and open with all their information but the government agencies only slowly, if ever, detail what information they have. This reveals an intrinsic conflict in the multiple roles these regulatory agencies have. From a food safety perspective, the agencies ought to share all information quickly as this helps define the scope and certainty of any problem. One assumes that they hold back because of their regulatory role and the possibility that charges might be filed against companies. They don’t want to “tip-off” the companies as to areas of concern for fear the companies might alter records or engage in a cover-up.

Beyond determining the extent of recalls and understanding the government’s trepidations, we are concerned that so little information is given out that we can’t evaluate the correctness or meaning of governmental claims. Labs make mistakes, they get false positives, they contaminate their own samples — the public is entitled to know what tests were used, who did the tests, what confirmatory tests were done, what safeguards have been put in place against contamination, etc.
Perhaps the most important lesson for the industry that comes out of this is the need to sit down with regulators and reassess how we go about figuring the extent of a recall. Dole is one of the largest and most sophisticated companies in the business. They have staff scientists, access to the best consultants, etc. Yet, here is what we learn about how the extent of a recall is determined:

In determining what action to take, we have to look at how far we should cast the net to protect the public. This is a very subjective call of what would make us comfortable, such as the number of days from now when no one is sick. When we think of how product was produced, that the problem involved 88 cases, or 500 bags and CFIA found one problem bag, that’s a pretty wide net.

When we talked about doing a recall in the U.S., it was a very subjective call; nothing scientific behind it.

And this is our problem as an industry. We know Eric Schwartz, we know a lot of people at Dole… nice folks, good values, definitely want to keep everyone safe. To us, we can accept that a judgment call is being made by these people.

Yet it strikes us that this is probably not going to be acceptable to the public at large. These people are strangers to consumers. Consumers in general and certainly the anti-business advocates who push for regulation are going to say there is just too much of a conflict of interest. Executives of a company, whose personal bonuses and wealth may be tied up in the financial success of that company are going to have too strong an incentive to narrow the scope of a recall.

These advocates will push for either a “rules based” system — in which some hard and fast rules are established such as in the event of a recall, you recall from the last sanitation of the line to the next — or they will push for a third party, such as the FDA to certify any recall plan as adequate.

It appears this recall was more than adequate — but the system for arriving there isn’t one that consumers will have faith in. We need to devise a system that will more effectively build consumer confidence.
There is something wacky with the testing in this industry. Dole isn’t speaking yet, but if CFIA tested 40 bags and found one positive — the odds are that Dole’s retention samples will be negative. This means one of four things: 1) The tests are giving false positives and there was never any E. coli 0157:H7 there to begin with. 2) The E. coli 0157:H7 may really be there but it got there through contamination in the lab or somewhere else. 3) Our washing systems are effective but not perfect and we can get rid of 99.99% but still have to deal with small residues 4) E. coli 0157:H7 is so episodic that one bird on one plant can put it on one leaf and we sometimes can’t wash it off. If, as an industry, we need to be looking for these kind of minor events, we have to rethink our food safety programs all together.
Yesterday we ran a piece that contrasted the FDA’s passivity with CFIA’s action. The FDA did nothing on its own initiative; the CFIA issued a public advisory not to eat the product. Although the point remains, the specific timeline was off. Eric Schwartz explains that CFIA waited for the confirmatory test before it published its advisory.
The reason that Senator Dean Florez and other opponents of the CLGMA border on the ridiculous is that nothing in any of their arguments would have prevented this incident. Dole signed the CLGMA and is being inspected and whatnot. Of course, upon learning that the romaine in the blend came mostly from Salinas and some from Colorado, the green leaf in the blend came from Salinas and the butter lettuce in the blend came from Ohio, a Salinas friend with a wry sense of humor immediately responded: “Clearly, it was the butter lettuce.”

Salinas has had its share of problems so we can scarcely begrudge him a joke, but if a lynchpin of credibility of the CLGMA is that state inspectors have the legal authority to walk on the property at any time once someone signs that agreement, well, the fact that this is missing in other states will make use of their product increasingly problematic. Especially if Arizona does launch a Leafy Greens Agreement — meaning that producers could use Marketing Agreement all year.
One wonders if the industry shouldn’t urge the positioning of USDA inspectors in fresh-cut plants. Although it is not likely to make any difference on food safety, it might build regulatory and consumer confidence. Maybe the cost would be worth the boost in public confidence.
This is all so troubling and so far afield from what we all want to do with our time that we yearn for a “kill-step,” yet Eric Schwartz says irradiation is not ready and, even if it was, it is problematic. We think we need to accelerate research in this area. Every month now, a new tropical fruit seems to be approved for import subject to irradiation. This frequency of use and a generally improved atmosphere for things nuclear as concerns over global warming sour environmentalists on fossil fuels may be opening a marketing window.