December 14, 2010
Stanford research follows spread of flu in high school
In order to better understand exactly how infectious diseases spread through real-life social networks, a group of Stanford researchers used wireless sensors to track everyone in one American high school during one day of last January's swine flu outbreak.
By Adam Gorlick
Flu season is here. And with every social interaction comes a game of chance: Does the person you're talking to, shaking hands with or kissing have a bug? And if they do, what are the odds you'll catch it?
Doctors and public health experts try to make mingling with the sick safer. They develop vaccines, promote the need for frequent handwashing and enforce other common-sense measures to keep coughs, sniffles and sneezes from spreading.
But in order to follow and better understand how infectious diseases spread through real-life social networks, a group of Stanford researchers used wireless sensors to track high school students, teachers and staff members throughout one day during the height of last January's swine flu outbreak.
"Do you know how many contacts you have with infectious people on a daily basis? Do you know how many contacts you have with anybody on a daily basis?" said James Holland Jones, an associate professor of anthropology and senior fellow at the Woods Institute for the Environment. "Very often, those are the things we know the least about because they're the hardest to measure."
Epidemiologists have always tried to answer those questions through pen-and-paper surveys, asking individuals to recall who they were in contact with on any given day. They've been forced to rely on shaky memories and vague recollections for their data.
Jones and his colleagues – led by Marcel Salathé, a former postdoctoral researcher at Stanford – used the wireless sensors to design a better method for tracking interactions in order to study how a flu outbreak might be headed off in a school. Their work is detailed in an article published today in the Proceedings of the National Academy of Sciences.
The researchers outfitted each teacher, student and staff member at an unnamed American high school with credit card-size gadgets that transmitted and received radio signals every 20 seconds during one day.
The devices logged more than 760,000 incidents when two people were within 10 feet of each other, roughly the maximum distance that a disease can be transmitted through a cough or sneeze.
"The enormous amount of interactions that occur in a single day is mind-blowing," said Salathé, who is now an assistant professor of biology at the Center for Infectious Disease Dynamics at Pennsylvania State University.
So are the chances to catch a cold.
After collecting the electronic tracking data, the researchers ran thousands of simulations of what would happen if there were a flu outbreak in the school.
They asked what would happen if there were enough of a vaccine to inoculate only a fraction of the school's population. Would it be better to vaccinate teachers or students? Would it make sense to vaccinate the more popular students, thinking they might have more interactions than their classmates who keep to themselves? Or would it be best to vaccinate a random sample of the population?
They found it hardly matters whom you inoculate, unless you are certain of how people are interacting with others.
"Almost nothing was better than the random strategy unless you measure who interacts with who and for how long in a typical day," Salathé said. "That flies in the face of what most people might think – that the super-popular kids with more connections than everyone else are more likely to spread more of the virus. But it doesn't matter if you're a teacher or a student or a staff member, or whether you're popular or not. Everyone's pretty much the same when it comes to transmission of the flu."
The information gleaned from the high school experiment could be helpful in putting the brakes on the spread of flu in a place like a school, where outbreaks sometimes lead to the closure of an entire facility. But Salathé stresses that authorities must consider the medical, social and ethical ramifications of doing what they did – tracking the movements and whereabouts of an entire population – on a larger scale.
Along with Jones, Salathé's Stanford co-authors on the PNAS article are biologist Marcus Feldman and computer scientists Maria Kazandjieva, Jung Woo Lee and Philip Levis. Their research was funded by a Branco Weiss fellowship, the National Science Foundation, the National Institute of Child Health & Human Development and the National Institutes of Health.