Posted: December 4, 2003
This page is categorized as:    link to Crisis Communication index  link to Pandemic and Other Infectious Diseases index
Hover here for
Article Summary In late 2003, an outbreak of hepatitis A in Western Pennsylvania provided a neat case study of pretty good risk communication (not perfect, but not bad) about a pretty serious problem (not huge, but not tiny). In this column, Jody Lanard and I use Pennsylvania’s hepatitis outbreak to illustrate four basic dilemmas in crisis communication – dilemmas that are sure to come up in bigger emergencies: preoccupation with panic; trust and secrecy; over-reassurance; and anticipatory guidance.

Practicing for The Big One:
Pennsylvania's Hepatitis A Outbreak
and Risk Communication

As the current hepatitis A outbreak in western Pennsylvania has evolved, we have, as usual, been doing kitchen tabletop exercises on its risk communication aspects. For us it’s a chance to assess how various risk communication principles work or don’t work in real life. And it’s a chance to practice on a significant, fast-moving risk event that is not as terrifying as a bioterrorist attack and not as hard to contain as smallpox. The health departments, the media, and the non-infected public in Pennsylvania are also practicing.

Infectious disease outbreaks originating in food contamination occupy a middle range among health and safety crises. This one is scarier than most, a very large outbreak with several deaths and a lot of morbidity – a more difficult risk communication challenge than a typical hurricane. (For one thing, hurricanes don’t leave thousands of people in doubt for months about whether they got hit or not.) Even so, this isn’t The Big One of everyone’s fears. So it’s a good opportunity to debrief how health officials communicate and how the media and the public react, so that lessons learned can be incorporated into planning for tougher challenges to come.

With some exceptions, participants have done pretty good risk communication during the outbreak so far; we have good examples to share of trusting the public with uncomfortable truths, avoiding premature reassurance, acknowledging uncertainty, and warning people about possible bad news to come. On average, communicators rate about a “B” grade. Below we will describe how some key risk communication techniques were used or not used, with links to more information about the techniques themselves.

Preoccupation with Panic

In risk communication terms, Pennsylvania’s hepatitis A outbreak, the largest single-source outbreak in U.S. history, is notable for an unusual reason: So far only one reporter has accused the public of panicking.

This bad example is too impressive a howler not to document. On November 15, Associated Press reporter Charles Sheehan wrote: “The nation’s biggest known outbreak of hepatitis A is causing such a panic that people are lining up by the thousands for antibody shots and no longer eating out.” His story appeared in dozens of online news sources, mostly under the headline, “Deadly Pa. Hepatitis Outbreak Spurs Panic.” Such a panic that people are lining up? If lining up for shots is panic, what do we call rioting in the streets and breaking into doctors’ offices?

During New York City’s 1947 smallpox outbreak, over six million people lined up to be vaccinated, and stayed on line for hours, sometimes returning day after day. This event is celebrated and studied in public health history as an example of an orderly, calm response to a very serious risk, the antithesis of panic and social breakdown. In her fascinating paper, “Public resistance or cooperation: A tale of smallpox in two cities,” medical historian Judith Leavitt compares New York’s smallpox vaccination lines to the 1894 smallpox riots in Milwaukee. Rioting is panic; lining up is not.

A county health commissioner in nearby Ohio also seemed excessively worried about panic. A couple of local food handlers were known to have eaten at the Pennsylvania Chi Chi’s restaurant where the outbreak originated. County residents were understandably concerned about the safety of the local restaurants where these workers were employed. Apparently false rumors began to circulate that the workers were infected, and the health commissioner responded … or over-responded. As the Columbiana County Morning Journal News reported on November 21, “[the commissioner] said the worst thing that can happen in situations like this is for public panic to set in, and his staff tried their best to keep that from happening, especially since such a reaction was unwarranted under the circumstances.”

Such a reaction – panic – is never warranted, of course. But social science research consistently shows that real panic is rare, at least in western cultures (although worst-case scenario planning must still take the possibility of panic into account). In this case, people were nowhere close to panic. We know, we know, it probably felt like panic to the officials dealing with the outbreak. But it wasn’t panic; it was anxious, even frightened, temporarily hypervigilant attention – the sort of attention that is rational and appropriate during an infectious disease outbreak … especially if county residents had reason to think officials might mislead them.

