2010 Guestbook
Comments and Responses

frank white comments:

Damn it, Dave – I was passively enjoying the dialogue and doing my best to stay disengaged, but I’ve been sucked in, I regret to say, by what I consider to be a misguided notion expressed by a few of your correspondents in one way or the other that somehow OSHA is a) a political annoyance that distracts us from our real work, b) an irrelevant anachronism, or, worst of all, c) the enemy of the profession that we must distance ourselves from in order to maintain our “true” identities. That’s just silly – again, in my not-humble-enough opinion.

Don’t get me wrong – I heartily subscribe to the line of reasoning advocated by [several dialogue participants] that our primary focus must be on engaging leadership at all levels of our organizations in the collective effort to assess and reduce, control and mitigate risk. I also want to second (or third) [one participant’s] observations about the need to do a better job of evaluating the root causes of near misses and similar pre-injury indicators.

But when did “compliance” become a discredited notion? Isn’t compliance with all laws and regulations embedded in the basic policy statements that establish our risk management systems, and doesn’t it complement the risk reduction focus?

PLEASE don’t reply that “compliance isn’t safety” or some such platitude – we know that to be the case, but it is an integral element of a safety and health management effort nonetheless. And our political earth-mother OSHA, for all her sometimes bizarre and counterproductive behavior disorders, is at the end of the day our ally, without which, candidly, most of us – individually and collectively – would have been much less effective in achieving the steady progress in improving worker health and safety than we actually have been.

It is surely true that the agency is, and has been for the past decade or so, a laggard in many of the areas – risk management, in particular – that we now view as essential to future success. But shouldn’t we all, whether from industry, labor or the profession in general, share some of the responsibility for that situation? There have, indeed, been institutional and political barriers to more progress by OSHA, but all of us should be engaging the agency, not distancing ourselves from it, and doing what we can to help it “catch up” and be part of the solution.

Finally, the notion that somehow, our environmental colleagues have succeeded in aligning themselves with some sort of “alternate universe” to EPA through focusing on Sustainability is ludicrous.

Organizations in general are much more attuned to compliance with EPA requirements than with OSHA mandates because EPA’s stick is so enormous relative to OSHA’s enforcement twig. Sustainability may have reached a tipping point of standing on its own merit, but not as a reaction to, or irrespective of, EPA’s influence, but because sustainability is increasingly grounded in principles that are becoming more successfully integrated into business strategies and processes.

But make no mistake, environmental professionals ignore compliance at their, and their organizations’, peril. The challenge for safety and health professionals is to find our own “sustainable” links that align us more closely to our organizations without needing to reject our legitimate compliance roles.

peter responds:

I fervently agree with Frank. U.S. companies spend a lot more money per life saved on environmental protection than on safety in large part because EPA has so much more clout than OSHA. I love his comparison of EPA’s “stick” and OSHA’s enforcement “twig.”

Another factor worth considering: U.S. environmental activist groups have a lot more clout than U.S. safety activist groups … all the more so as unions have lost ground and become more narrowly focused on job protection. Corporate environmental officials exist as much to cope with the demands of Greenpeace et al. as with EPA’s demands.

I wonder if it would be useful for Dave to ask respondents what role they see for safety activism. Is there any where they work? Would it help if there were more? Are there ways they – or the profession at large – could encourage more?

A related factor is what I call “outrage.” Corporate environmental misbehavior is widely seen as sinful, as evil; the victims (whether neighbors, creatures, or ecosystems) are seen as innocent and worth protecting. By contrast, on-the-job accidents and even worker exposure to toxicants tend to be seen as inevitable side-effects of the work itself. Affected employees may be viewed as complicit in what went wrong, as having assumed the risk voluntarily, and/or as having been paid for assuming it.

Dave might want to ask respondents what they think about the wisdom, efficacy, and feasibility of trying to arouse more safety outrage among employees, their families, and the general public. Arousing outrage is of course what activists do – which is why activists may be a more potent force for risk reduction than either regulators or risk management professionals.

tom lawrence comments:

On September 1, 2010, I begin my 41st year as a practicing safety professional. Just still learning.

I don't know anything about this radicalism you mention. How do you do radicalism in a professional role and who do you do it to?

Thanks.

peter responds:

I don’t think safety activism needs to be “radical” … not unless you think it’s radical to believe people should be outraged that workplace safety norms are as lax as they are. For those who think safety standards ought to be tougher, it’s not so radical to think a little outrage might help them get tougher.

But I agree that it’s a bit of a conundrum to figure out how safety professionals inside organizations could encourage such outrage without creating another sort of outrage altogether: their managements’ outrage at them for stirring up dissatisfaction and mobilizing pressure. Safety professionals working for advocacy organizations – unions, for example – might have more freedom of action in this regard.

The insufficiency of safety outrage/dissatisfaction/pressure does seem to be part of the problem. That’s why I thought it might be a productive topic for Dave to survey people about.

But mine is an outsider’s perspective, and perhaps an ignorant one. I am not a safety professional. I do spend a lot of time working with clients who are trying to respond to outrage, dissatisfaction, and pressure, the substance of which is far more often environmental risk than employee safety risk.

tom responds:

Let me see if I understand this. We need radicals to drive OSHA to impose more standards on the workplaces. Some say workplace safety and health standards are not tough enough. As a safety professional who has worked there and supports those who do now, I take strong issue with that.

We have forty years worth of OSHA standards that we deal with – thousands and thousands of requirements to manage. The radical’s view is that is old stuff. OSHA should be adding more standards and adding more requirements to those already required. Just pile it on.

Of course the radicals have zero responsibility and accountability for how those imposed requirements are met. But who does? Safety and health professionals have that responsibility and share the accountability with the management teams we work with and for.

So as soon as the ink is dry on the Federal Register for more imposed requirements, the radicals can retire to their Beltway latte parlors and celebrate their political agenda victory.

Meanwhile, back in the workplaces, the S&H Professionals are trying to find time in their already hectic schedules to get this new compliance requirement management piece in place. Oh, but we wanted to work on transformation issues for excellence in our workplace – but, there is no time. We have to spend all of our time on compliance – on basic safety and health – not excellence: transactional safety and health, not transformational. And thus, our profession is driven by OSHA and its politically driven agenda.

Dr. Sandman, do I have this process described correctly?

Would it be out of the question for at least some of us to feel that after 40 years of dealing with a politically obsessed and driven OSHA that we have had enough of it and want to separate our profession and its image from the Beltway OSHA process – except to try to protect ourselves and our workplaces from the radicals?

peter responds:

I don’t have an opinion – at least not a professional opinion – on whether OSHA is currently too strong or too weak or just right. I’m not an expert on safety or safety regulation. I get your point that many OSHA regulations impose onerous compliance obligations on management without having much actual payoff in terms of workplace safety. I don’t have the knowledge to agree or disagree. There are plenty of people on Dave’s list, you among them, who are far more expert on that point than I will ever be.

My point is simply that what I call outrage – stakeholder unhappiness / concern / anger / fear about the status quo – is a powerful driver of change. Or to put it differently: Professional problem-solvers gain clout when powerful institutions become convinced that the problem needs solving. And one important way powerful institutions become convinced that a problem needs solving is when stakeholders whose opinion matters are deeply dissatisfied with the status quo.

Of course there are other drivers of change as well. I’m not claiming that arousing more safety outrage is the only way to win more clout for safety professionals. But I do think the greater clout of environmental professionals is largely attributable to the greater outrage attaching to the issues they manage. Safety professionals who judge their own clout to be insufficient might find this a hypothesis worth considering.

As a risk communication professional (not a safety professional), I am often asked to help a client try to influence the level of stakeholder outrage about a specific issue so it is more commensurate with what my client considers the actual hazard – that is, to try to increase stakeholders’ outrage about serious problems that are receiving insufficient attention, and to try to decrease stakeholders’ outrage about small problems that are receiving excessive attention (in my client’s opinion, at least).

I don’t know which safety problems you would put in which category, nor do I know whether you’d say most safety problems are in one category or the other. I do know that managing stakeholders’ outrage upward or downward is a pretty effective way to motivate more or less management attention to a problem.

The process of arousing outrage about X (or diminishing outrage about Y) doesn’t have to be driven by “radicals,” or even by activists who aren’t radical. And it doesn’t have to be implemented via regulators. I am intrigued by the possibility that – perhaps – it could be driven by safety professionals and implemented by companies in ways that were more transformative than merely compliant.

tom responds:

Why do safety professionals need clout – provided by OSHA?

Line management and employees together own safety and health. My role is as a staff professional. If I as a safety professional am considered to “own” safety, that will ultimately be bad news. It is conceptually incorrect but it is also practically not achievable. There is not enough of me to go around. And my owning safety would separate it from the work processes – making it an off-line staff issue rather than a line issue. All of that would lead to poorer performance.

If clout means that I open my mouth and everybody does what I say, then they don't own safety, I do – and we have already premised that is bad. I don’t covet that kind of clout.

I want to earn respect and will work to do that – not just because I have a safety professional title. It comes from being a servant leader and demonstrating that my job is to help them do what needs to be done in their view. If changes to their view are needed from what I am seeing, we will have that discussion at some point.

The litany from some is that “management is the problem” in safety. Some of them are. I just have to work harder and more creatively with those. But I guarantee you that all of them are the solution – sustainable, lasting safety progress does not happen without them.

In my 40 years of experience, I can count on one hand the number of times I have invoked OSHA as my means of influence. To be sure, I help my clients understand the minimums – the requirements and the interpretations in the applicable OSHA regulations. That is knowledge that they want from me. And I definitely want them to know in that process that I am not OSHA or any extension of OSHA. I am working to help them accomplish their safety objectives, not because of some OSHA agenda that I represent or even my personal agenda. I am on their team and I let them know it.

That is not some unique-to-me skill. I talk about my approach because it is doable by any safety and health professional. We don’t have to put on the OSHA cloak as professionals to have influence for safety and health and for improvements in protection of people in the workplace.

comment:

I am currently writing an article on Obama’s crisis communication during the oil spill in the Gulf of Mexico. It will be a guest article published in an upcoming issue of a German monthly magazine on political communication.

The article will deal with Obama's strategy throughout the spill and analyze his successes and setbacks. Given your expertise in the field of crisis communication and your in-depth knowledge of both communication theory and the circumstances in the U.S., I would be happy if you could respond to the following questions:

• According to your approach, the two factors “hazard” and “outrage” determine the dimension of a crisis situation and should define the adequate reaction.
  • Where would you see the Deepwater Horizon case on your risk communication “map”?
  • What characterizes this kind of a crisis?
  • What are the critical reactions everyone involved in the crisis should take? (For example, is calming the public the first priority, or communicating every fact of the crisis, or what?)
• President Obama was often criticized by the media for reacting dully, without emotions, for not being personally involved.
  • How important is emotional/personal involvement in such a crisis situation?
  • What impact does the impression of being distanced have on Obama’s perception as a leader who can help the nation solve the crisis?
  • Would it have been better if Obama was seen more often on site, washing birds or talking to locals?
• With the first face-to-face meeting on June 16, President Obama met very late with BP executives at the White House.
  • Did Obama need the time to “get control over the situation”?
  • From a risk communication perspective, was it effective to attack BP as extensively as President Obama did?

peter responds:

I have posted five comments so far on the BP oil spill:

• “BP’s Communication Response to the Deepwater Horizon Spill” (May 3)
• “Communicating about the BP Oil Spill: What to Say; Who Should Talk” (May 30)
• “Why We’re Vilifying BP” (June 4)
• “Jim Joyce, Tony Hayward, and how to apologize” (June 5)
• “The ethics of risk communication consulting and the BP oil spill” (June 6)

But I haven’t posted anything since early June, and I haven’t posted anything at all about President Obama’s handling of the crisis. Let me take your three questions in order. I also want to add a comment on what seems to be happening now with regard to “good news” from the Gulf of Mexico.

What kind of risk communication does Deepwater Horizon require?

Deepwater Horizon is high-hazard, high-outrage – people are upset and right to be upset. That makes it a classic crisis, as opposed to the far more frequent low-hazard, high-outrage reputational “crisis.”

The difference is crucial. When people are more upset than the actual hazard justifies, the core risk communication task is to help them get the risk into perspective and calm down. But when the hazard is serious and people are right to be as upset as they are, as in the case of Deepwater Horizon, the core task is to validate their concerns, help them bear the situation and the way it makes them feel, and help them take (or support) wise rather than unwise actions to cope with what’s happening.

Corporate and government officials often think they must “calm the public” in a crisis. This is mistaken. Crisis management professionals know that panic is comparatively rare (at least when people are sober; a high percentage of actual cases of panic occur in nightclubs and sports stadiums among people who are not sober). Paradoxically, an overwhelming impulse to calm the public often leads officials to do things that exacerbate the public’s concern: withholding alarming information, giving false reassurance, expressing contempt for people’s fears, infringing on people’s freedom, etc. When people are rightly upset – a defining characteristic of a real crisis – the task is to help them bear what they’re feeling, not to try to talk them out of feeling it.

In any crisis, expressing compassion for the victims is also paramount; it’s a precondition for credibility in your efforts to validate people’s feelings and guide their actions. If your organization caused the crisis, expressing contrition is also paramount, and also a precondition. Even expressions of compassion aren’t credible from a “perpetrator” whose contrition isn’t clear. And if your organization is responsible for addressing the crisis (whether or not you caused it), expressing determination to stay the course is a third essential crisis communication requirement.

In the case of Deepwater Horizon, BP’s failure to express contrition made it impossible for the company’s expressions of compassion and determination to ring true. This is also the case for President Obama, though to a lesser extent. He hasn’t taken sufficient blame for the government’s failure to regulate the oil industry’s emergency preparedness adequately, and that has undermined his expressions of compassion and determination   as well as his expressions of anger at BP.

All of this, of course, is in addition to the rudimentary obligation to tell people what you know about what has happened so far, what is happening now, and what is likely to happen next. There are four usual pitfalls here:

• Over-reassurance – sources should always err on the alarming side, so later messaging takes the form of “it’s not as bad as we feared,” not “it’s worse than we thought.”
• Overconfidence – sources should always make sure their tentative messages are expressed tentatively, so the inevitable errors don’t come out sounding like lies.
• Refusal to communicate under conditions of uncertainty – sources should always do their best to fill the crisis information vacuum, so refusing to “speculate” until all the facts are known is almost as bad an error as speculating overconfidently.
• Outright dishonesty or secrecy – the discovery that you said things you knew weren’t true or refused to say things you knew were true is, of course, devastating to the credibility of a crisis information source … as it should be.

Obama’s leadership and emotional expressiveness

The conventional wisdom is that an American president must be not just Chief Executive but also Emoter-in-Chief – that he (and eventually she) must voice the public’s feelings. In the case of Deepwater Horizon, that means voicing both our sadness and devastation at what has happened to the Gulf, its residents, and its creatures and our anger at those who made it happen or let it happen.

Public emotional expressiveness obviously isn’t President Obama’s strong suit. He projects cool wonkiness, not emotionality.

The question is whether he should try to overcome this weakness or should play to his strengths. If I had been advising the President in the early days of the Deepwater Horizon crisis, I would have pushed him to be more visible, more personally involved. But I don’t think I would have pushed him to be more emotional.

He should have addressed the nation on the crisis earlier and more often. He should have met with BP leaders earlier and more often.

Above all, I think, he should have made greater use of his extraordinary ability to teach. When he’s at his best, President Obama can explain complicated and emotionally unsettling truths in ways that come across as genuine, insightful, and paradoxically calming – calming not because he is trying to calm us, but because he is calm himself as he explains something upsetting. He did it on race during the campaign. He needed to do it on the Deepwater Horizon spill. (He also needed to do it on the financial crisis, but that’s another matter.)

There are many things the President could have taught us about this crisis. Three that strike me as especially crucial are these:

• That Deepwater Horizon is simultaneously worse than we pretend and not as bad as we imagine – that it will be impossible to “put things right” completely (environmentally or economically) but that ecosystems and economies do recover and the sites of previously grievous oil spills give grounds for hope.
• That BP and its contractors are very much to blame but there is ample blame to go around – that the rest of the industry’s emergency response plans were just as inadequate as BP’s; that government (Obama’s government as well as Bush’s government) approved those patently inadequate plans; that we all made a foolish choice to develop oil where it’s most dangerous but least visible, far offshore, instead of opting for land and shallow-water development.
• That there are lessons that must be learned – lessons about alternative energy and conservation; lessons about domestic energy development; above all, lessons about taking worst case scenarios and emergency preparedness more seriously. Deepwater Horizon is the third inadequately anticipated risk management crisis of the Obama presidency. The other two were the financial crisis and the influenza pandemic (we got lucky on that one; it turned out very mild). We need to stop shrugging off dire possibilities and start taking more seriously the twin jobs of preventing them and preparing to respond to them.

No President can be successful if he comes across as unfeeling … or, worse yet, as insensitive to what the public is feeling. But Obama is never going to be a good Emoter-in-Chief. When he tries – as when he tried to excoriate BP – he sounds forced and insincere. He should play to his strengths.

Scapegoating BP

I’ve already answered this one. President Obama should have met with BP earlier and attacked it less vigorously.

Deepwater Horizon isn’t about a rogue company. Yes, BP has a safety culture problem – a problem that got worse during John Browne’s tenure as CEO and that Tony Hayward promised to correct, tried to correct, and failed to correct in time. There are other oil companies whose safety cultures are better than BP’s – other oil companies that might have cut fewer safety corners in the years before the Deepwater Horizon explosion and might have taken the warning signs more to heart in the days and weeks before the explosion. But there are also lots of oil companies less attentive to safety than BP, and lots of industries less attentive to safety than the oil industry.

The biggest failure revealed by Deepwater Horizon was the failure to accept that serious accidents can happen, that preparing to manage serious accidents is essential, and that if a serious deepwater oil spill cannot be prepared for adequately then the whole society needs to know that and decide together whether we want to take the risk or meet our energy needs a different way. That failure was shared by the whole oil industry and the government agencies charged with regulating the oil industry. Scapegoating BP doesn’t help us remedy that failure.

Good news in high-outrage situations

In late July and early August, the news from the Gulf got better – though maybe only temporarily; it’s too soon to tell. The spill was capped; the “static kill” was successfully implemented; the amount of oil (at least on the surface) started declining; some previously closed waters were reopened for fishing; some experts backed off their previous predictions of catastrophe.

President Obama and others in his administration hailed what the White House called on August 4 “the beginning of the end.”

Responses to this new tone of optimism – or at least reduced pessimism – was decidedly mixed. Some in the Gulf region saw it as evidence that the Obama administration was preparing to abandon them, just as they feel the Bush administration did after Katrina. Administration estimates of how much oil had been recovered, how much had naturally degraded or been eaten by microorganisms, and how much remained to do further damage were hotly disparaged as over-reassuring political spin.

This is perfectly typical. Outraged people do not like good news. When we’re angry and upset, our desire to be found right in our grievances becomes deeper than our desire to discover that things are better than we feared. Just as neighbors of a polluting factory normally dispute any evidence that the emissions probably aren’t going to give them cancer after all, Gulf residents and their sympathizers around the world are attached to what they endlessly call “the worst environmental disaster in U.S. history” and are extremely skeptical about anything resembling good news.

Of course this skepticism often proves prescient. The tendency of outraged publics to be highly skeptical about good news is matched by the tendency of crisis managers to be overconfident and over-reassuring – to over-emphasize good news even when the bad news is a lot more plentiful and reliable.

Risk communication professionals know something about how to cope with this phenomenon:

• Make the good news very tentative. In fact, put it in a subordinate clause: “Even though it looks like more oil than we expected may be degrading naturally, that evidence is very uncertain so far. We still don’t know how bad the spill’s impact may turn out in the end.”
• Express the good news as a hope, and explicitly deny that it’s a conviction. “There are grounds for hope that this awful spill may have less lasting impacts than we originally feared. But the experts are far from unanimous and far from certain. There may still be some very bad surprises in store for us.”
• Acknowledge the concern that the good news may be an excuse for reneging on your promises. “Some people naturally worry that our administration might pivot on this preliminary good news and back off our obligations to the people of the Gulf. That’s not going to happen.”
• Emphasize the known downsides. “Even the best-case scenario is no picnic. The economy, environment, and reputation of the Gulf region have all taken a big hit. We don’t know yet whether things are as bad as we initially feared, but they are plenty bad enough!”
• Go out of your way to disavow the “no-harm no-foul” position. “Regardless of the ultimate impacts of this spill, it will stand forever as a stunning example of irresponsible corporate risk-taking and inadequate government regulation. People’s righteous anger remains as justified as ever – anger at BP and the other companies involved, and anger at the Bush administration and my own administration.”
• Above all, subject the good news to as rigorous a quality control process as the bad news – maybe even more rigorous. Coming back later to say the situation is “worse than we thought” is a lot more damaging to credibility and morale than coming back to say it’s “better than we feared.” So err on the alarming side. Don’t lean very heavily on preliminary good news you might have to take back later.

Over the long haul, one important goal of risk communication in situations like the Deepwater Horizon spill is to validate and address people’s outrage. Only when the outrage declines will the public become more open to evidence that perhaps the hazard isn’t so horrific after all.

comment:

I am working on a story about little signs here and there that concerns about the health risks of cell phones might be gaining traction (SF’s labeling law, Kucinich planning to introduce a bill mandating federal labeling), and am wondering something that you might have some thoughts on: Why does the public not seem to care/believe that there might be a health risk?

peter responds:

Note to readers: Sharon’s August 5 article on cell phone risk quotes briefly from the response below, but focuses mostly on the endless uncertainties in the technical debate.

The first thing that comes to mind in response to your question is the role of uncertainty.

The evidence about the health risk of cell phones is exceedingly uncertain. After literally hundreds of studies, there is nothing approaching a smoking gun. Most studies have found no evidence of an impact. But a few studies have found weak indications of possible problems. Most experts look at this pattern and conclude that the risk is either small or nonexistent; if it were big it wouldn’t keep disappearing and reappearing from one study to the next. (Leave aside the important question of latency. If cell phones caused brain cancer after 30 years’ use, we wouldn’t have a clue yet.) A few experts look at the same pattern and find it worrisome.

People don’t need to read up on the studies to get the accurate impression that the experts simply aren’t sure. The question is how to react when the experts aren’t sure.

In determining whether people will take a risk seriously or shrug it off, uncertainty is a swing vote. When people are upset about the risk for other reasons, uncertainty adds to their outrage: “How dare you make me the unwitting subject of your experiment!” But when people are inclined not to worry, uncertainty becomes a reason for calm: “Even the experts don’t know whether it’s dangerous or not, so why sweat it!”

We rely on our cell phones – to get work done, to stay in touch with friends and family, to fill time when we’re bored or lonely. We really, really don’t want to learn that there are health reasons to restrict our use of this miraculous invention! Uncertainty gives us a reason to stay unconcerned.

Cell towers are a different story. A tower very close to your home isn’t essential to your wellbeing; your phone would work fine if the tower were further away. So people would rather worry about the nearby tower than the instrument clamped to their ear. From the outset, cell tower health and safety controversies have been a lot hotter than cell phone health and safety controversies. (Irony: The further the tower is, the harder your phone works to bring in a signal; cell phones may actually be less risky when the nearest tower is close.)

