Click here to get the PDF

October 17, 2007


The Honorable Tom Harkin
Chair, Senate Agriculture Committee
U.S. Senate
Washington, DC 20510

The Honorable Collin Peterson
Chair, House Agriculture Committee
U.S. House of Representatives
Washington, DC 20515

The Honorable Herb Kohl
Chair, Senate Agriculture Appropriations Subcommittee
U.S. Senate
Washington, DC 20510

The Honorable Rosa DeLauro
Chair, House Agriculture Appropriations Subcommittee
U.S. House of Representatives
Washington, DC 20515


Dear Senator Harkin, Representative Peterson, Senator Kohl, and Representative DeLauro,

In 2007 there have been 16 recalls of meat due to potential E. coli O157:H7 contamination, covering more than 29 million pounds of product, all of which carried the USDA stamp of inspection. USDA inspectors were in these plants, ostensibly every day, using the Agency’s “scientific” inspection and testing program. The USDA’s Food Safety and Inspection Service (FSIS) is now promising action. But this is just the latest round in a cycle of illnesses, recalls, and minor policy adjustments that has been going on for the last decade.

The massive 1997 Hudson Foods recall (35 million pounds of ground beef associated with 15 illnesses) and the 2002 ConAgra recall (19 million pounds of ground beef associated with nearly 50 illnesses) were both followed by public meetings and changes to the Agency’s policies on E. coli and recalls. The 1997 Hudson recall was followed by a public meeting on the Agency’s recall policy and six months later by implementation of the Hazard Analysis and Critical Control Point (HACCP) inspection system and a new E. coli O157:H7 testing protocol. The 2002 ConAgra recall was followed by an Agency announcement that it would ensure that all plants would take necessary steps to prevent H7 contamination and another public meeting on the Agency’s recall policies.

Now, in the wake of the Topps recall of 22 million pounds of ground beef, the Agency is announcing some minor changes to its policies and another public meeting to discuss its recall policies. Ten years later, the cycle continues, without any acknowledgement by USDA that more drastic changes are needed.

The Topps Recall

The largest of the 2007 recalls is illustrative of the ongoing problems with USDA policy on E. coli. The Topps recall, linked with 32 illnesses to date, involved the recall of a year’s worth of ground beef product, the closure of the nation’s largest producer of frozen ground beef patties, and the loss of employment for over 80 workers. Despite numerous statements made since the Topps recall, the Agency has not identified the specific violation of an Agency mandate that allowed the dangerous contamination to enter the market. If this is the case, the Agency’s regulatory approach to E. coli O157:H7 is clearly inadequate.

Equally disturbing is the fact that it took the Agency more than three weeks after officials were notified by a consumer in Florida about the first illness associated with Topps product before it even sent investigators into the plant to assess if there might be a problem. It appears that at least two victims might have avoided the contamination had the Agency acted sooner on the evidence before it. Agency policy is that unless there is a conclusive laboratory test linking an illness and an unopened container of product, there will be no recall. Due to this policy, the Topps recall occurred only because the New York State Department of Health established a conclusive link between a reported illness and an unopened package of product in a grocery store.

To date, the new policies issued by the Agency in response to the Topps recall are frustratingly weak:

• Over the years, the Agency has described, in general terms, many practices that plants might use to reduce the likelihood of E. coli 0157:H7 contamination. It has never recommended that all plants adopt them. Just like after the ConAgra recall, plants will again be required to reassess their food safety plans. This time the Agency will use a checklist to record which specific recommendations are adopted by each plant. But once again, there are no mandates or performance standards that plants must incorporate in their plan.
• The Agency is now proposing more aggressive follow-up testing after an E. coli 057:H7 finding “unless the establishment cannot continue to operate under that sampling frequency.” This caveat seems to offer plants a way out of being tested, which calls into question how effective the new testing program can be. Additionally, the sampled companies get so much advance notice that samples may not be representative of standard production and therefore may not be reliable.
• Once the Agency learns of product contaminated with E. coli O157:H7, the Agency will conduct a review of the plant’s food safety system. But the new policy allows 30 days for this to happen. This is no improvement over what happened in the Topps recall.
• The completed checklists from reassessing companies’ food safety plans will be analyzed to form the basis for “targeted” E. coli 0157:H7 testing, but the Agency has not demonstrated that it understands which of its recommendations, or combination of recommendations, will be most protective of public health. And given current resource constraints at the Agency, “targeted” testing likely means that some plants will receive minimal testing while others receive more. Previous Agency assumptions along these lines served as the basis for exempting the largest plants from testing – a policy that was implicated in the massive 2002 ConAgra recall.

