Dr. Richard Raymond, Under Secretary for Food Safety

Dr. Daniel Engeljohn, Deputy Assistant Administrator, Office of Policy, Program and Employee Development, Food Safety and Inspection Service (FSIS)

Dr. David Goldman, FSIS Assistant Administrator, Office of Public Health Science

Dr. Kenneth Petersen, FSIS Assistant Administrator, Office of Field Operations

UNDER SECRETARY RAYMOND: Good morning or good afternoon, depending on what time zone you are in. I appreciate the opportunity today to explain new actions we have taken at USDA to protect public health from the risk of E. coli O157:H7.

Since January there have been 15 recalls related to E. coli in beef this year, eight of those have been associated with human illnesses. In comparison, in 2006 there were only eight related recalls and none of those were related to human illnesses and in 2005 there were only five E. coli related recalls. So obviously something has changed.

We want to remind you that compared to 10 years ago, remember, the safety of the food supply is better than it has ever been. The number of recalls and the number of illnesses has declined in the last several years but as a result of what we are seeing this year we are going through a very comprehensive review of our entire inspection and recall procedures at this time and taking lessons learned from a number of recalls including the recent Topps recall. We have emphasized the need for us to do even more to strengthen our polices and programs.

Today we are going to provide an update on our stepped up efforts we initiated in the spring and summer of 2007 and also explain some new initiatives that we have either recently begun or soon will begin. We are announcing today that we will begin testing more domestic and imported ground beef components. FSIS will begin testing these materials that are used as components in raw ground beef, in addition to the beef trim that is already tested, which is the primary component of ground beef. FSIS is also requiring countries whose beef is imported to the U.S. to conduct the same sampling or at least an equivalent measure.

One area where I feel we have made progress is in getting recalls done more rapidly. FSIS now takes into account a broader, more complete range of evidence when evaluating whether to seek a recall or whether to take regulatory action. This gives the agency a credible approach to more rapidly taking action when certain types of evidence are available. In two recent cases, FSIS acted upon epidemiological evidence that linked illness to opened, FSIS-inspected product found in consumers freezers. There was more than 1 million pounds of ground beef recalled as a result of that and this would not have happened prior to looking at our recall procedures.

But we still acknowledge that we can and must do better. We are looking at training and staffing patterns to ensure inspection program personnel and supervisors are doing their jobs correctly. That they are held accountable, that they have appropriate workloads and appropriate supervision.

Based on the challenges posed by E. coli O157:H7 and what we have learned from recent recalls, I believe that we need to take additional time to strengthen our system and the data that supports it before moving forward with risk based inspection in processing.

We welcome the OIG report expected by the end of the year that is examining the data used in the development and design of risk based inspection in processing. We will take that report to further strengthen our system.

As one example, in the spring of 2007, concerns were expressed about the FSIS practice that allowed inspection program personnel to discard raw ground beef samples for product that was going to be diverted to cooking or destroyed, after testing positive by the industry. This essentially removed this raw product from the marketplace. This practice was followed because cooked or destroyed product did not pose a human health risk for E. coli O157:H7 when the plant had found it to be positive and either discarded it or cooked it.

FSIS implemented a new policy on September 28, 2007, under which FSIS inspection program personnel send samples to FSIS labs for testing irrespective of the company's test results. While the previous practice did not pose a human health risk, this new approach will allow us to increase the number of Pulse-Field Gel samples that will be entered into PulseNet.

Finally, I want to emphasize again, how important it is to me personally to base agency actions on the need to protect the public's health. As I have often said I did not move to Washington to oversee recalls, but instead to prevent food borne illnesses. Even one illness is too many. With the actions we are going to outline today for you, I believe we are on the right track to bringing those E. coli numbers back down to where historically they have been for the past two or three years. .

