Dr. Michael Osterholm on Incompetent Management of the Salmonella Saintpaul Tomato Investigation
We asked Mira Slott, Pundit Investigator and Special Projects Editor, to help find out Dr. Osterholm’s perspective on the tomato outbreak:
Q: What is your assessment of the tomato outbreak and FDA’s handling of the investigation?
A: One thing to help set the tone for this interview… Our group in Minnesota from 1975 to 1999 pioneered a lot of techniques used in outbreak investigations today. These techniques have led probes into many national outbreaks. We’ve been very involved in many national outbreak investigations, and I’ve published numerous reports analyzing the issues.
Q: Why is this outbreak investigation dragging on so long?
A: First of all, there is a common misunderstanding about how outbreaks get identified and evaluated in this country. Josiah Stamp, a famous British statistician, once said government statisticians are keen on amassing statistics. They add them, divide them, take the cube route and make wonderful diagrams. But you must never forget those statistics come from the village watchman and he puts down whatever he damn pleases.
In the first instance, all infectious disease is local and only as good as local surveillance. Fortunately, we’ve had states such as Minnesota, Oregon and Tennessee that have really been on the cutting edge of detecting outbreaks because of very good surveillance. Local and state surveillance only works when health laboratories are equipped with the capacity to quickly characterize the bacterial isolates from these patients.
For many states in the country, they do not have that kind of support. So you actually have cases either not getting reported to state and local departments or isolates not obtained, or they’re not characterized or fingerprinted until months after the outbreak is already over.
Q: Is this what happened with the tomato outbreak investigation? In the case of a highly perishable product such as produce, couldn’t the time lag alone be the nail in the coffin of an effective investigation to stop further illnesses?
A: An old case is categorized as two or three weeks ago in the produce world. Because outbreaks come out so quickly, you run into a phenomenon of many mini outbreaks. Perishable items are available 365 days a year, but on any one day where they’re coming from is a very short-lived event, maybe a crop is harvesting from one particular field for couple of weeks.
By the time we identify the food item, it’s already through the system. The timing of the public health response is critical to prevent any additional cases from occurring. If you don’t have a strong state and local health system, you’re trying to close the barn door after the cow is already out. State health departments will continue to be a problem as long as they are poorly resourced.
Q: In this case particularly, where did the system break down? New Mexico seamed to be quite aggressive and much more forthcoming in releasing details of its investigation.
A: New Mexico did a great job in this outbreak. On the whole, Minnesota, Oregon and Tennessee have led the way. The problem is that many states are not geared up to respond. In this outbreak, scores of cases are six weeks old or so. In Texas, many of the “new” cases are not new at all. The Texas Public Health Department is finally catching up in the lab with isolate characterization. The state only has two labs for this kind of fingerprinting.
By nature it’s slow. Imagine if every time you had a fire in the community you had to purchase a truck, get it inspected and do an oil change. Right there you have a problem.
Q: How does information from state and local levels flow to and between the federal agencies?
A: The whole system is flawed. The relationship between FDA and CDC is strained; it’s improved but still lacking. It’s not clear who is really in charge. Authority is split up in a way that decisions get bogged down and you don’t have anyone in command. You need that to effectively launch and oversee an outbreak investigation. I’m not sure who’s in charge. We have people identified as being in charge. We need someone that understands outbreaks, not just manages them. In this case, it appears that instead of experienced professionals, Outbreak Investigation Class 101 was doing the control study.
Also the influx of information from state and local authorities to CDC and FDA at best is very rocky, often not timely, and often not complete. Not getting isolate information for a case until weeks after the onset of the illness means it can’t be included or it is not included.
Q: FDA and CDC continue to report “new” cases piling in week after week, calling the outbreak “ongoing” with the numbers of illnesses appearing to escalate over time. Are you saying these are actually old cases slogging their way through the system? And if so, isn’t the manner in which FDA and CDC are releasing this information a gross misrepresentation of the state of the outbreak and current risk to the public?
