By Rick Weiss
Washington Post Staff Writer
Tuesday, September 12, 2006
The decline in U.S. air travel that followed the Sept. 11, 2001, attacks delayed the onset of that year's annual flu outbreak and slowed its spread around the country, according to a new analysis that could help health officials decide whether to impose flight restrictions in the event of a global flu pandemic.
The delay was modest -- flu deaths peaked about two weeks later than usual that winter -- and by the time the season was over, the same number of people had died from the disease as in a typical year.
But pandemic flu is expected to be much more deadly than conventional flu, so having a few weeks of extra time to deliver protective drugs or vaccines could make a big difference in total mortality, said the scientists who conducted the study.
"Flu spreads exponentially, and the more that's out there, the faster it spreads," said Kenneth D. Mandl, a physician with the Children's Hospital Informatics Program in Boston, who led the study with Harvard epidemiologist John S. Brownstein. "So anything you can do early on to reduce cases could have fairly large downstream positive effects."
The researchers and others also emphasized, however, that there are many reasons to think twice about imposing travel restrictions to stem the spread of a communicable disease, including economic impact, social disruption and even the ability to ship drugs to where they are needed most.
"When you look at the big picture, do you really want to shut down air travel?" asked Anthony S. Fauci, director of the National Institute of Allergy and Infectious Diseases and a key contributor to the U.S. pandemic preparedness plan. "What is that going to mean for the just-in-time economy, where you rely every day on shipments of medicines for hospitals?
"That's not to diminish the elegance of the study," Fauci continued. "You just have to be careful about how you interpret it."
The new analysis comes amid warnings that a pandemic of especially virulent influenza could emerge at any time, and as officials at the World Health Organization and the U.S. Department of Health and Human Services continue to debate the value of travel restrictions and other forms of quarantine should such an outbreak occur.
Plans developed by both those organizations call for travel restrictions only as a last resort. In part that is because researchers seeking to analyze the approach's effectiveness have had little to work with beyond computer models, whose relevance remains uncertain because they rely on many untested assumptions.
The Boston study -- published in the October issue of the journal PLoS Medicine -- reduces that uncertainty by taking advantage of an unanticipated trove of empirical data: annual statistics on the spread of conventional flu, gathered by the Centers for Disease Control and Prevention.
It was in the course of a routine analysis of the data concerning 1996 through 2005 that the Boston team noticed that the 2001-02 flu season did not fit the usual pattern.
While the national peak for flu deaths is almost always within two days of Feb. 17, the 2001-02 peak was not until March 2. And the amount of time it took the flu to spread around the country -- based on reporting from nine national regions -- was 68 percent longer than usual, or 16 days instead of the 8 to 11 days in the previous two years.
The obvious new factor that year was the 20 percent drop in U.S. air travel in the months after Sept. 11. That hypothesis was strengthened when the Boston team looked at the data from Europe, where air traffic was not reduced and where flu timing was unchanged from previous years.
Earlier research had already shown that the spread of flu within each region of the country was driven largely by workers commuting to their jobs. And computer models had assumed that commercial flights could influence the larger-scale spread of the flu among regions.
The new analysis confirms and puts numbers on those assumptions.
The study concludes that the number of international travelers flying into the United States during the month of September is a key determinant of when the U.S. flu season will peak. And it finds that domestic air travel during November -- perhaps especially during Thanksgiving week -- is a key determinant of how quickly flu will spread around the country.
"Eventually the virus gets in, but cutting back on air travel provides some lead time to get other measures in place," said Brownstein, who estimates that a 50 percent reduction in the number of passengers might delay the flu's peak by a month.
The researchers emphasized that while some people get sick from sitting near contagious passengers, the main benefit of curtailing travel comes from stemming the dispersal of sick people to unaffected communities, where they can launch new networks of infection.
But the researchers also conceded that many factors work against the idea.
"There are significant economic, social, legal and constitutional consequences," Brownstein said. "It's up to government to consider and balance these different factors."
Michael T. Osterholm, director of the University of Minnesota's Center for Infectious Disease Research and Policy, said the debate over competing flight-restriction policies is likely to prove irrelevant.
"If an influenza pandemic strikes, flight cancellations will not be a function of government -- they will be a function of people choosing not to travel, and of pilots and flight attendants not being there to work," Osterholm said. "We're not spending nearly enough on figuring out how to deal with realistic problems like that."