There are a number of key initiatives targeted to federally inspected plants that produce raw beef products, and I'm going to ask Dr. Daniel Engeljohn to further explain some of these actions. The Agency has outlined these actions for consumer and industry organizations and we have sought their support in working together to control this pathogen. I should also note that FSIS will also target outreach and training sessions around the country for small and very small raw beef processors, other stakeholders and FSIS inspection program personnel. This training will focus on new E. coli O157:H7 policies, as well as lessons learned from the recent recalls associated with E. coli O157:H7.

Now Dr. Engeljohn will you walk us through some of the steps we are going to be taking shortly.

DR. DANIEL ENGELJOHN: Sure, thank you. This is Dr. Dan Engeljohn. FSIS has determined steps were needed to ensure that inspection program personnel and the industry fully understand the nature of the challenge presented by E. coli O157:H7. The agency is ensuring that slaughterers and processors as well as FSIS will be able to identify an emerging problem as early as possible and to prevent contaminated product from entering commerce.

With that then, we are looking at testing and the analysis of trim. Based on preliminary data from the agency's beef trim baseline and scientific literature indicating that contamination of trim is related to contamination of ground beef, FSIS began trim testing in March 2007. We did not wait until the final analysis of the baseline. By testing earlier in the production chain to identify contaminated beef trim intended for ground beef, FSIS prevents this source from contaminating the ground beef available to consumers. This also gives the agency more data to analyze in determining and implementing the most appropriate actions to reverse upward trends.

FSIS also notified the beef industry that, as of November, all beef plants will be expected to verify that they are effectively controlling E. coli O157 during slaughter and processing. The agency also provided the industry with specific examples of minimum controls that would meet the minimum criteria for "well-controlled" processes. Identifying which establishments achieve the minimums, and which establishments do not, will provide FSIS the critical information on establishments with vulnerabilities.

With that then we've created a new checklist for verifying these controls. FSIS inspection program personnel will review both suppliers and process establishments based on this new checklist, once they complete specialized training beginning the week of Oct. 29. Data from the checklists will be completed in November and will be updated quarterly to help the agency more quickly identify potentially significant changes in an establishment's production process that may effect control for 0157:H7 and ensure the company takes more immediate action to support the changes through their own validation and ongoing verification. FSIS will analyze the checklist data and use it to adjust programs or policies as needed.

We are also looking at targeted routine testing whereby in January of this coming year, FSIS will begin routine targeted sampling for O157:H7 at slaughter and processing facilities. Currently, all plants have an equal chance of being tested. Under this new verification testing program, FSIS will test larger volume operations more frequently than in the past as well as those with a poor history or do not meet the minimum criteria for a well controlled process. Data from the checklists will be used to determine testing frequency for these establishments.

And now Dr. David Goldman will provide an update on keeping current with the latest science and our work with our public health partners.

DR. DAVID GOLDMAN: Thank you Dr. Engeljohn, this is David Goldman. And I want to start by saying among the recalls that Dr. Raymond mentioned just a minute ago, the eight recalls, so far in calendar 2007 that have been associated with ground beef products and human illnesses, four of those recalls were associated with frozen ground beef product. And while the agency does not know the significance of this fact it does provide me with an opportunity to remind consumers and your readers how important it is to cook ground beef products in particular to a temperature of 160 degrees in its internal parts and to remind consumers as well, for ground beef in particular, color is not a good indicator nor is the presence of clear juice that is sometimes reported. So I do want to emphasize that cooking to 160 degrees is really the only way to assure that the ground beef products that consumers are eating are safe.

It is important for FSIS as public health regulatory agency to continue to keep pace with scientific advances. Accordingly, in January 2008 we will begin using a newly developed test that will detect lower levels of E. coli O157:H7 contamination in products that we test. I want to remind people that the outcome of this new test use may be that we find more positive results than in years past once we begin employing this new test in January of 2008.

I want to turn now to the fact that one of our key efforts in protecting public health is always to work with and strengthen our communications with our public health partners, industry groups and consumer representatives as well as internally with our own inspection program personnel. I want to highlight a few ways we are doing just that with our partners.

