Tracking an Outbreak with Social Biosurveillance

Rapidly identifying and responding to outbreaks is a key function of public health agencies worldwide. With the advent of interconnected systems that contain both private and public health information, government and other agencies can now use automated monitoring to identify disease trends within hours of initial case reports. While there is a wealth of information available in electronic medical record systems, laboratory reporting databases, and state health agency records, obtaining this data can be difficult for outside observers due to privacy laws and technical interoperability issues. Harvesting public social data can provide an additional method for monitoring public health worldwide in nearly real-time, as most major local events quickly expand through social networks.

As part of the Now Trending Challenge, AgileMedicine has released Outbreak Watch, a social biosurveillance network. This system monitors the public Twitter feed for a number of keywords. When tweets match any of these words or phrases, they are analyzed against a natural language processing system to determine their health relevance and geolocated based on the tweet and user location. Real-time analytics are provided for each disease category, which allows public health officials to quickly determine what health concerns are trending in their geographical area as well as worldwide interest in each particular category. This data is useful for monitoring potential outbreaks, but can also be used for public health education.

As an example of the power of this system, we can look at the recent case of anthrax that occurred in Britain last week. The initial case was diagnosed in Blackpool and was reported by the Health Protection Agency the morning of August 17th. Within approximately 3 hours of the first media report, the Outbreak Watch system automatically detected a spike in worldwide tweets about anthrax. The ability to quickly and automatically find these potential outbreaks is a major improvement in worldwide disease surveillance. This efficient, cloud-based system demonstrates the power of big data and analytics in health care and public health response. By integrating other, official data sources, we hope to become a leader in biosurveillance for the public health community.

Anthrax-related tweets for 16 days prior to the case reported in Blackpool.