Home > News > FAQ Interregional spread of influenza through United States

FAQ: "Interregional spread of influenza through United States described by virus type, size of population and commuting rates and distance"

FAQs prepared for the paper, "Interregional spread of influenza through United States described by virus type, size of population and commuting rates and distance," by Viboud, et al., which appeared in Science Express (online; March 30, 2006) and Science (April 21, 2006). Here is the full text of the news release.

Does this study change current ideas about spread?

Our study shows that the key determinants of the regional spread of influenza throughout the United States are travels to work (workflows), that large and highly-connected population centers drive the epidemic, and that larger epidemics -- caused mainly by the more severe H3N2 influenza virus -- spread faster than milder ones. By regional spread, we mean spread between states or cities. It had been hypothesized that air travels were the key determinants of the regional spread of influenza, but we show that work-related travels, which are more "usual" travels, are more important. Travels to work encompass different modes of transportation, including car, train and air, and include both short-distance commutes (the majority) and long-distance commutes (1.5% of commutes are more than 100km).

Could you explain in lay terms the methods used and how they differ from previous methods?

The study relies on the comparison of long-term disease data for each state in the United States (30 epidemics of varying severity and viral circulating strain from 1972 to 2002) with statistics on human movements in the United States (including air travels, workflows, and long-distance trips). More precisely, epidemic timing and incidence are compared between states via the correlation statistics, which is a classical comparison method. Importantly, this study differs from previous methods or models in that it is the first study so far to compare human movements with extensive data on influenza epidemics.

Does your result suggest that spread by children is relatively unimportant?

Our study shows that adults are responsible for the regional spread of seasonal influenza epidemics, that is, between states or cities, because movements to work explain the spread of influenza throughout the United States. Both adults and children are affected by influenza, but adults travel farther and more frequently than children. Importantly, our study does not go against the theory that children drive the local transmission of seasonal influenza within a state or city, for instance, in schools and households. Our study is focused on the transcontinental/large-scale transmission of flu; and we have not studied local transmission.

What are practical applications of the findings for flu in general and avian flu in particular?

The study shows that the transcontinental spread of seasonal flu is rather rapid (between 5 and 7 weeks), and projections suggest that pandemic flu would spread even more rapidly (in the order or 2-4 weeks). This is because a flu pandemic virus is predicted to encounter little prior population immunity, and hence be more transmissible than a seasonal virus. As the speed of influenza spread between states is affected by work-related population movements and transmissibility of the virus, interventions which limit interregional routine travels might slow epidemic spread. However, evaluating the potential benefits of these interventions would require further modeling. Further, there is large uncertainty in predicting pandemic spread since we cannot predict exactly how an entirely new [pandemic] virus would behave.

No link was found between local transmission and population size and density. Why does influenza spread faster between more populous locations but not among the populous locations? Could the population size of children and the movement of children in different cities explain this difference at the local level?

Population size has a different effect on regional spread and local transmissibility.

Population size and regional dispersal:
Influenza spreads faster between more populous locations because population size is a proxy for how many people are traveling between locations; thus, there will be larger fluxes of people between populous locations. In other words, population size mirrors connectivity.
Population size and local transmissibility:
Within cities or states, population size determines the number of susceptible people, but not the transmissibility of a virus. There can be some variations between cities in the mixing patterns of children, which may affect local transmissibility, but there is no evidence so far (perhaps the differences between U.S. cities are too subtle).

Do you have any plans to apply this model to the global movement of people and the spread of influenza?

Our priority is to extend the model to study influenza transmission at a more refined geographical scale (within states or cities). On the global scale, it is possible that workflows do not drive the spread of flu between countries, since there are relatively few people who commute between countries. At this larger spatial scale, we would expect that air travel patterns may explain the spread of influenza better than workflows.

If you don't have the current version of Adobe Acrobat to read PDF files, please link to Adobe's site for the download.

Download for the MakeAccessible Plug-in for Adobe, to convert an untagged PDF file to a tagged one for purposes of accessibility.