For more on this topic, see the section on “Panic Panic” in “Fear of Fear: The Role of Fear in Preparedness … and Why It Terrifies Officials.”

Trust and Secrecy

It was probably his “panic panic” that led the commissioner to refuse to identify the restaurants where the possibly exposed food handlers worked, even though the public rationally wanted to know. This is a very common type of official secrecy; earlier this year, U.S. and state officials wouldn’t say which hospitals were treating possible SARS patients.

It isn’t irrational to deny the public this information. It is mistrustful – and it breeds mistrust in response. Although county officials took all the medically warranted precautions – immunizing the possibly exposed workers and all their co-workers, and keeping the possibly exposed workers off the job for a time – they still did not trust the public with information that would have enabled cautious restaurant-goers to choose to eat elsewhere for a while. Officials are at risk for future credibility problems if they are seen as more worried about local economic effects than about public health, and if they are seen as not trusting the public to handle complex or alarming information.

Given that at least 20 food handlers from other restaurants did get infected while dining at Chi Chi’s, it is not surprising (or irrational) that people became concerned about whether nearby restaurants might be a source of hepatitis exposure. Allegheny County (Pennsylvania) officials were more open than their Ohio colleagues when they told the public that three food court workers at Pittsburgh International Airport had been infected at Chi Chi’s. All 207 food court workers had been immunized, they said. They left it to the public to decide whether or not to avoid the food court for a while, and they provided enough information that people could make this adult decision for themselves. Ultimately, this kind of official trust in the public leads to more trust in officials by the public.

For more, see the section on “Candor versus secrecy” in “Dilemmas in Emergency Communication Policy.”


“Panic panic” and withholding information often keep company with over-reassurance, and they did in the case of the Ohio county health commissioner. “In rare instances [hepatitis A] can cause liver damage and even result in death for people with other health issues or lower immune systems,” the Morning Journal News reported. “[The commissioner] said that was the case with the three people who died.” It is unlikely the commissioner had inside information about the three people who had died so far; officials in Pennsylvania where the deaths occurred made no such claims. The same day, in fact, the Pittsburgh Post Gazette reported that two of the three patients who had died had no underlying relevant health issues, according to their relatives, who were quoted by name.

But this sort of over-reassurance, a common mistake in crisis communications, was uncommon during the hepatitis A outbreak. The Pennsylvania Department of Health, a key information source throughout the outbreak, did an especially good job of avoiding the premature over-reassurance trap. (Remember how many times Toronto and Hong Kong officials assured us that SARS was under control before they could possibly have known whether that was true or not.)

Officials can be pretty sure they are avoiding premature over-reassurance when they get a headline like this one from the November 20 Pittsburgh Post-Gazette: “Hepatitis outbreak in Beaver County running out of gas; State officials still won’t say danger has passed.” The article began: “State officials aren’t ready to say the nation’s worst hepatitis A outbreak is tapering off, but the numbers were doing the talking yesterday, as 10 new cases brought the total to 530.” This is one of the risk communication “seesaws” in action. Err on the alarming side, and don’t reassure people too soon – show you are still watchful and vigilant – and people will actually notice if the numbers seem to warrant less caution than you are displaying!

Even good risk communicators fell into one kind of subtle over-reassurance in the early weeks of the outbreak: one-sided risk rhetoric that paid too little attention to the reality that hepatitis A, though usually relatively mild and short-lived, can sometimes kill and can sometimes last up to nine months.

The Pennsylvania Department of Health, for example, released a statement November 5, when no one had died yet and only twelve Chi Chi’s employees and ten patrons were known to be sick. “People who develop Hepatitis A,” the statement read, “almost always recover from the illness without further complications. People may experience a range of symptoms, including fever, tiredness, weakness, loss of appetite, nausea, vomiting, abdominal pain and jaundice and a yellow discoloration of the skin and eyes. The disease may appear suddenly and last from one to several weeks.”

