We’ve carefully analyzed the cantaloupe crisis in pieces such as these:
When Elizabeth Weiss of USA Today broke a story — Listeria-linked Cantaloupe Farm Had Rated High In Audit — that detailed the fact that Jensen Farms had received a top score — 96% — in a Primus audit done just six days before the first person fell ill from these cantaloupes, we received many letters, which we will deal with in due course. One simple letter, from a very intelligent person who closely follows food safety, gives us a clue as to the issue the industry has to deal with regarding audits:
What is your feeling about the responsibility of the third party certifier of the safety of the melons ?? . . .
—Craig K. Harris
Department of Sociology
Michigan Agricultural Experiment Station
National Food Safety and Toxicology Center
Institute for Food and Agricultural Standards
Michigan State University
East Lansing, Michigan
Of course, anyone who works in this area knows that neither Primus nor any other auditor acts as a “certifier of the safety” of any produce item. There are no consumer websites urging consumers to buy audited product by these auditors and no corporate shield stickered on the cantaloupe or other items.
Although commercial buyers are free to set up their own standards — and the world is filled with organizations that have standards, and one can get an audit on one’s conformance with those standards — typically audits are simply scored against standard industry practices. They are not scored against optimal food safety practices at all.
That is what happened in this case. Jensen Farms did not have, and apparently Wal-Mart and others allowed it not to have, a Primus GFSI Audit. This was a standard Primus Good Manufacturing Practices Audit.
We have great respect here at the Pundit for Trevor Suslow of UC Davis, whose work we have featured many times including these pieces. Note that several specifically address cantaloupes and food safety:
We also deeply appreciate the work of Stephen Patricio, who is the chairman of both the California Cantaloupe Advisory Board as well as the Center for Produce Safety, and is president and CEO of Westside Produce.
Press reports lately, though, quote both Trevor and Garrett in a manner that mischaracterizes the state of the industry regarding standards on washing cantaloupes. The media is portraying it as if the Gold Standard practices of a half-dozen large California/Arizona firms are standard throughout the industry. They are not.
We would say the issue boils down to our original point that regional and local buying is interfering with food safety.
The issue is this: In the past Jensen Farms used a “hydro cooler” that had used recirculated water to wash and cool the melons. The recirculated water was treated with an antimicrobial. This season, it switched to a “one pass’ system that did not use recirculated water. It also did not use an antimicrobial.
We’ve been around this issue for more than 26 years and, in general, one-pass systems that do not use recirculated water are considered food safety advances. It is very hard to keep clean a dunk tank full of dirty cantaloupes with water being recirculated and a little chlorine is not a guarantee. Remember all the wash water during the spinach crisis was treated and that treatment did not prevent the crisis.
Still in the USA Today article, the quotes are clear:
Suslow, an expert on the post-harvest handling of produce at the University of California-Davis, said he was rendered ‘speechless’ at news that Jensen was using untreated water to wash its melons.
The problem, which Suslow called a ‘red flag,’ was a switch by Jensen to a new fruit-washing system in July 2011. According to the FDA report and Gorny, that month Jensen Farms purchased and installed a used potato-washing machine to wash its cantaloupe.
According to the audit done by Primus Labs in August 2010, it appears that Jensen Farms had previously used a ‘hydro cooler’ system to wash and cool the melons as they came in from the field, using recirculated water that was treated with an anti-microbial to kill bacteria.
For the 2011 harvest, the farm switched to a system in which cantaloupes were washed with fresh water that was not recirculated and ‘no anti-microbial solution is injected into the water of the wash station,’ the auditor, James DiIorio, noted on the first page of his audit.
‘You would flat-out never do that, absolutely not,’ said Suslow, who spent more than six years researching cantaloupe safety and handling. No matter how clean the source of water is, once it’s sprayed on ‘any kind of surface where you have multiple produce items rolling across it, you’re trying to prevent cross-contamination … so you always add something to the water.’
Suslow called this a ‘fundamental error with just tragic consequences. We can’t know that it absolutely made a difference, but I honestly think it could have prevented the scale and scope of what happened.’