On October 17, FSIS, along with the Department of Health and Human Services' Food and Drug Administration (FDA) and the Centers for Disease Control and Prevention (CDC), hosted a public meeting regarding E. coli organisms other than O157:H7 that may be related to food borne illness. FSIS expects that any future steps taken to reduce the prevalence of non-O157:H7 E. coli will be better understood by all food safety partners after such meetings are held.

Later this fall, FSIS will convene a meeting of experts and stakeholders to examine the current situation with E. coli O157:H7, in order to identify factors that may be leading to an increased number of positive test results, recalls and illnesses and importantly the additional steps that FSIS should be taking in addition to the ones that have been mentioned already and that the industry can take to lower the risk to human health.

Finally, this winter, FSIS will hold a meeting with its State and local public health partners, FDA, CDC, industry and consumer groups, about how to improve the effectiveness and efficiency of outbreak investigations and recalls conducted by FSIS in collaboration with these partners.

MODERATOR: Before we take questions, I want to remind everyone that an FSIS press release has been issued and is posted on the FSIS website at www.fsis.usda.gov

In addition, a comprehensive list of actions by FSIS to reduce E. coli O157:H7 is also posted on the FSIS website.

OPERATOR: We will begin the question and answer session. If you would like to ask a question, please press *1 on your touchtone phone. One moment as we await the first question. Our first question comes from Julie Schmitt with USA Today.

REPORTER: Thank you very much. Dr. Raymond, you said at first you guys will test for domestic and imported ground beef components. I guess that means trim if I'm understanding that correcting. Can you give us a rough idea of what percentage is tested now and what that will go to under your new plan?

DR. RAYMOND: Julie, thanks for the question. It will allow me to clarify what I meant, and evidently didn't say quite clearly enough is, the trim that we test is a new program that we started in March of 2007. Dr. Engeljohn, is 3,000 samples a year? -- 3,000 samples a year of trim will be tested. What I meant to imply was, we'll be testing other components that go into ground beef such as heart muscle, weasand meat and other entities that help, they comprise part of the ground beef that we have not been testing those particular products in the past, and we need to make certain that they don't have a high rate of contamination with E. coli than the trim does.

REPORTER: Okay. And so if you do 3,000 samples of trim, how many samples are you going to do of these other components?

DR. ENGELJOHN: We're going to initiate the program with 1,500 samples of the other components.

DR. RAYMOND: Then depending whether we find a higher or lower rate of contamination, we will adjust the sampling based on what we find.

REPORTER: So this doesn't really have anything to do with testing more imported product as it comes through the border then?

DR. RAYMOND: The different policy that's being written, directives going out we have informed our foreign trading partners that do ship ground beef and beef products to this country that we will assume, begin testing trim as it goes through our import houses. We are not currently doing that at all, but that will be increased to match our other testing programs at the import houses.

REPORTER: Thank you.

OPERATOR: Again, if you would like to ask a question, please remember to first unmute your phone and record your name so that we can introduce your question. No more questions?

MODERATOR: There's a few that are going to come back in; they forgot to unmute their mike.

OPERATOR: Our next question comes from Jeffrey Gold with the Associated Press.

REPORTER: Good afternoon. My question is regarding the Tops investigation in New Jersey. Has the agency concluded what led to the contamination? And has the agency made any findings as to the number of inspections the company has conducted prior to the recall in comparison to prior years?

DR. PETERSEN: For the conclusion as far as products perhaps became contaminated was -- I think we might have suggested this on some earlier calls, we've kind of got the facts nailed down now - when we look at that plant's production practices, they had a procedure where basically as they followed their procedure they would be grinding products that was tested and found negative for E. coli 0157.H7 first. So anything they received, largely trim, would be tested by typically the producing company. And then they would have a negative test result and grind that product.

And at some point in the last period of time they started, it appears, mixing and matching whether they did that, on some infrequent basis - but any basis of not receiving product that wasn't tested negative was not consistent with their program. So they had some product they were grinding that was tested negative and perhaps they would mix that with other product that was not tested at all. And so that doesn't mean that product that wasn't tested wasn't negative, but it means they don't know the status of it and so, over time, it introduces the opportunity certainly for contamination to go into their product. And when that kind of thing happens, we're going to have significant questions about obviously what they're doing. But that's the kind of situation that would lead you to have a one-year recall.

