Use of AI-Enabled Screening Tool Led to More Diagnoses of Heart Condition

A recent Mayo Clinic study found that the use of an artificial intelligence (AI)-enabled screening tool displayed the ability to assist providers while diagnosing low left ventricular ejection fraction, with high adopters of the technology diagnosing the heart condition more frequently than those who did not use the tool.

According to the Centers for Disease Control and Prevention (CDC), about 6.2 million adults in the US experience heart failure, which occurs when the heart cannot pump enough blood and oxygen to support organs in the body. Low left ventricular ejection fraction is a heart condition where the percentage of blood that pumps out of the heart is low. The condition can be caused by the weakening of heart muscles, issues related to heart valves, uncontrolled blood pressure, or damage that results from a heart attack.

According to the press release, Mayo Clinic researchers acknowledged that early diagnosis is highly beneficial when addressing this condition, as it can lead to a lower risk of heart failure and mortality. This led them to experiment with the use of an AI-enabled tool.

"AI decision support tools have the potential to be very effective in aiding the diagnosis of serious health conditions before the onset of usual clinical symptoms, and may outperform traditional diagnostic approaches," said David Rushlow, MD, a Mayo Clinic physician and chair of family medicine for Mayo Clinic in the Midwest, in a press release.

To test the AI-enabled tool, clinicians from 48 different Mayo Clinic primary care practices in Minnesota and Wisconsin took part in a randomized controlled trial. Additionally, the study involved 358 physicians, nurse practitioners, and physician assistants, 165 of whom were randomized to the AI arm.

Researchers then ran the AI algorithm on 22,641 patients, all of whom received an electrocardiogram (ECG) between Aug. 5, 2019, and March 31, 2020. Clinicians in the group that used the AI tool, known as the intervention group, had access to the screening report that displayed the AI-ECG screening as positive or negative. Clinicians in the usual care group did not have access to this report.

If the tool determined that a report was positive, clinicians received a recommendation to order an ECG. They also received an email alert if AI-ECG screening was positive, which indicated that the odds of the patient having a previously unrecognized low ejection fraction was high.

Researchers concluded that clinicians who used an AI-enabled clinical decision support tool had a much higher chance of successfully diagnosing low left ventricular ejection fraction compared to those who did not use it. It also led researchers to recommend further collaboration between specialty practices and primary care.

These findings add to research previously conducted by Mayo Clinic related to the use of AI when detecting heart conditions.

In September, the organization tested an AI-based screening strategy that aimed to identify new cases of atrial fibrillation following the evaluation of ECGs.

This study included over 1,000 patients who participated in continuous patient monitoring and over 1,000 who participated in usual care. Following the study, researchers concluded that the tool enabled increased atrial fibrillation detection, leading to the identification of high-risk patients.