Cloud-based biosurveillance ecosystem

Biosurveillance beginnings
Kiley said two things happened in 2009 that led to the BSVE’s creation.

One was the H1N1 swine flu pandemic that began in April, which, Kiley noted, could have been better coordinated as a public health emergency by everyone involved in the global effort.

The other, he said, was an October memo to military department secretaries by Andrew Weber, then-assistant secretary of defense for nuclear, chemical and biological defense programs, directing that emerging infectious diseases would now, along with traditional warfare agents, be part of the chem-bio defense mission.

In 2011 JSTO-CBD held a workshop with U.S. government biosurveillance practitioners to learn about common practices and gaps.

“Following the workshop, we did site visits with each of the participating organizations to observe their analysts during a traditional work day to see how they did things,” Kiley said. “That’s how we came up with the capability needs for the BSVE.”

System prototype
In 2012 JSTO-CBD kicked off a competition among three industry-led teams to design a system prototype. After the down-select process, Kiley said, biomedical and health software company Digital Infuzion of Gaithersburg, Maryland, became DTRA’s industry partner.

“The BSVE is designed to foster collaboration and data sharing but it uses a security model that provides for individual datasets and applications to be restricted to specific users, if necessary,” Kiley said.

A broader domain-level security structure is being used by DoD, DHS, other U.S. government agencies and international partners — Australia, Canada and the United Kingdom, he added.

“The BSVE is disease agnostic,” Kiley said, “so you can look for anything. It will look through the data sources that come into the system and alert you of anything anomalous.”

For example, Hannan said, in 2015 “we were testing diagnostic devices in Sierra Leone, not for specifically for Ebola but Ebola hit positive on a couple of the devices. Through what we call our device-to-cloud capability, the results were quickly transmitted to the BSVE,” which at the time was still being ramped up.

That wasn’t planned, he added, but it showed the ability in real time to rapidly learn about a disease moving into Sierra Leone.

“With this information, an analyst could then use BSVE analytic tools to correlate these data with other BSVE data sources such as social media or open-source information to gain additional context,” Hannan said.

Service in the cloud
The BSVE works as a dashboard-like service from the cloud, accessed through an Internet browser, Kiley said.

“The BSVE pulls in data sources and articles and then in real-time, using natural-language processing, highlights keywords of interest to analysts,” he added.

Rather than reading through entire articles, analysts can look for highlighted text and summarized results in a tabular or geospatial format.

The BSVE also lets analysts decide what confidence they have in each source and put a higher or lower ranking on this information, and Kiley said they’re working on a recommender service that, based on searches, will offer analysts related information to “help them locate relevant information or analytic capabilities more quickly.”

Kiley said that to date the scientists have focused mainly on bio, “but the definition the government has adopted for biosurveillance is ‘All Hazards,’ so we are actively looking to expand BSVE to support chemical hazards as well.”

The BSVE system is easily adaptable, Hannan said, and “other folks can generate their own applications that can easily be plugged in — that’s all been specified very well through a software developer’s kit. So part of the idea with BSVE from the get-go is that it would be user driven and sustained.”