Fong's Folly: Trying to Test Our Way Out
Opinion
U.S. Department of Agriculture Inspector General Phyllis Fong made her annual presentation to the Subcommittee on Agriculture, Rural Development, Food and Drug Administration, and Related Agencies Committee on appropriations on March 2. Most of what she said to this group of U.S. House of Representatives could only thrill the policy wonks that make up the group's membership.
Reading through Fong's prepared notes, Food Safety News reporter Helena Bottemiller caught something that should make the meat industry take immediate notice, though. Early in her presentation, Fong surmised that the deeply debated N-60 testing regimen, the one that was supposed to make detection of E. coli 0157:H7 much better, wasn't working.
With N-60 in its infancy, a comment like that by almost anyone else might be labeled as grossly premature and dismissed immediately. Quoting President George I, though, "It wouldn't be prudent."
Here is what General Fong said: "...we recently completed an audit that assessed how the Food Safety Inspection Service (FSIS) samples beef trim for E. coli, which can contaminate products such as ground beef. Currently, FSIS' inspectors take 60 samples from large lots of beef trim to test. We found, however, that this procedure does not yield a statistical precision that is reasonable for food safety."
Explaining herself, she said, "Although 60 samples may be adequate to detect widespread contamination, more are needed when E. coli is less prevalent. FSIS' current sampling methodology results in detection of E. coli less than half the time when it is present in 1 percent of a beef trim lot. Accordingly, we recommended that the agency place its testing process on sounder statistical ground by redesigning its sampling methodology to account for varying levels of contamination."
And to explain why dismissing her comment isn't a wise course of action, she concluded this part of her presentation by claiming, "FSIS generally agreed with our findings and recommendations. In related audit work, we have initiated a review of the agency's E. coli testing protocols to ensure that beef trim is effectively collected and analyzed. Together, our beef trim sampling and testing audits should help bolster public confidence that FSIS' tests are accurately identifying E. coli and ultimately preventing contaminated meat from being distributed and consumed."
What is most alarming is Fong's assertion that as the industry gets better at eradicating E. coli O157:H7, they need to ramp up their testing and look harder for an ever decreasing contaminant.
The first alarm bell that went off in my head was a comment made years ago by a food scientist friend. In the early days of the E. coli panic, shortly after the Hudson beef recall that started the stampede toward HACCP programs, I asked him about the validity of testing.
"Give me 2,000 pounds of beef trim and enough time and I'll find E. coli in there somewhere," he said.
Of course, that was a lifetime ago, when we were all just waking up to the threat of this pathogen and discovering how prevalent it was. Billions of dollars spent on research, millions of pounds of recalled product and the development of multiple kill steps, and we've all learned a lot about identifying and controlling E. coli. Today, my food scientist friend might literally test every ounce of that beef and come up empty.
But Fong's fear is that he might not strike out.
So let's ask this question - if 60 isn't enough, how many samples do we need to take?
I asked ex-USDA Under Secretary of Food Safety Richard Raymond if he had read Fong's comment. "Yep, I had read it," he said, "and I don't necessarily disagree and already have a blog working its way from my mind to my PC on it."
About the proper testing regimen, he said, "I am not a statistician, so I really cannot tell you what amount of testing reduces the risk to what level, but we all agree that the more you test, the more you will find."
"My question is, what will it cost to reduce the risk just a little bit more? Raw beef will still be a risk, it won't be risk free. If we cut the risk in half, what impact will that have on illnesses and deaths if consumers continue to not understand that raw beef must be handled and cooked with care?"
Dr. Jim Marsden, Regents Distinguished Professor of animal sciences at Kansas State University and Senior Science Advisor with the North American Meat Processors Association, agreed with Dr. Raymond. "The problem is that due to non-uniformity in distribution and the very low incidence of E. coli O157:H7, no amount of testing will provide an acceptable statistically significant confidence interval," he said.