The problems that were found at Jensen Farms are ‘Packing House 101,’ said Stephen Patricio, chairman of the California Cantaloupe Advisory Board. ‘Every common surface must be cleaned, rinsed and sanitized,’ he said. ‘These are all just known, recognized practices.’
‘It’s just disgusting to me,’ Patricio said of both Jensen Farms and Primus Labs. ‘I think of the damage that they’ve done to our industry as the result of this oversight. No, I won’t even talk about it as oversight, it’s abuse.’
Dr. Suslow and Mr. Patricio are, of course, free to attempt to persuade others that these are desirable steps, indeed that these are essential steps. They may well be correct. In fact we suspect they are correct.
It is worth noting, however, that they have not succeeded in persuading Wal-Mart or Costco of these facts. Neither retailer, nor any other retailer that we are aware of, has a specification on procurement of cantaloupes that all cantaloupes must be washed with an anti-microbial if the water is non-recirculating.
The FDA guidance on the matter makes no mention of any requirement for antimicrobial usage in single-pass or non-recirculating systems. See the highlighted and underlined section below:
Melon Cooling Medium
Melons typically are cooled by forced-air cooling or by use of a chilled water drench or flume immersion. Melon cooling with water, if done correctly, may reduce microbial loads on the outside surface of melons by 2-3 logs CFU. Microbial reduction on melon surfaces is dependent on disinfectant concentration and contact time. However, once present on the surface of a melon, human pathogens cannot be completely eliminated by washing.
Prolonged soaking of melons in aqueous solutions containing wash water disinfectants is not an effective means of eliminating surface microbial contamination of the melon rind and may actually aid in the infiltration of human pathogens into the edible portions by creating an infiltration driving force.
Melon cooling water also may be a significant source of microbial cross contamination if there is insufficient water disinfectant present. In addition, because melon cooling water is colder than the melons, infiltration of small amounts of cooling water may enter melons through the stem scar and rind. Forced-air cooling operations can avoid the risk of infiltration but also may spread product contamination if forced-air cooling equipment is not cleaned and sanitized regularly.
- Evaluating water quality of cold water used to cool melons to ensure that water is of sufficient microbial quality for its intended purpose.
- Evaluating and monitoring water disinfectant levels to ensure that disinfectant is present at levels sufficient to reduce the potential risk of cross-contamination when melon cooling water is re-circulated. If melons are fully submerged in water as a means of cooling, they are more likely to have cooling water infiltrate into the melons and consideration should be given to cooling water quality variables such as pH, soil (including organic) load, turbidity, and product through-put capacity, to ensure that the wash water disinfectant of choice is effective in reducing the potential for water-to-melon cross-contamination.
Using single pass (or one use) cooling water of sufficient quality for this intended purpose also may be used to cool product.
- Cleaning and sanitizing equipment on a regular basis to ensure that the potential for cross-contamination is minimized when forced-air cooling is used to cool melons.
There is no such requirement on other commodities either. For example, there are no requirements for antimicrobial usage in single pass or non-recirculating systems for tomatoes or in the rigorous California Leafy Greens metrics, which has this to say about post harvest systems:
Single Pass vs. Multiple Pass Systems
Single pass use – Water must have non-detectable levels of E. coli or breakpoint disinfectant present at point of entry
Multi-pass use — Water must have non-detectable levels of E. coli and/or sufficient disinfectant to insure returned water has no detectable E. coli (minimally 1 ppm chlorine).
If any one sample exceeds the acceptance criteria, then the water shall not be used for this purpose unless appropriate disinfectants have been added or until remedial actions have been completed and generic E. coli levels are within acceptance criteria:
Conduct a sanitary survey of water source and distribution system to determine if a contamination source is evident and can be eliminated. Eliminate identified contamination source(s).
For wells, perform a sanitary survey and/or treat as described in Appendix A Sanitary Survey.
Retest the water at the same sampling point after conducting the sanitary survey and/or taking remedial actions to determine if it meets the outlined microbial acceptance criteria for this use.