A: This is a bit hard for me to address because I come from a public health world, 33 years now; the point being many of the people at the state and federal level are friends and colleagues. When I was at the Minnesota Department of Health, we pushed on these issues. I owe it to the public health of this country to come forward with my knowledge and put on the record the dysfunction occurring at the federal level. A subsequent review by experts in food borne disease investigations would say this tomato outbreak investigation has bordered on incompetence.
Q: In what ways exactly?
A: First, it is not biologically plausible to have Mexico and Florida both at the source of this outbreak. It epitomizes the ineptitude of the investigation. FDA and CDC should have obtained tomato product consumption information from both outbreak cases and controls. They needed to conduct full fledged tracebacks all the way back to the source not only of the sick cases but of the control group as well.
The epidemiology conducted to learn what product is associated with this outbreak by interviewing those sick and those not sick on what they were eating is a prerequisite but the epidemiology needed to go further. It was critical for the investigators to learn the locations of where both the cases and controls ate tomatoes and trace back from there.
When doing the traceback, you may come up with a few possibilities of where the tomatoes came from. If you collect data over time of the cases and controls, using the same methods to make the product association, you may find 85 percent of outbreak cases trace back to a certain field or grower, or re-packer versus the control product. Then you begin to get a much better handle on where product came from.
Q: Why did the government forego this strategy?
A: FDA and CDC have not made this a priority. They will never find a silver bullet by just tracing back a product. They’re looking for a cluster that matches up exactly with the one gun barrel. They’re under the false impression that they’re going to get a clean, clear-cut answer by finding the magic cluster. They need to do this case/control approach.
Q: Could you define these terms?
A: Cases are those ill with the specific strain identified with their food consumption. The control group is matched with age, similar characteristics and geographic location, the same neighborhood, for example. The cases might have a 10-fold chance of getting sick on tomatoes compared to the control group.
Q: Is this case/control approach more effective in some outbreaks than others?
A: In this outbreak identification for tomatoes, it needed to be approached differently than the spinach investigation. When looking at bagged lettuce or spinach, you have the advantage of a barcode with very good specific information that can take you back to a field. With a product like tomatoes or cantaloupe, you don’t have a barcode, and product is further mixed and matched in packaging and repacking.
Tomatoes have a vibrant lifecycle. A harvest today may go to a primary source today, but a week later may go to secondary channel. In other tomato outbreaks I’ve studied, prime tomatoes were first shipped to key grocery stores, then, as the crop aged, they were shipped to different classes of trade. The same tomatoes, with the passage of time, can wind up in supermarket chains, restaurant chains, convenience stores, mom & pop restaurants, wholesalers, processors, etc. All the tomatoes, though, were harvested at the same time. If you don’t understand that and try and trace back, the investigation can get very confused, which is clearly illustrated in this case.
Q: Didn’t the original epidemiological investigation of New Mexico and Texas cases include a control group as well?
A: The case control study initiated by New Mexico was important and did provide critical information. It used interviews comparing foods eaten by ill and well persons to identify consumption of raw tomatoes as the likely source of the illnesses. I believe New Mexico did try to do some trace back to the source but didn’t have the information or authority to go all the way back to Mexico.
New Mexico should be applauded for what they did relative to what they could do. CDC supports the overall epidemiological investigation. I know for a fact, however, that the government did not trace back any product reported by the control group. We don’t have a traceback to the source using the information from the control group. That was not done.
You would need to check with New Mexico to see if they did any further work with the control group on where they ate the tomatoes. [Editor’s Note: According to Deborah Busmeyer, spokesperson for the New Mexico Department of Health, “The purpose of the case control study was to identify a hypothesis of what the food source was, and from that we determined it was tomatoes. Then we looked at single source exposure in the cases, and that’s how we were able to trace back to particular stores. At that point, we went to talk with the stores about where they got their tomatoes. With Wal-Mart it became more complicated because they had two sources, either Florida or Mexico. With Lowe’s and Bashas, it was Mexico.
With those two stores, we were able to go down the supply chain to determine the distributor. At that point it was really pointing to Mexico and we passed that information on to the FDA and CDC, but we really don’t have the ability to go into Mexico. The case control was more to determine the product. I don’t believe there was a trace back of the control group. New Mexico is realizing that the FDA is the lead in these kinds of issues. We’re realizing inadequacies in the state system to provide those tasks. We got on top of this quickly, but we have limitations as well and that has been frustrating to us.”]
