The U.S. Department of Health and Human Services on Thursday announced fresh funding to lend a hand improve the exploit and maintenance of artificial intelligence-enabled medical devices.
Studies have shown that machine-learning models used in clinical settings can degrade over time. The fresh HHS money is earmarked for the Advanced Research Projects Agency for Health, or ARPA-H, which will exploit it to innovate its operations to make AI tools more reliable for doctors and more beneficial for patients.
The initiative is known as the Performance and Reliability Evaluation of Continuous Modification and Usability of Artificial Intelligence, or PRECISE-AI.
WHY IS THIS IMPORTANT
According to ARPA-H, the U.S. Food and Drug Administration has approved more than 950 medical devices with integrated AI capabilities, a tenfold raise compared to 2018.
While fresh AI-based tools have the potential to transform what clinicians can do to care for patients, machine learning can degrade over time as input from clinical operations, data collection, patient populations, or IT infrastructure changes.
“The promise of AI-powered tools for healthcare is only as strong as the relevant real-world data that informs them,” Berkman Sahiner, PRECISE-AI program manager, said in a statement.
Current techniques require users to monitor and maintain the performance of AI-powered devices in real-world clinical settings.
“The inability to automatically monitor and maintain the performance of an AI-enabled medical device based on real-world operations creates risks for physicians and patients,” Sahiner added.
HHS said that through PRECISE-AI, it intends to create a repository of open-source tools that will support clinical decision-making using artificial intelligence by automatically identifying and correcting performance degradation – without the need for human oversight.
Root cause analysis tools must also provide clear and useful information about the sources of degradation to enable healthcare users to better interpret model uncertainty.
According to Sahiner, doctors can provide better care to patients by recommending “optimal approaches to detecting and mitigating the effects of degradation in the performance of the underlying AI model.”
The agency said the upcoming final call for proposals for the program will focus on five technical areas, and teams will begin by developing tools that will lend a hand establish the most right assessment of a patient’s diagnosis based on available evidence, or “objective truth.”
Teams will then develop autonomous methods to monitor AI tool performance against these basic facts, determine the root causes of degradation, and make any necessary corrections.
They will also create mechanisms and underlying data infrastructure to notify physicians, AI tool developers, hospital administrators and regulators about performance degradation.
Date of reply to solicitation The current date is January 15, 2025.
BIGGER TREND
As an independent arm of the National Institutes of Health, ARPA-H seeks to invest in ways to build stronger, healthier, and more resilient health systems.
Focused on building scalable platforms and strategies, from advanced mobile hospitals providing acute care in rural areas to securing open source software used in critical infrastructure.
As part of its Digital Health Security Initiative, DIGIHEALS, ARPA-H has partnered with the Defense Advanced Research Projects Agency on the Artificial Intelligence Cyber Challenge. In May, HHS announced $50 million in ARPA-H funding to lend a hand vendors patch ransomware vulnerabilities in networks and medical devices.
IN THE DOCUMENT
“PRECISE-AI addresses the growing gap in ensuring that AI tools used to make clinical decisions are accurate, safe, and robust in real-world settings,” ARPA-H Director Renee Wegrzyn said in the recruitment announcement. “In doing so, ARPA-H is establishing a foundation of trust between physicians and these AI tools, which will further expand AI’s potential to improve health outcomes for all Americans.”