Combined data from electronic health records, crowdsourced surveillance information, Google searches and Twitter posts can accurately track influenza outbreaks in real time, according to a study published Thursday in PLOS Computational Biology, Health IT Analytics reports.
Details of Study
For the study, researchers at Boston Children’s Hospital used “ensemble modeling,” which uses different sources of information and predictive analytics to determine the probability of an event.
The researchers used four major sources of data to predict flu symptoms for particular populations:
- Athenahealth electronic health record data processed in near real time;
- Crowd-sourced surveillance data from HealthMap’s Flu Near You website;
- Google search data; and
- Twitter messages.
Study Results
The ensemble model predicted results more accurately than models using only a single stream of data. According to Health IT Analytics, the ensemble model reached a 90% correlation with CDC’s two-week forecast for flu outbreaks (Bresnick, Health IT Analytics, 10/30). In addition, the model operated in real time and correlated almost exactly with CDC’s reports of actual flu activity.
Comments
Senior author and Boston Children’s Hospital Chief Innovation Officer John Brownstein said, “What have people in informatics, medicine and public health dreamed of for years? The ability to leverage all manner of data — historic, social, EHR and so on — to create a learning health system.”
The researchers said that while the model only tracks the flu on a national scale, they hope to expand it to operate within more-narrow geographical regions and for other diseases. They also hope to create a public tool for flu prediction (Boston Children’s Hospital release, 10/29).