Raymond talked about the inception of testing trim. "Just a few years ago, we were not testing trim at all," he said. "We began this process back in about 2006 or '07 - lots of testing, lots of dollars. What did that testing do to the number of positives in ground beef, in illnesses, etc.? Not very darned much. Seventy people died from E coli last year. Twenty percent of E coli outbreaks were linked to ground beef. The rest were produce, drinking water, person to person, recreational water."
"Should we really invest more money on testing or should we invest it on processing aids like low dose irradiation of whole carcasses and/or pre-harvest interventions (vaccines) that will reduce the risk more than more testing? $400 million will vaccinate every cow in this country, reducing not only E coli in ground beef, but also on the ground, in the rivers and ponds, and thus reduce the illnesses from petting zoos, recreational water, drinking water, lettuce and spinach."
Dismissing Fong's politically correct call to greater action with hard facts, Raymond suggested a different approach that promises a better outcome. "Demanding more testing is the easy way to look good for and to the consumer advocates. Demanding a more sweeping but debatable intervention is tougher sledding. And does not produce as many sound bites."
And it would be premature to take the step she's suggested. For the first two months of 2011, there have been 1,775 samples pulled at the federally inspected plants producing raw ground beef and none have proved positive. That's zero percent so far, although we can expect a few positives during the usually more active summer months. In 2009, FSIS pulled 12,797 samples and found 36 positives for an amazingly low 0.28 percent, compared with 0.44 percent in 2008.
The downward trend may be approaching the point where no amount of additional testing - other than N-100 percent - will find additional positives. In other words, ground beef, which has the potential to be a dangerous food if handled improperly, has become one of the safest foods on our dinner plate.
Dennis R. Johnson, a principal with the Washington, D. C. law firm of Olsson Frank Weeda and a specialist in food safety law and regulation, responded on behalf of several other people I contacted. Johnson represents large and small meat and poultry companies and trade associations before the U.S. Department of Agriculture's Food Safety and Inspection Service.
Responding to Fong's review of N-60 sampling, he pointed out that the meat industry "met Healthy People 2010 goals for E. coli O157:H7 in 2004 and 2009. In each and every year, the incidence was below the Healthy People baseline. We note that this is the only Healthy People pathogen reduction goal which has been met."
Agreeing with Dr. Marsden, Johnson said, "With a non-uniform, non-random organism present at a very low incidence rate, there would never be enough samples per lot to provide an acceptable statistical significance. It is simply impossible to test food safety into the system. Unfortunately, OIG focused on the effectiveness of N-60 solely as a lot release program, given the precise question posed by the Congress."
Johnson sees N-60 as a verification tool to measure process control. He said "In 2002, the annual trim incidence rate for some in industry (using a less robust sampling program) was 1.76%; in 2010, even using the more robust N-60, the rate was less than 0.15%. Beyond the historical data, establishments also view the results of the day to determine whether there is a basis to question any negative finding. These factors were not considered by OIG in its assessment of N-60, so its assessment is 'flawed.' "
Does N-60 work as intended? Johnson thinks it does. "Perhaps the most telling statistic," he said, "is that in the 17 foodborne outbreaks since 2006 involving E. coli O157:H7 with which I am familiar, in only one was raw material from an establishment conducting a 375 gram N-60 excision trim even in the blend suspected of causing illness."
"I am not a statistician," he said, "but the odds of 50% of the raw materials only appearing once in 17 events are extremely low. N-60 has a demonstrated benefit when measured by the statistic that truly matters - improved public health."
Moving the mark significantly forward might require tactics that aren't so attractive to the politicians who write the rules and regs because many segments of the public irrationally fear those tactics. For instance, the USDA, fearing public reaction, has been sitting on the AMI's petition to approve irradiation as a processing aid for chilled beef carcasses for 5½ years. The FDA approved its use for iceberg lettuce and spinach in 2008 and managed to weather the storm. The feds behind it and the produce industry survived. Surely the USDA can take the same path.
© Food Safety News
More Headlines from Opinion & Contributed Articles »Discuss (4)