Another way of framing the same explanation is in terms of the risk/benefit ratio. I don’t know the actual correlation between the risks and benefits of technologies. It’s probably positive, and probably pretty small. But people want the correlation to be big and negative. That is, people want to believe that technologies with high benefits are low-risk, and that technologies with high risks are low-benefit. We want to avoid the dilemma of a high-benefit high-risk situation: the uncomfortable choice whether to endure the risk or forgo the benefit.

So if we’re already convinced that something is dangerous, we resist learning that it’s valuable. If we’re already convinced that it’s valuable, we resist learning that it’s dangerous. On controversies from deepwater oil exploration to genetic engineering, people polarize, with almost nobody taking the “high-benefit high-risk” position.

Cell phones are obviously high-benefit. So people want them to be low-risk. As long as the evidence of risk is uncertain, most people will shrug off that evidence and hang onto their cell phones. (Earlier in the cell phone adoption curve, when lots of people loved them but lots of other people hated them and chose not to use them, the latter group tended to be more inclined than the former to take cell phone health concerns seriously.)

Cell towers are like living near a polluting factory: You face the risk (big risk or small risk or unknown risk) and somebody else benefits. Cell phones, on the other hand, are like microwave ovens, coffee, computers, Wifi, and smoking. You face the risk (again, big or small or unknown) but you also get the benefits. So uncertainty about cell tower health effects leads to concern and even opposition, while uncertainty about cell phone health effects leads to apathy.

What would change that? Obviously, strong, unequivocal evidence of a serious risk might do the trick. But it would need to be really strong and really unequivocal; look how long smokers held onto their wishful conviction that cigarettes might not cause cancer after all. A viable alternative technology would lower the resistance a lot. Imagine a new generation of cell phones that used a radically different technology. Should I buy one of those new phones or stick with my old one? Now, suddenly, evidence that the old one might be dangerous isn’t arousing risk-versus-benefit cognitive dissonance anymore; it’s resolving spend-versus-save cognitive dissonance.

comment:

I have been developing a training for officials on communication around climate change adaptation, and have found your writings on riskcomm and climate change helpful.

On another topic, I did have one question about the “hyping” of the H1N1 virus. Could the agencies involved be more in line for some credit for raising awareness and limiting the impact of the pandemic, rather than only being seen as crying wolf?

peter responds:

I certainly agree with you that it was a major service, not a disservice, to warn the public about the H1N1 pandemic.

If swine flu had turned out more serious than it did – which it might well have done – the warnings would have been very important. And if swine flu had turned out extremely serious, there would have been justified (and unjustified) criticism for the failure to warn more aggressively.

Did the warnings do actual good or actual harm, given the mild pandemic we have faced so far? I think the answer is a little of both, and not too much of either – except for the credibility crisis now facing the World Health Organization.

Most developed countries ended up buying more vaccine than they needed, for example, and didn’t have it available till the pandemic was already receding. So pandemic vaccination turned out more expensive than useful. (That’s not blameworthy. At the time various governments were deciding how much vaccine to order, they couldn’t possibly have known the future course of the pandemic or its eventual virulence.) Another negative: Some developing countries were distracted from endemic diseases of substantially greater importance to the health of their citizens.

On the positive side, there is documented evidence in some countries of increased hand-washing and increased flu vaccination uptake. Hand-washing presumably reduced the incidence of salmonella, shigella, norovirus, and other contagious diseases, whether or not it had any effect on swine flu incidence. Both increased flu vaccination and increased hand-washing may lead to reduced disease burden for years, if the behaviors can be successfully reinforced after the pandemic ends.

The proper standard for judging a warning isn’t how necessary and useful it turned out to be, but rather how likely it was to turn out necessary and useful, based on what was known when it was issued. Hindsight bias makes people segue from the accurate perception that the H1N1 pandemic has been surprisingly mild so far to the nutty charge that officials should have downplayed it from the start, and that their failure to do so constitutes hype.

Many national governments around the world have faced some criticism for responding more aggressively to the H1N1 pandemic than its later development (so far) has justified. But the charge of “hype” – accompanied by a charge of complicity in the economic interests of Big Pharma – has been leveled most insistently and effectively against the World Health Organization.

What most justifies that nutty charge is the failure of WHO to concede – even to this day – that the pandemic has been mild so far. Over the long term, the ability of officials to issue credible public health warnings depends on their ability to stand down from those warnings when the risk is found to be lower than initially feared, or when the risk is mostly gone (even if it may return – which is something else they can’t credibly warn about when they’re not yet saying it’s mostly gone for now).

It is as if hurricane forecasters were unwilling to say so when a hurricane veered off its likeliest course or decayed into a much milder storm.

When officials point to a serious or potentially serious risk, they’re warning, not hyping – even if that risk eventually becomes minor or never materializes at all. But when officials fail to downgrade their warnings as new data show the risk to be lower than it might have been, or lower than it looked at first, that’s a huge risk communication error. It might even be called hyping.

In talking about the status of an evolving risk, officials need to answer four main questions – and need to answer them over and over:

In many risk situations, including influenza pandemics, all four answers keep changing. (Even the first: Officials get new evidence that alters their assessment of past events.) And all four remain uncertain, often highly uncertain. So every update needs to be tentative. It needs to include the information that the update itself may turn out wrong, that the next update may be very different from this one. Ideally it should also include the information that the range of possible outcomes is even broader than the difference between the likeliest scenario and the worst case planning scenario – that the full range runs from anticlimactically minor to devastating.

For a hypothetical hurricane, the answers might go like this:

For the 2009-2010 H1N1 pandemic, the current answers should sound like this:

This is quite different from what WHO is currently saying.

The paradox is that accusations of hype make it harder for WHO officials to acknowledge that the pandemic has been mild so far and the risk is currently minimal. That’s no excuse, of course – and it doesn’t explain why WHO didn’t stand down much earlier on. Still, I can understand why WHO is disinclined to look like it is “caving” in response to pressure from its critics – pressure that combines far-fetched accusations of conflict of interest with valid assertions that the pandemic did in fact turn out much milder than the early justified warnings suggested, and that WHO should damn well say so.

comment:

I eagerly await your updated critique of BP’s crisis communication and actions in the Gulf disaster.

If you were helping BP as a consultant now, is there a point at which you would consider not lending the company your expertise on ethical grounds? Is this a question that practitioners should ask as they help organizations? If a client does not tell the truth or intentionally violates other core principles of risk communication, is there a responsibility to walk away?

peter responds:

I have posted four comments so far on the BP oil spill:

• “BP’s Communication Response to the Deepwater Horizon Spill” (May 3)
• “Communicating about the BP Oil Spill: What to Say; Who Should Talk” (May 30)
• “Why We’re Vilifying BP” (June 4)
• “Jim Joyce, Tony Hayward, and how to apologize” (June 5)

These four have far from exhausted the list of things worth saying about BP’s oil spill risk communication, even its oil spill risk communication so far.

For example, I was aghast at CEO Tony Hayward’s offhand comments that the environmental impact on the Gulf may end up “very, very modest” because it’s a “very big ocean.” Even if this turns out true, it is incredibly insensitive … rather like a child molester pointing out that the kid will probably get over it.

On the other hand, I was pretty impressed with BP’s reaction to the satiric comments posted on a fake BP Twitter account. Instead of asking Twitter to take down the imposter account, BP said it had no complaint about what it saw as a legitimate expression of people’s understandable anger and frustration.

The last six weeks have provided lots of examples of BP violating what I consider to be core principles of outrage management and crisis communication. There are also more than a few things BP has done that strike me as excellent outrage management and crisis communication. Most commentators have given BP extremely low marks overall – unfairly low, in my judgment. Any company that was responsible for a huge oil spill it couldn’t find a way to plug (at least so far) would be facing inevitable and justified public outrage no matter how good or bad its risk communication skills (or advisors) were.

On May 3 I told the BBC I would give the company a B in risk communication so far. Today I’d give it a C-.

Do I quit when clients do bad risk communication?

But even if BP were earning an F in risk communication, I wouldn’t necessarily feel an ethical obligation to walk away. (So far I have nothing to walk away from. BP hasn’t asked for my advice.)

My wife and colleague Jody Lanard is a psychiatrist. “If you’re not willing to work with crazy people,” she has often told me, “you shouldn’t be a psychiatrist.” Similarly, if you’re not willing to work with screwed-up organizations, you shouldn’t be a risk communication consultant.

My typical client accepts and implements around 15 to 25 percent of my advice. That means 75 to 85 percent of the time the client ends up doing something I consider less than best practice in risk communication. If I were going to walk away every time a client turned down some of my advice, I wouldn’t have any clients. Who would want to hire a consultant who was foreordained to quit the first time his or her advice was rejected?

The specific example you give is a client that “does not tell the truth.” I routinely advise my clients that withholding information from the public isn’t just unethical; it’s stupid. If the information comes out anyway and people feel they should have been told sooner, it does the client roughly twenty times as much harm as it would have done if the client had been candid in the first place. Mathematically, then, secrecy is a bad bet unless the client can sustain a 95% success rate at keeping its secrets … which I doubt any company or government agency in the developed world can sustain. (My “twenty times as much harm” figure isn’t based on actual data. It’s a guesstimate I use to drive home how foolish it is for clients to let upsetting information be belatedly revealed by third parties, instead of candidly and quickly revealing it themselves.)

But virtually every client I have ever worked for has at some point said, no, we’re not going to tell people that. My clients don’t often lie outright, as far as I know. But they very often craft their messages carefully to avoid sharing the part of the truth they’d rather people didn’t know. Both sides in risk controversies do this. If anything, “good guys” do it more than “bad guys,” in part because they feel more virtuous and in part because they’re less likely to get caught.

So no, I don’t feel an ethical obligation to quit when a client tells less than the whole truth … not even when a client says things I consider flat-out lies. I always advise my clients not to lie, and I always advise them not to shade the truth, and I try hard to persuade them why it is in their interests to follow this advice – at least their long-term interests, but usually their short-term interests as well. I usually succeed in part and fail in part. And I don’t believe the partial failure means I have to quit.

What ethical obligations do I obey?

What ethical standards guide my risk communication consulting? I have three I’m clear on. There may be others I’m acting on without quite conceptualizing, but these are the three that are codified in my mind:

So what about a client that simply isn’t taking any of my advice – a client that continues to do crappy outrage management or crappy crisis communication? I often stop working for such clients simply because the work isn’t useful and isn’t fun. It’s only remunerative. But I don’t consider that an ethical principle. It’s a self-indulgence. At the tail end of a successful career, I can afford to pass up work, even highly remunerative work, that isn’t useful or fun.

I have also consistently refused work for the tobacco industry. I sometimes tell myself this is an ethical principle; tobacco is the only legal product I know that, properly used, kills its best customers. But I’m not sure this decision isn’t more about cowardice than principle. Consultants who work for the tobacco industry are widely condemned. I get enough of that just working for polluters. It would be much worse if I worked for tobacco companies. (The World Health Organization, for example, requires contractors to sign sworn statements as to whether or not they have done any tobacco industry work; I don’t know how WHO would react if I checked the “yes” box, but at a minimum there would be some more hoops to jump through.) A year or so ago I was invited to be an expert witness for a tobacco company. The limited risk communication and risk perception points the company wanted me to testify to were valid, in my judgment. I declined anyway. I didn’t feel especially ethical about declining. I felt self-protective. Even though the company was in the right about the specific points it wanted me to make, I decided that helping Big Tobacco win a lawsuit wouldn’t be useful … and I sure as hell knew it wouldn’t be fun.

Ethics, of course, is intensely personal. Ethical questions often come up at my outrage management seminars. Half of my response is to explain what feels ethical to me and why. The other half of my response is to urge everybody at the seminar never to do anything he or she considers unethical just because Peter Sandman thinks it’s okay. Some people, for example, tell me they think it’s unethically manipulative for a company to be more responsive to its stakeholders when the company’s only reason for doing so is to reduce stakeholder outrage. I can’t see where they’re coming from – to me, wanting other people not to be angry at you seems like a pretty good reason for treating them better. But again, I advise seminar participants not to do it if it feels unethical to them.

The ethics of outrage management itself

Here’s the ethical challenge I hear most frequently from critics of my work. Helping truly bad companies do a better job of managing a crisis or a controversy isn’t improving the world, they say; it’s covering up the evil. The ethical standards I have articulated don’t include the standard my critics would recommend: not helping bad guys. And my critics think many companies – perhaps most – are bad guys that shouldn’t be helped.

Suppose a company is willing to follow my advice (at least some of it) on how to respond better to public outrage. It’s willing to be more transparent about its past sins and current problems; it’s willing to share more control with stakeholders and negotiate accountability mechanisms instead of demanding to be trusted; it’s willing to listen to criticism more openly and more empathically. But it has a bad safety record, worse than the records of most of its peer companies. Or it has a bad carbon footprint, spewing greenhouse gases into the atmosphere in ways it could easily (though expensively) reduce. Or it has other serious substantive blemishes: employing child labor in developing countries, purchasing raw materials from farms carved out of the tropical rainforest, buying and selling collateralized debt obligations, whatever.

It is my fervent conviction that helping such companies be responsive to their stakeholders – including the very activists who excoriate me for working with those companies – is a significant step toward substantive improvement. That is also my experience after 40+ years as a risk communication consultant.

Consider two hypothetical oil companies producing oil in a developing country. One is a publicly owned western company, the other a nationalized company controlled by a corrupt dictator. Activists are in a pretty good position to put pressure on the western company, largely by arousing outrage among regulators, shareholders, and the general public. Then I can try to show the western company why it should be responsive to that pressure, and how to be responsive to that pressure. It’s much harder for activists to find a way to put pressure on the corrupt dictator’s oil company. The dictator’s oil company is profoundly unlikely to hire me to help it address stakeholder outrage when it can afford to ignore stakeholder outrage.

The western oil company doesn’t just talk a better game than the dictator’s company. It is also more honest, more transparent, and more likely to make substantive improvements on everything from land stewardship to labor practices. Obviously the behavior of publicly owned western oil companies is sometimes pretty bad. Does anyone actually doubt that the behavior of nationalized oil companies controlled by corrupt dictators is worse?

The only organizations that seek my help are those that need to be responsive to outside pressure. I am part of an ecosystem that includes activists, shareholders, and regulators – an ecosystem that functions only because of my clients’ vulnerability to public outrage.

My clients’ vulnerability to public outrage doesn’t always result in significant substantive improvements. One way a western oil company can respond to pressure is by selling out to a dictator’s oil company – which may be a good outcome for the western oil company, but is unlikely to be a good outcome for the people who live near the oilfield.

And sometimes clients with serious substantive problems get better at managing stakeholders’ outrage without doing much of anything about the substantive problems. But as I invariably warn my clients, that’s a very temporary achievement. When there are serious hazards that need to be addressed, a company that learns how to talk better and listen better but doesn’t actually change its behavior is only setting itself up for higher outrage to come when its stakeholders eventually learn they have been gulled.

The situation is more complicated when the substantive problems are comparatively minor – in my terms, when the hazard is low. Then outrage management may be all that’s necessary. Or nearly all that’s necessary. Even in low-hazard high-outrage situations, a company that becomes more responsive in order to reduce the outrage usually ends up agreeing to some of its stakeholders’ substantive recommendations as well. So the high outrage gets lower, and the low hazard gets lower too. Still, most of the focus in low-hazard high-outrage situations stays on the outrage. I think that’s as it should be. But my critics don’t necessarily agree.

What my critics are often worried about, I think, is something like this. Suppose a company’s stakeholders don’t really understand what the company is doing that poses a significant hazard. There may be sizable stakeholder outrage (perhaps provoked in part by activist efforts), but it’s focused on comparatively minor substantive issues. So when I help the company address the outrage more effectively, the stakeholders calm down about those low-hazard issues … and still don’t know about the high-hazard ones.

I think this situation is more the exception than the rule, but it does happen. Some issues are easy to arouse outrage about, while others are harder, even though they may pose a more serious ultimate threat to health or the environment. Activists not unreasonably focus on the issues with the highest outrage potential. When I help a company manage that outrage, I am helping it defuse a situation that, if I hadn’t been there to help, might eventually have led to pressure to improve the serious problem.

Another valid criticism, I think, is the charge that outrage management is intrinsically reformist and therefore counterrevolutionary. I used to hear this language a lot more often in the early years of my career. It derives from Marxist theorists like Herbert Marcuse, who argued pretty persuasively back in the 1960s that powerful institutions often make minor concessions in order to relieve the pressure that might otherwise have forced real change. I help powerful institutions become more responsive. Marcuse was right that powerful institutions become more responsive mostly when they decide they must in order to hang onto their power. I think responsiveness under pressure leads to meaningful change – more meaningful than Marcuse believed, and more meaningful than many of my critics believe, and for that matter, more meaningful than many of my clients believe (they don’t always realize what they’re getting themselves into). But to those who think that incremental change tends to be illusory and “real change” needs to be revolutionary, I really am part of the problem, not part of the solution.

comment:

I’ve read some of your columns on how to apologize.

I wonder if you’d consider doing something on how well umpire Jim Joyce did this week when his mistake (calling Jason Donald safe at first with two out in the ninth) cost Armando Galarraga a perfect game. The comments I read on Facebook that night toward Joyce were venomous. It seemed like everyone hated him.

After the game, when he saw the replay and knew he was wrong, Joyce admitted it with no excuses, acknowledged the impact of his mistake, and asked to see Galarraga so he could apologize to him. The next day Galarraga brought out the lineup card before the game to publicly indicate that he didn’t have hard feelings toward Joyce, a classy gesture.

I think a lot of people still hate Joyce, but his response caused many to take a step back, think about how hard umpiring can be and how umpires put themselves on the line, and at least feel some empathy toward him.

peter responds:

I agree. Jim Joyce’s apology was as perfect as Armando Galarraga’s pitching.

A lot of commentators used the same word to describe both Joyce’s and Galarraga’s behavior: “class.” Galarraga didn’t explode in anger, and readily accepted Joyce’s apology. Joyce apologized not once but several times. He focused (rightly) not just on his mistake but on the damage it did: “I just cost that kid a perfect game,” he said, reportedly with tears in his eyes.

It’s instructive to compare Joyce’s apology with that of Tony Hayward, CEO of BP.

I don’t doubt for a moment that Hayward regrets the Deepwater Horizon oil spill in the Gulf of Mexico. He has tried to apologize repeatedly. But his most memorable apology itself required an apology. On May 30, speaking of the families of the eleven oil workers who died in the April 20 explosion, Hayward said – on tape – “We’re sorry for the massive disruption it’s caused their lives. There’s no one who wants this over more than I do. I would like my life back.”

It is human for Hayward to feel sorry for himself. I imagine Joyce does too; he would rather not go down in baseball history as the umpire who blew his most important call. But Joyce managed to stay focused on what he did to Galarraga. Hayward – like Exxon’s Larry Rawl after the 1979 Exxon Valdez spill – has seemed more preoccupied with the harm to himself and his company than to the Gulf and its residents (or to the families of the eleven who died).

This may be an injustice. I don’t know Hayward, and I certainly don’t know how he really feels. Maybe it’s just his technical training that keeps him from showing he feels awful … like the perpetrator, not the victim. Even when he says exactly the right thing – for example, when he acknowledged recently that BP hadn’t been sufficiently prepared for a deep-water blowout – he sounds more like a technocrat pointing out an interesting fact than a CEO confessing a grievous corporate sin.

Jim Joyce, I suspect, could do better.

Note: Five other pieces by me commenting on the Deepwater Horizon oil spill:

• “BP’s Communication Response to the Deepwater Horizon Spill” (May 3)
• “Communicating about the BP Oil Spill: What to Say; Who Should Talk” (May 30)
• “Why We’re Vilifying BP” (June 4)
• “The ethics of risk communication consulting and the BP oil spill ” (June 6)
• “President Obama’s handling of the Deepwater Horizon oil spill” (August 8)

comment:

I’m preparing a report, to be published by the Foundation for Public Affairs, about the role of public affairs professionals in enterprise risk management. I’d be interested in your expert thoughts on what role they tend to play now, what role you think they should play, what they can contribute to the risk management process, and whether risk management professionals and other executives tend to understand what that contribution could be.

The report will be published this fall and can be obtained online at www.pac.org/fpa.

peter responds:

Enterprise risk management has two principal components:

• Identifying things that might go wrong, and assessing their probability and magnitude; and
• Figuring out how to cope with each item on the list – ways to make it less likely to happen, insure against it, mitigate its magnitude if it happens, etc.

Too many organizations still see the role of public affairs in this process very narrowly. After a bad thing happens, public affairs people are expected to use their communications abilities in ways management hopes will mitigate the bad thing’s impact on the organization. Sometimes that means trying to minimize stakeholder and public awareness that the bad thing happened, or its connection to the organization. Sometimes it means “explaining” why the bad thing wasn’t so bad after all, or why it’s not really the organization’s fault.

In organizations that understand the distinction between spin and risk communication, the risk mitigation function of public affairs people is seen differently: acknowledging what went wrong; apologizing for the organization’s role in it; providing all the details, even the ones that are still tentative or that reflect badly on the organization; listening hard while people vent their outrage; collaborating with critics on steps to reduce the damage, compensate the victims, and prevent recurrences; etc.

But even this much-improved vision of how public affairs people can best mitigate organizational damage is still too narrow. What’s missing is a deeply embedded understanding that the very concept of risk has two components, which I term “hazard” and “outrage.” Hazard is substantive risk – how dangerous X is, how much actual harm it is likely to do. Outrage is cultural risk – how upsetting X is.

The correlation between hazard and outrage is exceedingly low; that is, people often fail to get upset about risks that seriously endanger them, and often get very upset about risks that endanger them far less. But the correlation between hazard perception and outrage is quite high; if people are upset about a risk, they’re very likely to think it is doing a lot of harm. Much is known about the factors that determine stakeholder and public outrage (control, trust, dread, responsiveness, etc.), and much is known about the organizational behaviors that exacerbate or mitigate that outrage.

Why is this central to enterprise risk management? Because public and stakeholder outrage is a hazard to the organization. In many cases public and stakeholder outrage is the principal hazard to the organization.

Most organizations have figured out that reputational risk is one of the main categories of risk they face. And most organizations intuitively understand that reputational damage isn’t highly correlated with actual, substantive, hazard-related damage. Sometimes organizations are responsible for awful outcomes and their reputations suffer only a little; sometimes organizations end up crucified for comparatively trivial infractions. (And sometimes, of course, organizations do enormous harm and rightly pay enormous reputational costs.) Finally, most organizations hold their public affairs people chiefly responsible for mitigating reputational damage.

What’s too often missing is a solid understanding of the dynamics of reputational damage – that is, the dynamics of outrage.

So public affairs people need to input an “outrage perspective” in every stage of enterprise risk management:

• When an organization is listing the risks it faces, public affairs people need to make sure that reputational risks (and reputational components of substantive risks) are on the list – and need to keep reminding management that under high-outrage conditions even small substantive problems can constitute huge reputational risks.
• When an organization is assessing the magnitude and probability of each risk listed, public affairs people need to insist on including reputational damage – especially the reputational damage outrage can do even when hazard is small.
• When an organization is deciding how to make enterprise-threatening risks less likely or less damaging, public affairs people need to argue on behalf of outrage-based public affairs strategies – for example, organizations that worry visibly and warn proactively about a risk encounter less outrage if that risk eventuates than organizations that ignore or over-reassure about the same risk.
• When an organization is planning how to respond to the risks on the list if they happen, public affairs people need to fight for plans that are responsive to the dynamics of outrage – for example, plans that put a premium on promptness, candor, and contrition.
• When bad things happen and these plans must be implemented, public affairs people need to manage the implementation in a way that prioritizes outrage mitigation as well as hazard mitigation – staving off the instincts of the legal department and top management (and perhaps even their own instincts) to circle the wagons instead.

