CREATION of A Secure Web Service for Visualizing Poison Center Data for Nationwide Biosurveillance

BACKGROUND: Currently, BioSense data are received and maintained in a data warehouse, pre-processed to map chief complaints and diagnoses, and analyzed using time series charts, line listings, and maps. The model is being transitioned to a secure, federated model that affords the ability to share data while minimizing data stewardship issues. In this context, the National Poison Data System (NPDS) serves as a near real-time data repository for 61 poison centers (PCs) with the capacity to share data with systems such as BioSense.  CDC and the American Association of Poison Call Centers (AAPCC) explored a collaborative, federated architecture initiative to augment the data available to the BioSense program and the analytical functionality supporting the NPDS.

 METHODS: The AAPCC enhanced NPDS with a secure web service to provide CDC aggregate clinical effect and case counts by location and time period. CDC leveraged data algorithms for time series detection based on a 28 day moving average with standard deviation, and used the Google Maps API and open source charting tools to create a web application named Quicksilver. 

 RESULTS: Quicksilver enables time series and mapping visualization of total call, human exposure, and clinical effect counts by time period, state or ZIP code. With this approach, one year of data on one clinical effect (nausea) was viewed and analyzed by state, with 13 significant (>10 stds above moving average) increases found.

CONCLUSION: Quicksilver is the first example of BioSense leveraging a distributed yet secure architecture to provide a shared method for visualizing biosurveillance information maintained outside of CDC. This benefits the public health community, by allowing the AAPCC to concentrate on data stewardship and provisioning through the NPDS, while allowing CDC to access and analyze important data without the responsibility of securing and maintaining third party information, thus decreasing data warehousing and processing costs.