Do you prefer the statistics about home foreclosures, or would you rather read reports about people losing their homes? Do you read the health care analysis diaries or case reports of disasters that hit the uninsured?
Both statistics and case reports tell a story, and they both impact decisions, but in very different ways. The same is true for stats about influenza and stories about children dying from influenza. Both scenarios were on display this week, both in Atlanta (CDC pandemic drill) and Hong Kong (seasonal influenza outbreak). Let's explore that a bit more.
Back in February, 2007 the Center for Disease Control and Prevention (CDC) was confronted with this mock scenario:
A 22-year-old Georgetown University swim team member just back from Indonesia eats dinner with his teammates but then develops a fever and doesn't accompany them to a meet in New York.
That is how a flu pandemic in the United States started.
It's a terrific thing, by the way, that these drills are taking place. Drilling and practice makes the actual response that much stronger, and while no exercise can be completely realistic, systems glitches will occur and be caught and improved on for the next time.
With outbreaks in multiple states, the exercise picked up where it left off a few months later, April 2007:
It's near the end of a 48-hour simulated flu pandemic. The Centers for Disease Control and Prevention's top brass are packed into a conference room glassed off from the hubbub of its Emergency Operations Center at CDC headquarters in Atlanta.
It's a chance to practice decision-making in case of a real pandemic. The gathered staff are discussing how to screen all international travelers for flu symptoms.
CDC director Julie Gerberding interrupts.
"I have a question," she says. "Just what is our containment strategy in Pea Ridge?"
She's talking about an actual place — Pea Ridge, Ark., population 2,346. But according to this fictional scenario, a dangerous new strain of Asian bird flu has found its way halfway around the world to Pea Ridge.
The scenario-writers have given Pea Ridge a small expatriate community from the Marshall Islands in the South Pacific. In the pretend pandemic, a Marshall Islander gets infected with the new flu in Indonesia and brings the virus back to Pea Ridge. That's the kind of thing that can happen when air travel is so fast that people can fly from anywhere to anywhere else before they develop flu symptoms.
In this recent CDC simulation, Gerberding needs to find out what's being done to stop the virus from spreading beyond Pea Ridge.
Flash forward to March 2008, where I joined the CDC exercise for two days as a (mostly) observer. That conference room is just behind the CDC's Emergency Operations Center, pictured above during this week's continuation of the exercise. Each of the 50 monitors is manned by someone coordinating one or another of the fuctions CDC needs to fulfill in a pandemic, such as collecting state data on hospital beds, or sharing their expertise and advice to local and state offcilas on infection control, quarantine and legal issues. There are liasons with key government partners like State and DoD (in the scenario, military personnel on an overseas ship are affected). It's now day 6 of the pandemic, according to the exercise, and the CDC Director, Julie Gerberding, is calling the Georgia State Emergency Manager to find out if Georgia schools are closing (the answer is: not yet) to assess the impact on CDC employees if their kids are sent home.
That's important information for many reasons. School closure (more properly termed student dismissal – the buildings remain open and might be used for alternate treatment centers or school lunch distribution) is a linch pin in the community mitigation strategy that CDC recommends in a severe pandemic (see this brief discussion on community mitigation from Dec 2006). However, the dependence on local decision making means that smooth execution of well-laid plans is not assured. For example, here's what happened in Atlanta this week (bolded added):
The CDC was very accomodating in allowing the observers access to senior staff and section officers on duty, including sitting in on an hour briefing/conference call with the affected states (whoever played those roles, the accents were perfect - Arkansas sounded nothing like Michigan). States had varying ability to update CDC with real time case numbers ("I'll get back to you on that" must have been the most commonly heard line of the day), and difficulty with adjusting on the fly to requests for extra personnel ("we need you to send 20 staff, varying qualitications, to support screening activities at your state's busiest airport") and rapidly changing policy requests. That was especially evident when the states were asked to consider community mitigation strategies including student dismissal. Some states pushed back on that, with an observation that it would be a) difficult b) disruptive c) expensive. In addition, in some states, there was no clarity as to whether decisions would be made at the local, county or state level (states with only a few cases were less ready to pull the trigger; Hawaii was quite worried about the effect of all this on the tourist industry).
[Remember, the above is role playing, but very realistic role playing. Some states might be more ready, some states less. If they wanted to make a point in the exercise that states were not ready to simply throw a switch and turn on non-pharmaceutical intervention, practice social distancing, close schools, etc, on a moment's notice on a call from CDC, that point was made.]
For more on local response, see the following from a Daily Kos diary written by oregondem in June, 2007 (bolded mine):
Lane County Public Health officials were "stunned" recently when a contagious man with measles went out on the town after he said he would stay put in his apartment.
County officials said they "had no reason" to think a young man might not curtail his active social life. They "take people at their word."
But what if the infection next time is an epidemic of bird flu that sweeps around the world - a pandemic?...
The Centers for Disease Control has stated that our only hope is to keep people from gathering in crowds - so-called "social distancing." This is, of course, what Eugene's measles patient refused to do two weeks ago.
The CDC stated in its February 2007 Pre-Pandemic Plan that schools, child care centers and universities should prepare to close for as much as 12 weeks to limit spread of the disease and save lives.
School closures are a key to effective action. In Seattle during the 1918 flu pandemic, Mayor Ole Hanson closed schools and theaters and banned religious gatherings. Seattle's superintendent of schools called the mayor "hysterical." The mayor replied, "We would rather listen to a live kicker than bury him." But public resistance was so strong that Hanson resigned and left town within a year.