Dozens of media reports followed this party line, with technically true but unbalanced quotes like this one: “Symptoms include fever, nausea, diarrhea, jaundice, fatigue, abdominal pain and loss of appetite. Hepatitis A usually clears up in about two months.” All it takes to place this encouraging but incomplete information in proper perspective is to put “Even though…” at the start of the second sentence – and then finish the sentence with the facts about prolonged or relapsing illness and the possibility of death from overwhelming liver failure. CDC Director Julie Gerberding is a master of this kind of two-sided risk rhetoric.

What happens when people aren’t adequately forewarned about the uncommon but not impossible worst cases? After three people had died, the Morning Journal News reported on November 21: “The three deaths have shocked western Pennsylvanians, because health authorities have been saying that hepatitis A is usually not fatal and normally runs its course in a few weeks after causing such symptoms as fever, jaundice, nausea and abdominal pain.”

The public has a right to the bad news as well as the good news. In fact, the public has an obligation to understand the bad news and bear it. How else can we decide whether to get vaccinations, or know how to vote on health department budgets, or have opinions about childhood immunization policy? How else can we consider whether public and employee bathrooms should be required to have faucets and doorknobs that can be turned with elbows instead of hands to reduce cross-contamination? How else can we decide whether and where to use our own elbows instead of our hands? Without full information – including the scary stuff – we are less likely to develop good individual hygiene habits, and far less likely to push for good public health policies.

For more on these topics, see “Don’t over-reassure,” “The seesaw of risk communication,” “Err on the alarming side,” and “Put the ‘good news’ in subordinate clauses,” all in “Anthrax, Bioterrorism, and Risk Communication: Guidelines for Action.”

Anticipatory Guidance

Telling people what to expect – especially if it’s bad news – is a good way of helping them prepare. Pennsylvania didn’t do this as well as it could have with respect to the most severe hepatitis A impact: death. People had absorbed the reassuring information about hepatitis A, and were not prepared for the deaths.

They may be even more shocked next spring when reports appear that some people are still sick. The CDC website states: “About 15% of people infected with [hepatitis A] will have prolonged or relapsing symptoms over a 6-9 month period.” As of late November, there had been virtually no anticipatory guidance from state or local health officials about this – in statements to the press or on official websites. Individual doctors are probably discussing this possibility with their patients, but the public at large is at risk of feeling misled and blindsided when they read stories next April about people still sick from a restaurant meal they ate in October.

On other topics, though, the anticipatory guidance has been good.

Richard McGarvey, a spokesman for the Pennsylvania Health Department, did well on the issue of secondary cases – that is, whether people could get hepatitis from others who got it from eating at Chi Chi’s. When two people who had not dined at Chi Chi’s developed hepatitis A, McGarvey was open about the likelihood that they might be secondary cases from the same outbreak, and that other secondary cases might emerge. “With 615 [primary] cases,” he said, “there’s probably going to be some cases out there. But we don’t have any confirmed yet.” State health secretary Calvin Johnson reiterated this anticipatory guidance: “This is the time period where we would begin to see secondary cases.… We have not seen any [confirmed] as yet. If we do see some, it would not be unexpected.”

For more, see “Do anticipatory guidance” in “Anthrax, Bioterrorism, and Risk Communication: Guidelines for Action.”

Nominated for Best Paragraph….

In the passage that follows, Richard McGarvey deftly combines anticipatory guidance with a two-sided approach that provides reassuring information without any over-reassuring rhetoric. Along the way he personalizes the outbreak by expressing wishes, acknowledges uncertainty, and implicitly apologizes for a prior error – all risk communication pluses:

Richard McGarvey, a spokesman for the Pennsylvania Department of Health, said that while further deaths were possible, officials hoped that the antibody inoculations given to more than 8,500 people in the weeks since news of the outbreak was made public would reduce the number of new infections…. “We first started the investigation looking at food handlers, but the numbers kept going up,” Mr. McGarvey said. “We had no idea the numbers would go up that high.” [New York Times, November 17, 2003]

Kudos to Richard McGarvey. And a fine example for the rest of us.

Copyright © 2003 by Jody Lanard and Peter M. Sandman

For more on crisis communication:    link to Crisis Communication index
For more on infectious diseases risk communication:    link to Pandemic and Other Infectious Diseases index
      Comment or Ask      Read the comments
Contact information page:    Peter M. Sandman

Website design and management provided by SnowTao Editing Services.