Picture this: A disoriented patient arrives at University of Vermont Health - Porter Medical Center, worried about a potential stroke. After an initial assessment, a physician's assistant orders a special CT scan called a computed tomography angiogram (CTA) to check for artery blockages in the brain.

Seconds after the CTA scan is complete, an artificial intelligence (AI) program named Rapid AI reviews the results and detects signs of a possible stroke in progress. Without delay, it notifies an on-call neurologist at UVM Medical Center, who is part of the 24/7 UVM Medical Center Stroke Team. The AI program even highlights the specific area of the brain that the neurologist should focus on.

"Sometimes we're actually able to look at the image before a radiologist,” explains Christopher Commichau, MD, director of stroke and neurocritical care at UVM Medical Center. “If there is an issue that clearly needs major intervention, the on-call neurologist gets the medical chopper on standby to bring the patient to UVM Medical Center for surgery and calls the ED provider to start an IV of 'clot busting' medications."

That is just one example of how artificial intelligence is improving health care. It can help in less critical situations as well. If a physician’s assistant at one of the UVM Health's hospitals requests a tele-stroke consult, a UVM Medical Center stroke specialist can review a scan or patient information remotely and help create a comprehensive care plan for the patient.

The ability to perform a CTA scan alongside a traditional CT scan is a game-changer, says Dr. Commichau. "The ability to identify large vessel occlusions early on has made a significant difference for our patients, especially in smaller communities where access to this type of imaging was once limited. It means less waiting when patients have a condition where time-to-treatment matters."

Artificial Intelligence in Health Care

Rapid AI was first introduced at UVM Health in 2018 as a cutting-edge technology to reduce the time from testing to treatment. Since then, other AI systems have helped improve patient care and streamline health care processes.

Iodine, a natural-language and data-mining program, scours inpatient medical records for potential diagnosis gaps. It reviews clinical notes, test results, diagnoses and medications, and flags areas where additional documentation might be needed. This extra precaution enhances diagnostic accuracy and saves time for medical providers.

AI is also helping tackle one of the major pain points in health care: documentation and charting. One of the most time-consuming tasks for health care providers is writing visit notes for the patient’s medical record. Often, documentation cuts into time seeing other patients, or is done late at night after a long day's work, leading to exhaustion and burnout.

Later this month, the UVM Health will begin testing an AI tool that, with a patient's consent, audio records the patient’s visit with their provider and converts that recording to visit notes. The provider will then review and edit the notes, if needed, before they are approved and added to the patient’s medical record.

“The less time they're spending entering data, the more time they can spend with their patients, says Todd Young, associate vice president of digital health services. “That's the whole reason they're here."

AI Is No Substitute for Clinical Judgment

UVM Health understands a healthy skepticism toward AI, as well as the importance of ethical AI usage and patient data privacy. So it has established a robust framework of policies and principles to govern AI implementation, ensuring safe, equitable and responsible use. AI can process vast amounts of data to gain valuable insights and efficiencies, but it is not yet at the point where it can or should replace clinical judgment.

“Our commitment lies in a thoughtful and responsible approach to AI,” says Douglas Gentile, senior vice president of network information technology. “By addressing ethical concerns, ensuring data security and promoting transparent governance, we’ve created a framework that safeguards patients and ensures that our teams use these tools to provide better care.”

Especially in clinical care, Gentile says, AI needs to be designed as a complement to medical expertise. “AI may generate a draft or an alert,” says Gentile, “but an expert clinician needs to review the output and then take appropriate action.”

The Future of AI in Health Care

As for the future, AI may soon assist in patient communication. It may also help streamline check-in processes.

Dr. Commichau sums up this incredible transformation, saying, "I finished my neurology fellowship about 25 years ago, and back then, I never imagined tools like these. Now it just seems like a natural evolution. This technology is just going to keep getting better and better."