This is our reply to a November 24, 2011 email we received from the president of Germany’s Federal Institute for Risk Assessment (BfR), responding to our August 14, 2011 website column, “Explaining and Proclaiming Uncertainty: Risk Communication Lessons from Germany’s Deadly E. coli Outbreak.”
During the month before we published our analysis, we wrote to BfR officials, including Dr. Hensel, with numerous questions and an interview request. An official from the BfR did write back during that period with a substantive answer to one of our questions (regarding the exact serotype of EHEC found on Hamburg’s contaminated cucumber samples from Spain), but we were unable to interview any BfR officials.
Our reply is organized in the order of the BfR’s comments. (There’s one exception: The BfR revisits and elaborates on its first point at the end of its response, and we have replied to both at once under #1 below.) The entire BfR response is also available.
We are very grateful for the BfR’s detailed comments, corrections, and explanations regarding our website column.
The text in sans-serif type on a blue background below is our summary of each BfR comment. Our reply to the comment comes after the summary and is indented.
1. The BfR says that our column did not address its efforts and the efforts of other German agencies to advise the public on hygiene measures to protect themselves during the outbreak. Despite the various criticisms in our column, the BfR points out, surveys showed that the “common dietary recommendations and hygiene rules” promulgated by German agencies (including the BfR) during the outbreak were well understood by the German public and were successful in protecting public health.
We agree that we did not focus on this. The column focused narrowly on the dilemma of explaining – and proclaiming – uncertainty, and particularly on the failure of various agencies to emphasize sufficiently their uncertainty about the source of the EHEC 0104:H4 in the food supply during the outbreak.
We strongly agree with the BfR that it was crucial to offer sound, actionable advice about what to eat and what not to eat during the outbreak. Advising consumers about food choices during a food safety crisis is even more important than explaining what you know and don’t know about the source of the crisis. The two tasks are obviously connected. But they are distinct. And we agree that German officials did a good job of advising consumers about food choices.
While this was not a focus of our column, we did touch on it:
As non-experts closely monitoring official and expert statements about the outbreak, we think it obviously made sense to advise consumers to avoid all raw tomatoes, lettuce, and cucumbers until the source was more definitely identified or the outbreak was over. And we think it also made sense (for the five days from May 26 to May 31) to advise consumers that Spanish cucumbers were suspected as the possible source – while still urging everyone to continue avoiding all raw tomatoes, lettuce, and cucumbers. In other words, we have no quarrel with the advice that German authorities gave to consumers. Our only risk communication criticism is their failure to proclaim their uncertainty.
We noted in passing that we had “no quarrel” with officials’ dietary recommendations, but we didn’t actually say that they were doing a good job of communicating those recommendations. It would have been fairer to have said so.
2. The BfR agrees with our argument that it would have been better if the Hamburg Institute had clarified how weak its evidence was linking the outbreak to cucumbers from Spain – especially after it became clear that the media were reporting the evidence as far stronger than it was. And the BfR says we were right that “speak with one voice” is an obsolete risk communication mantra today.
But the BfR feels that we shouldn’t have expected other agencies to correct the Hamburg Institute’s impression of overconfidence: “The consequence of scrapping the [‘speak with one voice’] dogma is that all involved institutions can reflect their own opinion on the situation in equal measure and that everyone should correct himself when needed. Specifically, for the EHEC outbreak it means that it is not the duty of other institutions to straighten out the opinion of the Hamburg Institute publicly.”
We don’t think the “speak with one voice” dogma has been scrapped; we think it should be scrapped!
We speculated in our column that this problematic but not obsolete dogma may be a big piece of the reason why German federal agencies (including the BfR) let stand the widespread misimpression that “Germany” had proved that Spanish cucumbers were responsible for the outbreak – when in fact a local agency in Hamburg had simply found some EHEC-contaminated cucumbers of an unknown serotype from Spain, cucumbers that had not been linked to the outbreak through epidemiology (by finding outbreak victims who had eaten cucumbers from the Spanish source) or through microbiology (by showing that the E. coli strain in the contaminated cucumbers was the same strain making people sick).
The Hamburg Institute misled the world, we wrote in the column, and for five crucial days – until it was proved wrong – far larger and more prestigious public health agencies stayed silent. (The BfR’s response shows us that we were mistaken in part about the silence. See #3 below.)
The values question here is whether or not it is the “duty of other institutions” to express their doubts publicly when a single institution has made overconfident claims that may well prove false, and when the media, European officials, and the public perceive those claims to be the collective consensus conclusion of all the institutions involved. (Headlines like “Spanish cucumbers may be E. coli source, Germans say” were legion in response to the Hamburg Institute’s announcement.) The BfR says it was not their duty to help correct such a misperception. We think it was.
