Q&A: Two views of the pandemic threat
http://www.computerworld.com/action/article.do?command=viewArticleBasic&articleId=9007058
Robert Mitchell Today’s Top Stories or Other IT Management Stories
January 08, 2007 -- Dr. Michael T. Osterholm, director for the Center for Infectious Disease Research & Policy, gives a sobering assessment of the avian flu pandemic threat and its potentially disastrous impact on business, IT organizations – and society. Martin Meltzer, a senior health economist at the CDC in Atlanta, acknowledges the reality of the pandemic threat, but says predictions about mortality rates, economic losses and other impacts should be taken with a grain of salt.
National correspondent Robert L. Mitchell spoke with both about what the possibilities – and probabilities – associated with a global pandemic.
. . . . . . . .
Dr. Michael T. Osterholm
Q&A with Osterholm: Why you should worry about pandemic
Dr. Michael T. Osterholm is a professor at the school of public health and director for the Center for Infectious Disease Research & Policy (CIDRAP) at the University of Minnesota. CIDRAP’s paper, 10 point framework for pandemic influenza and business contingency planning, is available at
www.cidrap.umn.edu/.
What do you see as the best-case and worst-case pandemic scenarios?
The lowest level of pandemic is one that occurred in 1968. If you looked at that today you would be talking 2 ½ to 7 million deaths worldwide, 30 percent of the population becoming infected and missing parts of a week or two, in terms of absence from the workforce. Even a mild pandemic will cause tremendous disruption in the work force.
On the other side you have the statement most recently made by the World Health Organization's Influenza Research at the Human and Animal Interface. In that document they very clearly delineate a potential pandemic that … could approximate what we’re seeing now [with H5N1 in birds], which is 65%. Obviously in that kind of worldwide pandemic it would be as catastrophic as anything we’ve ever seen or known. We’re talking 1 billion with a “b” or more deaths.
What steps can organizations take to protect employees?
We really have no data that supports the effectiveness or lack thereof of many of the currently recommended actions. Social distancing, what we call respiratory hygiene, even using certain types of protective devices is really at best inferences from other infectious disease situations but not directly related to influenza.
To do nothing is absolutely unethical but to provide the public with reassurances that these things will work is also unacceptable. We’re going to have a problem on our hands with the upcoming pandemic with making sure the public knows that these are common sense [precautions] but we don’t know how much impact they’ll really have.
How will the global economy be affected by a pandemic?
That impact … is difficult to predict. We can take any one impact issue, such as a lack of workers coming to work, or border closings or the shutdown of certain transportation systems. But for each one of those there is a domino effect. And the domino effect is [for example is when] not enough employees show up at the oil refinery plants. Even if you have some way of moving the petroleum, you can’t refine it. If you can’t refine petroleum you can’t provide gasoline or heating fuel or in some instances the petroleum you need for electrical generating plants. Once that happens you may have electrical grid shutdown and you’re sure to have other things such as transportation shut down.
Could we leverage the Internet and other communications networks to help in a situation like this?
Absolutely. The question is, will the Internet stay up? If we suddenly put a large part of the workforce on telecommuting status, what will that do to the last mile of the system? How will that affect the pipes? Those are issues that are unclear.
What role might other technologies such as e-learning and intranet portals play in helping to cope with a pandemic?
Many companies are counting on having Internet-based information systems for their employees and staying in touch that way. What we need to do is hope …that we’re able to maintain those.
CIDRAP’s 10-Point Framework for Pandemic Influenza Business Contingency Planning paper advises businesses to determine "trigger points," or thresholds, that dictate when to freeze IT systems and prohibit any system changes. Would those thresholds based on things such as having sufficient staff to keep the information systems running?
Yes. Or even the [workload] volume. Do you want to be making changes when your system is operating at 50% of capacity? It’s just common sense kinds of approaches.
Some organizations haven’t done any planning yet. What would you say to those who have yet to get started?
Pandemics are like earthquakes, hurricanes and tsumanis. They occur. Another pandemic is going to occur. It is not an option.
Assuming a pandemic started today, how would you expect that to unfold?
We just don’t know. In the last 10 pandemics over the past 300 years they started in all four seasons. Some have had two or three waves. Some of those pandemics had the first wave as the most severe, others had the second wave. Some, like the 1889/90 [pandemic], had a third wave.
All we can do is plan for a range of possible outcomes and one of those is a pandemic that could last 12 to 18 months or more, that could have a very high mortality and that the public, when realizing that there are many critical products and services not available, will potentially respond in ways we have never really witnessed.
Today up to 80% of all pharmaceutical products used in this country come from offshore. There will be instances if we have an impact on international trade and travel where people won’t be dying from influenza, they’ll be dying from a lack of pharmaceutical drugs that they count on every day for their well being. How will people respond to that?
Is there a precedent for the high mortality rates you’ve cited in the worst case scenario?
Nothing in modern times that would help us. Today if you want to create panic and fear in society give them some very threatening and widespread disease and we’ve seen time and time again that this often results in panic.
