Health Information Data Mining for Early Identification of Bioterrorism

Dynamics Technology, Inc. will build on the success of its automated anomaly detection processor (AADP) and develop a comprehensive system of medical surveillance and alerting for early warning of emerging infection's outbreaks and biological weapons attack. The AADP will use advanced signal processing technology-neural net clustering, dynamic change point detection, coupled with Self Organizing Map, Gaussian Mixture Models and Bayesian analysis. Data captured from clinical and veterinary information systems, poison center data, over-the-counter drug sales, public health laboratories, pharmacies, nursing homes, school health rooms, attendance records, and operational units and Navy ships will support real-time epidemiological analysis by the AADP for indicators of possible anomalous or threatening activity that are indicative of a biological event in the DoD military or civilian population. The AADP will automatically forward alerts to a central node for posting on a web-based geographic display and for in-depth analysis of anomalous or unusual health events. Our approach includes and leverages ongoing work for Intelligence and Civil Agencies to predict and assess infectious disease spread. The result is a novel software tool that provides automated situational awareness capability and supports improved dynamic health information data management without requiring users to specify threat behaviors or templates.