Computer Simulation Public health experts preparing for a feared avian flu pandemic are using computer models to get a glimpse of how the disease might spread if the bird virus were to start passing between people.
Based on their first models—of an outbreak in Southeast Asia—researchers affiliated with the Models of Infectious Disease Agent Study (Midas) research network have concluded that immediate treatment with vaccines and antiviral drugs, in combination with quarantines and other public health measures, could contain a pandemic.
“In the past, there hasn’t been much effort to work out predictions of how a disease might spread,” says Irene Eckstrand, scientific director at Midas. “These models have shown that if you’re going to contain the flu, you’ve got to do it really fast.”
Midas is funded by the National Institute of General Medical Sciences. Midas researchers in the United States and Britain are working along with other U.S. government agencies and the World Health Organization to outline how best to control a nascent flu epidemic.
The flu models are based in part on models developed in England in the 1990s to deal with an outbreak of foot-and-mouth disease among livestock. To model the avian flu, researchers used a cluster of 32 processors to simulate how 500,000 individuals in Thailand move and interact during a given period of time. The models simulate where each person is likely to go each day and how many people he or she will come in contact with. The model showed that a virus would spread “like ink blots all over the place,” Eckstrand says. The data revealed that an outbreak could take off quickly and would be difficult to contain unless health-care professionals and public health officials took action immediately.
Now the researchers are using supercomputers to create a similar model for the much larger U.S. population, says Ira Longini, a professor of biostatistics at Emory University and the leader of the Emory research group affiliated with the Midas project. The U.S. model requires supercomputers because it will simulate the movements of many more individuals than the original model, Longini says.
Public health experts agree that there will be another global flu epidemic, although they are not sure when or how severe it will be. They are also not sure whether the bird flu strain—known as H5N1—will be the origin of such a pandemic, although experts fear that the virus could mutate and start spreading among humans.
In the past, the spread of infectious disease was usually limited because people lived in smaller towns, and they didn’t travel much. “Now, with planes flying between major capitals, everyone is connected to everyone else,” Eckstrand says. “That’s why it’s potentially so important to create computer models that show how people are connected.”