Rutgers Professor Uses Games to Study How People Behave During Outbreaks
World of Warcraft online incident in 2005 provided lessons that resonate with Ebola response
Virtual worlds where real-life gamers role play as they come across deadly challenges may have much to teach us about confronting a global crisis. At least that’s what Nina H. Fefferman, a Rutgers professor, concluded after studying human behavior when a glitch occurred in the online game World of Warcraft in 2005. It led to panicked and some irrational behavior on the part of participants, after a mistaken incident of “corrupted blood” spread quickly through the game, killing off virtual characters.
Fefferman, an associate professor in the Department of Ecology, Evolution, and Natural Resources, uses mathematical models to study how diseases like Ebola spread. She has pointed to lessons to be drawn from the 2005 incident and hopes to continue to gain understanding through other virtual reality games that can be useful in preparing for future disease outbreaks.
The reason this incident is significant, is that scientists and public-health experts have become “startlingly good” at predicting the progression of outbreaks, like the Ebola crisis, if they know three factors: the traits of the disease, the health condition of the affected population, and the behavior of the population.
While the first two factors can be straightforward to determine, individual behavior is difficult to predict, since it depends on social and emotional influences as well as practical and perceived constraints on behavior, Fefferman said at a recent New Jersey Health Care Quality Institute event.
Predicting behavior is essential to understanding how a disease outbreak will progress. Thus, researchers access all sorts of information in order to draw conclusions, including studying historical accounts of earlier diseases; judging how animals like badgers or termites respond to outbreaks; and taking surveys of the public to see how they would respond to diseases.
However, “we’re just not very good at knowing what we would do in the emotional and social context at the time, and with the benefit of fear and adrenaline and shock and all of the other things that people go through when this is actually happening,” Fefferman said.
That points out the benefit of studying behavior in virtual worlds like World of Warcraft, where millions of players control characters.
In 2005, a new villain in the game named Hakkar was able to cast a spell called “corrupted blood” that spread through the characters much like an infectious disease. As the characters become sick from the spell, the players that controlled them didn’t die quickly or work to defeat Hakkar like the game’s designers expected. Instead, they ran to the nearest virtual cities to seek healing supplies, spreading the spell along the way.
“They brought the spell back from a remote jungle region to an urban center where not everybody had a lot of health to spare; instantly, low-level players just started dying in the streets,” Fefferman said.
World of Warcraft programmers tried to impose a quarantine, but it failed as players moved their characters to uninfected regions, spreading the spell before logging off of the game so that they wouldn’t have to deal with the aftermath.
“We have an experimental system for controlled outbreaks, looking at how people behave, and it’s a terrible version of the actual world, but it’s so much better than the other things we’ve got,” Fefferman said.
She and her students observed a variety of behaviors from the players.
“We saw altruism, which we kind of expect people taking the role of first responders, rushing in and trying to heal people; we saw compliance with public-health announcements and we also saw noncompliance,” Fefferman said, as well as suspicion that the game designers intended the devastation. There was also fear about the future; opportunism as stronger players took advantage of weaker players; and curiosity, a behavior that intrigued Fefferman.
“I myself was guilty of that and didn’t realize it,” she said, noting that she logged in after a student told her about the virtual outbreak, despite the fact that her actions put her character in danger. “As soon as that happened, it occurred to us that in all of the models that we built to predicted epidemics, none of them included well-meaning people rushing towards the outbreak, and then having to get out, who were not themselves medically trained.”
Fefferman noted that journalists do something similar.
“We need journalists -- we need governmental oversight in outbreak management -- all of these are people that we hadn’t accounted for, and really do change the course of an epidemic,” she said. “We were able to put new things in our models and learn new things about what we think the real systems would do that we hadn’t thought of before watching this behavior come out of real people doing it in response to an accidental plague in a game.”
Fefferman said the experience taught researchers the importance of being creative in learning from virtual behavior. She recently received a grant to study the behavior of teenagers in response to another game-based sickness.
In addition to these virtual models, Fefferman also observes the data from real-life incidents. She noted that public-health authorities could learn from experiences like the H1N1 virus, when so many worried, uninfected people filled emergency rooms in some areas due to their fear of the disease that the ERs were unable to treat those with heart attacks quickly, leading to deaths.
Fefferman and others take information from these various models to plan for disease outbreaks. She works with the federal Department of Homeland Security and the Centers for Disease Control and Prevention on outbreak preparedness planning.
She emphasized that complications arise from having many overlapping plans from different countries, states, regions, cities, hospitals, schools, businesses, and nonprofit agencies.
“If everyone has a plan, which plan wins (when they come into conflict)?” Fefferman asked.
She said this type of conflict was apparent in October as concern over Ebola grew.
“Even inside just the United States -- during panic that really shouldn’t have happened -- we have some interesting conflicts,” Fefferman said. “We had the CDC and the president on the one side versus the governors of New York and New Jersey on the other” deciding whether or not to quarantine people coming back from places where Ebola had spread.
“That’s terrible. If nothing else, it’s a mixed message to people who shouldn’t be making decisions about this, but it’s also a conflict for people who actively make these decisions,” said Fefferman, who has worked with the U.S. Department of Homeland Security on Ebola preparedness.
The professor “is approaching infectious diseases in an infectious way,” NJHCQI Vice President Linda Schwimmer said of Fefferman’s enthusiastic approach to describing her research.