For example, if a water sample for water used to clean food contact surfaces has detectable E. coli, STOP using that water system, examine the distribution line and source inlet as described in Appendix A Sanitary Survey, and retest from the same point of use. Continue testing daily for 5 days at the point closest to use, and do not use the water system until it consistently delivers water that is safe, sanitary water and of appropriate microbial quality (i.e. Negative result) for the intended use. If any of the any of the five samples taken during the intensive sampling period after corrective actions have been taken have detectable E. coli, repeat remedial actions and DO NOT use that system until the source of contamination can be corrected
The issue is really Best Practices vs. Standard Practices. Dr. Suslow and Mr. Patricio work mostly in California and Arizona, and this is the location of the world class facilities that handle high volumes. These desert cantaloupes are either field-packed — and thus not washed at all — or packed with sophisticated wash systems that Dr. Suslow and Mr. Patricio are talking about.
It is important to note that there are fewer than ten of these facilities in the country, yet there are hundreds of places, maybe more, that pack some cantaloupe at some point in the year.
When the FDA goes into a facility, it has the gift of 20-20 hindsight and unlimited time and resources to solve a food safety issue such as this one.
It is thus able to identify many ways in which this facility falls short of the ideal. It would be a terrible mistake for the industry, though, to think that this means that this was some horrid facility. It was not. It was a perfectly mainstream facility, better than many.
The flaws the FDA found at Jensen Farms could be found in the vast majority of produce packing facilities with a team there day and night swabbing and looking for days and weeks on end. But as with Jensen Farms vs. the California/Arizona industry, these packing facilities account for a small percentage of the nation’s produce.
Auditors Not At Fault
Auditors don’t write standards and don’t write POs. Any buyer has the right to go to Primus or any auditor and say, “We don’t want standard industry practices… We only want to buy world-class best practices.” We have no doubt Primus would gladly change its audit and fail people left and right.
To expect auditors to impose such standards on the trade without the support of either government regulation or buyer demand is to place bizarre weight on a very thin reed.
The government, media and academia should understand, though, that, almost by definition, everyone cannot be “best of class,” so demanding a “best of class” standard will mean flunking lots of producers, indeed most producers. It would mean Wal-Mart would, during the domestic season, buy from the half dozen or so largest packers. Of course, nothing is stopping Wal-Mart from doing that right now.
Over and over again, we have been repeating our story about Costco and its problems with carrots a few years ago in Canada. You can see the original piece here. The question we raised was how — in a company really seen as an exemplar of food safety practices — did a buyer decide it was better to buy carrots from an obscure Mexican grower through an intermediary in Los Angeles rather than buy direct from, in our example, Grimmway?
We harp on this because this is the actual food safety problem in produce. The key to food safety is to know who you are buying from and to have worked with that company over many years to achieve shared values. There must be transparency and mutual commitment to work together.
In our last piece, we pointed out that a food safety expert such as Frank Yiannas at Wal-Mart would almost surely have preferred that Jensen Farms buy brand new stainless steel equipment specific for cantaloupes. That is true, but only part of the story. After all, farmers are famous for their mechanical flexibility and ability to use and adapt economical used equipment. Sam Walton would have almost surely said let us drive costs out of the system and see if we can find ways to incorporate the used equipment to produce safe cantaloupes.
The problem is that because there is no mutual commitment, there is little transparency. A packer such as Jensen Farms won’t want to have a conversation with Wal-Mart because if Wal-Mart says “buy new equipment,” the farmer or packer has no assurance Wal-Mart will provide a price that justifies purchasing that equipment.
What To Do About Audits?
Recognizing that this “blame the auditor” frenzy is bizarre, we can say that the industry does need a different kind of auditing tool.
Right now, the system is that auditors go out, produce an audit and they give a score that is typically cued to industry standard practices and put comments in the notes. In this Jensen Farms situation, for example, the fact that no anti-microbial was being used in the wash water was specifically indicated in the notes as a possible area for analysis. It wasn’t scored against Jensen on the audit because it is not required in the FDA Guidance or by standard industry practice.
In theory, what is supposed to happen is that the audit will be submitted to Wal-Mart, a food safety expert will carefully review it and many discussions will be held about each of these notations on the audit. Wal-Mart, Frontera and Jensen Farms will all come to a conclusion about whether the additional safety gained by adding an antimicrobial to the water outweighs its cost and perhaps other negative impacts, say chlorine on the workers. They will also come to an agreement on which antimicrobial to use and in what quantity. They will request a follow-up visit by the auditor to confirm this problem has been appropriately resolved.