This requires two changes. First, public affairs people need to master the ways in which outrage management in high-controversy situations differs from traditional public relations. And second, public affairs people need to be centrally involved in every stage of enterprise risk management – not just implementing damage control at the final stage after earlier stages have gone awry.

There has been progress on both of these fronts over the past few decades. But there is still a long, long ways to go.

Consider some high-visibility risk controversies as I write this in May 2010: Did Goldman Sachs deceive investors in ways that exacerbated the world financial collapse? Did BP take insufficient precautions against a deepwater oil spill in the Gulf of Mexico? Did Toyota design dangerous cars and lobby against government efforts to make them safer? Did the Catholic Church cover up for child-molesting priests and let them continue to ruin the lives of children? I think these four organizations are very different in how much harm they actually did, and in how much responsibility they actually bear. But none of the four has shown much mastery of outrage management. And I would wager that public affairs people played at best a peripheral role in assessing the risks that now confront these four organizations.

comment:

In two recent RISKANAL (an Internet discussion list for risk professionals) postings, Peter Sandman has argued that CDC has been “intentionally misleading” in its discussions of the 2009 H1N1 pandemic. Throughout this comment I use Sandman/Lanard and “they” and “their” since Peter Sandman credits his colleague and spouse Dr. Jody Lanard in the development of the ideas in the links he provides.

I believe that Sandman/Lanard are being unintentionally misleading when they accuse the leadership at CDC of being intentionally misleading. Their assertion that the 2009 H1N1 2009–2010 flu season is mild depends on the word “risk” having a specific, single meaning. Like Humpty-Dumpty, “risk” means what they want it to mean, nothing more, nothing less.

I have read their posting “Why did the CDC misrepresent its swine flu mortality data – innumeracy, dishonesty, or what?” and the first 50 pages or so of their “Archive of Swine Flu Pandemic Communication Updates.” I find most (perhaps 95%) of what they write on the swine flu pandemic knowledgeable and reasonable, but will focus on the other 5%, because that is what leads Sandman/Lanard to criticize the CDC characterization of the 2009 H1N1 pandemic as intentionally misleading.

When Sandman/Lanard assert that the 2009 H1N1 flu season is mild, they assume and assume that the reader will share their assumption that “risk” in this context means the population mortality rate (PMR), which they define, reasonably enough, as the product of the case attack rate (CAR) and the case fatality rate (CFR). They use (in their 2 December 2009 update) CDC estimates from 12 November 2009 to calculate PMRs for children (7.49 dead per million children, age 0–17), adults (14.88 dead per million adults, age 18–64), and seniors (11.44 dead per million seniors, age 65+). They contrast this with the PMRs for seasonal flu: children and adults (13 dead per million non-seniors, age 0–64) and seniors (830 dead per million seniors, age 65+).

They note that CDC has been careful to emphasize the comparison of PMR for non-seniors in 2009 H1N1 and non-seniors in seasonal flu and careful to include the phrase “than in a usual flu season” when they claim the 2009 H1N1 is worse for non-seniors than seasonal flu. So, they are not accusing CDC of lying about the pandemic flu impact, but rather feel that the public misinterpret the claim and believe incorrectly that 2009 H1N1 is more dangerous (higher PMR) to children than seniors. They are concerned that misled seniors will not feel the necessity to be vaccinated when 2009 H1N1 vaccine is available to them. They are concerned about CDC and WHO losing public health credibility when the public and the media discover they have been “misled.”

This brief summary does not do justice to the subtlety and comprehensiveness of their argument.

Unfortunately, if you use different consequence measures than the CFR, the Sandman/Lanard argument falls apart. They recognize in their 2 December 2009 update that “risk” is determined using two factors (quoting):

• Probability (or frequency) – how often does it happen; how many people does it happen to?
• Magnitude (or consequence) – when it happens, how bad is it?

If instead of CFR as a consequence measure, we use the product of CFR and the number of years of life lost (YoLL), the 2009 H1N1 pandemic is almost certainly worse than the average seasonal flu season. Why should we prefer years of life lost as a consequence measure? Sandman/Lanard give us some of the reasons in their 15 December 2009 update where they write:

There are several coherent reasons for prioritizing children (and the youngest adults) even if they are less at risk than their parents and grandparents:

• Young people have more years of life expectancy left.
• Young people are our future.
• Young people may be more likely to transmit infections than the elderly.
• Young people usually have a better antigenic response to influenza vaccines than the elderly.
• Surveys show that most Americans would rather protect the young than the elderly.

I can add a few additional reasons we should use YoLL as the consequence measure for determining/understanding the “risk” of influenza and other diseases.

• We commonly realize that grief at the loss of a young parent or spouse or grief at the loss of a child is an order of magnitude worse than our grief at the loss of an elderly relative or friend.
• That realization is displayed in the success of cable news channels that exploit 24/7 crimes against children and young adults and diseases affecting children and young adults.
• Nature itself uses this consequence measure. Diseases and metabolic conditions that affect only those past the age of reproduction exert only indirect evolutionary selection pressure. For instance, evolution has provided an array of devices to maintain the integrity of the genome and prevent the development of cancer in the young. These protections gradually deteriorate as people age past capability of reproduction.

Sandman/Lanard combine their insistence on use of a specific meaning of “risk” with an insistence that CDC exhibit the sort of intellectual purity demanded by the far left of the Democratic Party and far right of the Republican Party. This shows up in their insistence that CDC management/spokespeople tell only the “truth” in all cases and push for only policies that are consistent with that “truth.” Otherwise, they predict disastrous loss of credibility.

An example is their use in the 15 December 2009 update of a 1996 Wall Street Journal article by Amanda Bennett and Anita Sharpe (winner of a 1997 Pulitzer Prize and available at www.pulitzer.org/archives/5997). Sandman/Lanard note that Bennett and Sharpe describe the U.S. government 1986 decision to conduct a societally broad-based anti-HIV campaign rather that a campaign targeted at high-risk populations. They quote medical ethicist George Annas (“When the public starts mistrusting its public health officials, it takes a long time before they believe them again.”) and former NY City Health Commissioner Stephen Joseph (“Political correctness has prevented us from looking at the issue squarely in the eye and dealing with it. It is the responsibility of the public-health department to tell the truth.”).

Sandman/Lanard say nothing about Bennett and Sharpe’s extensive discussion of why CDC and Health and Human Services chose a broad-based campaign, rather than the more targeted campaign that CDC understood was called for. In the ten years between 1986, when the policy was put in place, and 1996, when Bennett and Sharpe were writing, the levers of power of public health policy were fully controlled for all but two years by Republican administrations and/or Republican Congresses, including Senator Jesse Helms of North Carolina, who was adamantly opposed to, and able to block, any program targeting gays and intravenous drug users. CDC chose to get the policy they could, which had some benefit for the high-risk populations and “wasted” a lot of money on low-risk populations.

Sandman/Lanard say nothing about the courage of CDC managers in the early 1980s, who reprogrammed funds, without the knowledge of their Reagan administration superiors at HHS, to do the epidemiological research that established the mechanisms of transmission of HIV and established the basis for coherent policies to counter HIV infection. Other CDC managers were harshly punished a decade or so later for doing the same thing in the context of another health emergency.

peter responds:

First let me note that Jim Dukelow’s comment is nearly identical to one he posted a few days ago on the RISKANAL listserv. This response, on the other hand, is significantly expanded from the response I posted on RISKANAL, largely because my wife and colleague Jody Lanard collaborated on this one. To join RISKANAL, send the following email message to lyris@lyris.pnl.gov: SUBSCRIBE RISKANAL First_Name Last_Name

Jody and I have no quarrel with the YLL (years of life lost) calculation as an alternative to considering all deaths equally unfortunate and equally worth preventing … postponing, really. YLL certainly adds useful information.

But where was the YLL perspective for all those years when public health agencies were shouting from the rooftops that flu is a grossly neglected disease that kills 36,000 Americans a year, without mentioning that about 90% of those 36,000 are 65 or older, mostly a lot older? The 36,000-a-year factoid was deployed endlessly as a core argument for urging young people to get their flu shots.

Each risk measure points to something different. Population mortality rates (which we should have called “age-specific death rates,” clearer and more customary nomenclature), case fatality rates, attack rates, hospitalization rates, YLL calculations – they’re all useful. The question is whether public health agencies provide several such measures to help us understand ways that a particular disease outbreak is bad and ways that it’s not so bad … or whether the agencies cherry-pick the measure they can most easily use to convince us that “this one” is bad (or not so bad).

As for whether the CDC and HHS misled people (and misled state and local health agencies) about who was most at risk from pandemic H1N1, Jody and I do hold these two federal agencies responsible for the fact that lots of state and local agencies (and lots of journalists) got from the feds and passed along to the public the clear impression that kids and young adults up to age 24 were among those at particularly high risk. This is not a matter of our decreeing what the term “risk” ought to mean; we are simply reading what was said and written.

Here is one of many examples, a New York City health department press release from January 2010, a time when H1N1 vaccine was going begging. I have bolded the parts of this excerpt I think explicitly claim that young people are likelier than older people to get very sick or die.

“The best way to protect yourself or your loved ones from becoming very ill is to get vaccinated,” said Dr. Thomas Farley, New York City Health Commissioner. “People in priority groups are at higher risk of hospitalization and death if they get sick. So don’t take the risk – get the vaccine today.”

Last month, the Health Department lifted any remaining restrictions on H1N1 vaccine eligibility, while continuing to target those in high-priority groups. Those groups include pregnant women, anyone between 6 months and 24 years old, and adults with chronic health conditions, such as asthma, diabetes, or immune deficiency.

Examples like this one make a pretty unambiguous claim about age-specific death rates, not years of life lost.

I would have had no objection to a sentence like this:

Even though the pandemic is killing a much higher percentage of people 50–64 than any other age group, as long as the vaccine is in short supply we are prioritizing all children for vaccination rather than the 50–64 group, because children have their whole lives ahead of them.

I don’t feel qualified to question the vaccine prioritization decisions of the CDC’s Advisory Committee on Immunization Practices. They’re certainly debatable decisions. Very few (if any) other developed countries prioritized all children and young adults up to age 24. Canada, the United Kingdom, France, and Australia did not. But that’s not my field. My concern isn’t which target groups the CDC most wanted to vaccinate; my concern is what the CDC told people in order to get them vaccinated.

It is a truism of risk communication – of all communication – that widespread audience misunderstandings are by definition the communicator’s fault. In 2009 most Americans came to believe that swine flu was more dangerous to children than to their parents or grandparents. That belief was false. The CDC knew it was false. But that belief was conducive to getting kids vaccinated, so the CDC encouraged the false belief.

The CDC chose to take a short cut. Instead of emphatically offering an accurate rationale for prioritizing children (YLL, contagion, higher attack rate, higher vaccine efficacy, etc.), it implied a false rationale. And when local and state health officials, themselves misled, said flat-out false things about which age groups were most at risk, the CDC chose to do nothing to correct the error.

I believe this was not ethical, and in the long term I believe it was not wise either.

It was, however, successful. The percentage of all U.S. children vaccinated against pandemic H1N1 was far higher than the percentage of high-risk prioritized adults vaccinated.

In January 2010, the CDC published survey data about H1N1 vaccination uptake. Here’s what it reported:

By the end of December 2009, an estimated 61 million persons (20% of the U.S. population) had been vaccinated, including 27.9% of persons in the initial target groups, 29.4% of children, 11.6% of adults aged 25–64 years with underlying medical conditions, 22.3% of health-care personnel, and 13.9% of adults caring for infants aged <6 months.

In April 2010, the CDC published updated survey data about H1N1 vaccination rates through the end of January, and included this discussion about the lower rate in the high-risk 25–64 age group compared with the higher rate among children. By this time, the uptake in kids had risen to about 37%:

The 2009 H1N1 vaccination coverage rate among adults at high risk aged 25–64 years was lower (median: 25%) than the rate among children. Reasons for this might include a lesser emphasis on vaccination of this population compared with children, lack of preexisting relationship of state immunization programs with providers who serve adults at high risk, difficulty in implementing a risk-condition-based recommendation for persons in this age group (resulting in vaccination program implementation challenges), and historically low seasonal influenza vaccination rates in this population.

Jody and I were glad to see the authors consider the possibility that “a lesser emphasis on vaccination of this population compared with children” might account for the lower pandemic vaccination rate of adults with high-risk conditions than of children, even though adults with high-risk conditions were far likelier than children to suffer serious illness and death from pandemic H1N1.

But we were surprised to see “historically low seasonal influenza vaccination rates in this population” on the list of possible explanations. In recent history, adults with high-risk conditions have had a much higher seasonal influenza vaccination rate than children. The rate in children rose from 24% in 2008–2009 to about 40% in 2009–2010.

As the authors probably know already, high-risk adults 18–64 got the 2009–2010 seasonal flu vaccine at a much higher rate than they got the pandemic flu vaccine: 36% of high-risk 18–49-year-olds, and 45% of 50–64-year-olds, got the seasonal vaccine (about the same as in the previous flu season). So it’s not that high-risk adults were unaccustomed to flu vaccination. Rather, high-risk adults apparently saw pandemic vaccination as a high priority for their children but a lower priority for themselves.

Two important side notes:

First, I think Jim Dukelow rightly describes the Bennett/Sharpe work on the CDC’s decision in the 1980s to do broad rather than targeted HIV outreach. He does more justice to this decision than Jody and I did in the website post he quotes.

Second, there is one point about which I think we did unintentionally mislead Mr. Dukelow. He writes that our “assertion that the 2009 H1N1 2009–2010 flu season is mild depends on the word ‘risk’ having a specific, single meaning.” Because there were fewer pandemic H1N1 flu deaths in the U.S. than in average flu seasons and past pandemics, he says, we insist on calling the pandemic mild – and then we criticize health agencies for failing to adopt this single standard and failing to call the pandemic mild. A years-of-life-lost standard, he says, undermines this claim.

It does seem to be true – so far – that the number of deaths from pandemic H1N1, while lower overall than in the average flu season, is about three times higher in people under 65. Whereas about 90% of U.S. seasonal flu mortality is in people 65 and older, most of the estimated pandemic deaths in the U.S. were in middle-aged adults, and about 10% were in children under 18 … far more children than the average flu season kills.

But our assessment of the 2009 flu pandemic as “mild” is based mostly on years of official use of the term “mild” to describe previous pandemics (1847–1848, 1889, 1968). In “Swine Flu Communication Challenges and Lessons Learned,” Jody and I recently noted:

Because the H1N1 pandemic was killing more children than seasonal flu, it became anathema to call the pandemic “mild.” In the past, the CDC has routinely labeled certain seasonal flu years as “mild,” “mild to moderate,” etc. And public health officials have routinely called the 1968 pandemic “mild” or “relatively mild.”

Our use of the term “mild” is also based on criteria U.S. officials developed in advance of the H1N1 pandemic to characterize pandemic severity … and then decided not to use to describe the pandemic they actually faced. The proposed Pandemic Severity Index (PSI), developed by the Department of Health and Human Services, described a “Category 1” pandemic – the lowest category on the proposed severity scale – as one that would cause fewer than 90,000 deaths in the U.S.

Just as a Category 1 hurricane is still a hurricane and can still do considerable damage, a mild pandemic is not a walk in the park – especially for the most severely affected age groups. More widespread use of YLL can help clarify that this generally mild pandemic has had a comparatively severe impact on kids. But pretending that children are likelier to die than their parents and grandparents isn’t an appropriate way to “explain” that children’s deaths are especially devastating and therefore worth prioritizing.

comment:

Regarding pandemic preparedness, I have some strong feelings about how this issue was communicated.

From a behavioral standpoint, I know if we take a person, inject artificial noise and point to a behavior which mitigates this noise, we typically see an increase in adherence to that behavior. In June of 2009, it was clear to most public health experts that the H1N1 pandemic was not as severe as expected. However, the communication (in my opinion) continued to inject noise above that which was scientifically demonstrated (or as I like to put it – ethically guided).

I am interested in your opinions and insights. I know being a public health organization is tough. You communicate too little, you get hit. You communicate too much, you get hit like this. But I honestly feel (as a public health professional) that we lost our way on this one. In my opinion, our job is to communicate risk – and let people make the decision how to behave. However, on this one our communication caused excessive noise … leading to a behavior which has not yet demonstrated it has true value.

peter responds:

I share your judgment that from June 2009 onward, public health warnings about pandemic H1N1 were more alarming than the actual severity of the pandemic justified. As you suggest, that wasn’t so in the early weeks, when too little was known about its actual severity. But by June there was ample evidence that the pandemic was surprisingly mild so far. And to this date that assessment remains so.

But I don’t think you judge a warning by whether it’s commensurate with what’s happening, or with what actually happens later. You judge it by whether it’s commensurate with what might have happened. I think public health officials were justifiably concerned that the pandemic might become more severe over time, as some other flu pandemics have done.

That concern was justified in June when there was little individuals could do except wash their hands and cover their coughs, and it remained justified in the fall when vaccine began to be available in some developed countries. Some experts say it remains justified today, since there is no way to be sure yet if the new H1N1 virus might still return in a more virulent form.

When it comes to warnings and preparedness about potential catastrophes, it’s not “damned if you do and damned if you don’t.” It’s darned if you do and damned if you don’t. If officials had downplayed a pandemic that turned out severe (and ordered far too little vaccine), the criticism would have been much more vituperative than the criticism they face now for having hyped a pandemic that turned out mild (and ordering far too much vaccine).

So I am not critical of health officials for continuing to sound the alarm after June 2009, despite the increasingly clear evidence that the pandemic was much milder than they had initially feared. I am, however, critical of other things:

• Officials didn’t concede – and in many cases still have not conceded – that the pandemic was in fact very mild so far … even though it was deadlier to the young than the seasonal flu, which sickens all age groups (young people more than older people) but kills mostly the elderly.
• Officials didn’t explain clearly enough or often enough that their warnings weren’t about what had happened so far, but about what might happen next. They did little to help the public imagine the severe pandemic they feared, and to contrast it with the mild pandemic we were already experiencing.
• Officials didn’t share the dilemma of crisis preparedness, predicting (and hoping) that they were probably ordering more vaccine than we would end up needing, and explaining that that was a wiser error than running out if the virus took a turn for the worse.
• Officials didn’t suggest that seasonal flu vaccination was a lower priority than in most years, since one of the likelier scenarios was that the pandemic strain would supplant seasonal strains in 2009–2010. And officials rarely mentioned early signs that the worst seasonal strain still circulating in small amounts, H3N2, had mutated to become a bad match with the 2009–2010 seasonal vaccine strain.
• U.S. officials in particular didn’t adjust their messaging regarding the most vulnerable age groups when it became clear that adults (especially adults with preexisting conditions) were far more endangered by pandemic H1N1 than children.

For more than you want to read on these and other criticisms, see my index of “Swine Flu: Substantial Articles, Interviews, and Guestbook Entries” and my “Archive of Swine Flu Pandemic Communication Updates.”

I think continuing to warn was a justified public health decision. But I think the warnings could have been a lot more precise, more helpful, and more honest.

sean responds:

Your response really helped me identify just what I have been upset about for the past four months. The warnings could have been more precise, helpful, and honest. That sums it up perfectly. The cherry on the top of the sundae would be if organizations acknowledged this.

comment:

You ask: What Did Goldman Sachs Do Wrong?

It failed to avoid being scapegoated:

• For the ideologically-based policy failures of recent Presidential Administrations in their fore-doomed strong-arm promotion of universal home-ownership.
• For the exacerbation thereof by the unfortunate deflationary tendencies of a debt-based currency that was instituted with the formation of the Federal Reserve System itself.
• For the prior disappearance of Congressional (Glass-Steagall) segregation and protection of depositor funds amidst the effort to maintain the global competitiveness of American financial institutions.
• For the long-established general reliance upon the socialization of large-institution financial losses amidst an emerging climate of bureaucratic deregulation and of enhanced opportunity for private profit.
• And for shamefully failing to incompetently but equitably share in the general distress induced by the senior institutions identified above.

Show trial, anyone?

peter responds:

Well yes, various arms of the government acted in ways that caused or exacerbated or permitted the economic collapse. So did Goldman Sachs and the rest of the investment banking industry. So did most of us, one way or another. But government bears the lion’s share of the responsibility simply because preventing economic collapse is one of the government’s core tasks. It’s not my task, or your task, or Goldman Sachs’s task.

And yes, your sarcastic last bullet point is sound: Part of why people are outraged at Goldman Sachs is that it has survived and even flourished, while so many others have suffered and are still suffering. People and institutions that profit from the misfortune – or imprudence – of others can expect a fair amount of heat. That’s true even if they’re not actually causing the conditions from which they profit. Betting against the general welfare isn’t likeable.

So what should Goldman Sachs have done to avoid the outrage, and what can it do now to mitigate the outrage? Suffer more? Avoid being smarter than the government and the rest of us? Restrict itself to profit opportunities that don’t come at the expense of the overall society?

I think there are more generic outrage management lessons to be harvested here. After all, John Paulson is the character in this drama who most clearly profited from betting correctly that the subprime housing market was about to tank, while most of the “real experts” were betting that it would stay healthy. Goldman Sachs just facilitated Paulson’s gamble and went along for the ride. And yet there’s not nearly as much public outrage at Paulson as at GS.

Scapegoating is a kind of projection. We project our own guilt, shame, and responsibility onto the scapegoat, making it bear the weight of everyone else’s sins as well as its own. The risk communication question here is: What could Goldman Sachs have done to “stay out of the projective field” – to avoid being the lightning rod for all this projection, the most attractive potential scapegoat around? And what could it do now to “get out of the projective field” – to allay some of people’s outrage, help people accept our collective responsibility, and persuade people to prefer the benefits of sound regulation to the satisfactions of punishing the scapegoat?

I think the principles of outrage management offer some answers.

comment:

This is in response to your email exchange about Goldman Sachs.

I really liked seeing how opinions can change and shift as a result of research and discussion. That’s how it should be so I was glad there weren’t any accusations of “flip-flopping.”

My specific comment has to do with your fourth question – are we learning the right lessons? I don’t think we are.

I think a key lesson from the whole mess is the importance of regulation. As you say, corporations have a duty to their shareholders to do everything legal to make money. I tell my students, if it is not illegal to dump waste irresponsibly, arguably a corporation is violating its fiduciary duty to its shareholders if it spends more to “do the right thing.” A certain amount of going beyond the law can be justified in the name of public relations, but if we, the people, think something is damaging to society, we need to make it illegal.

Instead, there is a perception that regulation and capitalism are opposites. People think that the way for capitalism to flourish is to eliminate regulation. And that seems to help in the short term – but didn’t work out so well in the long term for the financial markets. (Or so I understand. I don’t fully understand how the deregulation in the ’90s led to the mess, but it seems to be a contributing factor.)

