The Centers for Medicare and Medicaid Services (CMS) wants to make patients’ medical data more accessible in real time, so they’re better informed about their health and care options, according to its deputy administrator and chief of staff.
Speaking Tuesday at NVIDIA GTC in Washington, D.C., Stephanie Carlton said the Trump administration is working with the healthcare industry to improve data accessibility rather than change regulations.
Carlton was introduced to electronic health records as a labor and delivery nurse, when the hospital she worked at began adopting one in 2002, but the process wasn’t smooth. This is one reason the sector has lagged others in adopting new technologies. A notable exception is artificial intelligence; in this area, the healthcare industry is a leading adopter, as the tech can improve work-life balance.
“That’s life changing for clinicians that are burnt out at this point,” Carlton said.
Click the banner below to start implementing smarter security.
Addressing Clinician Burnout With Agentic AI
A 2021 American Medical Association-led study found that 1 in 5 physicians and 2 in 5 nurses said they intend to leave practice altogether, in part due to burnout.
The age of agentic AI brings with it the promise of automating many of the jobs clinicians don’t want to do, such as clerical and back-office work.
“We have this opportunity now to leverage this technology to help healthcare be more human again,” said Dr. Shiv Rao, CEO and cofounder of Abridge.
Abridge predicts that low-acuity care could be automated within the next decade, but the rest of healthcare will not be. That means that when people get very sick, conversations with doctors and nurses will drive workflows, Rao says.
The Abridge tool uses generative AI and ambient listening to capture and summarize clinicians’ conversations with patients into detailed notes. These are high-value artifacts because they are paid based on care documented rather than care administered.
Building Clinician Trust in AI Healthcare Tools
“I think we’re ripe with all that data that exists,” said Dr. Kent Thielen, CEO of Mayo Clinic in Florida.
Mayo Clinic boasts troves of data, dating back to the Mayo brothers more than a century ago, including digitized slides of biopsies from every pathology patient in the Mayo Clinic Platform. These data sets are used to train AI-enabled tools to do everything from detecting pancreatic cancer more quickly to calculating the “biological age” of a patient’s heart. In turn, such tools, and the lives they save, generate increased interest in using AI in healthcare.
Providers “make a decision fairly quickly about whether these tools work for them,” Thielen said.
Additionally, Mayo Clinic is working with health AI startups to accelerate its work while validating the tools it is developing. This further builds trust in the technology.
For its part, Abridge is building on that trust. The company allows end users to audit their audio and has published papers on its benchmarks for transparency. While Abridge is scaling quickly, it publicly self-grades its IT infrastructure.
“Trust is the only currency that matters in healthcare,” Rao says.
Click the banner below for the latest federal IT and cybersecurity insights.
Ensuring Patient Consent and Confidentiality in the Agentic Age
As CMS works with industry to expand medical data access with AI, it’s ensuring that patient consent and confidentiality remain part of the process, Carlton says.
The agency plans to let industry build AI applications while it fosters an environment where they’ll thrive, establishing foundational safety criteria in the process. What those criteria look like remains to be seen, however.
“We want to promote innovation because we feel it’s important to improving health and longevity,” Carlton says. “But we want to balance that with safety.”