Q: What is your explanation for the latest onset of illness extending now to June 11?
A: You could have tomatoes through the system that long if you look at the normal growing season. We still don’t know the mechanisms behind the contamination. It could be contaminated water in a washing area and it is plausible tomatoes continued to go through that. Potentially the problem could be with a re-packer or region or irrigation source.
To narrow down the problem, it was paramount for FDA to obtain the same standard of completeness in investigating product from the control product group. Where did you eat tomatoes, at a restaurant, a grocery store, then you trace those channels back. All those factors should be known for both control and case. You learn it would be 15 times more likely to eat a tomato from this place, than the control. Now all you get is what appears to be inconsistent information.
Q: You place much weight on the importance of the control group. Was this the ultimate undoing of the investigation?
A: FDA never talked about tracing back product from the control group. It was a near fatal flaw from the get-go, and if they don’t change their approach they will never find the source.
The second problem — all cases with tomatoes typically have only lasted about 6 weeks. Any growing region of the world pretty much exhausts the product in that time frame.
This is not like a factory where every day widgets go through it. When you harvest that entire area, it’s done. Even if the product goes to a re-packer, if the onset of illness started mid April, even by end of May it’s pretty much the end of system. .
Yet here we are almost to the end of June. There are all these old cases in the system and the Health Department has done little to correct this.
Look at the curve. Reporting the most recent onset of illness is as late as June 11 is dramatically different than saying the outbreak is still going on. In the case of a tomato outbreak, you won’t have an ongoing continued source of contamination that goes on and on and on. Even if you look at the spinach outbreak, it’s not like the peanut butter outbreak, where leaking contaminated pipes would result in continued product contamination as long as that plant was producing without fixing the problem.
The FDA has done a very poor job of helping the public understand how outbreaks in the produce industry occur. No one can say with certainty when this tomato outbreak ended. But there is overwhelmingly strong evidence this outbreak peaked many weeks ago.
Look at the re-washing scenario. Product production shifts and moves to another part of the country. Look at past outbreaks. These outbreaks end on their own. They were over by the time we figured out what the problem was, if we figured out the problem at all.
If FDA and CDC took Mexico and Florida cases and worked the controls, I think this would have traced back to Mexico a long time ago. It was a flawed investigation. Why were some people getting sick and others not? They needed the control group to get to the answer.
The other problem is faulty analysis from mixing early and late cases. You may very well also be confusing the source of those tomatoes because tomatoes when first harvested may go to one area. For other buyers of the tomatoes, it may depend on the ripeness of where and when they’re sold, and that hasn’t been accounted for.
Q: A lot of this information on harvest cycles and distribution is readily available through produce industry executives. Isn’t this produce 101?
A: The final piece on this outbreak debacle: several of us have done a lot of work on outbreaks. We went to a wise group in the produce industry with no horse in the race. They looked at the patterns and they knew this market so well. This simple task of eliciting help from industry experts was not done by the investigators to any meaningful extent. We do very a poor job of bringing in the industry to help solve these outbreaks. What does this mean? That straightforward questioning didn’t happen in a way that would have been most helpful.
Q: Doesn’t this apply to the infamous FDA-approved source list? If the list was based on growing areas that were not in production at the time of the outbreak, wouldn’t that just involve a phone call or two to the produce trade organizations?
A: Right from the get-go, it was unclear how FDA was coming up with that list; why certain states and countries made the list while others were excluded. If any place not producing in late-April/early-May is safe, why the confusion and piecemeal additions to the list day by day? It appeared there was no rhyme or reason why a state or country was on the list. The logic was lacking. If FDA believed states and countries not in production at time of the outbreak were safe, they should have included all states and countries in that category.
On the other hand, if FDA has the view that a re-packer could be involved, which it now says it does, then it means FDA’s list would have been invalid and in their view putting consumers at risk. Either way, it’s not based in logic. Is it by dates of production, or is it not by dates of production because product could come through a re-packer?