The thing that’s important about regulation is doing it correctly. In environmental regulation, we’ve found that telling people exactly what they have to do doesn’t work well. It’s better to tell them what we will and will not accept and let them figure out how. That’s where capitalism shines – figuring out innovative ways to reach goals more efficiently. So I worry that our financial regulations will end up outlawing the specific instruments that are implicated in this crisis, but not saying what we will and won’t accept. If that’s the case, we’ll just do it all again in a few years, because you can’t write laws as fast people can come up with new ideas.

Unfortunately, I don’t know enough about the stock market to know what’s acceptable and what’s not. It seems to me that just betting whether stocks will go up or down without actually providing capital to get things done should not be part of where people’s pensions get invested. If people want to do that on the side, fine. But it would take someone who understands the system to know how to write things like that into law.

Wouldn’t it be nice, though, to be able to have that debate on a national level, rather than who should get the blame?

peter responds:

I agree with both of your points: (a) that corporations are impelled to maximize profitability, which means we need regulation to make sure profitable but socially harmful corporate behaviors are illegal; and (b) that regulation works better when we tell companies what outcomes to achieve rather than how to achieve them (in engineering terms, setting “performance specifications” rather than dictating “technical specifications”).

You wrote: “A certain amount of going beyond the law can be justified in the name of public relations.” That relates to where outrage (and thus risk communication) fits in.

Part of the answer, clearly, is that outrage is the other main constraint, besides regulation, on what companies may do to make money. If Goldman Sachs had been outrage-savvy, it might have realized that its high-visibility, master-of-the-universe arrogance and its astronomical profits made it an ideal target for public outrage. It might have judged that painting a target on its own back was especially inadvisable while it was heavily engaged in financial transactions (like going short on much of the U.S. economy) that ordinary people would find extremely unappetizing. It might have predicted that if the economy collapsed (or even contracted significantly) in ways that were arguably related to the new and poorly understood conventions of investment banking, the government and the public would be actively searching for suitable scapegoats, and GS would be a prime candidate. It might therefore have decided to make more of an effort to explain how the business of banking had changed, and less of an effort to conduct that business without regulatory constraints.

But part of the answer, too, is that outrage makes it much, much harder for people to learn the right lessons. The more outraged the public becomes, the stronger the pressure on government to punish instead of regulating – and to regulate in ways that exacerbate the punishment instead of ameliorating the problem. We have just been through a bailout that was profoundly unpopular because it helped big banks and other institutions widely seen as undeserving. The government did a pretty poor job of explaining – regretfully – that it was necessary to bail out (rescue) the banks in order to rescue (bail out) the economy. Now people are in a mood to punish the banks, even if the punishments we settle on might also punish the economy rather than protecting it. Unfortunately, we’re in a mood to prefer tech specs to performance specs. We want the government to micro-manage the investment banking industry, not because we really think Congress understands investment banking, but because we will get pleasure from seeing investment bankers tied up in knots.

So the better job Goldman Sachs does of managing the public’s outrage – outrage that is richly justified but partly misdirected – the likelier we are to end up with regulations that work instead of regulations that punish.

My wife and colleague Jody Lanard adds the following:

Professor Selby writes: “A certain amount of going beyond the law [in the direction of voluntarily refraining from legal, profitable, but harmful corporate behaviors] can be justified in the name of public relations, but if we, the people, think something is damaging to society, we need to make it illegal.”

If one company unilaterally refrains from the legal, profitable, but harmful behavior, that might give it a public relations benefit, but it still harms its shareholders relative to similar companies that continue doing the legal, profitable, but harmful behavior.

This reality is an argument for corporations – not just activists and “we, the people” – to lobby for regulation that restrains or prohibits the harmful but profitable behavior. This kind of lobbying is “precaution advocacy” risk communication, as opposed to “outrage management” risk communication.

Corporations often want a level playing field of regulations, especially when their behavior is better for the world but worse for their own shareholders (in terms of short-term profits) than the behavior of competitors. Corporations sometimes have to beg regulators to “stop all of us from sinning,” in order not to lose customers or shareholder value as a result of voluntarily behaving well.

I would add that my corporate clients often hesitate to do the kind of precaution advocacy Jody is recommending, even though they know she’s right. I once had a paper industry client that had built a new mill with state-of-the-art dioxin controls, anticipating that the regulations would soon require such controls. When regulatory fervor declined, my client was left with an unusually expensive paper mill. The CEO acknowledged to me that it made economic sense to lobby for more stringent dioxin regs – regs his new plant would already meet. But, he said, it just felt wrong to lobby for tougher regulations.

What kind of game-changer might it be if Goldman Sachs were to lobby for tougher regulations on investment banking?

comment:

Thanks for all your insightful and considered commentary about the H1N1 pandemic. I recently came across your forensic dissection of the then Australian Health Minister’s speech about pandemic preparedness back in 2005. (We have had a change of government since, and Mr. Abbott has recently become the opposition leader.)

I wonder whether you have monitored the current Australian government response to the H1N1 2009 pandemic.

Although initially emphasizing the seriousness of the pandemic, by June 22 government messaging was that “the disease is mild in most, severe in some and moderate overall,” and that the aim of public health policy was to protect the most vulnerable. In fact the government modified its initial phased pandemic response and created a new “PROTECT” phase which acknowledged that the existing phases were not appropriate given the mildness of the pandemic.

The official government website says: “PROTECT recognises that infection with pandemic (H1N1) 2009 is not as severe as originally envisaged when AHMPPI [pandemic management plan] was written in 2008, and that although this new disease is mild in most cases, it can be severe in some. PROTECT puts greater focus on treating and caring for people in whom the disease may be severe.”

So far, it looks to me like they were getting the message about right.

However, we have been studying the community response to the management of the pandemic and it seems that the perception that the government overemphasized the seriousness of the pandemic is quite common. Do you have any thoughts about what might have got lost between the official messaging and the community perception?

peter responds:

My wife and colleague Jody Lanard collaborated on this response.

The main recent flu vaccination issue in Australia has been a shortage of seasonal vaccine. The southern hemisphere’s winter is fast approaching, vaccine demand is up (largely because of swine flu), and vaccine supply is down (for a range of reasons: the priority given to the swine flu vaccine, manufacturing problems, shipping holdups because of Iceland’s volcano, etc.). You’ve also got a breaking story about possible side effects in some young children who received the 2010 seasonal flu vaccine produced by the Australian company CSL.

But the issue you raise is an interesting and important one. Australian journalists and the Australian public had no trouble figuring out that the swine flu pandemic was pretty mild – certainly a lot milder than the bird flu pandemic for which health officials in developed countries like Australia had been trying to prepare. But a lot of reporters and a lot of ordinary citizens didn’t quite get it that the Australian government had noticed too, and had changed its policies in response to the unexpected mildness of the pandemic.

The question, then, is how can a government agency talk about crisis management policy changes in a way that sinks in?

The Australian government certainly tried. Your quote from the Department of Health and Ageing website is impressive. Before swine flu, Australia’s pandemic phases were labeled ALERT, DELAY, CONTAIN, SUSTAIN, CONTROL, and RECOVER. The CONTROL phase would have included broad-based H1N1 vaccination as soon as there was sufficient vaccine available. Because swine flu turned out mild in most cases, Australia explicitly abandoned that phase. Instead, it implemented a newly created PROTECT phase that focused on “treating and caring for those more vulnerable to severe outcomes.”

Vaccination in the PROTECT phase was similarly aimed at “people at increased risk of severe outcomes.” Instead of struggling to vaccinate everyone, or triaging limited vaccine to protect people needed to keep critical infrastructure up and running, Australia decided that most people could probably weather a fairly mild swine flu pandemic without medical help. So it decided to concentrate on treating – and vaccinating – the most vulnerable. And it said so.

By contrast, many other governments refused to acknowledge until very late in the game that the pandemic was looking rather mild, compared with their initial fears and compared with previous flu pandemics. Some have yet to make this acknowledgment.

Many countries, especially in the developing world, put far too much emphasis for far too long on containment, which sent two false signals: that the pandemic could be contained and that it had to be contained – that it was stoppable and severe. Others, like the U.S., said from the outset that containment was unachievable, but still emphasized how dangerous the pandemic was, even as their own emerging data showed it to be less dangerous in many ways than an ordinary flu season.

The mere use of the word “mild” came to be seen as a sign of callousness toward those who had died, ignoring the fact that for years the World Health Organization and national governments around the world had routinely referred to the “mild” or “relatively mild” pandemics of 1957 and 1968.

Ultimately, WHO paid the highest price for this reluctance to acknowledge the ways in which the pandemic was mild. Largely because it failed to integrate severity – low severity – into its pandemic messaging, WHO was accused of having created a “fake pandemic” that unduly frightened millions. National governments around the world were similarly castigated for having wasted precious healthcare funds on pandemic vaccine nobody wanted.

Australia’s government was less criticized than many – and rightly so. But as you point out, even in Australia many people got the impression that the government had made the pandemic sound more severe than justified.

So what could Australia have done better? The list that follows makes some of the same points I made in my February 2010 column, “Telling People You Got It Wrong.” I recommend that column as a companion piece to this response.

rebecca responds:

Overall, I think that the government has been saying most of what you recommend.

Health Minister Nicola Roxon repeated the “mild in most, severe in some, moderate overall” theme every time she talked about swine flu. However, I think from our work it is possible that the message would have got through more clearly if it had been delivered by someone other than a politician.

I have just one thing to note about your appraisal of the government’s website. I agree it says mostly the right things. The problem is that very few people in our surveys (less than 5%) reported using the internet to get pandemic information, and even fewer preferred it as a source. We have also done focus groups and interviews, and these suggest that people would go to these websites only if directed or if specifically looking for something. (Some of our focus group participants had obviously visited the websites to bone up on swine flu before the focus group.)

The government recently launched a new advertising campaign called “The Facts about Swine Flu,” which is designed to encourage people to have the vaccination by highlighting that H1N1 hit younger people more severely, including pregnant women, and more people ended up in ICU than in a normal flu season. However, I don’t think it’s working that well because at the moment we seem to have some expert or other in the media nearly every day talking up the possibility of a return of H1N1 this winter. It feels coordinated.

Of course, the government paid for 21 million doses of vaccine, of which only around 8 million have been ordered so far. So they obviously want to use it up. On the other hand in the last few days there have been stories saying that H1N1 will dominate seasonal flu and that probably 30% of kids had swine flu last year and with the 20% or so of adults that have been vaccinated we might have already got herd immunity. So, as with most of the swine flu communication it is quite confusing. What the general public who are probably not taking much notice are expected to make of it all, I really don’t know.

I do think the idea that swine flu was not even as bad as normal flu has taken hold. Since vaccination rates for seasonal flu are only about 40% anyway (mainly in older and at-risk people), it’s unlikely that most people would see it as necessary to get the swine flu vax. Also there is a mismatch between rhetoric and action. The government is saying everyone should get the swine flu vaccine, but at the same time saying H1N1 is mild, and not providing the seasonal flu vax for free (which now includes H1N1 as well). So for your average person there’s a contradiction (given they think that swine flu isn’t as bad as normal flu). Also I noticed today that information websites and media stories about “how to tell if you have swine flu” say something along the lines of the symptoms are similar to seasonal flu. So little wonder if people are a bit confused by the messaging.

So the problem for the Australian government is:

• People don’t think H1N1 is bad enough to be vaccinated against.
• We have not a shortage but rather a surplus of H1N1 vaccine.

From a risk communication perspective I guess this is a case where precaution advocacy is required. However, I think it is a bit of a losing battle.

Last year’s experience has already led most people to consider the virus as mild and that’s exactly what the government said. Now they are trying to convince people that: “Well yes we did say that it was mild but it’s still worth getting the vaccine because some people were severely affected and that could be you, even if you are not in a high-risk group, even though last year you would have heard a lot that the people dying had underlying conditions and we were focusing our efforts on only the most vulnerable. Now we know (now that the “facts about swine flu” are in) that around 30% of people that were hospitalised and who died had no underlying condition and that they were on average much younger than normal seasonal flu victims. And we expect a second wave to happen any day and it could be worse than last year.”

Of course, this is all presented in the context of the many doses of vaccine that have already been paid for. I’m sure this leads many people to think that the government is now trying to talk the pandemic up just to get rid of the vaccine, not necessarily to make sure that the community stays healthy.

So it seems to be a bit of a low-hazard, low-outrage scenario, which probably wouldn’t call for any risk communication at all – except that there is excess vaccine, and the situation could transform into a high-hazard, low-outrage or even high-hazard, high-outrage situation if the virus changes or a second wave comes that is worse. I wonder whether some dilemma-sharing might be the best way to encourage vaccine uptake and counter complacency?

peter and jody respond:

The Australian government must feel like it can’t win! It was more candid than most governments about the mildness of H1N1 so far, but the Australian public still got the impression that it was hyping the pandemic. Now it’s trying to convince the public that getting vaccinated against H1N1 is a good idea as an extra margin of safety, and people are saying, “But you told us it was mild.”

It sounds to me like the Australian government’s main messages right now should be these:

• So far, the swine flu pandemic has been mild, milder than we initially expected. For most people who have gotten sick, it’s been pretty much like the seasonal flu.
• We keep saying “so far” because the story isn’t over. Swine flu could come back in a more severe form. We have no way to know if this will happen. There are no signs of it yet.
• There have been some important differences between swine flu and seasonal flu so far. The seasonal flu kills mostly the elderly, but swine flu has killed a wider range, including a lot more children and young adults than the seasonal flu. Most of the swine flu victims so far have had preexisting health problems, but about 30% were previously healthy.
• Back when we worried that swine flu might turn out more severe than it has been so far, we made the decision to buy 21 million doses of swine flu vaccine. We have a lot left, and it would be a shame to waste it. We think it makes sense for people to get vaccinated against swine flu. The H1N1 vaccine is free (except for your doctor’s consultation fee) and readily available.
• Even though the odds that swine flu will do you serious harm are pretty low so far, the odds that the H1N1 vaccine will do you harm are much, much lower. We can’t honestly say it’s a top public health priority to get vaccinated against swine flu – not compared with making sure your children are up to date on their childhood vaccinations; not compared with wearing your seatbelt, or quitting smoking, or taking precautions against sexually transmitted diseases. It isn’t a top priority, at least not right now. But it is sensible. Why risk the small chance of serious illness and the bigger chance of an unpleasant week in bed when a simple jab can cut both risks substantially?
• As for the seasonal flu vaccine…. [Here’s where it gets a bit complicated. Presumably getting the seasonal vax makes more medical sense than getting the swine flu vax, since it offers protection against H1N1 plus two other flu strains. But at the moment there’s plenty of swine flu vaccine in Australia and it’s free except for the GP consultant fee; by contrast, there’s a shortage of seasonal flu vaccine, people in low-risk groups have to pay for it, and there’s some question now about its safety for young children. It seems to me that that’s exactly what the Australian government should be saying, sharing the policy dilemma regarding tradeoffs between the two vaccines.]

comment:

You describe your career as “muddling” into risk communication, a discipline that did not exist formally when you started out. What combination of education and training would you recommend for a risk communicator starting out today?

peter responds:

The first national conference with the phrase “risk communication” in its title took place in January 1986 in Washington DC. Today there are too many risk communication conferences to count, let alone attend. I now actually meet people from time to time who tell me they’re in risk communication and then ask me, “and what do you do?”

Even today, most risk communication practitioners learned on the job. They studied something else in college (and perhaps in grad school): risk assessment, mass communication, psychology, public health, health education, etc. My own academic background, for example, is in psychology and mass communication.

I think this is still a viable career path, especially for people who have already finished their schooling and aren’t enthusiastic about doing it again. Read a lot of risk communication – in journals, on websites like this one, etc. Read the news with a risk communication perspective in mind. (This week’s question: What would you have done differently if you were helping European authorities talk about the air travel problems caused by Iceland’s volcanic eruptions?) Start getting your risk communication opinions into other people’s hands: Write letters to the editor, trade journal articles, unsolicited emails full of unpaid advice, blog entries, even comments on this Guestbook. Then look for a job where you’ll actually get paid for your risk communication opinions.

But the time is fast approaching when new risk communication practitioners will need formal risk communication credentials. Now that risk communication exists as a field, I suppose the proper education and training for that field is education and training in risk communication.

More and more academics are taking an interest in riskcomm, and some have managed to cobble together undergraduate and graduate programs. I don’t know most of these programs well enough to assess them. There are also a large number of individual risk communication courses, which can be housed virtually anywhere, from journalism to public health.

The best way to judge these programs and courses, I think, is to look at course reading lists. I would focus especially on two questions:

Here are some examples of some risk communication course syllabi that are available online with their reading lists:

• http://www.grady.uga.edu/syllabi/JRMC_8170_FreimuthV_FA09.pdf
• http://training.fema.gov/EMIWeb/edu/docs/syllabi/University%20of%20Pittsburgh%20-%20Risk%20Communication.doc ( document)
• http://www.jasonswrench.com/courses/450/syllabus.pdf

In spite of what I said in #1 above, a solid academic program in risk communication should push students to take courses in collateral fields as well. The three most important, in my judgment, are these:

• Social psychology and persuasion theory (and perhaps even advertising, the discipline that takes applied persuasion most seriously)
• Risk assessment (so you can understand what your employer or client is telling you about whether something is dangerous or not)
• Mass media studies (including social media and public relations)

If you’re in a risk communication program that lets you get through without much social psych, risk assessment, and mass media, supplement the program on your own. If you’re considering entering such a program, look elsewhere.

A doctorate is a very different animal than an undergraduate or master’s degree. What matters most when studying for a Ph.D. is the individual faculty member you hope to work closely with – TA his or her courses, RA his or her research, help with his or her consulting, and do your dissertation under his or her guidance. The rest of the program matters, but the mentor matters more – so much more that it’s not crazy to jump through the hoops of a program that’s not really your cup of tea in order to work with the person you most want to work with. To find the right mentor, you need to read the work of university profs until you find someone you can imagine modeling yourself on for a few years. Then start a correspondence. Then ask what sort of doctoral program might be arranged.

The other side of the argument: Your may change your mind. The prof you came to study with may leave, or retire, or turn out to be a jerk. It’s nice if there’s a decent program to fall back on. A third side of the argument: You may be more interested in the credential than the mentoring, and mostly looking for a place that won’t make you jump through too many hoops. Still, in a small field like risk communication, I think picking a Ph.D. program is mostly picking a mentor. You want to end up feeling and functioning more like an apprentice than a student.

I have three jaundiced, undocumented, and I hope incorrect prejudices about why it’s difficult to get really useful education and training in risk communication.

• I think too many “risk communication” practitioners are really just relabeled PR people or science educators – two fields that I consider very different from risk communication (and from each other).
• I think “risk communication” educators tend to be closer to the mark than practitioners in understanding what’s new and different about risk communication. But too many riskcomm educators are excessively focused on theory-building and data-dredging, as opposed to what practitioners actually need to know. I wouldn’t advise studying with anybody who isn’t out there consulting in the real world.
• I think the exceptions, the people who are good risk communication practitioners/consultants themselves and should be spending some time in colleges and universities preparing the next generation of practitioners/consultants, are having too much fun (and making too much money) trying to help solve big problems instead. They don’t get around to teaching a lot of courses.

I should add that the “you” to whom this answer is addressed isn’t Rusty Cawley, who posted the question. Rusty is a 20-year veteran PR man who has made the transition to risk communication and now works as communications coordinator for the Integrative Center for Homeland Security at Texas A&M University. He is a fan of my work, and I’m a fan of his. Check out his risk communication website, “ColdCrisisTV.”

comment:

This comment relates to a Guestbook post from September 2009 regarding measuring the success of your risk communication principles in an experimental setting. You suggested that in the real world focus groups and possibly a survey would be the only practical way to assess communication effectiveness. However, I think there may be some occasions where a randomized design would be ethical and appropriate. I am specifically thinking of a cluster randomized design which compares standard practice (i.e. the communication strategy that would be used anyway) with one based on your risk communication principles.

To be ethically acceptable randomization should only be used when there is equipoise – when there is genuine uncertainty about which course of action is more effective. It seems to me given the lack of empirical evidence about risk communication outcomes that equipoise would quite often exist. (Arguably proponents of each approach might believe that their approach is better, but if that is based on nothing more than essentially gut feeling equipoise would still exist – that is, they genuinely don’t know which is better.)

Something like this would probably be best tested in a situation where the hazard is not extremely high. It’s difficult to see a research ethics committee agreeing to a study where a failure of the risk communication strategy could be seen to lead directly to deaths, although logically, if we don't know that the standard communication plan will work then in practice this is allowed all the time. Ideally, the study might involve a situation where the desired outcome is a behavior change and where probably precaution advocacy is needed and can be leveraged by using your theory of increasing outrage. If this was successful, and with the lessons that were learned in undertaking the study, you might then move on to use the same methodology to try to reduce outrage.

Practically, I can see this working in more closed settings rather than necessarily at a community level – for example in a health service, where the units of randomization would be individual administrative units (e.g. hospitals or community health centers). Ideally, you would want a setting where people wouldn’t be likely to discover that the communication method was different for different randomized units. It might also be possible to test the theory with community members in a general practice setting (i.e. primary health care physicians), where individual practices could be randomized but the outcome would be measured in the patients. I’m sure there would be similar administrative units in areas outside of health care.

Of course there are specific methodological challenges in using cluster randomized designs (the sample size needs to be adjusted because the individuals in each unit that is randomized are more similar to each other than to individuals in other units). But there is a growing methodological expertise on how to address these challenges. These kinds of designs are being used, in health care at least, to assess complex packages of care and particularly behavior change interventions.

Ideally, focus groups and surveys could be embedded in a study like this to explore some of the more complex responses and barriers/enablers to uptake of the various communication methods. This would strengthen the findings of the study.

peter responds:

Together with my wife and colleague Jody Lanard, I have long fantasized about creating a foundation that would fund research only if advocates of competing hypotheses applied jointly, stipulating in advance that both sides would abide by the results. So I very much like your point that if you haven’t got a clue which works better, X or Y, there’s no compelling ethical reason not to test the question in a real-world experiment – whereas if you were already pretty confident that X was preferable, you’d arguably be mistreating the part of your sample you assigned to Y.

And I accept your counsel that a cluster randomized design is usable (and used) to get around the problems of real-world randomization. Because you randomized, your conclusions can be a lot more confident than conclusions drawn from focus groups or surveys, or from case studies, or from simulation experiments.

Nonetheless, let me say a word in defense of focus groups, surveys, case studies, and simulations … and even less rigorous methodologies than those. Many of my clients despair of being able to do methodologically state-of-the-art research on the behavioral impacts of alternative messaging strategies. They lack the requisite expertise, budget, or time (or all three) to design and implement a cluster randomized experiment. In fact, they may lack the expertise, budget, or time to field a fully professional focus group or survey.

Too often they conclude that there’s no point in doing any research at all. So they go with standard practice, or their intuition, or my intuition.

Rather than doing no research at all, I’d like to see them ask a few key questions of a dozen people on the bus or in the supermarket. (We can call this a “sample of convenience” so it doesn’t feel so amateur.) Of course a carefully done focus group or survey is far more reliable than this kind of slapdash, informal “study” – and a true real-world experiment is far more reliable – though less rich – than a focus group or survey. But even an extremely casual (and totally unpublishable) piece of research is a huge improvement over no research at all.

