Open source text based biovigilance

−Timely detection of disease outbreak events is of paramount importance for the defense against infectious diseases and biological threats. Internet-based communications can provide good situational awareness for countries where public data collection is inadequate, unreliable or missing. The key challenge is to sift through this vast amount of unstructured text to identify relevant reports and to extract disease related information into a structured format suitable for analysis. In this work, Natural Language Processing (NLP) techniques are used on data from news feeds, websites, and medical publications to extract key biological event data. We developed the Threat Assessment Dashboard (BioTHADTM) in order to improve detection and monitoring of biological events. We demonstrate that disease outbreak incidence and timing can be effectively extracted from open news sources using NLP. The BioTHADTM application could serve as a model for tracking not only infectious, but chronic diseases and other types of events worldwide.