In light of these inadequate proposals from USDA, we feel that Congress must step in and facilitate an independent and public review to get straight answers about the Agency’s policies regarding E. coli O157:H7 and product recalls. Specifically we urge you to look into following areas:

FAILURE TO REQUIRE SPECIFIC PRACTICES

Under the Agency’s current HACCP inspection regulation, very little is mandated or specifically prohibited by Agency regulations. As long as companies provide documentation that indicates that its particular processes, techniques, or methods will likely be effective in minimizing food safety hazards, the Agency applies its stamp of approval to the company’s product.

This “flexible” approach has several problems:

• These company plans are often insufficient to ensure a safe product. While consumer groups have often advocated that written company plans receive prior Agency approval, the Agency has rejected this approach. Reviewers, such as the Government Accountability Office and USDA’s Office of Inspector General, have documented the inadequacy of company plans . Companies can change plans as often as they want and are not required to notify USDA inspectors of any changes, even though plans can be hundreds of pages long. This is especially difficult for inspectors assigned to multiple plants, each of which might have 10 plans or more.
• While the Agency claims to require “verification” that the company techniques are working, the ConAgra, Topps and other recalls demonstrate that the Agency’s requirements for proof are not sufficiently rigorous.
• The Agency’s inspection system now consists of little more than making sure that the company follows its own plan. Prior to 1996, inspectors enforced specific mandates and prohibitions. For example, it was mandated that fecal contamination be trimmed from beef before it entered commerce. This approach was discarded and replaced with HACCP. Under HACCP, all inspection tasks focus on determining whether a company is following its own plans. Inspectors do not even need to determine for themselves whether a company is following its own plan. Often, they check a company’s records of its own employees’ findings. As long as the company’s records indicate that its employees found no problems, the inspector documents that s/he performed an inspection and that the company was in compliance.

The Agency maintains that it is sufficient to rely on “guidance” documents for the industry rather than requiring specific practices. Because the guidance is discretionary, the market pressure to avoid the cost of additional safeguards is increased. It is only in the unusual case, when very infrequent Agency tests discover contamination in products headed for market or the even more rare situation when tests can demonstrate a conclusive link between illnesses and a product, that the inadequacy of the Agency’s policy is revealed.

USDA inspectors have reported that plant management has rejected their suggestions for practices to deal with E. coli, such as limiting or even rigorously documenting the use of rework (using the remainder of yesterday’s production in today’s product), rigorously documenting the specific source of the raw products they combine into ground product, or slowing line speeds so that all fecal contamination can be detected and removed. What’s worse, inspectors have been told by their supervisors not to make such recommendations because it is the plant’s responsibility to devise its own food safety plan.

FAILURE TO FIND THE SOURCE OF CONTAMINATION

E. coli O157:H7 contamination occurs at the slaughter plant when fecal matter comes into contact with meat. Unless the Agency identifies those slaughter plants that are not preventing this deadly contamination, it cannot use its regulatory authority to make these plants do a better job. Less than 150, 000 pounds of the nearly 30 million pounds of product recalled in 2007 was recalled from plants that slaughter cattle as well as grind beef. That means that the vast majority of product was recalled from plants that do not slaughter – and could not have been the origin of the contamination. Regulatory action at those plants will have little if any effect ON the slaughter plants which originally contaminated the meat.

While the Agency has said that it does follow-up at supplying plants, some grinders involved with previous recalls told us that Agency officials have not been interested in learning about their suppliers. The Agency has said that it reviews the records at plants identified as suppliers who contributed to a contaminated lot, but Agency officials admitted to consumer representatives that it was unlikely those records would reveal if the plant was the source of the contamination. Further, while the Agency has conducted some sampling at supplier plants, plants are given one or more days advance notice and therefore test results may not reflect normal conditions.

The Agency seems now to be proposing to do extensive follow-up sampling at supplying plants (although the procedure is not particularly clear). But, in keeping with its policy to date, the Agency will not be attempting to sample product from the same lot of product that might have contributed the contamination, which would enable it to identify more potentially contaminated product that reached the market through other channels. The Agency will not even begin sampling immediately at supplier plants. Instead, a monthly review by FSIS will identify plants that have provided raw supplies for product eventually found to be contaminated.

Finally, the new Agency policies again avoid any discussion of continual problems with fecal contamination at slaughter plants, the source of E. coli contamination. Inspectors continue to complain about their lack of authority to get large slaughter plants to slow line speeds, which would enable workers and inspectors to spot more of the visible fecal contamination than they can currently. Additionally, the increased use of chemical rinses, designed to kill microorganisms, may sometimes have the opposite effect. Rinses designed to kill pathogens in very minute or invisible contamination are used instead on large areas of contamination, spreading the danger over large areas of the carcass and sometimes rendering it invisible. Spray chillers in coolers can also spread fecal contamination until it is no longer visible. However, it cannot be assumed that the absence of visible fecal material indicates that these rinses are effective in removing all dangerous contamination. The fecal contamination found by inspectors in plant coolers, after all rinses have been used, confirms that even visible fecal contamination is making it off the slaughter floors. It must be assumed that invisible contamination is also.

