You’d think new viruses start spreading first in big cities, with their packed urban streets and large international airports. But a new study challenges this idea. Researchers have found that during the last influenza pandemic, the virus first gained a U.S. foothold in four smaller cities in the Midwest, the Southeast, and California’s Central Valley. From these hubs it engulfed the entire nation, eventually sickening more than 60 million people and killing more than 12,000.
The findings suggest those seeking to combat big epidemics need to look beyond the major population centers, says epidemiologist Dennis Chao of the Institute for Disease Modeling in Seattle, Washington, who was not involved with the study. “Anything we can learn about forecasting [flu’s] spread is useful.”
The idea that major cities and their airports are important drivers of worldwide contagion makes intuitive sense, and it’s supported by mathematical models. But models don’t always reflect reality. Researchers prefer real-world data to track how an epidemic spreads. Yet for influenza, fine-grained epidemiological data were always hard to come by.
That changed with the 2009 pandemic. For the first time, researchers had access to geographically tagged electronic health claims to insurance companies that showed where people got sick at the local level.
When Stephen Kissler, an epidemiologist at the University of Cambridge in the United Kingdom, and his colleagues analyzed these data and used them to build a new model of how flu spread, they found that packed metropolises such as New York City and San Francisco, California—home to 8.6 million and 1 million people, respectively—weren’t where the pandemic picked up steam. Instead, four cities that are home to between 100,000 to 600,000 people each were the hubs that “seeded” the U.S. epidemic, they report in a paper in press at Epidemics: Grenada, Mississippi; Albany, Georgia; Stockton, California; and Omaha, Nebraska. Three-quarters of all U.S. cases had its origins in one of these cities.
“The ‘Hollywood’ notion of a highly infectious pandemic [starting in a major city] is not the way things happen in reality,” Kissler says. The results mean that factors beyond population density and travel are important, he says. For instance, previous work has hinted that school start dates might play a role—young kids often serve as viral incubators in their community—or the degree to which a population has prior immunity from recent infections. Sheer coincidence likely plays a big role as well.
Sean Moore, an epidemiologist at the University of Notre Dame in South Bend, Indiana, praises the new study. He says future work should also look at the genetics of the virus itself, as a way to track it from place to place. “I’m sure the reality of the pandemic was more complicated than can be measured with the [current] model.”
Kissler next wants to examine the years immediately preceding 2009 to gauge whether the annual waves of seasonal influenza take off from smaller cities as well, or whether that was something unique to the 2009 pandemic. “Pandemics can be very surprising things,” he says. “We need to be prepared for the unpredictable.”