Every time I have managed to convince a client to spend an hour or two “testing” a proposed communication strategy with a handful of miscellaneous people, the strategy has improved markedly. Or at least it has changed markedly. To be certain the change was an improvement, we’d probably need a cluster randomized experiment.

comment:

I have found the material on your website very informative and helpful in making sense of risk communication.

I am presently completing my master’s degree which is risk-based. I am studying the conflict that arises between the management of the needs of victims’ relatives and the requirement to gather evidence at disaster scenes – i.e., keeping determined relatives away from disaster scenes.

I am trying to address this conflict from a risk communication perspective and would be grateful for any guidance which may assist.

peter responds:

Anyone who watches cop shows on television is familiar with the problem. The relatives of the victims of crimes and natural disasters rush to the scene and get in the way. The authorities have to keep them from interfering with rescue and recovery efforts, from contaminating evidence, and from getting hurt themselves.

It is of course natural for the relatives to want to get to the scene, see for themselves, look after the bodies and possessions of their loved ones, and begin the process of grieving. It is also natural for those in charge of an already chaotic situation to want to get the relatives out of the way with a minimum of additional effort – especially if they’re feeling anxious and stressed themselves. Maybe they can spare a minute or two to explain why it’s so important for the relatives to leave them alone. No more than that.

But brief, one-sided “explanations” are not what we need most when we’re coming to grips with a devastating and totally unexpected loss.

So the situation is a setup for conflict. People in these sorts of situations aren’t at their best. That’s obviously true of the relatives, but it may also be true of the authorities, if coping with catastrophe isn’t their daily fare. Under stress we often regress; we become more childish, self-centered, demanding, and unreasonable. We look for a scapegoat. And so the people in charge may project their feelings of discomfort and inadequacy into anger at the relatives. And the relatives may project their grief into anger at the people in charge. The relatives are “interfering” with essential work. The authorities are “insensitive” and “mismanaging” the crisis.

Many hospitals, airlines, morgues, police and fire departments, and similar organizations have crisis management protocols that include a brief section on communicating with the relatives of victims. Some are available online. I don’t know if there has been a systematic review of the recommendations contained in these protocols. (I didn’t find one.) If not, that would be a valuable project.

The protocols cover a wide range of communication issues: how to handle the initial notification; how to conduct interviews aimed at eliciting necessary information; how to arrange for crisis counseling when needed (and how to tell if it’s needed); how to manage the formal identification of bodies; etc. I’m sure some of the protocols must have something to say about how to keep the relatives out of the way – and maybe even how to do so empathically.

Here are my thoughts on that last task:

Bottom line: The best way to keep the relatives of victims from interfering with your efforts to manage the crisis and collect the evidence is to try to respond empathically to their needs. Some needs you can’t meet (and you should try to say so empathically). Meet the ones you can – especially their need for another human being who talks to them and listens to them.

comment:

I’m a big fan of your site. I’ve been combing through it for a few years now, ever since I saw your name in Freakonomics.

I was wondering what advice you would give the Catholic Church for the current pedophilia scandal. In recent news articles, it sounds like the Church is taking the “hazard” approach, logically stressing that such abuse happens everywhere, and priests are no more likely to commit such acts than the average person.

The media have of course focused on the “outrage” portion – the outrage that someone as trustworthy as a priest would commit such acts, and then the organization covers it up quietly.

I think this would be a case where inviting some people “inside” would be a way to show they are opening up (like when you suggest chemical companies can appoint oversight boards from the local community to get people involved).

Your thoughts?

peter responds:

There’s too much to say about the communication issues raised by the Catholic Church’s pedophilia scandal to encapsulate fully in a Guestbook response. I will focus on three aspects: contrition, dilemma-sharing, and accountability.

But first, let me echo what you said about hazard versus outrage. Child molesting is a huge outrage, among the hugest. When the molester is a priest, an agent of God entrusted with children’s moral and spiritual wellbeing, the outrage is exacerbated. When the Church has turned a blind eye for decades, the outrage is exacerbated further.

And when Church officials emphasize that other institutions also have a child molesting problem, and complain that it is unfair to single out the Catholic Church for criticism, they are, as you say, unwisely focusing on the hazard instead of the outrage. “Everybody does it” is never an effective defense; it’s not a defense a priest would accept from a misbehaving parishioner or child. But that’s not the point. The point is that “defense” isn’t what you do when people are rightly outraged at you. You apologize.

Contrition

The most obvious and most important failure of the Church’s communication has been the insufficiency of its expressions of contrition. The Church can’t protect its moral authority by understating its sins. It must regain its moral authority by acknowledging them wholeheartedly.

The irony here is that no institution in Western society understands the dynamics of contrition and forgiveness better than the Catholic Church. Both Catholic doctrine and secular experience make forgiveness conditional on the following steps:

1. Admit you did it.
2. (In the secular process, shut up while we berate you – a step not required by God but absolutely essential to secular forgiveness.)
3. Apologize – expressing regret, sympathy for the victims, and moral responsibility/fault.
4. Explain what failings in your organization let it happen.
5. Make it right – in particular, compensate your victims and improve your organization so it’s less likely to happen in the future.
6. Do a penance.

I discussed this process in more detail in a 2001 column entitled “Saying You’re Sorry.” The only step I’ve added since then is explaining what failings in your organization let it happen. That step has its own column (also written in 2001), “The Stupidity Defense.” Note that the stupidity defense is grounded in my conviction that organizational misbehavior is usually more stupid than evil. Acknowledging you were stupid is a good way to keep people from unfairly assuming you were evil. But covering up pedophilia to preserve the Church’s reputation really was evil – even if it was stupid too, since the cover-ups were bound to emerge eventually and lead to the crisis the Church now faces. To be forgiven, the Church will need to own up to more than stupidity.

These six steps on the road to forgiveness must be taken in the right sequence. You get no credit for apologizing until you have admitted precisely what it is you’re apologizing for, and until the rest of us have had a chance to yell at you for what you did. And until you have apologized sufficiently, explaining what went wrong sounds like you’re making excuses, and compensating your victims feels like you’re trying to bribe them.

It’s fairly clear that the Church has yet to go through some of these steps with regard to pedophilia. Obviously, it took the Church far too long to own up to what was going on. But it has done so now, at least to some extent. And it has certainly been berated, and continues to be berated. Where the Church is most obviously falling down is in Steps (c) and (d). It hasn’t been nearly apologetic enough (c), and it hasn’t been candid about the organizational failures that contributed to the problem (d). As a result, it gets very little credit for its efforts to make it right (e) – to compensate the victims of priestly pedophilia and to implement changes that will make future abuses less likely to happen and less likely to get covered up.

Instead of focusing on its own sins, the Church has taken offense at its critics. A high-ranking Vatican official compared the widespread criticism to anti-Semitism. Many other Church leaders attributed it to anti-Catholic sentiment. And the Pope himself lashed out against the news media, especially The New York Times, asserting that the Church would not be “intimidated by the petty gossip of dominant opinion.”

This is exactly the opposite of contrition.

Undoubtedly, the outrage of Church officials is getting in the way of their ability to express remorse, perhaps even their ability to feel it. It is also getting in the way of our ability to hear the Church’s remorse, even when it is expressed (as in the Pope’s letter to the Irish clergy). We hear mostly the Church’s defensiveness, which inevitably increases our outrage.

Of course the outrage of the Church at its critics has some foundation. It is true that many good priests have been damaged by a broad-brush indictment; it is true that in some cases specific allegations have been bandied about without convincing evidence; it is true that other institutions that work extensively with children face similar problems. There are valid points to be made in the Church’s defense. But defense and counterattack have interfered with what should be Church officials’ top communication priority: contrition.

Importantly, what most requires expressions of contrition isn’t the pedophilia itself. Child abuse is horrifying, but that sin is individual, not institutional. The Church’s sin is choosing not to see – and when forced to see, choosing not to tell.

Far too often, Church officials averted their eyes from instances of priestly pedophilia. Far too often they ignored, or ridiculed, or paid off those who dared to come forward with accusations. Far too often they took no action, or inadequate action, to prevent recurrences. The sin of willful obliviousness has been substantially corrected. The Church is no longer oblivious to child abuse; if anything it is now hyper-aware. But that sin needs to be acknowledged more often, more dramatically, and more contritely than has been the case so far.

And even when the Church has taken appropriate action to protect children from priests it had reason to know were pedophiles, it has often done so secretly. This seems to have had more to do with protecting the reputation of the Church itself than with trying to save the reputations, vocations, and souls of the offenders. Its preoccupation with secrecy too often kept the Church from doing what it should have done to aid the recovery of victims and to identify additional victims so that they, too, could be helped. I believe it is as important for the Church to acknowledge, atone for, and correct its self-protective secrecy as its obliviousness.

We can feel some empathy for senior Church officials. Their reluctance to express contrition for their obliviousness and secrecy regarding pedophilia is fueled by more than outrage at their critics. I would guess it is also fueled by self-loathing. Some Catholic leaders must find it unbearable to contemplate what they did and failed to do, the extraordinary damage their obliviousness and secrecy caused to thousands of children … and to the Church itself. Facing up to their role in perpetuating the horror must be agonizing. But it must be done.

And it must be done again and again. Even for offenses far smaller than these, it is rarely enough to apologize once or twice – and never useful to respond to continuing criticism by pointing out that you already apologized. I am certain that many Church leaders feel they have abased and humiliated themselves sufficiently; I am certain that some believe the apologizing has gone on too long already. But “let’s put that behind us” should always be the victim’s line, never the perpetrator’s. It is not for the Church to decide when no further expressions of contrition are required.

The Church will also need to punish those who contributed most egregiously to the obliviousness and the secrecy. Punishing individual wrongdoers is less important to forgiveness than acknowledging and correcting institutional failings; the core problem here is with the barrel, not a few apples. Nonetheless, punishment is part of “making it right”; it is also part of making institutional contrition credible. I think the Church has made substantial progress in taking action against pedophiliac priests, and I think it has largely stopped trying to keep the problem secret. Where the Church has made the least progress is in punishing the bishops and archbishops who sustained the willful obliviousness and the cover-ups. The Church’s pedophiles are being dealt with a lot more aggressively today than those in the Church who allowed them to flourish. I think the scandal will not subside until some of these higher-ups are punished.

Is the Pope himself one of the higher-ups who must be punished? I don’t know. I have read the allegations that in previous Church positions he was at least complicit in a pattern of cover-up. Given how widespread the pattern was, and how powerful Cardinal Ratzinger (now Pope Benedict XVI) was, I assume he was part of the pattern. How could it be otherwise? I have also read that the Pope (when he was Cardinal Ratzinger) was instrumental in rebalancing due process for accused priests against protection of children, reversing centuries of Church law and decreeing that insufficient evidence was no longer sufficient justification for doing nothing.

My impression is that the Pope was part of the obliviousness and part of the cover-up ... but saw the light more quickly and took action more responsibly than many in the Church. As the head of the Church he must now express not just the Church’s institutional contrition, but also his own personal contrition. He needs to say more about his role in the scandal over the decades, and how he sees that role today. I am not a Roman Catholic, but I am told that public papal expressions of contrition – and of sinfulness – are neither unprecedented nor contrary to doctrine.

Now would be an excellent time for Pope Benedict to acknowledge publicly his sins and the sins of his Church with regard to ignoring and covering up priestly pedophilia, and to begin working his way through the steps Catholic doctrine says are required for forgiveness. I don’t know what penance is most appropriate for the Pope, or for other Church fathers who tolerated the intolerable for decades. But I am sure the Pope will be able to think of some appropriate penances, and so will the public.

On April 15, just as I was completing this response, the Pope referred publicly (though obliquely) to the pedophilia scandal for the first time since his March 20 letter to the Irish clergy. He talked about penance:

I must say, we Christians, even in recent times, have often avoided the word “repent,” which seemed too tough. But now, under attack from the world which talks to us of our sins, we can see that being able to do penance is a grace and we see how necessary it is to do penance and thus recognize what is wrong in our lives.

The Pope saying that Christians should recognize the value of doing penance isn’t the same as the Pope publicly doing penance for the Church’s tolerance of pedophilia. And doing penance for this sin would be premature. Penance is the last step in forgiveness, and the Church is still stuck at the third step; it hasn’t apologized enough yet. The Pope’s April 15 statement is nonetheless a good sign.

Dilemma-sharing

Eventually, the Church needs to make more use of dilemma-sharing.

I say “eventually” because in situations like this contrition must precede dilemma-sharing. An institution that hasn’t sufficiently acknowledged that it often ignored the problem or covered up the problem is disqualified from explaining that it’s a tough problem to solve. Dilemma-sharing without contrition just looks like another way of making excuses.

But contrition without dilemma-sharing is sterile. Part of the forgiveness process is to “make it right.” It’s not enough to regret the problem; the Church needs to make a determined effort to solve the problem. It dare not leave the misimpression that priests will stop abusing children as soon as the Church owns up to its history of obliviousness and cover-up. There are more scandalous pedophilia revelations to come – both of past horrors and of new horrors. The Church should say so, often – not matter-of-factly but with anticipatory anguish and sorrow. Any implication that the scandal is over will backfire badly when new chapters emerge.

The essence of dilemma-sharing is this: The Church needs to help us see that there are no easy answers.

In particular, Church leaders need to talk about the awful tradeoffs between protecting children from a priest against whom an accusation has been made and protecting priests from false accusations. Complicating this conflict is the reality that children who have been molested at home do sometimes accuse someone else; taking such accusations at face value abandons both the priest and the child. False accusations can result under other circumstances as well – such as when a malicious adult persuades a malleable child to invent a story or embroider an essentially innocent encounter.

I don’t know what percentage of pedophilia charges against priests are false. Surely there are more genuine cases that go undetected than false accusations. Nonetheless, the dilemma posed by the risk of false accusations is real.

In each individual case, proof of child abuse is extremely difficult to come by. One solution is to wait until a pattern has emerged. But waiting for a pattern means allowing a suspected priest to keep doing what he is doing – an incredible injustice to the second or third child whose complaint proves the pattern. The alternative is to expose, or punish, or at least isolate a suspected priest against whom the evidence is scanty. That too is an injustice.

For a long time, the Church has “solved” this dilemma with secret investigations. It looked into pedophilia accusations quietly, informally, in a way that protected the suspected priest’s reputation and due process rights. The fatal flaw here was that the Church was also attending to its own reputational concerns (not to mention legal and financial concerns). If the Church could convince itself that the evidence was insufficient, it could make the problem go away … at least until another victim came forward. And so a legitimate worry about false accusations slid easily into a pattern of cover-up.

The Church could wash its hands of the responsibility for policing priestly misconduct, and simply refer all complaints to the cops. This is what many critics have recommended, but it has its own drawbacks. The legal system is far more strictly bound by the rules of evidence than a Church investigation; a pattern sufficient to persuade a bishop to remove a priest might still be insufficient to persuade a district attorney to file charges. (And should a priest acquitted because of insufficient evidence continue in his pastoral duties?) Moreover, many children and parents who can find the courage to tell one priest what another priest has done would simply refuse to come forward to the police and the courts.

As it shares this dilemma, the Church needs to acknowledge that thoroughly justified outrage at pedophiliac priests makes us all understandably uninterested in fairness to suspected offenders. But no good would come if the Church were to lurch from cover-up to witch hunt. The Church has been there before too, and Church leaders should say so.

Another possible approach is simply to keep priests away from children as much as possible. This is the solution taken by many schools and daycare centers today. Kindergartners no longer get to sit in the teacher’s lap, and a child who seems troubled is unlikely to be invited into the cloakroom for a private conversation. (There is now a camera in the cloakroom.) We pay the price of reduced innocent intimacy rather than risk the horror of child abuse … and, for teachers and schools, the horror of a false allegation of child abuse. Think back over some of the hallowed Hollywood movies of priests helping kids – and try rewriting the scripts without touching or privacy.

I don’t know the best way for the Church to cope with the child abuse issue. I’m not sure the experts know – and I’m not an expert. I do know that the Church has failed to make clear that the answers aren’t obvious. And I know that it cannot do so until it has first apologized more fully for its longtime failure to see the problem, own up to it publicly, and start looking for answers.

For the foreseeable future, every dilemma-sharing effort will need to be accompanied by a reminder of the Church’s contrition: “Notwithstanding the genuine dilemma of what to do with unprovable charges of pedophilia, it is clear that even when the evidence was substantial, the Church has too often ignored or even covered up that evidence instead of acting to protect the children in our care. Now that we are addressing the problem at last, we are coming to realize that there are no easy answers. But this does not in any way excuse our failure for so long even to look for the answers.”

Although I am not an expert in pedophilia, I would bet my mortgage that the priesthood attracts more pedophiles than, say, plumbing – for two reasons:

• The obvious reason – priests get to have a lot of close, intimate, emotionally charged contact with children.
• The less obvious reason – some people who are struggling with unacceptable sexual urges seek out churches (and priestly celibacy) in an effort to control themselves and avoid wrongdoing. Sometimes that effort succeeds, but sometimes it fails.

Unless it has persuasive evidence that this is false, the Church needs to acknowledge that it is true, or probably true, and needs to think through publicly its implications. This is yet another dilemma that deserves public discussion.

Accountability

You said it perfectly when you posed the question: The Church should fashion a role for outsiders – and especially for critics – in its response to pedophilia ... just as the oil and chemical industries have had to fashion a role for critics in their response to pollution.

Insularity is one of the most vivid characteristics of the Catholic Church. It is clearly both a strength and a weakness. But in coping with pedophiliac priests, the Church’s insularity has been a weakness. I doubt there are many sex abuse experts in high places in the Vatican hierarchy. Or many women. In an April 12 cover article in Newsweek, Lisa Miller writes: “The cause of the Catholic clergy's sex-abuse scandal is no mystery: insular groups of men often do bad things.”

Remember, the Church’s institutional sins are obliviousness and cover-up – not pedophilia itself. A secretive, self-protective, inward-looking, unworldly culture (that is, an insular culture) is especially vulnerable to obliviousness and cover-up.

Even if the Church’s inwardness and isolation were not part of the problem, accountability to critics outside the Church – or at least outside the clergy – should certainly be part of the solution. When I try to convince companies to establish Community Advisory Panels (CAPs) in environmental controversies, I usually stress the importance of the CAP as a source of credibility. “When outsiders are allowed inside,” I tell my corporate clients, “they get to see how difficult your environmental problems are, how hard you are working to try to solve those problems, how much money you’re spending, how much progress you’re making. These are not things you can easily explain to people who are disposed to mistrust you. When they see these things for themselves, they slowly come to trust you more.”

Though I typically make the point less emphatically, the Community Advisory Panel is also a source of wisdom. Outsiders not only learn about all the good things you’re doing (and discover some bad things they hadn’t known about before). They also point out some additional good things you ought to do. Some of their advice isn’t feasible, of course – and some that might be feasible you’re simply unwilling to take. But every CAP I have ever watched operate managed to come up with suggestions the company was able and willing to implement. The result wasn’t just improved corporate credibility. It was also improved environmental protection.

I am confident that the Catholic Church has much to learn from its critics about how best to address its twin problems: pedophilia itself, and the culture that too often ignored, tolerated, and covered up pedophilia.

Letting outsiders in isn’t just a way of showing the world that the Church is changing. It is a way of changing the Church.

comment:

Thank you very much for the very interesting presentation at the technology risk forum. It was very nice to understand the underlying logic behind the “Risk = Hazard + Outrage” equation.

As part of my internalizing it, I can clearly relate the “hazard” part to a loss distribution which would capture the expected value variable, or the probability times impact. On top of that we have the “outrage” variable, which as we have seen in the recent past to be at least just as important.

While you mostly spoke of “outrage” in terms of the public at large, a similar effect happens in smaller specialized groups such as financial market makers. In fact, it is a big part of the explanation of how contagion spreads – your fellow dealers stop dealing with you because your reputation has taken a knock (rightly or wrongly), counterparties rush to terminate trades and unwind positions, investors want out, and suddenly you have a crisis which can first fail you and then spread to other market participants. And all of this happens outside the “hazard” variable.

The points you made about communication are just as valid even in this kind of a situation because even though the audience is slightly different, the same dynamics apply as in the “town hall meeting” example you gave. When every other weekend was bringing news of a large institution’s collapse, those who survived had as much to attribute to their communications as they had to their economic fundamentals. Communication is an important part of risk management in that sense.

Now my question is – if this be so, why is it that all financial risk management and regulation is focused solely on expected values, a certain percentile (as in VaR) of loss? Why isn’t some of the “outrage” concept being baked into the Basel Framework, or the regulatory responses we have seen since the beginning of the recent crisis? (In fact Basel II goes to the extent of defining operational risk as specifically excluding reputation and strategic risks.) We also do not see any references to the “outrage” factor (regardless of the verbiage used) in the recent papers on liquidity risk and stress tests as put out by the BCBS, or the FSA in the U.K. In the U.S., the supervisory stress tests done in the early part of 2009 were solely based on math and “shock factors” applied to economic variables, and completely ignored the impact of the “outrage” (though admittedly the manifestation of outrage among market participants is different from that of people in a town hall meeting).

Why would someone like you not try to propose including this in a rational and consistent way when these regulatory bodies come out with proposals, consultation drafts and discussion papers? Surely this is worthy of being addressed in the interest of our financial system and the economy at large. Have you ever made proposals to the Fed, the FSA, BIS or others? If so, what has their response been?

peter responds:

I thoroughly agree with you that both individual responses to a trader’s reputation and overall market responses to an economic situation are hugely influenced by outrage.

And as far as I can tell, you’re right that the dynamics are pretty much the same as in my “angry town hall meeting” paradigm. Outrage affects hazard perception, and hazard perception affects precautionary behavior – and in economic systems (micro- or macro-), precautionary behavior affects market conditions and thus the cycle (or spiral!) continues.

This isn’t exactly news to economic theorists. Behavioral economics isn’t a hot new iconoclastic subfield anymore. It’s a fundamental part of economic thinking. Theories grounded in the “rational man/woman” assumption have been, if not discredited, at least identified as accounting for only a part of economic behavior. Princeton psychologist Daniel Kahneman won the Nobel Prize in economics in 2002 for his work exploring the ways various heuristics lead economic decision-makers to “wrong” conclusions. Kahneman’s work since 2002 has made emotion more and more central.

I’m not nearly knowledgeable enough about market regulation to have strong opinions about how outrage should be integrated into regulatory standards. I barely know what the Basel Framework is. The closest I have come previously to writing about the questions you’re raising was a March 2009 Guestbook entry on “Credit default swaps, financial meltdown, and risk communication.”

But it does seem to me that if a government is developing “stress tests” aimed at assessing the likelihood of a repeat of the market meltdown of 2008–2009, it would want to pay some attention to outrage.

Certainly I push my industrial clients to think about the ways in which their behavior could trigger outrage widespread enough to damage their reputations and depress the price of their shares. The example I gave at the presentation you attended was an oil company that aroused human rights outrage over its investments in an oppressive developing country, leading the socially responsible investment community to dump (and badmouth) its stock. Either the company should have avoided such investments or it should have taken seriously the task of mitigating the outrage those investments were likely to trigger.

