How Twitter Can Support Early Warning Systems in Ebola Outbreak Surveillance

Background: The Conceptual Framework of Public Health Surveillance and Action recognizes the need for active surveillance during re-emerging rare infections such as Ebola. The framework also demonstrates how outbreak surveillance can be linked to action through data information messages. The current study reveals how tweets can be collected and analyzed to support early warning systems. 

 Methods: Tweets mentioning Ebola were collected during the current outbreak from July 24^th - August 1st, 2014. Keywords used for the identification of data included #Ebola, #EbolaOutbreak.  Data elements collected for tweets included contents, time stamps, latitude and longitude codes.  To evaluate temporal patterns of information dissemination, tweets were characterized by date.  Content was examined through natural language processing and time series analysis.

 Results: A total of 42,236 tweets mentioning Ebola were posted and disseminated to 9,362,267,048 people.  A time series analysis using an exponential smoothing algorithm identified rate trends of tweets mentioning Ebola (number of disseminated tweets = 520,441*Minute of Time + -2.17785e+010, P<0.0001, SE 7.80976e+006). Our model explained a 520,441 increment of disseminated tweets every minute. Tweet content captured public knowledge and identified information needs.

 Conclusion: An extended definition of public health outbreak surveillance is needed as social media content can be used to support early warning systems.  We demonstrate how Twitter can serve as a cost-effective tool to supplement traditional systems.  Our findings indicate how rapid exchange of Ebola information through Twitter supports early epidemic detection for public health action and informs data information messaging during outbreak surveillance efforts.