There is a major issue with transparency. Why would you say something is or isn’t involved? First, the formation of the FDA list is disingenuous. It was based on those that screamed the loudest, and from a public health standpoint, that’s not right.
Q: What steps do you recommend at this juncture?
A: We need a serious examination of how this outbreak was handled from the first instance of illness. By the time tomatoes were identified, it was the end of May and the vast majority of product was through the food system already.
Another area that must be addressed is the poor job that’s been done in defining the risk.
As the Perishable Pundit pointed out so aptly, in a 40-day period, the public consumed three billion servings of tomatoes. We need to provide a better perspective. I don’t see anyone putting this in context. I’ve spent my career trying to understand and stop outbreaks, and I in no way want to minimize the dangers attached to an outbreak. This comes from a pioneer on getting outbreaks on the map. I gave a talk in 1988 that produce would face problems with outbreaks, and I was laughed off the stage.
The media has been just as guilty by not better defining the risk and through sensational and often inaccurate reporting has magnified the problem. Lou Dobbs, for example, doesn’t have a clue what he’s saying on this tomato outbreak and it doesn’t lead to solutions that are helpful. The risk of getting Salmonella from these tomatoes speaks to the safety of produce. What we don’t read about is how many times we don’t have outbreaks. Whatever the number of illnesses turns out to be in any outbreak -- 400, 4,000, 14,000, 24,000 -- that sounds terrible and it is not a good thing for a nyone to be sick. But we’re so fixated on the numerator we never talk about the denominators.
The take-home messages are that the food borne investigation system is flawed. This outbreak was inherently flawed from the beginning. New Mexico deserves credit, but the lack of state and local lab capacity and data collection and integration at the federal level can bring the industry to its knees. We must clarify risk.
Timeliness is everything. We must question how many cases we could have prevented through our public health action as opposed to those old cases in the works after the warning went out. If we’re four to six weeks into the outbreak, the cow is already out of the barn. For the FDA to then shut the barn door with all the authority of Public Health behind it, does it make any difference what we’re doing in the name of food safety?
Q: Could you address industry concerns that no matter how diligent a company is in implementing the highest food safety standards, when an outbreak occurs, it seems all companies are reduced to the lowest common denominator in the government’s eyes?
A: We clearly hurt the industry. The way we have healthy measures is to have a healthy industry. Our job is to work with people. If the industry is heavily stressed financially, they’ll have no option but to take shortcuts and cut corners. You want a viable, profitable industry that has the resources to invest in food safety. If every time we have an outbreak, the entire industry is implicated and takes a big hit. It doesn’t do anything to help industry invest in food safety. The industry is still suffering on post outbreak debate and discussion. Again, my interest is not to protect the industry; but a financially strong industry is part of the way to get viable public health.
The industry also needs to be part of the solution. Of course, “March of the Killer Tomato” makes for great news media but doesn’t help consumers. I’ve seen operators in the food service industry eliminate all tomatoes from the menus questioning the safety of the entire category when areas had been cleared. That is a copout. No tomato or any produce item can ever be 100 percent guaranteed safe. This is a time when retailers and food service operators should be going out of their way to support their suppliers.
Public Health would be faulted if the outbreak was still ongoing and it allowed contaminated product on the market. But what we can do is pinpoint the source a lot faster and stop the spigot there. Even if we don’t understand how the contamination happened, at least you identify where it came from. We don’t have that priority on speed and efficiency. I see the problem as a combination of lack of leadership and lack of understanding of how to work these outbreaks.
Q: Did you catch any of the FDA/CDC press briefings? The information relayed was often fraught with ambiguity. And reporters seemed to have a difficult time getting clarity.
A: A lot of people in the media are not experts in health or produce industry matters. Their coverage is often sensational, incomplete and tends to miss the overall perspective.
I read transcripts of the press conferences and I believe that the media did a poor job of holding FDA’s feet to the fire. The level of questioning has been very soft. They need to provide the checks and balances. I was particularly saddened when I heard the traumatic news that Tim Russert had died. I’ve always been a fan of his. He asked the hard questions. No one should have a free pass. It’s not about ‘Gotcha’ either; it’s just about trying to do the best job possible. We either take a political view or do nothing and go on to the next one.