It’s not too great a jump to wonder what sort of behavior by a banking company might lead to a similar loss of reputation and share price … and what sort of behavior by the entire banking field might trigger another loss of faith in the market itself.

Has anyone asked me to consult on these issues? No, never. I have had banking clients over the years, but always on a peripheral issue: how to prepare for a possible pandemic, for example, or how to address outrage at the bank’s investment in a controversial industrial facility. It’s worth noting that you heard me speak at a meeting of banking business continuity and information security professionals. The people who are worried about restoring the reputation of your industry and lessening the public’s hatred haven’t called.

Like the Fed, the FSA, BIS and others, I doubt they know there’s a field there, a field I call “outrage management.”

comment:

Thanks for your presentation on risk communication. It was insightful, and it was great to hear your approach firsthand.

I was wondering if you have any thoughts about whether it is ever a good strategy for outrage management to rile up a group instead of the responses you mentioned?

The reason I ask is the stratagem from Sun Tzu’s Art of War: “If your enemy is angry, irritate him.” It seems clear this would at best be a risky strategy for a public company, although maybe there are examples where it might be effective.

peter responds:

Of course increasing outrage is the core task of what I call “precaution advocacy,” when people are insufficiently concerned about a serious risk.

But that’s not what you’re asking about. You’re wondering whether I think it ever makes sense to “irritate your enemy” (as Sun Tzu suggests), presumably because an irritated enemy makes more mistakes.

While this may make sense in war, I don’t think it makes sense in outrage management, for two reasons:

• The goal in outrage management isn’t to provoke opponents into acting unwisely so you can defeat them. Rather, it is to ameliorate their outrage enough that they can accept a win-win or a compromise, an outcome that works for you both … and may even view such an outcome as tantamount to defeating you. The assumption is that defeating opponents is extremely difficult and extremely costly; “defeated” opponents too often rise like a phoenix from the ashes, or their defeat inspires an ever-larger cohort of new opponents. It’s better to aim for a settlement.
• When opponents are unwilling to settle and insist on continuing the battle instead, the fallback in outrage management is to allow them to marginalize themselves. This is best accomplished by persuading attentive stakeholders (potential opponents) that you have made major concessions to your opponents, that the serious issues have largely been resolved, and that your opponents are unreasonably raising ever-more-extreme and ever-more-peripheral new points of contention, Acting provocatively toward your opponents – goading them into unreasonableness – is likely to leave these attentive stakeholders hostile to you for being provocative instead of hostile to your opponents for being unreasonable.

Bottom line: A conciliatory approach via outrage management aims at reconciling opponents who are reconcilable, and at isolating and marginalizing opponents who are not. This is incompatible with intentionally irritating opponents into further outrage.

I have sometimes advised clients to try to make sure the rudest, most obstreperous opponent comes to the meeting, so calmer, more courteous opponents (and attentive stakeholders) can watch, think to themselves “I’m not like Susan,” and become more open to reconciliation than they might otherwise have been. But that is fruitful only if the client is visibly treating Susan with respect. There’s a seesaw at work here. If the company is respectful toward Susan, others are free to roll their eyes and complain that she’s hijacking the meeting. But if the company becomes angry or contemptuous, or if the company does things to goad Susan into greater irritation and unreasonableness, then everybody else’s outrage will be directed at the company instead of at Susan.

comment:

I read your October 2009 comment on “Mandatory vaccination for health care workers” with great interest.

Even though I was once militant that HCWs should be mandated to receive influenza vaccinate, I have been influenced by the ethical arguments put forward by George Annas and others.

However, the one problem I have with your response is that in my experience there are about 30% of HCWs that have a belief system about influenza vaccine that causes them to reject it. You can get 40% without much trouble and maybe another 30% with mandatory declination, but then you have the group that won’t get vaccinated no matter what you do. I worked in a hospital for 20 years and can tell you I did everything I could to educate this recalcitrant group to no avail, including one-on-one education. As you know, hundreds of studies have been published about methods to increase uptake among HCWs. I have come to the conclusion that mandating vaccination is the only thing that will work for this last group of holdouts.

That said, I don’t like the “get vaccinated or wear a mask” approach either. To me it is disingenuous because if hospitals were really worried about HCWs spreading influenza, they would have had HCWs wearing masks throughout the H1N1 pandemic before either seasonal or H1N1 vaccine was available, but they didn’t. It seems punitive and when pushed, most hospitals will admit they use it as a “stick” to get HCWs vaccinated.

The issue of mandating vaccination has some real legs now. I think it will be helpful if you and others can speak to this issue more.

One thing I find interesting is that most hospitals have mandated immunity to measles, mumps, and rubella (and more recently varicella) for a long time. I think the difference is that most HCWs were vaccinated or had these diseases when they were children and didn’t have to make the vaccination decision for themselves (every year in the case of influenza). HCWs have accepted this requirement, which is usually in place for students entering health care fields.

The psychology is interesting.

peter responds:

I appreciate your point that if hospitals were really all that worried about HCWs spreading flu to patients, they would have required masks before there was an available vaccine. So making HCWs choose between vaccine and mask is arguably hypocritical.

Of course it’s possible for a hospital to argue something like this:

Masks are uncomfortable, inconvenient, potentially hazardous (as fomites), and expensive. Making HCWs wear masks would protect some patients from influenza, but the collateral damage is too high. Vaccines have much less collateral damage, so we push vaccination on HCWs in situations where we don’t normally push masks. But if an individual HCW is resistant to vaccination for whatever reason, then it makes sense to offer masks as a less desirable option, an accommodation to that HCW’s desire not to be vaccinated. Requiring masks in situations where there’s no vaccine has too big a downside; requiring vaccination for HCWs who fear the vaccine has a different but still unacceptably big downside. But requiring the choice – with vaccine preferred but masks available for conscientious objectors to vaccination – is reasonable.

I’m not sure I buy that argument, but it’s not obviously foolish or hypocritical.

The bigger question for me is the rationale for requiring HCWs to get vaccinated against the flu.

If it’s for the HCW himself/herself, then it’s unconscionable coercion. Making employees do things for their own good is pretty obviously wrong. We don’t (yet) make other people get vaccinated against flu. Why coerce HCWs for their own good more than we coerce people in other jobs? When officials tell HCWs “this is for your own good,” I think they’re undermining their own case.

If it’s for the hospital, aimed at reducing absenteeism and thus the cost of health care, then one wants to see the data. How much is actually saved? Are there bigger savings available with less collateral damage that the hospital isn’t pursuing? Is the hospital including morale issues in its cost-benefit calculation? Does the benefit justify the coercion? Moreover, in a unionized setting battles between what’s good for the employer and what’s good for the employee are the classical venue of labor-management negotiation. It would save the hospital money to pay HCWs less, too, but that’s not enough reason to countenance unilateral pay cuts. If vaccination is for the sake of the hospital, it ought to be a contract negotiation issue.

If it’s for the patient, the rationale for mandatory vaccination is stronger. Hospitals are entitled to regulate employee behavior for the benefit of patients. But here we really need data. My impression is that there are pretty good data that HCW flu vaccination reduces hospital costs, but not very good data that HCW flu vaccination reduces hospital-acquired flu in patients. Patient health is the strongest rationale for coercing HCWs, but only if the evidence is strong. Is it? And as you pointed out, if HCWs really give lots of patients the flu, you’d expect different hospital mask policies too. So officials end up trying to argue that the impact on patients is enough to justify making HCWs get vaccinated, but not enough to justify masking them when there’s no vaccine (or when the vaccine is a bad match). That’s a pretty narrow window. Similarly, why aren’t hospitals requiring visitors to prove that they have been vaccinated? Unvaccinated family hang around the patient all day with impunity … but the orderly has to get vaccinated?

Sometimes my clients get into fights with their employees (or other stakeholders) that started out over a real substantive issue (usually a fairly small one) … and morphed into something that’s really more about power and ego. I wonder how much of that is playing out in the HCW vaccination battle. “Whose hospital is it anyway?” “How dare someone without an M.D. question my judgment that the vaccine is safe?” “If we let them win this fight, what other policies will they decide to flout?” Of course the same could be true on the other side of the battle lines. When HCWs insist on their right to go unvaccinated, they may be bringing to that fight animus that comes from other labor-management issues, from pay to parking.

kathleen answers:

You hit the nail on the head with your question about whether vaccinating HCWs is a patient safety issue. A few years ago I would have said yes, absolutely, because there were a couple of studies that seemed to support this view (Carman and Potter). However, now I’m not so sure the data are as strong as I might wish for, particularly in settings other than nursing homes. Thomas Jefferson who has done Cochrane reviews on this subject doesn’t seem to think the data are there. However, he appears to be a polarizing figure and seems to be a zealot about RCTs, so it’s hard to know what to make of some of his more public comments on this issue.

I think what most mandatory vaccination proponents think is this: Influenza vaccination will prevent influenza in most vaccinated HCWs, and infected HCWs have been associated with nosocomial transmission of influenza (that is, transmission in a health care setting). Hence vaccinating HCWs will make it less likely for patients to be infected by a HCW. There certainly are reports of nosocomial outbreaks of influenza caused by HCWs. But are there enough data to mandate vaccination? I don’t know.

This seems like an issue where for many people the train has left the station and it doesn’t really matter what the data are.

I’m surprised by how I’m starting to think about this issue. Ten years ago I thought there should be mandates. I still think mandates are the only thing that will bring vaccination rates to 90–100%, but now I wonder if the juice is worth the squeeze. I think most people are newer to thinking about this topic, still in the process of thinking through a lot of different issues.

As for your argument for mask versus vaccine, I’m not sure I would buy it either, but at least it’s honest.

comment:

I took your class on risk communication in April 2008, and have been wondering how the principles would be applied to a case facing the local community currently, an interesting scenario of high hazard and high outrage.

Fairbanks has recently become a non-attainment area for PM2.5 in the ambient air. Most attribute this to increased wood burning for home heating, which drastically increased after the 2008 oil price spike. The outrage in this case hasn’t primarily been from those outraged about the air pollution, but instead from a very vocal minority that has opposed proposed limits on wood burning, leading to some rather remarkable allegations that the EPA (through the state and local governments) is trying to limit our freedom, etc. People obviously feel squeezed on their heating bills, and don’t like what they perceive to be federal intrusion on their rights.

There is some outrage about the air quality problems amongst the general public, but it seems more diffuse. Only people that are right next to smoky problem areas get really steamed, but they don’t tend to be as vocal as the aforementioned outraged wood burners.

I guess my main point is that the outrage has been more significant so far against the attempts to solve the problem (bad air quality) than the problem itself.

The local government’s attempts to pass nuisance rules to clamp down on the worst offenders thus far seem to have increased the outrage, while most people in the community seem to like or at least feel okay with the proposed new rules. The current approach developing seems to be an education campaign, which doesn’t seem like it will reach the outraged constituents.

Any quick advice that you could offer in terms of analysis or proposed directions?

peter responds:

Outrage about regulation of wood burning has a long history. I first encountered it personally in the 1970s (I think it was), when regulators tried to shut down a famous Seattle restaurant that featured wood-smoked salmon for violating air pollution regulations. I’ve encountered the same issue several times since, especially in areas where wood burning for heat was widely practiced and strongly cherished, not just for economic reasons but also as emblematic of a close-to-the-land eco-friendly lifestyle. The idea that getting your heat from a utility company might actually be less damaging to air quality than chopping and burning your own energy supply simply didn’t compute.

Regulations against leaf burning used to raise similar protests for similar reasons. I grew up loving the autumn smell of everybody’s burning leaf piles. My middle-aged children have rarely smelled that smell and don’t miss it.

Conceptually, this is not about high-hazard, high-outrage scenarios. A true high-hazard, high-outrage situation is a crisis. People are both endangered and upset by whatever’s happening, and the key risk communication task is to help them bear their outrage and choose wise rather than unwise emergency response actions. A factory explosion, for example, simultaneously endangers and frightens people.

In the case you’re describing, the high hazard is the particulate emissions from incomplete combustion of wood. The high outrage isn’t about the effects of wood burning at all; it’s about interfering with people’s desire to burn wood. As you rightly point out, what’s arousing people’s outrage is the precaution, not the risk.

This is quite common. When industrial employees won’t wear their respirators, eye gear, and other PPE, for example, the problem could be insufficient outrage about workplace risks. But it’s likelier to be excessive outrage about the PPE itself; the respirators itch, the glasses look geeky, etc. I was interviewed on this problem in a 2005 article entitled “Getting Workers to Wear PPE: Communication Is Key.” I argued that precaution advocacy aimed at increasing outrage about workplace threats to the lungs and eyes would be less effective than outrage management aimed at reducing outrage about the respirators and the glasses.

The usual outrage management strategies are your best bet: sharing control with your critics, acknowledging their sound arguments, giving them credit for changes you make in response to their concerns, etc. (See especially “Reducing Outrage: Six Principal Strategies.” )

Educating people about the hazards of PM2.5 pretty obviously won’t help assuage the outrage of those who hate losing a cherished piece of their lifestyle (and a good way to save on heating costs) at the hands of an intrusive government bureaucracy.

Working to increase the overall community’s outrage about PM2.5 may be worth doing for a variety of reasons, especially to increase their support for the proposed regulation and their compliance after it’s in force. But you have to decide whether you think mobilizing the majority’s outrage at the minority’s intransigence is a desirable path forward. It has powerful pros and powerful cons, but I’d certainly think twice before trying to split the community. The collateral costs are high even if you succeed. And you might not succeed; instead, those who are okay with regulating PM2.5 might rally to the support of the old-timers who want to keep their home fires burning.

I suspect I’d look to ameliorate the opponents’ outrage rather than trying to make them the object of everyone else’s outrage. One thing I’m sure I’d do is to express regret at what’s being lost. You need to be visibly saddened that a venerated piece of Alaskan culture and a way to live inexpensively off the bounty of nature has turned into a threat to the ability of asthmatics and others to breathe free.

comment:

I am an internal auditor in Turkey. I want to prepare a report for my office on ERM (enterprise risk management).

First of all I want to measure risk perception of our office workers (managers, engineers, inspectors, doctors and the other lower level staff). I want to learn how they perceive the risk (and hazard). So I want to measure this perception by using survey methods.

I need an example survey form. Will you please help me?

peter responds:

It’s a little hard to advise you on survey methodology or to suggest possible survey designs and questions you could borrow from without knowing a lot more about what you’re studying and why.

The most generic way to study enterprise risks is to present people with a “risk matrix.” Typically, there are two dimensions, probability and consequence. (These are the two components of conventional definitions of risk.) Each dimension typically has four or five gradations.

Here’s a sample risk matrix borrowed from Wikipedia. The rows are probability, ranging from “certain” to “rare.” The columns are consequence, ranging from “negligible” to “catastrophic.”

Participants are asked either to assess a preexisting list of risks according to these two parameters, or to identify risks on their own and locate them in the matrix. The assessment can be done individually and the results tabulated, or groups can debate their views and try to reach a consensus assessment. Whichever procedure is used, it’s useful to keep track of the range of opinion. Don’t average away the views of outliers.

The outcome of a risk matrix exercise is information about which risks to the organization participants consider more or less likely and more or less damaging. This is often seen as a way of assessing enterprise risk itself. But I think it’s better seen as a way of getting a handle on perceived risk. Depending on the expertise of the participants, their perceptions may or may not tell you much about the actual probability and consequence of various risks to the organization. Mostly it tells you what your participants think.

There is considerable debate over the pros and cons of risk matrices as a way of identifying and assessing enterprise risks.

Certainly risk matrices are a handy way of finding out what your people think your organization’s biggest risks are (the very-high-consequence ones that aren’t vanishingly low probability, and the very-high-probability ones that aren’t vanishing low magnitude).

But there’s a lot of evidence that people don’t necessarily interpret adjectives and similar verbal risk descriptions in the same way, even if you try to help them with definitions. Picking labels for your categories is also a far-from-obvious set of decisions. Consider the Wikipedia risk matrix, for example. Is something that’s expected to happen about once every decade “likely,” “possible,” or “unlikely”? And what are we to do with risks whose consequence strikes us as worse than “marginal” but not really “critical”?

Most people also respond differently to risk adjectives than they would to actual quantitative estimates of probability and consequence. When going from risk numbers to risk words, we tend to dichotomize. If the number strikes us as fairly small, we pick a less serious verbal risk descriptor than the number justifies; if it strikes us as fairly big, we pick a more serious verbal descriptor than the number justifies. People asked to translate from words back to numbers do the same thing: they go from not-very-alarming descriptors to very small numbers, and from alarming descriptor to very big numbers. Thus: “Only once a decade? That’s unlikely.” And later: “Unlikely? That’s like once a century.”

Risk matrices are extremely clear … often much clearer than they ought to be. In other words, they tend to “over-communicate” – leaving people with the impression that they know more than they know.

In a 2008 article in the journal Risk Analysis, Tony Cox makes a case that risk matrices can often provide misleading information leading to poor risk management decisions.

If you know my work, you know I have another objection to risk matrices: Neither axis of the risk matrix explicitly addresses what I call outrage – how upsetting the risk is. In enterprise risk management, stakeholder outrage is arguably a component of the magnitude dimension. A “minor” glitch in a safety system, for example, may have very little direct impact on the organization, but if it terrifies neighbors into wanting to shut you down, its indirect impact could be catastrophic. But there’s nothing in traditional risk matrices to push participants to consider this kind of impact. I’d be happier with a three-axis matrix that asked participants to assess separately the consequence, probability, and outrage potential of the risks they considered.

Many have proposed replacing the phrases in risk matrices with quantitative estimates. This has different problems. Unlike adjectives, numbers are precise (even if they're precise guesses) – but people often haven’t a clue what they mean. Words are vaguer and people disagree about what they mean (often without realizing it), but at least they think they know!

And if you’re going to use words, a matrix helps you think about two (or three) aspects of risk at the same time, which has got to be better than thinking sometimes about consequence, sometimes about probability, and sometimes about outrage without ever putting them together.

Of course risk matrices are only one of many ways of measuring enterprise risk and enterprise risk perception. I’m not at all confident they're the best one for your purposes (since I don’t really understand your purposes), but they might be a place to start.

comment:

I’ve long been an admirer of your body of work and I’m thrilled you’re considering contributing to my inaugural “Three Tough Q’s” feature on my blog, www.jamesjdonnelly.com. I may edit your responses there to the key concepts, but will certainly link back to your full responses here.

Q: Over your lifetime of study, contribution and counsel – any professional regrets?

peter responds:

My biggest regret by far is that I have not managed to create a cadre of people who do what I do.

On the one hand, “risk communication” is now a recognized field. (I had to laugh when I asked my airplane seatmate a year or so ago what field she was in, and she said “risk communication.”) But most of what goes under that name seems more like conventional public education or public relations to me. The principles of precaution advocacy, outrage management, and crisis communication that I have been trying to articulate for years are still very far from conventional practice. And I don’t really see a new generation of risk communication consultants emerging to push the rock further on up the hill.

Maybe this is my own biased, narcissistic perception of what “counts” as good risk communication. Maybe my approach is just too idiosyncratic (or just plain wrong). There are plenty of people coming out of universities with at least a little exposure to risk communication … and even to my approach to risk communication. (I sometimes meet young people at seminars or consultations who tell me they studied me in college and seem surprised to learn that I’m not long dead.) Still, when clients ask who else does what I do – because I’m busy the day they want me or too expensive or whatever – I don’t have a lot of names to give them.

Back in the 1970s, when I was briefly an “environmental communication” professor at the University of Michigan, I had graduate students – young people who spent a year or two focused on trying to absorb everything I could teach them. Then I moved to Rutgers University, where I taught almost entirely undergrads, who took at most a course or two on what I was coming to call risk communication. And then I moved to independent consulting, where my “students” were practitioners and I was lucky to get them for as much as two whole days.

There are thousands of people out there who have had a day or two of Sandman, but few if any who have had a week or a month or a year. Insofar as what I do is unique and valuable – which isn’t for me to judge – I have been unsuccessful in replicating it. My best “student” by far is my wife and colleague Jody Lanard. Jody has taught me as much as I have taught her. But she’s not much younger than I am.

I am doing what I can to try to remedy this defect. I’m trying to get as much of my approach as possible onto my website, www.psandman.com – and I’m negotiating with a university to host the website and sustain it when I’m not doing so any longer. I was about to launch a “master class” (asking participants to commit a week a year for at least two years) when the economy collapsed – and I plan to restart that effort when the economy recovers sufficiently. I’m also looking for commercial partners – ideally a worldwide public relations agency and a worldwide management consulting firm, both of which would undertake to let me train their key people (and would send a few champions to the master class).

If I had it all to do again, I would make “legacy” a higher priority earlier. Maybe I’d affiliate with a university and have graduate students again. Maybe I’d set up some kind of apprenticeship program. If I’d launched the master class a decade ago, it might have a couple of hundred alumni by now, and maybe they would be linked to each other in some kind of intranet.

Very few people are fortunate to change the world even a little. I have had more impact than I ever expected. And I know all impact is transient. Still, my biggest regret is not finding a way to keep my approach to risk communication going and developing after I’m not doing it anymore.

Is it possible that is already happening, and I just don’t get to see it? I hope so.

comment:

I’ve long been an admirer of your body of work and I’m thrilled you’re considering contributing to my inaugural “Three Tough Q’s” feature on my blog, www.jamesjdonnelly.com. I may edit your responses there to the key concepts, but will certainly link back to your full responses here.

Q: Can organizations mitigate public panic – and a current zeal for instant information – when hazard levels are unclear?

peter responds:

Let me start by saying that panic – real panic – is quite rare. Disaster management experts define panic as what happens when people are so terrified that they just can’t stop themselves from doing something they would otherwise know to be useless or harmful. People in horrible situations often feel panicky, but they almost always manage to control themselves and act sensibly. For more on how rare panic actually is, see my 2005 article with Jody Lanard, “Tsunami Risk Communication: Warnings and the Myth of Panic.”

In crisis situations and even in ordinary risk controversies, officials and the media often mistakenly think people are panicking, a phenomenon I have sometimes called “panic panic.” But what you’re calling “public panic” isn’t panic at all, I think. It’s rationally high levels of concern about a situation that may be genuinely dangerous. Organization shouldn’t want to “mitigate” that concern. They should want to guide it.

Let’s take a concrete example. In late April 2009, a novel influenza virus emerged in Mexico and California that looked likely to launch the first flu pandemic of the twenty-first century. It did in fact launch such a pandemic, but so far it has been a very mild one. But nobody knew that in the early months. So there was reason to be concerned – in two senses: Concern was justified by the real possibility of tough sledding ahead; and concern was useful because it motivated people take precautions, and to seek information that would help them determine which precautions to take.

Many people were insufficiently concerned and therefore not taking any precautions – so one job for pandemic risk communicators was to try to arouse this audience to greater concern. Some people were appropriately concerned and very anxious to take precautions – so another job for pandemic risk communicators was to try to help this audience figure out which precautions were feasible and useful and which were not. And a few people were excessively concerned (perhaps even feeling panicky) and inclined to take really inappropriate precautions – so a third job for pandemic risk communicators was to try to reassure this audience. Even for this third audience, the smallest of the three, the job wasn’t to “mitigate” their distress, at least not in the sense of trying to convince them to stop worrying; rather, the job was to help them bear their distress and put it into perspective, so they could make wiser choices about how best to protect themselves and their loved ones.