If people read the worst case scenario you just laid out, they may just throw up their hands and say why should I bother if that’s the case?
We’re working very hard on this issue because we believe that there is much that we can do [but] we don’t know how much difference it will make. A good example is if you have a life threatening illness, i.e. someone who is diabetic, you’re going to want to make sure you can get insulin for potentially 60 to 90 days if there is a shortage. Today how many people are in a position to be able to do that - potentially stockpile a drug that’s critical? Because their health plan won’t let them or the system is unable to provide that much of a drug. There’s an example of something someone can [change].
I know this is daunting. But on the other hand, there are things that we can do, that we should do.
. . . . . . . .
Q&A with Meltzer: Some pandemic predictions on shaky foundation
Martin Meltzer is a senior health economist at the CDC in Atlanta. Meltzer co-authored the paper, The Economic Impact of Pandemic Influenza in the United States: Priorities for Intervention.
Some estimate that an H5N1 avian flu pandemic could kill 1 billion or more people worldwide with a 60-65% mortality rate. What’s the probability of that?
There are many extrapolations of the potential death rates out to the whole population. There are two things you need to know when you see an estimate like that. One is how is that number constructed, what the “engine” is. The second is, what is the probability of any given estimate. That is commonly never discussed.
But if H5N1 becomes the next strain could it be as fatal and lethal as the current strain is now?
Sure, it could be. What’s missing is the probability of that happening. Recently my colleagues published a paper that said they reassorted an H5 avian [influenza virus] with other strains and couldn’t get it to transmit among ferrets [Lack of transmission of H5N1 avian-human reassortant influenza viruses in a ferret model, Proceedings of the National Academy of Sciences, Aug 8, 2006 ]. They tried to make a new strain from the H5N1 strain and tried to make it transmittable by mixing it with human adapted strains. As much as people say it’s possible, it’s not so easy for a virus strain to adapt to humans and it’s not too easy to produce the lab results you need to prove that a bad thing can happen.
Extrapolations aren’t very reliable. Extrapolations from smidgeons of data can be very unreliable.
A recent WHO report stated that, if the H5N1 pandemic mutates directly into a human-transmissible virus as opposed to combing with another virus to gain transmissibility, the mortality rate could remain at current levels, which are 60-65%.
So a more lethal virus might infect fewer people?
As influenza adapts to humans it becomes less virulent. That’s the conventional wisdom. A virus that kills off very quickly without allowing the vertebrate to replicate isn’t going to survive very long. [But] because the incubation period is so short [48 hours] and you can shed virus before become symptomatic, some of that doesn’t hold.
The truth is the vast majority of people who got ill in 1918 survived. So why would H5N1 be different?
The 1918 [outbreak] was virulent but does that mean H5N1 will be more virulent if it mutates? All three pandemics [in this century] had the same overall clinical attack rate, between 25 and 35% as near as we can tell. The difference is what happened to people once they became ill. In 1918 they were more likely to die than in 1957 or 1968.
How likely is H5N1 to become the next pandemic?
If H5 adapts itself to humans – that is a big if. H5N1 has been in Asia since 1997. It’s a region of the world here there’s a large population, with large numbers of chickens and pigs, and it still hasn’t mutated to the point where it has sustained human-to-human transmission. This should tell you that these must be quite rare events.
I’m willing to bet that if we had a crystal ball we could find a strain that adapted to humans but didn’t take off. If it was easy to adapt a strain in nature to maintain its lethality and attain human-to-human transmission we should have seen it by now. These are fairly rare events, once every 10 to 50 years.
I’m not saying that there’s no concern. But as an economist my sense is to wait until I see a little bit more data from the real world.
What would a pandemic’s economic impact be on business?
We did a study in 1999 where we looked at the economics - medical care and loss of life -based on a 1968 pandemic. We estimated $70 to $170 billion, based on modern costs of medical care, without loss of business disruption to society.
It’s simple to put these things down in a model but if you don’t have enough data, the model results are not very accurate predictions of the future.
A large number of data points come from the SARS experience. $10-$30 billion was the economic impact. But SARS was a very different disease than influenza. Its impact might be relevant and it might not.
What does all this that mean for business contingency planning?
Somewhere along the line we’re going to have an influenza pandemic. It could be next year or five years or longer from now. The uncertainty of when it could occur isn’t necessarily taken into account when planning. So what happens if the pandemic doesn’t occur for two years? Will everyone go home and stop planning? That would be a complete disaster.
. . . . . . . .
The WHO just released a workshop [report] earlier this year that talks about the current state of things [see Influenza Research at the Human and Animal Interface]. They do mention that it could be lethal but they don’t know the probability.
Concerning the potential high lethality, could the spread be sustained? Some mathematical models suggest that if you have a very lethal strain of flu it might be difficult to sustain transmission. The key phrase is “difficult to predict.” We really don’t have enough data points for that. Surprisingly small data sets are used to run these models.
