WEB EXCLUSIVE

Supercomputer vs Superflu

Statisticians simulate an avian flu outbreak.

By Amos Kenigsberg|Monday, April 10, 2006

Unchecked, an outbreak of avian flu in America might well strike 151 million people, according to a new study, potentially causing thousands or even millions of deaths. But the paper also bears some encouraging news: the nation can keep the flu's fall-out to a tiny fraction of that figure if we use the tools at hand effectively.  

Three researchers at Los Alamos National Lab and one biostatistician from the University of Washington made the projections using a detailed simulation run on a Los Alamos supercomputer called Pink. The researchers programmed the model to account for the specific ways people can transmit disease rather than averaging out the nation into unrealistic uniformity. They broke up the country into 2,000-person communities and modeled the spread within those communities – in homes, schools, offices, and even malls – and then between them, largely through travel.  

The simulation took into account a wealth of factors that would influence the spread of disease, such as limited quantities of antiviral drugs, travel data from the Department of Transportation, and the five percent of people figured not to take prescribed antivirals. The most important single factor is the basic reproductive number of the virus, or how many people each patient will infect.  

Against the background of the virus, the people, and their communities, the researchers examined how potential responses would fare. For the most benign virus studied, public-health officials could easily mount a successful defense. (The researchers consider success to be reducing the number of people affected to 10% of the population, approximately the fraction that gets hit during a normal flu season.) A program of "dynamic vaccination" – administrating vaccines as scientists continue to refine them – would limit the infected population to .7%, while targeted use of antivirals would allow only .06% to become sick.  

With even slightly more virulent strains, a successful response become much more complicated. Even so, the researchers say it's do-able. For the most virulent flu modeled, the model suggests that a program of judicious antiviral use, closing schools, and encouraging people to stay home or close to home will limit the disease to 2.8 percent of the population.  

Some of the study's other important findings: 

- It is about as effective to close schools as it is to dynamically vaccinate children, so we can most likely avoid closing schools, which exacts a big cost on society.  

- Travel restrictions will not by themselves really change the number of people affected but could slow the spread of the disease and give scientists more time to make vaccines and antivirals.  

- A vaccine designed for a different strain of flu can help dramatically even though it won't work perfectly. 

- In current-day America, it's all but impossible to stop a pandemic from covering the country.

Images and Quicktime video of the computer simulations are available at www.lanl.gov/news/images/avianflu.shtml and http://www.nigms.nih.gov/News/Results/FluModel040306

 An abstract of the research article is available at http://www.pnas.org/cgi/content/abstract/0601266103v1

Comment on this article
ADVERTISEMENT

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

ADVERTISEMENT
ADVERTISEMENT
Collapse bottom bar
DSCDecCover
+

Log in to your account

X
Email address:
Password:
Remember me
Forgot your password?
No problem. Click here to have it emailed to you.

Not registered yet?

Register now for FREE. It takes only a few seconds to complete. Register now »