Uncertainty about the extent of the hazard was inevitable to the early months of the pandemic. We had only partial and uncertain information about how bad it was; and we didn’t have a clue how bad it might get. (We still don’t, really; we only know it hasn’t been too bad so far.)

Uncertainty about the extent of the hazard is intrinsic to most evolving risk situations, and especially in the early stages of a crisis … or a possible crisis.

The effect of uncertainty on what I call outrage – on people’s inclination to get upset – is binary. If outrage is low, uncertainty keeps it low: “I won’t worry till the experts are sure there’s really a problem.” If outrage is high, uncertainty makes it higher: “How dare they expose me to this contaminant when they know so little about its long-term health effects!” You can see this dual response to uncertainty in a lot of risk controversies, from global climate change to industrial air emissions.

For the people who are frightened or angry about a particular risk, then, uncertainty exacerbates their feelings – adding to what you call “a current zeal for instant information.” I think the solution is to provide as much information as you can as quickly as you can, while trying to highlight the most important and actionable information.

Obviously, that doesn’t mean providing more information than you have! Some of the most important information you give people in an uncertain situation is information about the uncertainty itself:

• How much you don’t know.
• What you are doing to learn more, and when you expect to have some answers.
• How much you won’t know for years, if ever – and why some questions are difficult or impossible to answer.
• How awful it is for everybody – for you and for your stakeholders – to have to endure so much uncertainty, and to have to make uncertain decisions knowing that they may turn out mistaken in the end.
• What sorts of decisions you are making in the face of your uncertainty, and what sorts of decisions you are advising your stakeholders to make.
• Your guiding principle for decision-making in the face of uncertainty: erring on the alarming side (“better safe than sorry”), but not in such an extreme way that over-preparedness or over-reaction will do as much damage as the hazard itself.
• Your other guiding principle for decision-making in the face of uncertainty: staying flexible, reconsidering prior decisions as you learn more and as the situation changes.
• Your advice to your stakeholders to adopt the same guiding principles: to err on the alarming side but within the bounds of reason, and to be ready to gear up or gear down as appropriate in response to new information.

Of course you should also tell people what you do know – and not just what you’re certain about. Acknowledging uncertainty is crucial to good outrage management and to good crisis communication, but acknowledging uncertainty doesn’t mean throwing your hands into the air and claiming total ignorance. Nor does it mean confining yourself to the few things you know for sure. When you brief top management, you say things like this: “We’re certain of A. We’re pretty sure about B. C would be a surprise, but not a shock. D is extremely unlikely. E is impossible. We haven’t a clue yet how likely F is. We’ll probably never know about G.” Be similarly discriminating about levels of uncertainty when you brief the public.

Bottom line: The best response to people’s zeal for instant information is instant information – especially information about uncertainty.

For more on how to talk to upset people in uncertain situations, see my 2004 column on “Acknowledging Uncertainty.”

comment:

I’ve long been an admirer of your body of work and I’m thrilled you’re considering contributing to my inaugural “Three Tough Q’s” feature on my blog, www.jamesjdonnelly.com. I may edit your responses there to the key concepts, but will certainly link back to your full responses here.

Q: Has the rise of online social networks changed outrage management (low hazard, high outrage), for better or worse?

peter responds:

Asking someone over 60 to comment on the impact of online social networks is pretty risky. Asking me is very risky. I’m trying, but I’m not really a netizen.

Still, it’s obvious even to me that an ever-larger segment of the population gets most of its information online from peers. If you want to tell people about your organization, if you want to know what others are telling people about your organization, and if you want to be part of the dialogue about your organization, you have to be on Facebook and Twitter and YouTube … and on whatever may be about to supplant Facebook and Twitter and YouTube as next week’s online social networks.

There is one less obvious impact on outrage management I think is pretty important. To make sense of it, I first have to outline a distinction that’s key in outrage management – the distinction between publics and stakeholders.

Publics are large groups of people peripherally interested in your activities. Collectively they can have a huge impact on your reputation and your operations, of course. But since they’re not about to come to a meeting or even read a newsletter, they have traditionally been reachable only via the mass media.

Stakeholders, on the other hand, do come to meetings and read newsletters; some stakeholders hold meetings and write newsletters. The traditional media matter much less in reaching stakeholders. Traditionally, stakeholder relations is retail, not wholesale; the best stakeholder relations is one-on-one.

Outraged people – people who are angry, frightened, skeptical, distrustful, or otherwise upset about something you’re doing – are stakeholders. So outrage management is more a kind of stakeholder relations than a kind of public relations. In fact, good outrage management often pays a public relations price. To address their stakeholders’ outrage, organizations need to do a lot of acknowledging and apologizing with regard to prior misbehaviors and current problems. Most of what they’re acknowledging and apologizing for is stuff their publics don’t know and don’t especially want to know, and stuff they’d really rather their publics didn’t find out.

Cluing in your publics to your organization’s defects is collateral damage when you’re responding to stakeholders’ outrage about those defects.

The distinction between publics and stakeholders has never been airtight, of course. Your hostile stakeholders do their own outreach to publics they think have the potential to join them as stakeholders; they’re revealing your defects to those publics whether you choose to do so or not. You may (rightly) see a meeting with critics as an opportunity for stakeholder relations and outrage management. But your critics (rightly) see it as an opportunity for recruiting via the media. And reporters who cover the meeting (rightly) see it as a story.

Still, the public/stakeholder distinction has been very important in my outrage management work.

Thanks to online social media, this distinction matters less and less. People who learn things about your organization via online social media are somewhere in the middle between publics and stakeholders. They’re not the passive, barely attentive audience of the mainstream media, glancing at a story about you before moving on to something else. Nor are they active, committed, and comparatively well-informed the way stakeholders are.

Two aspects of this new intermediate space strike me as especially relevant to outrage management:

• Online social media facilitate participation without requiring serious interest. It’s easy to add an offhand comment to somebody’s blog, to forward an interesting online article to groups of friends, to retweet something you saw on Twitter, etc. The mainstream media provide a much more passive environment, while traditional participation – going to a meeting, for example – requires much more effort. People who don’t really care very much about you, certainly not enough to attend a meeting or write a letter to the editor, may nonetheless add their two cents to a discussion about you in the social media or email the discussion to others. And that, of course, may trigger greater interest and more involvement down the road.
• Online social media put a premium on opinion rather than information. Or at least the ratio of opinion to information is much higher than in the mainstream media. And the information that is passed along hasn’t been vetted by reporters and editors who have a stake in avoiding inaccuracy. So the ratio of false and misleading information to sound information is also higher than in the mainstream media. Moreover, the role of information is different: A large percentage of participants in online social media are expressing a viewpoint, and selectively marshalling information to support that viewpoint. All this is done in a casual, expressive tone. Conventional definitions of credibility put a lot of stress on authority and expertise. But credibility in social media (I would bet) is grounded much more in emotional expressiveness, and in coming across as similar (in feelings and values) to the other participants in the dialogue.

I don’t fully understand the implications of these changes for outrage management, but I sense that they are important and going to be more so. Many organizations have learned to interact with their outraged stakeholders in ways that are participatory, two-sided, responsive, human, emotionally expressive, and empathic. But when targeting less involved publics, those same organizations tend to churn out the usual one-way, one-sided, just-the-facts “public education.” Now they’ll need to learn to talk to relatively uninvolved people via online social media in ways more like the ways they have learned to talk to stakeholders.

comment:

Your January 17th update on the H1N1 pandemic is splendid.

I wonder if all current plans to prepare for the next phase of the pandemic could be recast as just that – preparations. So, in the broader context of increasing knowledge and better methods, “we” are always progressing toward doing it better. The lessons learned in the past six months will inform the preparations for the next six, and so on.

Thus the question “What have we learned so far?” would have elicited an earlier “confession” that the morbidity and mortality hasn’t been high, i.e., the effect of the pandemic has been “mild.” Earlier statements of consensus on the evident facts would allow earlier discussion about what is to be done next.

In that light, perhaps you may find a bit of time to reflect on the consequences of vaccine purchasing contracts in the USA and Australia.

May I tease apart your comments on viral mutations, as they apply to influenza viruses? I think you may have presented two different concepts.

One, you often suggest that the dreaded mutation of the virus to a highly virulent form has not happened yet. This lays the foundation for a mistaken belief that “mutation” is a rare event. The word itself conjures up visions of the horrible, anyway, so it’s a pity another word couldn’t be used.

Then, two, you do say that mutations are the way of this particular virus. In fact, as I understand it, the reproduction of the viral genome is an imperfect process, and many variants may be churned out of just a single infected cell. Some of the changes in bases may be of no consequence, some may cause that individual virus to “die,” and occasionally a change occurs that has dramatic consequences. A good place to see this variability in action is at the repositories where the maps of individual strains of virus have been laid out for comparison. I am sure a virologist could guess at the likelihood of viral isolates from different times and/or places being identical in every respect – that is, in genetic sequence. I guess, though, there are tight constraints on the resources available for performing the sequencings, so “identity” may be difficult to prove. That aside, the mappings are things of some beauty and they may persuade you about the limitations of the use of “mutation” as a helpful descriptor.

peter responds:

I thoroughly agree with your first point. In any emerging crisis, it is crucial for governments to keep updating their assessment of the situation. It is crucial for them to update their messaging to match. And it’s crucial for them to keep telling the public that that’s what emerging crises are like, that people need to expect the unexpected … and in particular that people need to expect changes in their assessment and therefore in their actions and recommendations.

As you say, this kind of focus on “what have we learned so far” would have made it easier for authorities to say that H1N1 was looking milder in June than had been feared in May – and to couple that reassurance with a strong warning that the virus could turn more virulent at any time.

Good crisis communication is usually provisional – uncertain about what’s happening and extremely uncertain about what’s going to happen. Authorities who box themselves with overconfident descriptions and predictions may hesitate to announce new information, fearing that people will become confused or mistrustful if they change their story too much. Thus they get stuck in outdated messaging. They may end up reluctant to say that the pandemic is turning out mild so far (WHO); or reluctant to say that elderly people are more at risk than was initially believed (CDC); or reluctant to say that containment efforts have done all they can and now we must switch over to mitigation strategies instead (public health agencies in many developing countries).

I’m not certain what you think the implications of this are for the vaccine purchasing contracts of the U.S., Australia, and other developed countries. (Developing countries had higher priorities for their limited health budgets.)

But the implications for talking about vaccine purchasing contracts are pretty obvious. Governments had to guess how many people they should be ready to vaccinate. And governments had to guess how much vaccine would be needed per vaccinee. As they guessed, they rightly judged that overestimation was better for public health – and for their careers – than underestimation. I have no quarrel with the way most governments of developed countries (Poland excepted) decided how much vaccine to order.

But these governments rarely put out appropriately provisional messages about their vaccine purchase decisions: “We’re guessing that we’ll probably need two doses per person; we’re guessing that the pandemic will probably be severe enough that most people will want to get vaccinated; we’re guessing that the vaccine will probably be available soon enough that there will be time for most people to get vaccinated before they’ve been sick already or the pandemic is waning. If we turn out wrong about some of these guesses – which we very well may – we will end up with extra vaccine. We believe that’s a lot better than turning away people who want to be vaccinated.”

We all have a tendency toward thinking; we tend to think that our governments should have purchased exactly as much of every resource as they ended up needing. That’s how the U.K. public and the U.K. media could have simultaneously blasted the U.K. government for two failures: buying too much pandemic vaccine (when the pandemic was turning out mild) and buying too little road salt and grit (when the winter was turning out severe).

I know of no way to prevent this natural-but-perverse style of thinking. But the best way to minimize it is to anticipate it: “If we guess wrong we could end up with too much or too little vaccine. We are choosing to order more than we may end up needing; that will reduce the risk of running out. We’re unlikely to get it exactly right, and we would rather be criticized for wasting money than for endangering the lives of our people.”

Your final point is excellent. And I can only plead guilty. I get it that influenza viruses are constantly mutating; that most of the mutations are minor and most of the major ones are dead ends; and that only once in a while does a major change take hold and meaningfully affect the character of the disease we are fighting. By invoking the word “mutation” only when I’m talking about such a major change – especially an increase in virulence – I am perpetuating the widespread misunderstanding that all mutations are horrific. I’ll try to do better.

comment:

I have gone through all three of your “Swine Flu Pandemic Communication Updates” addressing your criticism of the CDC for misrepresenting its data about swine flu severity:

• It’s Official (sort of): The Swine Flu Pandemic Is Mild So Far
• Update on the December 2 Update
• What the CDC Is Saying about Swine Flu Severity

Very interesting reading for sure.

Speaking just as an observer of your commentary, I am left wondering after reading your comments: Is this a case of willful negligence by the CDC (which at times I hear you saying) or innumeracy (which is what I think it probably is)?

I have seen enough statistical misrepresentations of data in research findings to make me skeptical about most of what I read or hear relating to risk and risk communication. I understand how crazy that sounds but honestly I don’t know many people that understand probability very well.

A recent article of the journal Science tackles this very issue. In the article, Victor De Gruttola, the Chair of Biostatistics at the Harvard School of Public Health, argues that statistical measures such as p values are misunderstood by far too many scientists and journalists that report on research studies. Okay, that is a long way of saying I believe there are a lot of misconceptions concerning the application of probability and statistics.

From reading your analysis I agree that it appears that the CDC got the statistics perfectly wrong with the populations of children and those over 65. However, as I mentioned earlier I am not surprised by what appears to me as a miscalculation. Without having people like yourself checking the stats and showing your work we are destined to create meanings about data that are unreliable. As a result, our risk communication becomes dubious as well.

peter responds:

I find it easy to imagine that CDC communication professionals could construct messages that convey the impression that children are most at risk for complications or death from the H1N1 pandemic without realizing that the CDC’s own data say otherwise. After all, they’re communication professionals, not flu experts.

Similarly, many local health officials, many state health officials, and even the President of the United States have received and reiterated the message that all the prioritized groups are at high risk of severe cases of flu, despite the clear statement of the CDC’s Advisory Committee on Immunization Practices that children were prioritized because they were more likely to catch and spread the flu, not because they were at higher risk of complications. These people are not flu experts either, and I don’t hold them responsible – or at least not very responsible – for the misleading claims they passed along.

And I will grant that it is just barely imaginable that even CDC Director Tom Frieden could reiterate these claims without realizing that they contradict the CDC’s own data.

But it is inconceivable to me that the epidemiologists and other subject matter experts who do infectious disease risk assessment 24–7 for the CDC could undertake the complex calculations required to estimate age-specific pandemic H1N1 cases, hospitalizations, and deaths without also doing the incredibly simple additional calculations required to derive estimated age-specific case attack rates (CARs), case fatality rates (CFRs), and population mortality rates (PMRs). These parameters are fundamental in the training of infectious disease experts. It is virtually automatic to convert infectious disease data – even preliminary, tentative estimates regarding an ongoing pandemic – into these familiar parameters. That’s how experts determine how severe an outbreak is so far, and who is most at risk.

Moreover, some of the technical experts who do influenza work at the CDC have in fact calculated estimates of the pandemic’s age-specific CARs, CFRs, and PMRs. A prime example: On November 30, about 43 minutes into a webinar, CDC expert Martin Cetron presented preliminary PMR estimates for various age groups, covering the period from April through November 14, 2009. The webinar itself is online, but when Jody Lanard and I sought permission in mid-December to reproduce one of its tables on my website, Dr. Cetron declined, in part because the data were not yet “validated and ready for prime time.” (However, when we requested the raw data from which such estimates are derivable, for the period from August 30 through mid-December, the CDC promptly arranged to send us a spreadsheet, for which we are grateful.)

An easier-to-access example, from several months earlier in the pandemic, is an August 12 presentation by Sonja J. Olsen, head of the CDC’s Global Activities Team, Division of Emerging Infections and Surveillance Services. Slide 8 reports age-specific population mortality rates based on early confirmed deaths in four countries, including the U.S. The title of the slide is “Deaths from Novel Influenza A (H1N1) Most Common in Adults” – not children.

In addition, several state health departments have openly reported age-specific estimates of H1N1 population mortality rates, using laboratory-confirmed hospitalization and death data from high percentages of hospitals in the state.

Florida, for example, reports age-specific deaths in an easy-to-understand table. The currently posted update clearly shows that Florida’s pandemic death rate has been highest for people 50–64 years old, and lowest for the group for which vaccination was universally prioritized: the 5–24-year-old age group. (Most of the deaths reported here occurred before significant vaccination immunity could have been a factor.) Here is Florida’s table:

On January 22, 2010, the CDC itself (as opposed to individual CDC experts) published age-specific population mortality rates for the H1N1 pandemic … for the first time as far as I can determine. That week’s Morbidity and Mortality Weekly Report (MMWR) includes this paragraph:

During August 30–January 9, a total of 1,779 deaths associated with laboratory-confirmed influenza virus infections were reported to CDC through AHDRA. The 1,779 laboratory-confirmed deaths are in addition to the 593 laboratory-confirmed deaths from 2009 H1N1 that were reported to CDC from April through August 30, 2009. Since August 30, cumulative deaths associated with laboratory-confirmed 2009 H1N1 infection per 100,000 population were 0.31 for persons aged 0–4 years, 0.26 for 5–18 years, 0.38 for 19–24 years, 0.60 for 25–49 years, 1.03 for 50–64 years, and 0.65 for ≥65 years. For the period August 30–January 9, the median number of states reporting laboratory-confirmed deaths per week through AHDRA was 34 (range: 23–38).

When Jody and I converted the CDC data into a bar chart, the age-specific PMRs looked like this:

Note that the lowest population mortality rates are in the three age groups for which universal vaccination was prioritized (except for children 0–6 months, for whom influenza vaccine is not approved).

PMRs that are calculated based on laboratory-confirmed deaths are of course underestimates, since not every swine flu death gets lab confirmation. They are underestimates in other ways as well: The state totals omit hospitals that didn’t send in their reports; the nationwide totals omit states that didn’t send in theirs, or reported using a different metric.

That’s partly why the CDC developed the whole-population estimates on which I based my articles. These broad-based population estimates are preferable to the incomplete tallies of laboratory-confirmed cases for assessing how mild or severe the pandemic has been overall (so far) – as well as for other purposes for which a wide-angle view of the pandemic’s total impact is what’s needed. But the two sorts of data show very similar patterns with regard to which age groups have been most at risk (so far). And the laboratory-confirmed data are tabulated with a larger number of narrower age categories, which may make them more useful for many purposes, including vaccination campaign decisions.

Do you believe that most people over 65 (and their doctors) realize that they are at higher risk than children, and at about the same risk as people 25–49, many of whom are considered “young adults”? Do you believe that most people 50–64 (and their doctors) realize that they are the group at highest risk? I do not.

But it is clear that the CDC’s flu experts have known these facts for months – early on from the data on laboratory-confirmed deaths, and later from the CDC’s estimates of total deaths as well.

Another reason I am convinced that the CDC is well-aware of the population mortality data: Most CDC statements about the vulnerability of different age groups to swine flu have been misleading rather than flat-out false – for example, eliding from the accurate claim that the pandemic has killed far more children than seasonal flu usually kills to the false implication that it is deadlier to children than to older people. If the CDC didn’t realize what its data meant, it would have stumbled often into overt falsehood. Instead, it has crafted messages that carefully mislead without lying. (State and local health officials, misled themselves, are unable to craft careful messages, and do stumble often into overt falsehood, as you can see if you go to the links in the second paragraph of this response.)

I offered a number of detailed examples in the three articles you mention at the start of your comment. Here’s one from my “Update on the December 2 Update,” on a December 4 Q&A between Bob Roos of the University of Minnesota’s Center for Infectious Disease Research & Policy (CIDRAP) and CDC Director Thomas Frieden.

Bob Roos: Thanks for taking my question. I had a question about the kind of the overall severity of the pandemic and case mortality rate. The CDC recently estimated the case fatality rate at 0.018%. This is much lower than the seasonal flu epidemic. I wonder if you can comment on that.

Tom Frieden: The key point is really age specific case fatality rates. We’ve reported that H1N1 has not affected the elderly significantly. So while some elderly have gotten it and those who are infected sometimes become severely ill and that’s why we have emphasized the importance of prompt antiviral treatment of the elderly and others with underlying conditions who are severely ill. What we’re finding really is this virus is a much worse virus for younger people. The number of people, not just children, but young adults under the age 50, who will get severely ill or die from this virus is much higher than from seasonal flu. The fact is that with 210 laboratory confirmed [pediatric] H1N1 deaths, we are really before the beginning of flu season, we don’t know whether there will be much more H1N1 or not. Already three times the number of deaths among children than we would [see] in a usual flu season.

At the time of this December 4 dialogue, here is what the age-specific population mortality data available to the CDC – but not to reporters or the public – looked like, converted into a bar chart. These are lab-confirmed cases; the CDC estimates Roos and Dr. Frieden were discussing showed a similar but less detailed pattern.

As the update discusses in greater detail, Dr. Frieden’s answer first bridges away from Roos’s question about overall pandemic severity to age-specific severity. Then he makes some broad claims that “H1N1 has not affected the elderly significantly” and “this virus is a much worse virus for younger people.” These are misleading but legalistically defensible statements. It is true that the pandemic has been less deadly to the elderly than the seasonal flu usually is; and it is true that the pandemic has been less deadly to the elderly than to adults 50–64, who are technically “younger people.” It is decidedly not true that the pandemic has been less deadly to the elderly than to children. Perhaps wishing to imply the latter without actually misstating the facts, Dr. Frieden bridges again, this time to the key fact that seems to support his two propositions, that the pandemic isn’t mild and that it is especially horrible for children: The pandemic has already caused “three times the number of deaths among children than we would [see] in a usual flu season.”

As I wrote in the update:

This is true. The pandemic is (a) killing more children than the seasonal flu kills, though it is (b) killing fewer children (as a percentage of the total number of children or the number of children with swine flu) than adults and seniors.

Dr. Frieden’s response showcases the first fact (a), which supports the CDC’s vaccine prioritization groups, and avoids the second fact (b), which doesn’t.

Let me add two other pieces of evidence that the CDC’s failure to give the public age-specific breakouts of population mortality rates (until the January 22, 2010 MMWR) was intentional.

• At its press briefings, the CDC routinely tries to correct inaccurate perceptions about vaccine risk and other issues. The national pandemic influenza communication plan recommends that officials “[p]romptly address rumors, inaccuracies, and misperceptions.” But as far as I know, the CDC has never tried to correct publicly the thousands of news stories stating or implying that everyone between the ages of 0 and 24 is at increased risk (compared to other age groups) of complications and death from swine flu.
• On January 8, Sharon Begley, science editor of Newsweek, wrote a blog post on my criticisms and calculations. Before doing so, she asked Beth Bell (associate director for epidemiologic science at the CDC’s National Center for Immunization and Respiratory Diseases) to go over my numbers. Dr. Bell said, “The basic calculation is right.” But the agency’s messaging didn’t change. National Influenza Vaccination Week happened; HHS officials held press conferences to announce new public service announcements (PSAs) aimed at Indian Country and African-Americans. You can scroll through page after page of swine flu PSAs without finding any targeting Americans 50–64 or 65+.