Then your second question if you could repeat that?

REPORTER: Has the company reduced the frequency of testing at some point over the past few months?

DR. PETERSEN: When we last looked closely at this plant - and "closely" means I think as you've seen in some of the reports, we have inspectors in the plant every day, and then I have on an occasional basis more specially trained investigators go in and look more comprehensively at what exactly the scientific design of these plants' programs are.

And the last time we had one of those specially trained individuals in the plant was about two years ago. And when he looked at the operation at that time, the plant was doing a higher level of their own testing. They were receiving products, all of which was tested negative at the time. And in addition to that they were also doing testing of incoming product as a separate check. And they were doing that roughly monthly.

And somewhere between two years ago and now they dropped to a lower level of testing. And then they also opened avenues, as I said earlier, for grinding material that had not been tested. So there was some modification to at least the execution of their program. And we've observed that-we started identifying this kind of thing really several months ago where plants over time, many times for good reasons, would make adjustments to their process. But if you do an adjustment now, an adjustment six months from now, another adjustment a year from now, then you may be putting yourself in a position for not having as effective a system as you would like. And so we've already put in place procedures - again this goes back several months - to begin capturing those kind of changes in a plant program. And there will be indicators to us that, well something's changed, and is that change something that causes us concern. That will give us a little bit of an early warning for these kind of changes over time.

REPORTER: Going back to the contamination source, you discussed that they at some point shifted to this mix and match system. But have you determined where the contaminated beef came from? Or is that investigation still continuing?

DR. PETERSEN: We're following up, but say on a typical day - and this is typical of any plant - most plants are not grinding a single source of material from a single supplier. They may grind from four, five, six or 10 different suppliers. And so all we would know is, perhaps one of them or more, but one of them at least, was the source. And so what we do is follow up to each of those sources. We look at what they're doing today, and we also do follow-up testing at each of those sources. And that's what's going on for suppliers of Topps as we speak.

Then of course initially with Topps there was three days of interest that really had three different genetic fingerprints of E. coli. That's consistent with what I said earlier for multiple, for an error that happens over time these, you know, these untested trims getting three different genetic fingerprints is unusual. But that indicates a breakdown in their system.

DR. PETERSEN: I think we're ready for the next question.

OPERATOR: Our next question comes from Julie Schmitt with USA Today.

REPORTER: Thanks. I'm sorry, but Dr. Raymond could you explain again what you're doing differently as of September 28? I didn't quite follow that about the cooked product contaminate. Could you explain that again? You said something was different --

DR. RAYMOND: Okay. I had to stop and think; there's so many different dates. I had to put the 28th to figure out which one we're doing, the 28th. Prior to the 28th we would pull a sample in a plant of ground beef and if the plant sampled of the same lot at the same time, we would hold that sample until the plant got their results back.

If the plant result was positive for E. coli 0157.H7, they would either destroy the product or they would cook it, and we would discard the samples because there was no threat to the public's health so we didn't need to know that it was positive. It was pointed out to us by some of the people we work with on a regular basis trying to get input as to how we can do our job better that that gave us a falsely low number of positive E. coli for a particular plant that had its own testing regime. And that was one of the criticisms of our risk-based inspection program, that our database was flawed.

To try to improve our database, we now pull that sample and we send it to our laboratories for testing. And if it is positive, of course we hope that the company has held and will either destroy or cook the product. If they didn't hold the product, we will recall it. And that's about half our recalls are for.

But even if the company does their own testing and destroys that product, we have now entered one more positive into the Pulse Net database at CDC which somewhere down the road may help us in our epidemiological tracing of outbreaks or illnesses. But it also gives us a better, more accurate score, databank, on that particular plant so that the inherent risk or the ability of the plant to control risk will be better reflected in our data sets.

