March 3, 2014
This article was reported by Business Standard.
Business Standard reported on a study that used real-time social media to track HIV incidence and drug-related behaviors as a means of determining HIV outbreaks. Researchers from the University of California, Los Angeles tested the hypothesis that it may be possible to use data from social media for remote monitoring and surveillance of HIV risk behaviors and potential outbreaks.
The researchers used data from more than 550 million tweets between May 26 and December 9, 2012, to formulate an algorithm to locate words and phrases in the tweets suggesting drug use or potentially high-risk behaviors and drug use. The algorithm found 8,538 tweets indicating risky sexual behavior and 1,342 suggesting stimulant drug use. The researchers then located the tweets' places of origin on a map and tested statistical models to determine whether these areas had reported HIV cases, using geographical data on HIV cases from AIDSVu.Org -- an interactive online map illustrating HIV prevalence in the United States.
The findings indicated that the majority of general and HIV-related tweets were from California, followed by Texas, New York, and Florida. When the researchers linked tweets to HIV data, results showed a significant relationship between tweets concerning risky behavior and counties with the highest numbers of HIV cases.
The researchers concluded that "big data" on real-time social media like Twitter could be used to collect information on sexual and drug use behaviors, to map the tweets' places of origin, and to use the information to understand, and possibly predict the location of HIV and drug use.
The full report, "Methods of Using Real-Time Social Media Technologies for Detection and Remote Monitoring of HIV Outcomes," was published online in the journal Preventive Medicine (2014; doi: 10.1016/j.ypmed.2014.01.024).