Automated Real-Time H1N1 Monitoring

On October 28th, GE Healthcare announced that they had received a grant from the CDC to perform “near real-time” monitoring of the spread of the H1N1 influenza virus.  This project pulls daily reports from patient data already being collected in GE’s Centricity EMR system to send to the CDC.  The advantage of this system is that GE is able to leverage the medical data that they are already collecting on an estimated 14 million patients.  The obvious disadvantage is that their system is specifically limited to data from hospitals and clinics that are already using GE’s system for their EMR.

For the past few years, Pangaea Information Technologies has been working on a project in collaboration with Rush University Medical Center called GUARDIAN (Geographic Utilization of Artificial Intelligence in Real-Time for Disease Identification and Notification) which takes a different approach to solving this problem.  Like GE’s system, GUARDIAN is currently generating daily reports on likely and confirmed H1N1 cases.  However, the GUARDIAN system is not tied to a particular EMR.  During its development, GUARDIAN has interfaced with EMRs from EPIC, Picis and Cerner using the industry-standard HL7 messaging protocol.

GUARDIAN uses a variety of techniques – from simple code lookup-tables to advanced natural-language processing algorithms – to transform and standardize these disparate data sources into a relational, hierarchical data structure that is optimized for infectious disease modeling. Additionally, GUARDIAN has a flexible and modular architecture in addition to a real-time web-based user interface.  This allows GUARDIAN to be used for monitoring of any set of infectious diseases in real-time (in addition to the daily and weekly summary reports that it generates).

GE’s contribution to the CDC’s H1N1 surveillance is a major improvement over the analysis of 2-week-old data that they were previously limited to.  However, we at Pangaea believe that the ability to interface with any medical record system and the flexibility to quickly add/update disease-modeling profiles will make the GUARDIAN system a more powerful and useful biosurveillance system for the next major outbreak.