Study Suggests Twitter Data Can Help Track Opioid Crisis in Real Time

Research is first to verify that Twitter posts from North Carolina can be used to statistically predict future opioid overdose deaths

RESEARCH TRIANGLE PARK, N.C. — A new study by researchers at North Carolina A&T State University and RTI International, a nonprofit research institute, suggests that data from Twitter can be used to track the opioid crisis in real time and inform public health strategies to combat it. 

The study examined tweets from North Carolina over a nine-year period from 2009 to 2017 and compared the content of the posts to opioid overdose deaths figures in the state. Results showed that the content of the tweets tended to track with the three-phase progression of the opioid crisis, with tweets containing keywords related to prescription opioids emerging first, followed by tweets related to heroin and concluding with a surge in tweets related to synthetic opioids starting in 2016.

“Right now, the monitoring of overdose deaths relies on data that lags between 12 to 18 months behind real time,” said lead author Mohd Anwar, PhD, who is a professor of computer science at North Carolina A&T and an RTI Scholar. “Our research suggests that Twitter and perhaps other social media platforms could help public health officials detect trends in a timelier manner.” 

While there was no significant relationship between the number of tweets about prescription opioids and prescription opioid deaths, the number of tweets related to heroin and synthetic opioids were significantly associated with overdose deaths of the same type. 

On average, each additional tweet related to heroin corresponded to 0.13 additional heroin overdose deaths the following year, and each tweet mentioning synthetic opioids corresponded to an additional 2.68 overdoses caused by synthetic opioids. Moreover, heroin tweets in a given year significantly predicted heroin deaths better than lagged heroin overdose deaths alone.

The researchers concluded that opioid-related Twitter posts constitute a potential alternative data source to monitor the opioid crisis that is far more available and timelier than mortality data, which can lag months to years behind real time. Based on these findings, further research exploring the use of Twitter as a timely indicator of opioid overdose mortality is warranted, they noted.

The study was published in JMIR Public Health and Surveillance. To view the full manuscript, click here.