AI’s rising tide: Transforming government healthcare

By analyzing vast amounts of data, AI and machine learning are helping government healthcare agencies accelerate critical research, improve decision-making insights, and enhance outcomes across various health-related initiatives. But in a new FedScoop video interview with leading government and technology experts, it’s also evident that agencies need to take deliberate steps to have the right strategies, infrastructure and training in place to realize those gains.

One way the Department of Health and Human Services has embarked on supporting AI advances across its agencies is by establishing an AI community of practices and sharing use cases online at HHS.gov/AI, according to Steven Posnack, HHS Principal Deputy Assistant Secretary for Technology Policy.  

“We’re also working on additional strategic plan activities and internal policies to help enable and encourage innovation and entrepreneurship. I like to describe it as being AI adventurous,” he said. “HHS is a big department…and in an AI space where you have a high pace of change…everyone’s trying to keep up with the new changes in each of the new models being released and how we can apply them to our missions.”

Among other AI use cases showing particular promise is an initiative called Trail GPT, developed by researchers from the National Institutes of Health (NIH), designed to speed up matching potential volunteers to relevant clinical trials. “We can use technology in this case to help match people to the clinical trials that they’re best fit for and help give that information to clinicians on the front lines,” Posnack explained.

While AI has drawn much attention for its potential to accelerate research and medical diagnoses, experts see many other opportunities. “Applying increasingly advanced technologies specifically for fraud detection is a promising opportunity to maximize government savings but also control increasing healthcare costs for beneficiaries,” noted Dr. Colleen Kummet, epidemiologist and director of Health Analytics at GDIT.

Kummet also stressed the importance of having “humans in the loop,” describing a tiered approach. “For high-risk AI solutions, we may design a system where a human is still fully in the loop. In a medium-risk scenario, we might have implementations of AI where humans are sampling or auditing the AI performance, which can be used as continuous feedback for the AI system. And then, in low-risk scenarios, particularly where speed is important, and risk is low, a nearly full automation [approach] with periodic audits by human reviewers may provide the largest return on investment.”

She also described the work GDIT has been doing for many years to help agencies, insurers and other health partners analyze and protect different forms of patient data. “All of that ecosystem has to be extremely secure to protect patients.”

AI is enabling agencies to imagine tackling larger and more complex challenges more quickly than ever before, added Matt Doxey, executive lead of federal health research at Google Public Sector. “If there’s a challenge or solution that agencies are having, there’s likely a way that we can leverage AI or data-driven solutions to help overcome that problem.”

“The first thing we always talk to agencies about [to leverage AI] is, let’s get the fundamentals in place before you start thinking too far down the road. This is a socio-technical revolution we’re seeing right now. If you don’t have governance policies in place…the leadership strategies and…the right mentality in place…and those fundamental digital modernization tools that need to be in place,” it will be hard to marshal the promise of AI, he said.

Learn more about how GDIT and its partners are helping government health agencies capitalize on AI and modern technology.

This video panel discussion was produced by Scoop News Group for FedScoop and underwritten by GDIT.