November 21, 2022 — Artificial intelligence can improve diagnosis and treatment for patients, but first AI-enabled clinical tools must be readily available and used.
New research from Mayo clinic finds that clinicians who were high adopters of an AI-assisted clinical decision-making tool were twice as likely to diagnose low left ventricle ejection fraction as low adopters of the tool. The study, published in Mayo Clinic Procedures, found large differences in the adoption rate of AI recommendations. Physicians who adopted a lot generally had less experience in dealing with patients with complex health problems, but age, gender, years of experience and the number of patients cared for were not significant factors.
“It was surprising to see the significant difference in the rate of diagnosis between high adopters and low adopters,” he says David Rushlow, MDa Mayo Clinic physician and president of Family Medicine for Mayo Clinic in the Midwest. “The tool is extremely useful, but we didn’t expect to see a full doubling of the diagnosis rate of low ejection fraction compared to low adopters.”
Ejection fraction measures the percentage of blood that leaves the heart each time it contracts. A low ejection fraction can be caused by heart muscle weakness, such as cardiomyopathy, as well as heart valve problems, uncontrolled high blood pressure, or damage caused by a heart attack.
Early diagnosis and treatment in patients with a low ejection fraction is critical to reduce the risk of symptomatic heart failure, hospitalization and mortality. “AI decision support tools have the potential to be highly effective in helping diagnose serious health problems before the usual clinical symptoms appear, and can outperform traditional diagnostic approaches,” said Dr. Rushlow.
Physicians from 48 Mayo Clinic general practices in Minnesota and Wisconsin participated in the randomized controlled trial, which involved 358 physicians, nurse specialists and physician assistants, of whom 165 clinicians were randomly assigned to the AI arm and included in the current acceptance analysis. The AI algorithm was run on 22,641 patients with a electrocardiogram (ECG) performed between August 5, 2019 and March 31, 2020. The clinicians randomized to the intervention group had access to the screening report, which showed the AI-ECG screening as positive or negative; the clinicians randomized to usual care had no access.
When the report was negative, further testing was not recommended, but when it was positive, the recommendation was “consider getting a echocardiogramThe clinicians also received an email alert when the AI-ECG screening was positive, indicating that patients had a high likelihood of previously unrecognized low ejection fraction.
“Clinicians most likely to follow AI decision aid recommendations were generally less experienced in dealing with complex patients,” says Dr Rushlow. “This demonstrates the importance of AI systems that integrate seamlessly into clinicians’ workflows. Given the technical nature of AI in healthcare, it is often initiated and developed in academic specialty practices. To maximize the benefits of AI, more collaboration is needed between specialist practices and primary care.”
Mayo Clinic has a patent on the AI technology and can derive financial benefits from it, but it will not benefit financially from using it in the care of patients at Mayo Clinic. Co-authors Itzhak Attia, Ph.D., Paul Friedman, physicianand Francisco Lopez-Jimenez, MD, may also receive financial benefits from this agreement. The other co-authors report no conflicts of interest.
The study was supported in part by Mayo Clinic’s Robert D. and Patricia E. Kern Center for the Science of Health Care Delivery.
For more information: https://www.mayoclinic.org/