Twitter Tracks Disease Outbreaks
New research from Brigham Young University says such posts on Twitter could actually be beneficial to health officials seeking an early warning of outbreaks.
The study examined 24 million tweets from 10 million unique users. The study revealed that only 15% of tweets contain accurate location information. The data was obtained from user profiles as well as tweets that contained GPS data. This could be a crucial mass for an early warning system that can detect terms such as "fever", "flu" as well as "coughing."
"One of the things this paper demonstrates is that the distribution of tweets is about the same as the distribution of the population , so we have a fairly accurate representation of the nation," said BYU professor Christophe Giraud-Carrier. "That's an additional good reason to be able to prove that it's valid, particularly if you want to look at issues like spreading diseases."
Professor Giraud-Carrier (@ChristopheGC) and his students in computer science from BYU discuss their findings in the latest issue of the Journal of Medical Internet Research.
Researchers found that Twitter's feature for location-tagging, which allows tweets to be tagged with the location was not as useful as they anticipated. They found that just 2 percent of tweets had the GPS info. This is a lower rate than the results of a survey conducted by Twitter users.
"There is this gap that is well known between what you think you're doing and what you're actually doing," Giraud-Carrier said.
User profiles are typically used to discover and interpret information about location. Of course some people use this location field to create an amusing joke, i.e. "Somewhere in my imagination" or "a cube world in Minecraft." Researchers have confirmed that the data provided by users was accurate 88 percent of the times. Apart from the jokes, a portion of the errors are due to people tweeting while they travel.
The net result is that public health officials could capture state-level info or better for 15 percent of tweets. This is a good sign for the possibility of a Twitter-based health monitoring system to augment the evidence-based data from sentinel clinics.
Scott Burton, a graduate student who was the lead author of the study said that "the first step is to search for symptoms tied to actual locations" and then plot the points on maps. "You can also find out if people are discussing actual diagnosis or self-reported symptoms such as "The doctor told me I'm suffering from the flu.'
Two BYU health science professors collaborated on the project with the computer scientists. Professor Josh West claims that Twitter's primary benefit for health professionals is speed. All About Video Games
"If people in a certain area are reporting similar symptoms on Twitter health officials might issue a notice to providers to gear up for something," West said. "Under circumstances like that, it can be extremely useful."
Kesler Tanner, a BYU student, is the co-author of the study. He was the one who wrote the code that allowed him to access the data via Twitter. He will be graduating in April, and then heading to graduate school to pursue the pursuit of a Ph.D.