This is the way current audits are designed to be used. If, instead, a secretary receives the audit and checks it off on a compliance spread sheet as the secretary would the receipt of a commercial license, then files it to never be seen again, the audit is not being used properly.
The problem is that this ideal procedure is almost never followed. Melinda Dwyer at Costco is passionate about food safety and tries to do this, but she can’t possibly do it for all of Costco’s vendors, and most companies don’t even have one person working on this.
This is one important reason why our references to an aligned supply chain as the key to food safety are so important. It takes a lot of resources to vet and stay on top of a supplier. A company must constrain its supply chain because it can’t possibly do the right job on a large number of vendors.
As a practical matter, what has happened is that audits are consistently misused. So when questioned about food safety, many buyers at retail and wholesale may just say “hey, these guys got a 96 on their Primus audit” and never even look at the paper.
So, with these facts and this horrid situation, we would lay out four key issues for the trade regarding audits:
1) What Standard Does Each Retailer Want?
There are many standards out there, but most audits focus on standard industry practices. That is what the whole movement toward harmonization of audits almost intrinsically leads to.
The first question a retailer has to address is whether it is satisfied with that standard. Wal-Mart, for all its pushing on food safety, didn’t demand a different standard of audit. It didn’t demand a GFSI audit, and it didn’t ask Primus to fail the facility unless it operated under all “best practices”.
The trade associations like to suggest that we can have a unified standard. We can, but, by definition, all God’s children can’t be above average.
2) Who Is Going To Determine “Best Practices”?
Ok, we now all realize that the audit is based on industry standards and, surely, many buyers will say that industry standards are too low and they want to audit firms for best practices.
Sounds good. But who decides what is a best practice? Is it dry harvesting of cantaloupes? Or is it washing with an anti-microbial? Which anti-microbial? In what concentration? Is just washing with anti-microbial enough, or should it be a one-pass system with anti-microbial?
These are not just details. Decisions such as these are the essence of food safety.
3) How Can We Make Audits Easy To Use?
It is clear that for the produce industry, the optimal outcome would be that a passed audit would symbolize a lot more than conformance with industry standards. To do that the comments would have to be moved into the body of the audit and scored against the company. Many more companies would get low scores or fail the audits. Is the industry prepared for this?
What about government? We would estimate that more than 80% of the packing houses in America would fail a “best practices” audit. Not only that but most cannot meet the standard ever. This is because they involve things like building pre-cooling facilities. Is Washington D.C. going to be OK with putting so many people out of business?
4) Frequency Of Audits Or Surprise Audits
One of the likely reasons that the Primus audit showed the facility in a good light is that they knew Primus was coming and gussied up the place.
There is nothing wrong with having a first audit on a scheduled basis. Then, intrinsic in every audit regime, has to be a follow-up audit to verify that the audit has been reviewed, non-conformities resolved, all comments addressed.
If something is important enough to put on an audit, then it is important enough to make sure it is being done.
Audits still can only verify conditions on a specific date and time. If the next day the facility head doesn’t like the change made to the pension plan and decides to neglect his job or actively do something evil, then the audit won’t catch it.
It is hard to do surprise audits, especially in foreign countries. Even domestically, word gets around quickly that the auditors are in town.
Still, the possibility of putting on a show for auditors necessitates more frequent visits. Certainly world-class foodservice firms such as Darden and McDonald’s have staff visiting facilities quite often.
Auditors could visit more frequently, but quarterly visits would cost almost four times what an annual audit does. Is the world ready to pay this price?
Auditors didn’t cause this problem. Yet it would be a great thing if any wholesaler on any terminal market could hit a button on a computer and say, “This is OK. It is audited to Best Practices.”
The system now is obtuse. It is as if a teacher gives a student a 96 because on the curve of other students he is doing great – but then writes a lot of comments saying that the student can improve in many ways.
Many won’t even look at the comments because the grade is expected, in school, to speak for itself.
It doesn’t in audits, and that is a big industry problem.