St. Louis, Mo., closed its schools on Oct. 5, 1918, within two days of the first reported case. Philadelphia's mayor ignored pleas from health officials, saying that a patriotic parade planned for Sept. 28 was too important. Within days, the flu spread like wildfire there. The quick actions in St. Louis and Seattle resulted in lower death rates in 1918 than the skyrocketing deaths in Philadelphia. Seattle had one-third as many deaths as Philadelphia.
Local government and public-health officials have been slow to make their plans for a bird flu epidemic available to the public. As a consequence, Eugene-area citizens have no basis for confidence that local officials today will move quickly.
The truth is that while state and local officials are empowered to make school closure decisions, there's every reason to believe they will hesitate when the time comes. In the CDC drill, and in the planning, the trigger for school closure will be an increasing number of cases (click for bigger picture):
You don't want to be disruptive to family life, education, commerce, etc. and close schools too early. At the same time, schools need to close within 10 days or so of cases beginning to spread in order to do any good, or it will be too late. The idea is for schools to close early in the acceleration phase. But you need data to know that's where you are, and early on, surveillance at the state and local level is not so sharp that federal and state planners will have the data they need to make decisions about school closure based on statistics and graphs, not until all states improve their surveillance and reporting mechanisms.
But there's another way schools close. In Hong Kong this week, schools closed because of a severe case of seasonal flu that led to the deaths of a few children. At first, the health authorities figured to ride it out (Wednesay, 12 March; my bolded, and note the ubiquitous response about local closure):
But Secretary for Food and Health York Chow Yat Ngok said there was no need to have a territory-wide closure of schools. He urged people to stay calm.
"We are closely monitoring the situation now and we do not see the need to close all schools at this stage. But if the outbreak of the flu-like disease becomes more serious, we may consider that.
Dr Chow said it was up to individual schools to make the decision whether to close or not.
Well, after a few pediatric deaths, that didn't last. The schools abruptly closed on Thursday, 13 March:
More than half a million primary and kindergarten students in Hong Kong have been ordered to stay home from school for two weeks after a outbreak of the flu.
The school closures, which started Thursday, were ordered after three children died within the last week. The deaths are all suspected to have been caused by the virus...
CTV's Steve Chao, reporting Thursday from Hong Kong, said many parents only found out about the sudden decision to close schools Thursday morning, as they brought their children to school.
"It sparked a degree of panic as many hospitals as well as doctors' offices soon filled up with parents taking their children who had coughs or a degree of fever in."
Chao said memories of the SARS outbreak in 2003 remain strong in the minds of many in Hong Kong.
As reported and translated at Flu Wiki Forum, here's what happened:
Coincidentally, there was apparently a meeting of heads of schools in the Tuen Muen school district. Some of the schools became very concerned, and many parents started calling in to say their kids were sick, just so they could keep them home. Some parents who were walking their kids to school turned around after they started hearing about school closures and as the news of the latest death spread. One parent said "I have only one son, I don't want anything to happen to him."
At the same time the Centre for Health Protection was holding a series of territory-wide public meetings to address the concerns of teachers and school heads, who repeatedly asked to be given better 'metrics' or criteria as guidance for school closure. There was a lot of confusion and no consensus as to when schools should close.
Reporters and public officials sometimes misuse the term "panic" when they mean "appropriate concern" or confusion. In NYC on September 11, for example, New Yorkers didn't panic, they did the smart thing and got the hell out of the immediate area as quickly as they could. People aren't robots and they're not stupid. If there's a deadly infectious disease in the schools, parents will force school closure by applying the precautionary principle. The HK officials did the right thing. It doesn't matter if the parents (and teachers) are 'right' or 'wrong' about the decision; we've seen that with MRSA and trace asbestos enough times to know that pressure on officials from parents will close schools just as surely as word from health officials, and maybe faster than if those same officials are far away.
That fact is not lost on planners, but school closure is a state/local and not a federal task (nor is CDC in charge of every aspect of pandemic planning. That is a joint task with HHS and DHS and coordinated via the White House Homeland Security Council, following this framework).
The question remains, in a real pandemic will the schools close because of the surveillance and statistics or because of the demands of parents? Either way, they will close. The real questions for the planners are which reason for closure will come first, which will be the more orderly and which will turn out to be the more timely. Data on the aggregate cases may lag behind an early news report of pediatric mortality, or it may be the other way around. But if the schools do not close in a timely manner, and deaths ensue, the point of closing them may be lost altogether. And the fact is that whether it's an Atlanta exercise or a Hong Kong flu outbreak, the data and statistics that officials would ideally want to use to make these decisions will be in short supply.
And speaking of exercises, everyone ought to consider their proper role in (and preparations for) an emergency such as a pandemic, and that includes blogs. After all, you never know where and when the next natural disaster will hit, and sharing information is something we do really well. Certainly, this blog kept up to the minute coverage of the California wildfires going (see explanation of crisis blogging), and we did our share during Katrina. And while we won't be setting policy we can certainly keep track of things like school closures in a way that shares information with everyone (including any web-surfing officials who want more up to date info than "I'll get back to you on that").
Of course, the internet needs to stay up for us to be of use. But that's a different story for another day.
For more background, see Pandemic Flu Preparation and the Role of Internet Communities, Pandemic Challenges For Hospitals and Flu Stories: SARS And H5N1 — The Precautionary Principle.