NIH, State Leaders Call for Smarter Federal Data Sharing

Federal agencies are struggling to unlock the full value of their data as siloed systems, fragmented governance structures and privacy constraints continue to slow mission delivery and drive up costs, officials said Wednesday at the GovCIO Media & Research Federal IT Efficiency Summit in Reston, Virginia.

“The pace at which technology is moving is so fast. It would be really nice if … in the next year, that government IT folks can agree on a set of parameters, guardrails and infrastructure to facilitate the breakdown of these data silos to be more efficient with data sharing across the board,” said National Institutes of Health’s National Heart, Lung and Blood Institute Scientific Solutions Delivery Branch Chief Sweta Ladwa. “There’s a lot of data that sits within different agencies … that could be brought together to rapidly bring more therapeutics to market.”

The Cost of Inefficiency

For decades, agencies have treated records as proprietary, localized assets rather than enterprise resources, Ladwa said. This created fragmentation that keeps critical information behind administrative barriers. This structure generates an “inefficient gravity” that locks assets within individual subagencies, she added.

The private sector approaches information through a different financial lens, said Tom Guido, vice president at Salesforce. In commercial environments, organizations treat internal information systems as core advantages that must be integrated to remain competitive, he said.

“The public lens is often clouded by compliance. I think that’s where you see these pockets of data and silos being held there really tightly,” Guido said. “In the private sector, data is monetized, and it’s a competitive advantage for companies that can’t afford for the silos to persist.”

Strategic Enablement Over Strict Compliance

Breaking down silos requires a significant shift in internal communication and relationship management, said Courtney Winship, CDO for diplomatic technology at the State Department. She said that State operates offices, embassies and bureaus in highly decentralized global environments, so institutional reform requires extensive coordination across its components.

“We’ve gone through a big reorganization, which is allowing us to centralize some of our costs and the platforms we’re using to make data more readily available,” Winship said. “I see it as my goal to make data more accessible and usable for the [State Department staff] across the world to handle those crisis situations.”

NIH is reshaping the traditional view of the institutes as isolated data centers, Ladwa said. This makes data more accessible to improve research, collaboration and mission efficiency while reducing the time researchers spend searching for information.

“We’ve invested substantially in data quality and making sure that all the different data we collect are interoperable [and] semantically aligned,” she said. “That’s not only for our own researchers, but also to enable the use of data collected across other systems and research domains throughout NIH, and hopefully across much of [the Department of Health and Human Services].”

Data leadership requires an understanding of localized business problems and demonstrating to individual teams that enterprise data governance directly supports their operational goals, Winship said.

“We’re … thinking of it as part of the strategic enablement, as opposed to compliance only,” Winship said. “Another way is thinking of governance in action, so we’re making sure that it’s applied to high-priority projects and solutions or processes.”

Navigating Privacy Hurdles and Retrospective Challenges

While agencies move toward more open integration, they continue to face challenges with historical data collections. Decades of retrospective data contain complex and varied permission structures.

NIH data systems host massive repositories of highly sensitive medical records, making subject privacy a critical concern. Combining clinical, imaging and genomic datasets across silos can increase the identifiability of individual research participants, requiring careful oversight and governance.

NIH is constructing secure governance frameworks without directly processing raw, restricted-access records during initial model training. Human expertise and automated workflows are both critical to making data accessible while navigating health data privacy requirements, she added.

“Our research domain experts and physicians are in the lead. They are helping us identify what concepts map to what terminologies,” Ladwa said.

Technology, Collaboration and Change Management

Tight agency budgets are forcing IT officials to maximize the value of existing data storage and accessibility investments, Winship said. Strategic partnerships with industry are critical to extending those capabilities, she added.

“We’re all in a money-strain situation,” Winship said. “We can partner with industry around [questions like] ‘How do we best leverage the technologies and solutions we have that we’ve already invested in?’”

Guido added that change management challenges are rarely driven by technology alone. Instead, organizational cultures are often slow to adapt to new approaches for data storage and accessibility. Modern user expectations are shaped by highly integrated and responsive private-sector digital platforms, creating a growing experience gap for the federal government that demands faster operational execution, he said.

“It’s a human change management challenge,” he said. “I think resetting expectations internally in your organizations, having the conversations that Courtney was referencing, it’s hard. But that’s where we solve this.”