REPORTER: Okay, I see. So the plant would test the product, it would be positive, but then you guys wouldn't retest it, so the result never went to Pulse Net?

DR. RAYMOND: We would have a sample pulled and we would destroy that sample; and it did not go to Pulse Net, that's correct, because the plant was cooking or destroying that product. So there was no threat to the public health. It's not a public health issue, but it is an accuracy-of-data issue.

REPORTER: Yes. I see like in about the recalls, I mean the product goes out before the FSIS test comes back. Are companies supposed to be able to do that?

DR. RAYMOND: Julie, no, it doesn't always happen that way. Last year, for instance, we had 20 positive samples and we had eight recalls because of positive samples. So that would indicate the other 12 had been held by the company.

Now it is an elective, a decision of the company at this time whether they hold or not. We do not require them to hold. It's an issue we will be looking at, just because the number of recalls is consistent the last few years of about 8 recalls because of tested product. And of course that's product that has gotten out into the supermarkets and grocery stores and refrigerators at the homes. So we will work with industry to figure out ways to further encourage holding while we test product. But it's not more than 50 percent of the product that tested positive was held by the company.

REPORTER: All right. Then just a final follow-up on the Topps. Have you determined with any more clarity why the FSI inspector who worked in the plant notice the reduced testing?

DR. PETERSEN: The reduced testing, I mean it's important, but I think for us as far as the actual recall and other actions this inconsistency in following the program was much more significant, where they have a program that says, gee, we're going to receive a negative test from everybody, and then we expect them to follow that.

And then the design of that program is something that we looked at, we're looking at it now. In fact we're already pushing out some new training. Did we here in headquarters sufficiently communicate to those inspectors that, here are the kind of things you should be looking for when you look at these programs? You know, when a plant says, we have these controls in place, what kind of documentation exactly should they be looking at?

And so in relooking at our training, we're giving them better clarity on what they should exactly be looking for so that these things don't get missed.

Now the plant was receiving other information from their suppliers-you know, information about, the HAACP plan that it was produced under. So there was some documentation, but it was documentation that we thought was full and complete.

And so the inspectors in looking at their training we wanted to make sure they understood that there may be different facts that they need to get with each shipment, and they just need to understand what those facts are.

Now we've been testing this plant roughly monthly like we do for virtually all ground beef grinders, and going back four or five years now we've had a string; all our tests have been negative, our monthly or 10-times-a-year tests at the plant. Looking back in time here I've got about 51 negatives. And so our independent verification was showing something quite different than it turned out at the end of the day.

REPORTER: Thank you.

OPERATOR: Our next question comes from Caleb Hellerman with CNN.

REPORTER: Hi. I was in mute there for a second. There's something you could expand a little bit more on the targeted routine testing and how different that's going to be. Are the total number of inspections likely to change? How much different will it be, that it will be more focused in that way?

DR. ENGELJOHN: On the targeted testing, specifically right now roughly each plant is tested on average seven times a year if you look at the amount of testing that we do for raw ground beef and the number of establishments producing ground beef.

Under the targeted program, a lower-volume establishment likely would be tested at least four times a year, and a high-volume, poor-performing establishment may be tested up to 11 or 12 times a year. So the likelihood of being tested is reflective of the volume of production and the performance of the plant. So rather than seven times a year, it will fluctuate between the four to approximately 12 times a year. Does that help?

REPORTER: Yes. That certainly helps give me a sense of it. And what would be the criteria for being a bad history plant? Is there a specific trigger for that or specific figure as far as volume or anything like that?

DR. ENGELJOHN: The criteria will be refined over time as we analyze our data, because data will inform our algorithm, our process that we would use to schedule these tests that would be conducted at each establishment. But volume being a primary indicator, the amount produced in a given day or over the course of a year has one factor because that relates to the exposure of the public to the product.

But other factors would be related in part to the checklist that we issued last Friday for our employees to complete and send in the month of November.

We identified best practices that we believed at a minimum would demonstrate that 0157 is being sufficiently controlled if in fact the best practices for the operation were in fact put in place.