Or maybe “duty” isn’t the right category. We think it is simply foolish for prestigious scientific agencies to stand silent (or merely whisper – see #3 below) while a much smaller agency makes a scientifically unsupported claim that media, officials, outside experts, and the public are attributing to German officials and agencies as a group. Staying silent, or nearly silent, puts at risk both the public’s short-term understanding of the situation and the agencies’ long-term credibility. If the “speak with one voice” principle is causing agencies to avoid correcting the record in such situations, then scrapping that principle is a high priority. If it is bureaucratic silos that are causing agencies to avoid correcting the record, then those silos need to be drastically reorganized.
3. The BfR points out that we were mistaken in claiming that other German agencies left the Hamburg Institute’s overconfident claim completely unrebutted. In fact, in an opinion entitled “EHEC: Consumers to continue to refrain from eating tomatoes, cucumbers and green salads raw” the BfR itself stated that the type of EHEC found on the Spanish cucumber samples was not yet known. Importantly, this opinion is dated May 26, and was presumably posted on the BfR website on that date – the same day as the overconfident Hamburg Institute press release linking the outbreak to Spanish cucumbers.
In its response to us, the BfR writes:
By the way, the BfR published the following opinion on 26 May 2011 explaining exactly this situation: “The detection of EHEC in cucumbers in Hamburg, which were, amongst others, imported from Spain, resulted in several warnings through the European Rapid Alert System. It has not yet been proven that the EHEC subtype on the analysed cucumbers is the same as in the stool specimens of the patients….”
We missed this May 26 BfR opinion entirely when we were researching the column.
We speculate that its title (“EHEC: Consumers to continue to refrain from eating tomatoes, cucumbers and green salads raw”) did not draw our eye, since we knew that the media were covering the dietary recommendations well. The opinion’s title gave us no hint that its text deviated sharply from the title’s focus, and from the Hamburg Institute’s strong implication that the source of the outbreak was almost certainly the Spanish cucumbers.
The headline on the BfR’s May 31 opinion , which we did find and quote in our analysis, is much more on point: “EHEC pathogen not yet typed….”
Still, the fact that the EHEC pathogen hadn’t been typed was present in a BfR opinion on the BfR website five days earlier, and we missed it.
Given that journalists love to report on expert disagreement – but didn’t report on this one – we are virtually certain that the major media also missed the BfR’s May 26 opinion.
Had we seen it, we would certainly have discussed it. We would have pointed out that the BfR had in fact gone on the record – on the same day as the Hamburg Institute news release – with the key information that the Spanish cucumbers hadn’t yet been shown to contain the same strain of E. coli as the strain that was making people sick throughout much of Europe. (Ultimately, of course, they were shown to contain a different strain entirely.)
Our main use of the May 26 BfR opinion (if we had seen it) would probably have been as a further illustration of the need to proclaim uncertainty, not just explain it. There is no better example of that need than the fact that the BfR’s May 26 opinion did not make the news during a firestorm of erroneous headlines and news stories about Spanish cucumbers causing the outbreak. The BfR did point out that the source of the outbreak was still uncertain. But it failed to proclaim that uncertainty loudly, repeatedly, and insistently. It didn’t put it in a news release (though the BfR does do news releases). It mentioned it once in a website “opinion” on another topic.
And the media missed it.
In fact, we have yet to find any English-language news medium that reported it. Nor have we found any indication that the BfR did anything to bring its May 26 opinion to the attention of reporters during the next five days before the Spanish cucumber connection was definitively disproved on June 1 – five days during which the media widely reported the Spanish connection as the consensus conclusion of German officialdom.
Now that the BfR has pointed us to the May 26 opinion, we have Googled its language. We found nothing in English until after June 1. But we did find a few Web sources that picked up on the German-language version before June 1. Among them:
Like the English-language media, the major German-language media missed the BfR’s May 26 opinion. Or at least they missed the passage that mentioned the absence of evidence linking the contaminated Spanish cucumbers to the outbreak.
So did German officials, EU officials, and national officials in other European countries. Official comments on the outbreak throughout the five days from May 26 to May 31 clearly reflected the same misimpression as the one the media got: that the connection between the Spanish cucumbers and the outbreak had been pretty firmly demonstrated. (Even some Spanish officials questioned only where the cucumbers got contaminated, not whether they were linked to the outbreak.)
We are certain that BfR officials were well aware that officials of other German and EU agencies were being quoted in front-page articles between May 26 and May 31, repeating their misperception that Hamburg had identified the outbreak EHEC strain on a Spanish cucumber on May 26. They knew it was the big news; they knew it was mistaken; they knew an opinion on their website said so, but wasn’t being quoted and apparently wasn’t being noticed.