INADEQUATE RECALL SYSTEM

Had it not been for the tests by the New York Department of Health which conclusively linked the strain of the pathogen making people sick with the strain found in an unopened box of product, there would have been no Topps recall. The Agency tested some product from a Florida victim’s freezer in the beginning of September but considered the match between the sample and the victim as inconclusive because the package was open. According to officials, the only other investigative testing performed by the Agency to identify potentially contaminated product in commerce was to test one box of unopened product sent to the Agency by Topps. Thirteen tests of that one box of Topps product were all negative.

The Agency could have taken a number of other actions early on. As soon as the Florida consumer identified Topps product as a potential source on August 31, FSIS managers could have reviewed the Agency’s computerized database of inspection findings from the plant. They could have reviewed the results of their E. coli O157:H7 and salmonella testing at the plant. According to the Agency, they did not do this until they had confirmation of several illnesses in several states with suspected links to Topps product. They could have immediately contacted the inspector in that plant or the local supervisor to determine if there were problems with the plants’ system. To our knowledge, this was not done. They could have sent in a team to take a close look at the company’s food safety system. The problems with the company’s program that eventually necessitated a recall of a year’s worth of product should have been evident at that time. But the Agency didn’t do this until the New York authorities established a conclusive link between illness and product.

In addition to the time wasted due to Agency policies, there is the more basic issue of the Agency’s lack of authority to order a product recall. The recent situation involving salmonella contaminated frozen potpies produced by ConAgra offers a vivid example of how much time can be wasted because FSIS does not have mandatory recall authority. Despite evidence linking illnesses in two states to the product, the company refused for several days to issue a formal recall, despite being asked to by state health authorities from two states that had been trying to find the source of several illnesses since the spring.

WEAK TESTING PROGRAM

Although the Agency conducts a routine E. coli O157:H7 testing program, it does little to identify contaminated products on the market. This year, less than 50, 000 pounds of ground beef was recalled because of Agency testing, but over 29 million pounds was recalled after consumers became ill and contacted health departments. A more aggressive response to all findings of E. coli O157:H7 by a plant or the Agency would likely lead to the identification of more contaminated products which should be pulled from commerce, before consumers become ill. Relying on consumer illness to indicate that a plant has a food safety problem is unacceptable, and indicates that the Agency’s current E. coli testing program is broken.

It is unclear what role the generic E. coli testing program really plays, even though it is supposed to serve as an indicator of food safety process effectiveness in slaughter plants. It remains to be seen if this year’s surge in E. coli recalls was hinted at in the results of generic E. coli testing done by companies, or in the Agency’s random E. coli 0157:H7 testing.

INSPECTOR SHORTAGES

Agency officials have stated that its inspector covering the Topps plant was also responsible for inspecting four other plants on a “patrol” assignment, which means that it was only possible for the inspector to be there for a portion of any given day. The reality of how little time is spent in each plant on a patrol assignment is best described by a seasoned FSIS inspector:

“[I]f an inspector has 5 plants in an 8 hour day, they are probably driving 3 hours a day and have only 1 hour [at] each plant. They would arrive, get dressed for the processing floor, set up their laptop computer and only then go to the production floor. Management has saddled us with an unworkable satellite hookup for the computers, so retrieving a 5K message can take up to five minutes ...if the computer doesn't lock up entirely. Failing to check the computer at least twice a day for orders has resulted in disciplinary actions and you are required to enter your [inspection results] data for each plant. I would bet that the Topps inspector has at most 40 minutes per plant of actual observation of conditions. Heaven help you if you have to write an NR. [Non-compliance Record – an official notification of a regulatory violation. Inspectors are instructed to write NRs for each violation, but many have told us this is impossible, because each NR requires approximately 1 hour of research time.]

In addition to the five plants the Topps inspector was officially assigned, chronic inspector shortages and vacancies often mean that inspectors are required to cover additional assignments – they complain of being routinely “doubled and tripled up.” We have learned of situations in which inspectors have been assigned to cover nearly 20 plants a day, despite the fact that the Agency admits that this would be impossible. Like USDA’s Office of Inspector General, we have repeatedly urged the Agency to allow inspectors to code their online report of inspection findings to indicate when they haven’t performed an inspection task because of such time shortages. The Agency has repeatedly refused, although this was a standard practice, prior to 1998.

