Mid-Pandemic, CDC Looks to Upgrade Its Biosurveillance Database

As the coronavirus spreads across the U.S., the Centers for Disease Control and Prevention is renewing its partnership with Johns Hopkins University’s Applied Physics Laboratory to pull data from hospitals across the country into the government’s biosurveillance program.

APL’s Electronic Surveillance System for the Early Notification of Community-based Epidemics, or ESSENCE, was developed for the government’s biosurveillance program, which tracks key indicators from doctors and hospitals across the country to give researchers and officials a view into the spread of diseases.

APL’s ESSENCE is the "primary syndromic surveillance tool” for CDC's BioSense platform. It provides the raw data that is then filtered through BioSense’s other applications, which include research, analytic and visualization tools.

The ESSENCE application itself is “a data-agnostic tool that can process any data that contains a date and something you want to count,” Sheri Lewis, manager of APL’s Health Protection and Assurance program area, told Nextgov in an email Monday.

“There are many goals for many different users. Public health users at the CDC, state and local levels use the data for situational awareness of the current disease burden in their communities, as well as early event detection, and numerous analytics to help understand ongoing situations—like COVID-19,” she said. “It is often used for population-based surveillance more so than individual surveillance. Understanding the potential spread of a disease, understanding the results of interventions, understanding current trends in a timely fashion before finalized ‘gold-standard’ datasets are available are all things for which the system is used.”

In this case, the contract for upgraded system capabilities is with the CDC’s National Syndromic Surveillance Program, which runs the BioSense program.

“In the National Syndromic Surveillance Program, ESSENCE is used to analyze emergency department and urgent care center visits, laboratory orders and results, mortality records, [Defense Department] visits, Veterans Affairs visits, weather information, air quality information, and [Assistant Secretary of Preparedness and Response–Disaster Medical Assistance Teams] visits when available,” Lewis said.

But the program has the potential to get more granular. 

“What do you do when you’re sick? Think about it,” Wayne Loschen, an APL software engineer on the ESSENCE project, said during a talk in 2017. “Your kids might not go to school. You might go to the pharmacy and get some over-the-counter drugs. If it’s really bad, you might end up in emergency rooms. … You might buy more orange juice. Maybe you’ll buy more Kleenexes. The real fun one: You might flush your toilet a little more often.”

The ESSENCE project pulls data from a wide variety of sources to give epidemiologists a broad view of the situation as a pandemic develops.

“If you can understand the behaviors, you can start to look for the datasets that contain those behaviors,” he said. “And if you get those datasets, then you can analyze them, visualize them, help detect what’s going on, potentially, so that you can have better situational awareness.”

The base datasets include information collected from local sources, including emergency room visits, poison control calls, reportable disease data, death records, school absences, over-the-counter drug sales and other data from local health departments, Loschen said.

The program has an authority to operate, or ATO, from the CDC—a certification that the app meets a baseline for cybersecurity. However, the level of data anonymization is determined by the data source, according to Lewis.

“The level of anonymization depends on each jurisdiction’s version of ESSENCE and the data they put into the system,” she said.

The program has been used many times in the past, including recent global outbreaks of Ebola and Zika, and even to combat America’s opioid addiction.

Under the new contract requirement, CDC wants APL to upgrade the system, including new filters, an improved user interface and analytic tools, and a focus on cloud infrastructure.

Along with ongoing access and maintenance of the system, the contract calls for APL to add new functionality to ESSENCE. The new requirements include:

• A tool that can identify hospitals that have disruptions in their data feed and return a “yes/no” value for whether hospitals have “been consistent” or “has disruptions.”
• A tool to better track data quality measures for states.
• Making the alerting system more actionable at the local level.
• Exploring machine learning natural language processing improvements to allow users to tag records and develop new connections and search methods.
• Developing a “US_Census_Region” filter.
• User interface improvements, such as interactive data visualizations.

CDC also wants APL to look at improving the architecture and data flow of ESSENCE, with a focus on using cloud services and cost control.

“These tests should consider at least two major cloud service providers,” according to the work statement, which suggests Amazon Web Services and Microsoft Azure, “in order to determine if there are advantages or disadvantages associated with different providers or if in general the services provided across vendors are essentially offering equivalent options that could be deployed in any major provider environment.”

The last two major tasks of the work order include updating end-user documentation and APL’s online training courses to focus on the current pandemic and get new users—including state and local participants—up to speed on the system.

The work statement also suggests an additional, optional task for APL: adding modules to enable more data sources, such as lab tests and mortality rates.

The contract includes a one-year base period, with an additional one-year option.

While the first new capabilities are scheduled to roll out three months from the start of the contract, work is not set to kick off until September 25, according to the contracting documents.

RELATED PODCAST

Editor's Note: This story has been updated to make it clearer that this is an ongoing partnership.