Predicting The Next Epidemic: Like Finding A Needle in a Haystack

Courtesy Photo | Current BSVE applications.  

FORT BELVOIR, VA, UNITED STATES

12.14.2016

Courtesy Story

Defense Threat Reduction Agency's Chemical and Biological Technologies Department

Today’s digitally connected world is filled with clues about the next epidemic. People across the globe post and share real-time information that includes indicators of potential disease, but piecing together key bits of data amidst the terabytes of tweets, news reports and blog posts is similar to finding a needle in a haystack. In an ongoing effort by the Defense Threat Reduction Agency’s Joint Science and Technology Office, scientists are addressing this challenge by developing the Biosurviellance Ecosystem (BSVE), an open-sourced software tool that will improve warfighter decision-making capabilities.

A cutting edge technology, BSVE is an open-architecture, cloud-based service that analyzes public information from social media, news reports, diagnostics tools and health organizations to predict health concerns for warfighter and civilian safety.

A novel element of the BSVE program is that it allows third party developers to create applications to increase the utility of the program. These apps provide analytic solutions such as anomaly detection, geospatial analysis, forecasting, prediction and data fusion. JSTO and the EcoHealth Alliance, a global environmental health nonprofit organization dedicated to protecting wildlife and public health from the emergence of disease, collaborated to develop several new, unique applications to enhance the BSVE platform.

One application created is the Global Rapid Identification Tool Set. The application, also known as GRITS, analyzes textual data sources, such as online news outlets, ProMED reports and blogs by identifying, extracting and succinctly mapping critical public health information. GRITS translates non-English sources before searching for case-counts, symptoms, pathogens, transmission types, hosts, dates and locations to suggest possible outbreaks of infectious diseases. Providing a timeline and map of potentially related diseases based on probability, GRITS allows analysts to visualize potential threats before they occur.

Another application, the Novel Infectious Agent Monitor, or NIAM, is dedicated to understanding whether or not a news article is reporting on an emerging disease occurrence. NIAM scans for mentions of diseases in online news feeds and indicates a disease’s prevalence by measuring the time since the prior mention. The application then compares the current frequency of mentions to historical averages. Using NIAM, analysts can view disease trends over weekly, monthly or annual periods and correlate a disease’s trend strength.

The Flight Risk Tracker, or FLIRT is another new tool in the BSVE suite. FLIRT enables users to examine transportation and travel networks to identify where infected travelers and contaminated goods are likely to travel. Increased globalization means that an infectious disease can quickly become an epidemic. Consider the 2009 H1N1 outbreak. In just 90 days, the disease spread throughout 70 countries and was declared a pandemic. FLIRT’s interactive interface allows users to query a database of more than 1.6 million commercial airline flights departing and arriving from all major international airports. By better anticipating epidemics through travel patterns, analysts and local communities can prepare for an outbreak and take preventative measures, reducing the overall cost and burden of mitigating disease spread.

Finally, the Emerging Infectious Disease Repository, or EIDR, allows analysts a historical context to view potential diseases. EIDR includes a curated, expansive and transparent repository of information on past emerging infectious disease events.

GRITS, NIAM, FLIRT and EIDR all currently assist surveillance analysts in anticipating the next outbreak. Additional applications will be developed to increase the accuracy and usability of the BSVE toolset. Tackling the problem of disease surveillance is daunting and no single analyses or application can do it all, but increasing prediction accuracy is critical to warfighter health and troop safety.

POC: Dr. Christopher Kiley; [e-mail]

• Date Taken: 12.14.2016
• Date Posted: 12.14.2016 11:30
• Story ID: 217669
• Location: FORT BELVOIR, VA, US