Emergency medicine physicians from The Permanente Medical Group in Northern California now have the ability to use an augmented intelligence (AI) risk tool that is seamlessly integrated into electronic medical records to improve care of patients with heart failure.
People with heart failure often have multiple medical problems and take a number of medications. Consequently, it can be difficult for emergency medicine physicians to fully assess these patients’ risk of experiencing an adverse event.
Dana R. Sax, MD, MPH, a research scientist at the Kaiser Permanente Division of Research and an emergency medicine physician with The Permanente Medical Group, knew this difficulty firsthand—and embarked on a decade-long research study to develop a way to identify the low-risk patients who might not need to be hospitalized and could continue receiving care outside of the hospital.
That’s where the capabilities that AI can offer came into play.
“We sought to develop a novel tool to risk-stratify these patients,” Dr. Sax explained during a recent episode of “AMA Update.”
“The tool works in real time at the point of care and it uses about 60 different variables from our electronic medical record and a machine learning model to predict risk of an adverse event. The scores are presented to the treating provider at the point of care, streamlined into the workflow of the emergency provider,” she said.
The Permanente Medical Group is a member of the AMA Health System Program, which provides enterprise solutions to equip leadership, physicians and care teams with resources to help drive the future of medicine.
While the AMA had no role in the development in this AI risk-stratification tool for heart failure, you can learn more about how the AMA is fighting to make technology work for physicians—seeking to ensure that in areas such as AI implementation and EHR usability, tech is an asset to doctors and not a burden.
A tool to help physicians
Dr. Sax said the risk tool considers the patient’s historical data—comorbidities, medications, most recent ejection fraction, and their recent emergency department, hospital and intensive care unit visits and stays.
It also takes their acute physiology into account—their vital signs, renal function, and key laboratory values, as well as their weight and blood pressure in the emergency department compared with baseline.
“It pulls in all of these and then calculates a risk class in real time,” Dr. Sax said.
Through a series of research studies, Dr. Sax and her colleagues developed and tested the tool and then studied it retrospectively and prospectively. They also piloted the study in two Kaiser Permanente Northern California emergency departments. Their most recent research, published in JACC: Heart Failure, set the stage for the tool to be made available to emergency medicine physicians in all 21 of the integrated health system’s Northern California hospitals.
Stay up to date on AI
Follow the latest news on AI and its applications and effects for health care—delivered to your inbox.
“We found that among patients that the tool identified as low risk and who were discharged home, there were no 30-day adverse events, suggesting that the risk estimates and the guidance to consider discharge was a safe recommendation,” Dr. Sax said. “We also found the tool gives some guidance on medical management for these patients, and increased the proportion of patients who are on guideline-directed medical therapy.”
This work gives time-pressed emergency physicians “one additional tool at the point of care,” Dr. Sax said. “There’s a wide range of data that the AI tool pulls in. It would often be very difficult in a very busy emergency department to look at all that information in an efficient manner.”
Dr. Sax said she and her colleagues will continue to evaluate the risk tool. They also expect to further improve the tool based on physician feedback. Going forward, Dr. Sax said, their team may also look at the potential of using the risk tool in the inpatient setting or when arranging transitions of care for patients going home.
Continued growth of AI
More tools like the one Kaiser Permanente is using are expected to help physicians in the coming years as AI use begins to soar. Among physicians the AMA surveyed, 66%, reported using health care AI in 2024. That’s a 78% jump from the 38% of physicians who said they used it when the AMA conducted a survey in 2023.
The AMA is fighting on the legislative front to help ensure that technology is an asset to physicians and not a burden and it has developed advocacy principles (PDF) that address the development, deployment and use of health care AI.
Learn more about applying AI to transform health care solutions with the “AMA ChangeMedEd® Artificial Intelligence in Health Care Series.”
“AMA Update” is your source for physician-focused news. Hear from physicians and other experts on trending public health concerns, practice issues and more—because who’s doing the talking matters. Catch every episode by subscribing to the AMA’s YouTube channel or listen to all AMA podcasts at ama-assn.org/podcasts.