I conclude that the CDC’s decision not to report estimated age-specific CARs, CFRs, and PMRs when it reported its age-specific estimates of cases, hospitalizations, and deaths was a conscious decision – a decision it has made three times so far in three separate reports. (The latest report and links to the earlier ones can be found at http://www.cdc.gov/h1n1flu/estimates_2009_h1n1.htm.) As I mentioned above, the CDC did publish age-specific population mortality rates in the January 22 MMWR. But as of February 2 there have been no CDC pandemic update press briefings since January 7, and this information has not hit the media.

And I conclude that the CDC’s decision not to change its vaccination priority groups after the age-specific data started to emerge – which occurred prior to the start of the vaccination campaign – was also a conscious decision. Finally, I conclude that its decision to continue to justify the vaccination priority groups with misleading public statements about the relative risk of different age cohorts was also a conscious decision.

Note that I do not object to the decision to keep the vaccination priority groups unchanged – only the decision to keep the rationale unchanged. There were still good (though debatable) reasons to focus on vaccinating children and to deemphasize vaccinating seniors, even after it became clear that seniors had a higher pandemic PMR than children (and after it became clear that older adults, 50–64, had a higher pandemic PMR than children or seniors). The CDC could have explained those reasons. It chose instead to continue to signal that children and young adults were most at risk, leaving the public and outside officials to infer incorrectly that “most at risk” meant “most at risk of complications and death.”

If your hypothesis were right – if no one at the CDC realized what its own data meant – that would be truly scary. Health officials who occasionally mislead the public are less dangerous (and more correctable) than health officials who are unable to assess basic health metrics. Forced to choose between dishonest leaders and incompetent leaders, I’d have to go for dishonest.

I don’t know why the CDC chose to mislead the public, but it’s easy to speculate. Some possible reasons are simple and even compelling. For example, it’s easier and more “credible” (if you don’t get caught) to stick to one rationale for a policy than to change rationales in the middle of the policy’s implementation. Other possible reasons are prosocial-but-Machiavellian. For example, a successful pediatric flu vaccination program could be a platform for wider public acceptance of vaccination against childhood diseases.

The likeliest hypothesis, in my judgment: The CDC probably felt that more kids would end up vaccinated if it conveyed the impression that children faced the greatest risk than if it said truer but less inspiring things, such as the fact that children catch influenza at a higher rate than older people (which is true for seasonal flu as well as pandemic flu), and that vaccinating them is thought to be among the best ways to protect their parents and grandparents.

My wife and colleague Jody Lanard collaborated on this response.

robert answers:

Well, at this point with the current evidence I see this ultimately as an empirical question. Is there any way we can determine if my contention (innumeracy) or your contention (dishonesty) is correct?

I find it interesting that we are both cynical but about different things.

Ultimately, it sounds like you are saying CDC wanted more children to be vaccinated so they were dishonest in their communication about vaccinations. I hope this is not true. As you astutely say in your commentary, one can be clear in risk communication about risk levels and still get parents to have their children vaccinated. My guess, without having the evidence to know for sure, is what you mention earlier: that conventional wisdom suggests children are high-risk for the flu and that idea directed CDC’s risk communication. That is why I mentioned innumeracy earlier in my first post.

Risk communicators, journalists, and anyone else must do their due diligence and make sure their research findings and risk communication are congruent. Our common sense and everyday ways of knowing often get us in trouble – particularly when we have empirical data that suggest otherwise.

For these reasons my evaluation suggests the innumeracy/miscalculation/misinterpretation conclusion for CDC’s risk communication over the determination that CDC is dishonest. However, it is an interesting case of risk communication and one that all risk communicators can learn from.

peter responds:

I do think the CDC has been dishonest about which age groups are most at risk from the swine flu pandemic. Having said that, I should instantly add that I am not accusing the CDC of lying, only of misleading. I think it tries hard not to lie.

Does the distinction between lying and misleading matter? Corporate communicators have long since learned that proving they didn’t actually lie when they misled the public may save them in a courtroom – but it doesn’t protect them from public outrage when their dishonesty is discovered.

On the other hand, the U.S. media and the U.S. public love to uncover and excoriate corporate dishonesty. But similar dishonesty on the part of public health officials seems to go comparatively unremarked and unpunished by journalists and citizens alike, as long as it falls short of outright lying. This may be because the dishonesty of public health officials is motivated less by self-interest; when they mislead us, they usually do so with our interests at heart. It may be because we depend on public health officials and we want to trust them.

In four decades of risk communication consulting, I have learned that “good guys” are actually likelier than “bad guys” to construct misleading communications. (My 2009 Berreth Lecture to the National Public Health Information Coalition addressed this conclusion and gave many examples.) I think there are three reasons why this is so:

It’s a potentially lethal combination. It leads good guys – officials at the CDC, among others – to take uncompromising stands on behalf of misleading claims they have talked themselves into considering the unvarnished truth, with blowback too infrequent to disrupt the pattern.

By contrast, corporate communicators find it easier to remember that their critics have a point; easier to understand that when their critics are wrong about most things it is still useful to concede the ways in which they’re right or partly right; and easier to worry that even minor deceptions could get them crucified.

Despite these differences, bad guys often mislead. So it shouldn’t be surprising that good guys mislead a lot.

But I believe – perhaps naively – that even the good guys get caught in the end. And when a credulous public and credulous journalists finally realize that the good guys have been less than punctilious about some uncomfortable facts, we feel betrayed. We pile on. Suddenly even minor peccadilloes become big controversies.

This is what has happened in recent months to climate change scientists and activists, including the Intergovernmental Panel on Climate Change. Since the email scandal widely referred to as “Climategate” (which wasn’t minor), nobody is giving IPCC and its affiliated scientists a free pass on anything.

And this is what is happening in Europe right now to the World Health Organization (WHO), in the controversy – much hotter there than in the U.S. – over whether WHO promoted a “fake” pandemic in cahoots with the pharmaceutical industry. The charge is ludicrous, but it has legs because WHO has been less than candid about the mildness of the pandemic (so far), and about changes in how it treats “severity” in its descriptions/definitions of what constitutes a flu pandemic.

But that’s another story.

comment:

I have been reading articles on how the WHO is defending itself against the “fake pandemic” charge.

They do not seem to have taken any communications training. In their statement, they note, “The description of it as a fake is wrong and irresponsible.” How foolish.

They set themselves up for this and are failing to acknowledge the confusion they facilitated by not addressing the severity part of the pandemic equation from the beginning. Much of the world is having to deal with this critique because countries never fully explained as a foundational component of their communications strategy (if they had one) the distinction between the geographic term pandemic and the self-evident need to provide a severity context for their actions.

They said they were going to do this early on, and even looked at the U.S. Pandemic Severity Index (PSI) as a model in May of last year and gave it no credence. But then again, the U.S. chose not to use its own PSI, developed over years as a consensus document for virtually all of U.S. public health and a foundational component of virtually all state plans. The PSI was not perfect, but was an excellent foundation.

I hope everyone uses these lessons to improve their actions in the future. We now have WHO in a reactive, defensive mode they didn’t need to be in, defending their strategic approach while we are still in a tactical mode. How foolish for them and for the Council of Europe committee attacking them. They very likely will adjust their tactical messages to the world as a result – to public health’s detriment.

These are classic mistakes easily foreseen and even warned against and trained for. Human beings seem to never learn from history.

peter responds:

I agree with you that severity is obviously relevant to flu pandemic response policy. The World Health Organization (WHO) agrees with you too – see for example section 3.2.5 of the April 2009 WHO pandemic influenza guidance document, entitled “Roles and Responsibilities in Preparedness and Response.”

WHO even gave some thought to the creation of a pandemic severity index – or at least coming up with criteria for labeling pandemics as mild, intermediate, or severe. The question of how to describe severity during a pandemic was repeatedly discussed prior to the issuance of the April 2009 pandemic guidance document, though no conclusion was reached. In a May 2008 presentation on “Updating WHO Guidance on Pandemic Preparedness,” for example, WHO’s top influenza official, Keiji Fukuda, included a slide that starts: “Update #5 Incorporate Pandemic Severity Assessment & Reporting.”

But WHO left severity out of its April 2009 definitions of the pandemic phases, and its descriptions of how it would decide about ratcheting up from one phase to the next. Previous guidance documents had considered severity in their discussions of the pandemic phases. And then, instead of acknowledging this difference and forthrightly discussing why it believed the 2009 pandemic was severe enough to satisfy the earlier definitions/descriptions/discussions, WHO chose instead to defend itself by asserting (accurately but irrelevantly and misleadingly) that the formal definition of the generic term “pandemic” has nothing to do with severity. This was neither an honest nor an effective defense.

Jody Lanard and I wrote about this controversy at some length a couple of days ago.

In a nutshell: Most WHO documents describing what influenza pandemics are like, and how to respond, have talked about severity in various ways. But a streamlined flu pandemic guidance document that didn’t address severity in its revised descriptions of WHO’s pandemic phases was hurriedly published in late April 2009, shortly after the appearance of novel H1N1. Then the 2009 pandemic turned out milder than most experts expected or feared – although there was absolutely no way to know that when the novel flu virus began its inexorable global spread, and even today there is absolutely no way to know whether it will remain the case in 2010.

WHO would be wise to argue that by June 2009 swine flu was already novel, widespread, and severe enough to qualify as an influenza pandemic, though it has turned out to be a mild one so far.

WHO is much less wise to imply that severity has nothing at all to do with whether a novel, widespread influenza outbreak ought to be called a pandemic. Hiding behind a narrow, generic definition of the word “pandemic” that focuses exclusively on geographical spread is a disingenuous and ineffective way to respond to the false charge that WHO intentionally hyped an outbreak of normal flu (“une grippe tout ce qu’il y a de plus normal”) in order to enrich Big Pharma.

That said, the tougher question is how to integrate severity into our definitions/descriptions of influenza pandemics.

The U.S. Pandemic Severity Index (PSI) pretty much assumes that as long as a novel flu virus is spreading widely, even an extremely mild pandemic is still a pandemic … though this is exactly the issue at the core of the “fake pandemic” controversy. The lowest level of the U.S. PSI (Category 1) is defined as a pandemic with a case fatality rate (CFR) of less than 0.1% – that is, a pandemic less deadly than the average seasonal flu. The CFR of the swine flu pandemic in the U.S. is around 0.02% so far – one-fifth as deadly as the average seasonal flu. Is it a Category 1 pandemic, or not a pandemic at all? What would we do with a novel influenza virus spreading worldwide and killing almost nobody? Would that be a Category 1 pandemic too, or not a pandemic at all? It would still need to be watched closely, because a very mild novel influenza virus can mutate or reassort in ways that dramatically increase its virulence. But as long as it remained really, really benign, would we want to call it a pandemic at all?

In 1977, a 1950 version of a then-seasonal H1N1 virus emerged, probably as a result of a laboratory accident. It soon spread throughout the world. But most experts don’t include 1977 in their lists of flu pandemics, because the 1977 outbreak caused virtually no excess mortality, and most of the people who caught it at all were under age 23. In its 1999 flu pandemic guidance document, WHO refers to this as a “benign pandemic.” But usually WHO and most other experts do not consider the 1977 Russian Flu to have been a pandemic at all. The mildest flu pandemic we routinely call a pandemic was 1968.

Well, if the 2009 pandemic is nearly over – something we don’t know yet, obviously – then it’s turning out milder than 1968 but more severe than 1977. So it either replaces 1968 as the mildest pandemic on record or it replaces 1977 as the poster child of an outbreak too mild to be called a pandemic. (Most experts seem to be voting for the former, given the elevated death rates in people under 65 compared with the seasonal flu.) A pandemic severity index would illuminate this controversy, but it wouldn’t resolve it. The U.S. PSI, which doesn’t really address pandemics less deadly than the average seasonal flu, wouldn’t even illuminate the controversy.

Part of the problem is that there are lots of kinds of severity. The U.S. Pandemic Severity Index defines severity almost entirely in terms of case fatality rates – the percentage of those sickened who die. It assumes a (symptomatic) case attack rate of 30%. This leaves a lot of factors unaccounted for. What about a worldwide flu outbreak that sickened, say, 60% of the world’s population – disrupting everything from education to manufacturing to shipping to policing for weeks at a time as it rolled around the world – but killed almost nobody? We’ve never seen a flu outbreak like that … but as the experts keep telling us, influenza is endlessly surprising, and things that have never happened before happen all the time.

Or consider a more immediately relevant example. Although the two estimates are grounded in different methodologies, so the comparison is a long way from airtight, the swine flu pandemic has so far killed far fewer people in the U.S. than the average seasonal flu. That’s why I keep insisting on calling the pandemic mild. But even so, the pandemic has killed far more children (and non-elderly adults) than the average seasonal flu. In the U.S., CDC models and estimates suggest that the average seasonal flu kills tens of thousands of elderly people and a few hundred children. The CDC’s estimates of U.S. swine flu mortality so far: around 1,360 elderly people and around 1,180 children. Are we just counting deaths here, or should we consider who dies? Doesn’t it matter that children have more years of life left to live? Doesn’t it matter that far fewer children die in non-pandemic flu seasons?

And suppose a novel influenza outbreak sends millions of people to the hospital, burdening intensive care units to the breaking point. Isn’t that a kind of severity too, even if they nearly all survive?

When the World Health Organization thinks about pandemic severity, it is also appropriately preoccupied with the immense differences in physical condition and medical care among the world’s peoples. What do we say about a flu outbreak that the developed world can take in stride but the developing world experiences as incredibly deadly and debilitating? And how will we get accurate measures of how deadly and debilitating the outbreak is in countries whose health surveillance is as deficient as their health care?

These are tough questions – but the toughest question of all is the timing question.

Unless it is very severe from the outset, a pandemic’s initial severity isn’t its most fearsome aspect. What worries public health officials most about a pandemic is the possibility – but not certainty – that it could turn much more severe later. So one of the most important reasons to label a widely spreading flu virus a “pandemic” is to send a signal: “Watch out!”

This is largely a new capability. The 2009 pandemic was the first influenza pandemic the world’s experts were able to monitor from close to its inception. The others were “discovered” after they were well underway. Thanks to improvements in surveillance, we can now give people some warning of an emerging pandemic: a chance to implement social distancing strategies, to prepare to cope with absenteeism problems and hospital overcrowding, to teach people the right way to cough, and above all to begin the laborious process of manufacturing and administering a new vaccine. The more warning time, obviously, the better.

But pandemic severity isn’t something we can expect to know early, at least not with any confidence. Preliminary data from Mexico suggested that the swine flu case fatality rate was going to be around 0.4% – four times as bad as the seasonal flu. By the time WHO issued its pandemic declaration on June 11, we already knew that number was too high … so far. But the 1957 pandemic and even the disastrous 1918 pandemic also started with mild first waves. There was no way to know whether the 2009 pandemic would follow the same pattern: a second wave far deadlier than the first. It didn’t. In most northern hemisphere countries the second wave was mild too; in the southern hemisphere and some northern hemisphere countries, there hasn’t been any second wave yet. But WHO couldn’t have known that when it declared the pandemic.

And what are the odds that the northern hemisphere will see a third wave of swine flu in 2010 that’s far deadlier than the first two? Who knows? There were three pandemics in the twentieth century, and none of them had two mild waves followed by a whopper (unless there was a wave so mild that the much less sophisticated surveillance of the time missed it altogether). How confident does that make you that it won’t happen this time?

Here’s the essence of the dilemma:

1. Now that we’re capable of warning people about emerging pandemics in time to do some good, it would be insane not to use this ability. Early warnings are invaluable.
2. But when we issue those early warnings, we won’t know how severe the pandemic is with any certainty, and we won’t have a clue yet how severe it will get.
3. Still, it’s not foolish to think there ought to be some minimum level of severity to justify calling a flu outbreak a pandemic at all, even though the need to monitor the new virus and prepare for the worst would be unaffected by this labeling decision.

So WHO has two options, neither of them ideal:

Option One: WHO can declare a pandemic early, realizing – and acknowledging – that the pandemic’s eventual severity is unknowable, and might turn out so low as to make the precautions being urged excessive. That means defining a flu pandemic without a severity criterion. There should still be a pandemic severity index, of course, but it would have no floor.

Option Two: WHO can issue its early warnings but frame them in terms of a “possible pandemic” – and later on, when the disease is waning and the data are in, make the determination of whether it turned out to be a severe pandemic, a moderate pandemic, a mild pandemic, or a false alarm. That means having a pandemic severity index with a floor, below which it’s not a pandemic at all.

In its 2009 revised pandemic guidance plan, WHO partly resolved this dilemma by making the actions prescribed for Phase 5 (when it looks like a pandemic is probably imminent) exactly the same as those for Phase 6 (“It’s a pandemic”). That’s fine for the technical aspects of pandemic response.

But the language for communicating an emerging pandemic to the world will need to be flexible, depending on what most people already have in their minds (if anything) when they hear the word “pandemic.” In 2009, the shadow of avian influenza and 1918 loomed large in many non-experts’ (and many experts’) “mental model” of a flu pandemic. A pandemic was a very frightening prospect to contemplate. The widespread impression that the 2009 pandemic has turned out too mild to deserve to be called a pandemic resulted directly from the sort of pandemic people had been conditioned to worry about. WHO’s refusal to validate that the pandemic we got wasn’t that pandemic contributed to widespread skepticism and mistrust, and ultimately to the emergence of cloud-cuckoo-land conspiracy theories.

Next time, the world’s mental model may well be the extraordinarily mild pandemic of 2009-2010 (if it stays extraordinarily mild). And so communication strategies will have to take that very different pre-existing impression into account.

In 2009, WHO did in essence declare a “possible pandemic.” That’s what Phases 4 and 5 are all about. Its reluctance to declare an actual pandemic – Phase 6 – sooner than it did was grounded chiefly in some Member States’ concerns that the novel H1N1 virus hadn’t spread widely enough yet to qualify. There was also concern expressed back in May and June (as WHO pondered ratcheting up to Phase 6) that maybe the novel H1N1 virus wasn’t deadly enough either. But in keeping with its 2009 guidelines, WHO ultimately chose to escalate from possible pandemic to pandemic on the grounds that the virus was spreading efficiently in at least two WHO regions.

The virus was looking pretty mild by June, though certainly not harmless. When it declared the pandemic, WHO did not cite any severity criterion for doing so. In the months that followed it usually referred to the pandemic as “moderate” rather than “mild,” partly because it had so little data on the effects of the disease in most developing countries.

Would it be better if a Phase 6 declaration required specified evidence of severity? Should WHO wait till a novel influenza virus that’s spreading efficiently from person to person has killed enough people? Should WHO wait till the virus shows evidence that it’s probably going to kill enough people? Is it sufficient that the virus might get more severe and then kill enough people? Or is the notion of killing “enough people” the wrong way to define a flu pandemic in the first place?

I’m not sure it matters a lot. I’m okay with Option One: calling swine flu a pandemic that’s mild so far but could get more severe. And I’m okay with Option Two: calling swine flu a near-pandemic so far that could still turn pandemic if it gets more severe. There is precedent for Option Two. In 1977, WHO officials kept an eye on the spreading Russian Flu in case it took a turn for the worse, but they did not in real time call it a pandemic.

Option Two would mean a change in established definitions of the word “pandemic” and conventional descriptions of what constitutes an influenza pandemic. Most of the experts would be happier with Option One, I’m sure. But if a sizable segment of the public has learned (from the experts) that flu pandemics are supposed to be pretty bad, and if that convinces WHO to tweak its next pandemic guidance document so the ones that aren’t so bad yet are called “possible pandemics” or “emerging pandemics” or whatever alternative nomenclature emerges, I won’t consider that a horrible outcome at all.

What matters, it seems to me, is that WHO figure out the best way to warn people that something unusual is happening that might or might not get really bad, and that now is the time to start gearing up in case it does:

• It was important to say something in late April telling people that swine flu wasn’t widespread yet but was starting to spread and looked really deadly.
• It was important to say something in June telling people that swine flu was spreading wildly, and that it looked less deadly than at first but that could change.
• And in the months that followed, it was important to keep people informed about ongoing changes in swine flu’s spread and in its deadliness, and to keep warning them that more changes – including sudden changes for the worse – were among the possibilities for which it was important to prepare.

WHO’s big mistake, I think, had nothing to do with whether or not it considered severity part of the definition of an influenza pandemic. WHO’s big mistake was failing to tell people clearly enough that what was happening so far wasn’t looking remotely like the devastating pandemic the experts were worried about. In places were surveillance was good, it didn’t even look like a “moderate” pandemic. It was either a very mild pandemic or a not-quite-pandemic (I don’t care which) that might turn into the devastating pandemic the experts were worried about … or might not. So far it hasn’t.

Weather forecasters often need to warn people that a devastating storm might be headed their way. But they’re careful not to imply prematurely that the devastating storm has already arrived, or even that it’s a sure bet. They don’t want people looking out the window at gentle squalls and scoffing, “That’s what they consider a devastating storm!” And they don’t want people feeling angry and misled afterwards if the storm dies down or goes elsewhere. They want people preparing for the worst, hoping for the best, and knowing both are possible.

However it integrates severity into its pandemic definition, WHO needs to do a better job of integrating uncertainty about severity into its pandemic communications. I’m comfortable thinking of swine flu as a mild pandemic so far. But I could get comfortable thinking of it as a near-pandemic so far. What matters most is for the public to understand that nobody knows whether or not swine flu will morph into a severe pandemic in the months ahead. The public needed to understand that back in June when WHO issued its pandemic declaration, and the public still needs to understand that today.

My wife and colleague Jody Lanard contributed to this response.

comment:

We had a shooting on campus last night. What do you think of this paragraph in terms of risk/crisis communication from the university president?

Acts of violent crime are always unsettling. We’re fortunate to have a relatively low crime rate, especially for an urban campus, but statistics are rarely comforting when members of our community are the victims of crime. Rest assured that the University takes public safety very seriously. In recent years, we’ve increased the size of our police force and increased digital surveillance, with 1,800 cameras around campus. Through continued persistence, investment, and community awareness, we will provide a safe environment for students, faculty, staff, and visitors.

I think it’s good except for the “rest assured.” That sounds patronizing to me.

peter responds:

I agree that “rest assured” sounds patronizing. Apart from being a stuffy, antiquated phrase, it basically says “we’ll handle it and you shouldn’t worry about it” – and it instructs people to feel “assured” at a time when they should be told instead how natural it is to feel agitated.

I also agree that the rest of the paragraph is pretty good, for a single paragraph. The “statistics are rarely comforting” passage is especially nice – a very empathic way to frame the statistics.

But I wouldn’t advise any client to promise to “provide a safe environment.” That’s okay as an aspirational goal, but not as a promise; it can’t be done … especially since “safe” without any qualifiers implies zero risk to many people. Promising to work to increase safety is a lot wiser.

In a longer statement, I would expect some exploration of whether there are any lessons to be learned from last night’s shooting – or if it’s too soon for that, a promise to look for possible lessons to improve campus safety efforts.

And of course if there are things that went wrong, or if there are rumors of things that went wrong, or if there will soon be rumors of things that went wrong, now is the time to address them. “You may have heard that [we were warned about this guy and did nothing; our system for a campus lockdown and automated cell phone alert failed; whatever]. Here’s what we know so far….”