Managing Intelligent Decision Support Networks in Biosurveillance

Introduction

Bio-terrorism and rapidly spreading naturally occurring public health threats are one of the biggest challenges the biosurveillance community faces.  The need for automated disease surveillance systems was identified in the late 1990s, and it became critical after the 2001 anthrax attack.  These systems are capable of processing large amounts of data and alerting users of possible outbreaks in their community at the onset of the outbreak.  However, these systems produce a large number of false alerts.  Situational awareness enhancements and integration of Electronic Medical Records (EMR) within public health surveillance systems require complex information fusion based on Intelligent Decision Support Network (IDSN) models. 

Objectives

The objective of this project is to build a generic software framework for supporting any complex fusion model (IDSN). Intelligent Decision Support System (IDSS) is the framework that empowers the users of the system with flexibility, transparency of the information, and common structure for defining and managing these models.

Methods

IDSS consists of a set of models like Belief Networks that compromise the Intelligent Decision Support Network (IDSN), a configuration file that describes the IDSN and its behavior, a manager application for processing the IDSN, and databases for storing and retrieving the outcomes of the models. 

Results

IDSS was successfully deployed for Water Quality Sensor data and Health Information fusion models.   Also, it effectively processed a disease severity score estimator model from EMR data.

Conclusions

IDSN creates a collaborative environment where experts from different domains can set up their own models.   The outcome of these models can be passed to the other models in the network.  Within IDSN a model's decision making is not only based on a data stream but also from other models' decisions.