We are unaware of any efforts the BfR might have made to correct this official misimpression, other than posting the May 26 opinion on its website.
4. Our column criticized a June 1 BfR news release for not being emphatic enough about what was certain and what remained uncertain, once the Spanish cucumber link had been disproved. But the BfR’s response to us says that it published a parallel FAQ page that covered this ground.
We haven’t found this FAQ page yet on the BfR website, in English or in German. But we may have missed it again. We’ll keep looking. For sure there’s no link to it on the news release.
We did find a June 15, 2011 BfR FAQ entitled “Questions and answers on EHEC infections caused by vegetable foods.” It’s labeled as “Updated,” so perhaps there was an earlier version posted on June 1.
5. The BfR points out that we identified Ilse Aigner in the column as Germany’s Health Minister. She is actually Germany’s Minister of Food, Agriculture, and Consumer Protection.
The column quoted a Der Spiegel story that got this wrong. We are sorry that we failed to fact-check Minister Aigner’s title. We continued to refer to her as “Health Minister” throughout the column.
6. The BfR states that our column is written as if risk assessment, risk management, and risk communication were interconnected tasks. But as the BfR explains, it is organized as an independent scientific agency, charged with assessing risks and communicating its assessments, without regard to the more political risk management decisions that other agencies must make.
We understand this distinction, though the BfR response concedes that it is not ironclad: The BfR does derive recommendations (presumably risk management recommendations) from its scientific conclusions, which it then shares with other agencies and the public.
We believe that risk assessment, risk management, and risk communication are integrally connected, and should be. But we see why an agency focused on scientific risk assessment might be wise to avoid expressing risk management opinions that go beyond the science.
So what should such a risk assessment agency do when other agencies, more focused on risk management, are making public statements and risk management decisions that show they misunderstand a key scientific fact? (The misunderstood fact in the case at hand was the fact that there was no evidence linking the Spanish cucumbers to the outbreak.)
- Should it say nothing as other agencies misstate the evidence?
- Should it publish the accurate information in an obscure location on its website and then say nothing?
- Should it give the other agencies a private heads-up but say nothing publicly?
- Should it proclaim the accurate information aggressively, doing all it can to correct the error?
- Should it go beyond correcting the error and publicly criticize the agencies that got it wrong?
We probably wouldn’t go all the way to #5 (except in situations where 1-4 hadn’t worked and the error was killing people), but we’d certainly recommend #4 in this case – not just #2, which the BfR apparently considered sufficient … until May 31, when it finally posted an opinion on its website with a strong headline stating that the E. coli serotype on Hamburg’s Spanish cucumber samples had not yet been identified.
We imagine that the BfR would have done more than it did between May 26 and May 31 if it had felt that the safety of the public required doing so. The public communication responsibilities of scientists should always include speaking out to protect public safety when normal bureaucratic and political channels aren’t working. But this is a tougher case. Consumers were already avoiding the right suspect vegetables.
Still, for reasons detailed in the column we think it was profoundly unwise to leave the Hamburg Institute’s overconfident speculations unrebutted (or invisibly rebutted). If it wasn’t the BfR’s proper role to proclaim uncertainty about what food was responsible for the outbreak, then some other German agency needed to step in and get it done.
But we think that probably was the BfR’s proper role. We get it that the BfR is not the agency that decides whether to ban or not ban a food in Germany, nor is it the agency that regulates food imports from Spain to Germany. Those are risk management jobs. But why shouldn’t the BfR be the agency that aggressively clarifies what is and is not known about the source of Europe’s worst food poisoning outbreak in decades?
On the BfR’s home page, on the left-hand navigation bar where we found “Communication and Public Relations” (where press releases are posted), “Publications” (where opinions are posted), and “FAQ,” we also found – at the very top of the list – “Risk Communication.”
The BfR’s “Risk Communication” link leads to a description of the agency’s risk communication principles:
In its risk communication activities, BfR upholds three principles in order to raise the confidence of all the stakeholders in the risk assessment process:
- Greatest possible openness
And the stakeholders? Here’s how the BfR describes them:
The stakeholders of BfR and, by extension, its risk communication are wide-ranging. They include
- federal government and federal state ministries, agencies on the municipal, regional and federal levels
- consumer associations and their interest groups
- scientific institutions
- national and international agencies and organisations
- business community and trade associations
Nearly all of these seem to have missed the BfR’s May 26, 2011 opinion noting that the serotype of EHEC on the Hamburg Institute’s Spanish cucumber samples had not yet been identified.
That is why we so strongly emphasize the need to proclaim, not merely explain, uncertainty.
Copyright © 2011 by Peter M. Sandman and Jody Lanard