I also wondered why the industry wasn’t pounding away with its own questions.
Industry often takes a back seat, in some ways like the Stockholm Syndrome. We all need to be held accountable for our actions.
This outbreak investigation has failed miserably. I’ve talked to many people about this in both state and federal agencies. There should be a call for a formal investigation on what happened and why. I really believe this outbreak is virtually over. The point is not to go back and punish people. The point is to learn so this doesn’t happen again. Outbreak investigations are not just to prevent cases during that time but into the future.
Q: Due to the nature of produce outbreaks, some industry executives think likening the goal of finding the exact source of an outbreak to the Holy Grail is over-glorified. Not only is it unlikely in most cases, but also unlikely to be the true salvation to fixing food safety problems in our industry. Do you have examples through your work where discovering the source of the problem led to corrective actions alleviating future outbreaks?
A: A large Salmonella outbreak in ice cream centered on a home delivery system in Minnesota. It turned out the trucks were hauling back raw egg to the plant and then contaminating the pasteurized ice cream. Lysteria in ball park franks in the 1990s was traced back to an environment where Lysteria was dripping from ceilings and creating condensation. Another outbreak investigation in the beef industry found that a new process in meat packing resulted in people eating thyroid glands mixed into hamburger, and people developing this horrible thyroid disease. Neck trimming is now banned. .
Q: Do you have an examples related to produce outbreaks?
A: No unfortunately. There are inherent risks with fresh raw produce; this is a risky business with the interface with nature. When FDA can’t find the smoking gun in an outbreak, it tries to find logical things to fix. But, of course, those fixes centered on the particular outbreak would need to be applied across the board if they are logical and based on science.
Clearly one of the options for the produce industry at this point is irradiation. Look at what’s been done with pasteurizing almonds. A lot of work needs to be done in this area. But in the end it’s a very important food safety solution for the produce industry
The whole reason I’m heading up the Fresh Express research mission is to find out what we can learn to take back to the production area to make a difference in food safety. I’m not speaking for Fresh Express here, but I use this example only to make a point as it relates to food safety in this industry. Here’s a company that in 25 years has never been implicated in an outbreak and has a very high percentage of the market. Negative data is a powerful tool. We get these high visibility outbreak investigations that result in limited public health and safety, but take up a lot of our time and get a lot of publicity.
The FDA and CDC folks already know I’m going public with these views and they have to get their ship turned around. I can walk into the mouth of the lion and say what I say because I worked there. Being out there to speak the truth on these issues isn’t always the easiest thing to do, but when public safety is at risk there is no alternative.
Dr. Osterholm is director of the Center for Infectious Disease Research and Policy (CIDRAP), Associate Director, National Center for Food Protection and Defense, and a Professor in the School of Public Health at the University of Minnesota. He has written papers published in numerous journals, including The New England Journal of Medicine, Foreign Affairs and Nature.
Dr. Osterholm has testified before Congressional committees and is the author of an important book: Living Terrors: What America Needs to Know to Survive the Coming Bioterrorist Catastrophe, which Dr. Osterholm wrote with John Schwartz. Dr. Osterholm is also a noted speaker. He is a member of the of the Institute of Medicine of the National Academy of Sciences.
He is a recipient of the Harvey W. Wiley Medal from the FDA, the Wade Hampton Frost Leadership Award from the American Public Health Association and five separate major research awards from the National institutes of Health and the Centers for Disease Control and Prevention.
The Milken Institute relies on his expertise, and PBS turns to him as an authority. Even Oprah Winfrey calls on Dr. Osterholm, and all this is on top of 24 years in various roles at the Minnesota Department of Health, the last 15 of which he was State Epidemiologist and Chief of the Acute Disease Epidemiology Section.
Dr. Osterholm has done much work on produce, including writing smaller pieces such as Lessons Learned Again: Cyclosporiasis and Raspberries, which appeared in the Annals of Internal Medicine, and an important investigation, published in Epidemiology and Infection entitled, Outbreaks of Salmonellosis Associated with Eating Uncooked Tomatoes: Implications for Public Health. Dr. Osterholm also has done important work with irradiation and food safety.