For the past decade, people have been getting sick and dying from a virulent form of avian influenza - H5N1.

This strain has killed untold millions of poultry and several hundred people. And many health officials around the world worry that this virus will mutate into a form that's easily passed from person to person. In many countries, health officials are planning how to react if H5N1 does become a pandemic. So, they're looking at what strategies were successful at reducing the rate of sickness and death during previous pandemics.

Ira Longini is a statistics professor at the University of Washington in Seattle. He and his colleagues wrote a computer program that simulates what would happen in Chicago - a metropolitan area home to some 8.5 million people - in the event of a pandemic. In this model, every person is followed every day in terms of where they mix and whom they would mix with.

"And if they are infected, we track who they infect and we simulate pandemic spread with this model," Longini says. "Then we simulate the potential impact of various control measures."

Longini says there are a number of things public health officials would ask the public to do in the event of a pandemic. The most important recommendation would be for people to practice something called 'social distancing'.

"One (form) is closing the schools ? since children are extremely important spreaders of influenza and they are highly susceptible as well," Longini says.

"Other examples are to isolate severe cases of influenza-like illness and to do a partial voluntary household quarantine of exposed people. Or simply asking people to not congregate in large groups, cancel large events which would bring people together? so, essentially trying to limit the contact people have."

Longini says their computer simulation found that if cities practice a minimum of two of these social distancing strategies, the chances of flu spreading are greatly reduced.

The more social distancing, the better.

"We had looked at a lot of these strategies individually, and some of them were moderately effective and some not very effective at all. So the big surprise was really when you combine strategies you get a synergistic effect," Longini says.

"You get a much larger effect than you would expect, from just adding the strategies? you get quite a reduction in transmission."

The strategies Longini's team used in their computer model came from responses to the Spanish flu pandemic of 1918.

His paper appears in the Proceedings of the National Academy of Sciences.