Data-Sharing is Essential for the Federal COVID-19 Response

Comprehensive data-sharing programs recently launched by the Department of Veterans Affairs (VA) and the Department of Health and Human Services (HHS) have boosted the response to COVID-19 outbreaks and limited the spread of the virus.

The pandemic alerted congressional leadership to the necessity of well-funded, modernized IT infrastructure across agencies that manage healthcare and emergency response. Congressman Frank Mrvan, chairman of the Subcommittee on Technology Modernization within the House Veterans Affairs Committee, said the VA’s modernized electronic health records (EHR) system helped ensure patient information is stored and easily accessible while providing essential care at GovernmentCIO Media & Research’s Digital Services: Health IT event.

“We’re especially focused on the electronic health medical modernization program,” he said at the event. “We are working on two fronts. The transition of the electronic medical records — making sure we do it effectively, efficiently, making sure that we have effective and accountable reporting, making sure that the infrastructure is in place, that the experts who are working on this are moving in a direction of teamwork, and the training and schedule scheduling is on time — and that we’re doing it efficiently and effectively.”

Against the backdrop of ongoing modernization, federal agencies mobilized a large-scale data-sharing initiative to respond to a nationwide pandemic and provided support to hospital networks managing an unprecedented level of strain while balancing COVID-19 triage.

First among these adaptations was widespread COVID-19 screening at VA medical facilities, which allowed the VA to protect healthcare staff from viral exposure.

“We did things like changing the way our medical kiosks, which at the time were the first checkpoint when people entered, were engineered to ask questions about Coronavirus exposures and symptoms so that we could make sure they had a mask on them right away if they had symptoms,” said Dr. Sophia Califano, deputy chief consultant for preventative medicine at the Veterans Health Administration (VHA), at the event. “This was even before widespread masking. Then in about March [2020], we moved to symptom screening at every entry point of controlled entry, and new tools came out of that as well. We had simple questionnaires to get your go-ahead screen to pass into facilities to make sure folks were symptom free when they were coming in, and to keep that flow so people can keep coming to work.”

This process of using COVID-19 testing data and health information to prevent viral spread extended to intra-agency projects designed to better understand the virus, including collaboration between the VA and HHS with critical assistance from the Department of Energy’s advanced computing projects.

“In the very beginning, we worked closely with our VA colleagues and the Department of Defense, as well as the Department of Energy,” said Kristen Honey, HHS chief data scientist, at the event. “One of the first things that we did as an interagency group was to put massive amounts of data on DOE supercomputers. We were able to do agreements to get electronic health records, this really high value data on the health of people who had COVID to learn more about virus susceptibility, and that was all because of interagency data sharing, collaboration across HHS, VA, DOD and the DOE.”

This focus data analytics assisted with hospital scheduling and triage management, a process that incorporates artificial intelligence (AI) to alleviate repetitive tasks for overtaxed healthcare workers.

“One of the biggest challenges early on — and actually still today — is scheduling inside of a COVID ward,” said Ben Cushing, Red Hat’s federal healthcare practice lead, during the event. “You have the patients and the providers working in a really dynamic environment, and the providers themselves are often becoming patients because the virus is so contagious. So the outcome of that is that you need to change your scheduling behavior. Usually a schedule gets assembled over a few days, usually three to four days to put out a schedule for the next month. And schedules change rapidly when you have the workforce itself getting sick and not being able to come in. So this is an area where AI is really suited, because all of the constraints are known. That’s something we went ahead and piloted, and we’re able to actually turn around a full hospital schedule within less than an hour.”