Conclusion

All of these issues add up to consumers not being adequately protected from E. coli, even after years of debate and pledges of action by the industry and the Agency. We urge Congress to investigate the Agency’s actions during the Topps recall and policies on E. coli so that the response to this outbreak is not just a repeat of previous cycles. At a minimum it seems clear to us that Congress and the USDA must make the following changes:

• The Agency must be given mandatory recall authority. Allowing companies to decide when and if a recall is required wastes valuable time when consumers could be prevented from eating contaminated product.
• The Administration and the Congress must recognize that there are not enough inspection personnel slots currently funded to do the job properly. The Agency has been forced to concoct unreasonable inspection assignments that are putting the public’s health at risk. Additionally, USDA should instruct processing inspectors to record when they are unable to perform inspection tasks because of staff shortages.
• USDA must more aggressively pursue efforts to connect reported illnesses to products. The requirement that an illness be linked to an unopened package of product is unreasonable and wasted time in the Topps recall during which more consumers may have been exposed to risky product.
• USDA must immediately act when illnesses are reported and can be linked to a particular plant. The Agency should immediately look at company testing data, reassess their HACCP plan, and conduct more testing at a plant that has been linked to an illness.
• USDA should approve company HACCP plans and require them to include minimum requirements. Additionally, the Agency should require notification of USDA inspectors whenever a plant changes its HACCP plan.
• USDA must redesign its microbial testing program. Specifically, the Agency should require that company testing samples be selected in front of an inspector and original lab results be provided to the Agency; end the practice of keeping positive samples out of Agency databases if the product is removed from the market; increase the amount of government testing for E. coli 0157:H7; and require companies to institute “test and hold” procedures under which products are not released for sale until test results come back negative.
• When a processing plant is linked to illnesses or a recall is initiated due to testing results, the Agency must immediately begin to trace back to the plant’s suppliers. The Agency should identify all possible suppliers and do targeted testing of these suppliers to try to determine the cause of the contamination. The Agency must then identify and remove any other potentially contaminated product from that supplier from the market.

Attached to this letter is a list of questions we feel should be asked of USDA about their response to the Topps recall and their overall policies on E. coli as you look further into this issue.

Thank you for your attention to this important public health issue. Should you have any questions about this issue, please feel free to contact Tony Corbo in our office at (202) 797-6548.


Sincerely,


Wenonah Hauter
Executive Director


Cc:

Senator Chambliss, Ranking Member, Senate Agriculture Committee

Senator Bennett, Ranking Member, Senate Agriculture Appropriations Subcommittee

Representative Goodlatte, Ranking Member, House Agriculture Committee

Representative Kingston, Ranking Member, House Agriculture Appropriations Subcommittee


Attachment
Food & Water Watch Letter on E. coli
October 17, 2007


We urge you to ask USDA for the following information as part of an investigation into recent recalls of meat due to E. coli 0157:H7.

FAILURE TO REQUIRE SPECIFIC PRACTICES

• A report on the cause of the problem triggering the recall for each facility. Identify how many plants currently employ each of those practices identified as problematic and/or written Agency policies that regulate the use of these practices.
• The noncompliance reports, food safety assessments, and any other enforcement actions taken at these facilities since October 2002, when the last revision to E. coli regulations went into effect in the wake of the ConAgra recall.
  

FAILURE TO FIND THE SOURCE OF CONTAMINATION

• What actions FSIS has taken since 2002 to identify suppliers of components used in the production of all lots of ground beef determined to be contaminated; and any actions FSIS has taken at those supplier plants.
• Information regarding all impediments to the identification of supplier plants and taking action at supplier plants.
• How many noncompliance records were written this year for fecal contamination found in coolers (after final inspection)? What actions does Agency policy mandate in such cases and in what document(s) does this policy appear?
  

WEAK TESTING PROGRAM

• Results of microbial sampling at the plants with 2007 recalls and a comparison to results from other ground beef plants.
• How much generic E. coli testing has been done in 2007.
• Has USDA determined that the generic E. coli testing performed by plants in 2007 did not indicate any increase in the amount of E. coli contamination? If the levels of company generic E.coli findings were comparable to previous years, has the Agency determined that there was some shift in the proportion of E. coli that is pathogenic?
  
• Has the Agency looked into the possibility that company testing did indicate a problem but that companies didn’t act on it? Has the Agency pursued any investigation into what company testing programs can tell us about the wave of E. coli recalls this year?
  

INSPECTOR SHORTAGES

• Was the inspector assigned to the Topps plant ever doubled or tripled up, and if so how often? How many plants were then part of that assignment?