And so we would have an idea of which establishments are following best practices or combinations of elements from the best practices, and then use that to identify which establishments who have poorer control would in fact be the ones we would hope to bias the testing such that it would be targeted at those with poorer control, with no establishment being exempted but those that we believe need to have more attention, getting more attention to our testing program.

REPORTER: One quick follow-up, the total number of inspections likely to stay the same - it will be divided up differently, depending on sort of the risk criteria?

DR. ENGELJOHN: Yes. In this particular aspect, which is just one aspect of what inspection actually entails, this is related strictly to collecting samples and sending them to the laboratory. Our routine type of activity would be the inspections that occur on a daily basis, and this doesn't affect that specifically although the inspector would be pulling samples instead of performing other food safety activity.

But as well, poorer performer establishments, those that perhaps would have positive results, also would be put into the queue for prioritization for our comprehensive food safety assessments. Those are the in-depth reviews of the actual design and execution of the food safety system. And that's what really tells us whether or not the underpinning validation, the data used to support that system, is actually sufficient to control the organism. That's how we really find out whether or not there's a risk being presented by that establishment.

And so the data collectively will inform us as to whether or not we would perform that type of activity as well. So the data particularly from our testing program helps feed other indicators that tell us we need to get into an establishment sooner rather than later.

REPORTER: Thank you.

OPERATOR: Our next question comes from Michael Newhouse with Juneau Food Service.

REPORTER: Yes. I have a question in that what is FSIS doing differently at the originating source slaughter plants, versus other processing plants that do not have the opportunity to introduce the pathogen?

DR. ENGELJOHN: Much of our activities will in fact put more attention at focusing on the slaughter and the manufacturer trim operations. There are controls that can be implemented there more effectively. We're making sure that in fact they are. So part of the issue with our checklist was to find out what practices are employed at the slaughter and the fabrication operations.

Interventions can be applied at those points more effectively than perhaps on the ground product, but I would say that regardless the grinders, the individuals who are procuring these materials and combining them into making ground product also have an obligation to ensure that they're either purchasing from operations that do employ best practices or they themselves are seeking interventions that can actually be applied to their products. And there are interventions for ground products as well as temperature control.

So there are a number of things, but I will say our focus does move us more upstream to look at the slaughter and the trim manufacturing operations.

DR. RAYMOND: I want to just add on to what Dr. Engeljohn said to remind the listeners of something I said earlier, and that is that we began testing trim in March of 2007. That is, doing testing at the slaughter houses to try to keep that contaminated product from getting to the processors, number one. Number two, another new venture that we just started on Friday, when a processing plant or grinding plant does have a positive sample come back or is linked to an outbreak or an illness, if the provider of record - and there may be many as Dr. Petersen has mentioned - but if the provider of record for that day has been noted as the provider of record on any other day in a 120-day period, in other words twice they provided trim to a processing plant that was tested positive, we do not know that it came from that plant but that slaughter plant will get 16 samples pulled in the next couple months to get a better assessment of their policies and procedures if they are effective also.

And then one other thing, we've kind of concentrated on slaughter processing here. Dr. Goldman touched on what the individuals can do to try to make sure that hamburger is safe and when it's being fed -- particularly the frozen patties need to be cooked thoroughly. But we also are very interested in what our scientists are going to tell us about the ecology of the bug itself. Is there more E. coli 0157 coming into the slaughter plants because of changes in weather patterns out there? Is it because of changes in feed patterns; is it because of turnover in employees? The response from pulled hides can possibly cause contamination? They're looking at vaccines you can give the cattle to reduce the load. They are looking at bacteriophage, and of course they're always looking at different rinses and washes, etcetera.

So there's a lot of activity being done kind of to reduce the load ever getting into the slaughter house, and that of course would be our ultimate goal.

OPERATOR: At this time I show no further questions.

MODERATOR: Okay. Thank you. I want to remind everybody that the press release will be on the website. It's WWW.FSIS.USDA.GOV . Thank you.