Citizen Portal
Sign In

Researchers tell House veterans committee AI tools can identify and help treat sleep disorders common among veterans

2845084 · April 2, 2025

Get AI-powered insights, summaries, and transcripts

Subscribe
AI-Generated Content: All content on this page was generated by AI to highlight key points from the meeting. For complete details and context, we recommend watching the full video. so we can fix them.

Summary

University of Arizona and other witnesses described AI and machine-learning tools to detect sleep apnea and support adherence to therapy; witnesses emphasized implementation science and the need for clinician workflow integration to realize benefits.

Dr. Sairam Parthasarathy, chief of pulmonary, allergy, critical care and sleep medicine at the University of Arizona, told the committee that sleep and circadian research yields cross‑cutting benefits for veterans and that implementation science is necessary to bring algorithms into routine care.

Parthasarathy said many veterans have undiagnosed sleep apnea and that AI/ML algorithms embedded in electronic health records can flag patients at high risk and prompt clinicians to order diagnostic testing. He also described patient-facing interventions, funded in part by VA research and PCORI, that use automated prompts and behavioral supports to improve adherence to CPAP and other therapies.

"We have refined machine based algorithms that are developed by various researchers, including at MIT, that embed into electronic medical records, within the system of the University of Arizona health care system, which can identify individuals with high likelihood of sleep apnea and alert health care providers to the fact that they have it and enable them in a facile manner to play diagnostic tests and enable treatments to be brought to fore," Parthasarathy said.

Why this matters: Sleep disorders increase risks of heart attacks, strokes, accidents and worsen PTSD and traumatic brain injury recovery; better case-finding and adherence support could reduce hospitalizations and deaths. Parthasarathy urged that implementation planning consider provider workload and local clinical capacity, and that VA and its university partners are well positioned to test and scale these approaches.

Ending: Witnesses and lawmakers framed AI-enabled case finding and adherence tools as promising but said successful rollout requires protected time for clinicians, integration with EHRs and reliable staffing to follow up on alerts.