Last Updated: 10 February 2025
A recent study published in Nature Medicine from Lund University, with funding from the Heart Lung Foundation, has revealed that AI is significantly more effective than human analysts at detecting dangerous heart rhythm disorders.
The research suggests that integrating AI into clinical practice could lead to faster and more cost-effective diagnoses, potentially improving patient outcomes.
Atrial fibrillation, the most common heart rhythm disorder, carries serious risks such as stroke, dementia, heart failure, and premature death. Early detection using long-term ECG monitoring has become a key tool in identifying these conditions before severe complications arise.
"Our research shows that AI was much better than human biomedical analysts in finding dangerous rhythm disorders. It missed 14 times fewer diagnoses than people did".
Linda Johnson, associate professor at Lund University
The study involved 14,606 patients and assessments from over 50 cardiologists worldwide, making it the largest of its kind for long-term ECG interpretation.
The AI algorithm demonstrated impressive accuracy, missing only 0.3 percent of serious arrhythmia cases compared to 4.4 percent for human analysts—a remarkable improvement despite a modest increase in false-positive results.
These findings not only underline the potential of AI to transform diagnostic processes but also pave the way for more streamlined and accessible healthcare services. By reducing diagnostic errors, the implementation of AI could alleviate the burden on clinicians and accelerate treatment decisions.
“We hope that these research results will lead to faster, cheaper and better diagnosis of rhythm disorders such as atrial fibrillation, which can reduce the risk of stroke and save lives"
Kristina Sparreljung, Secretary General of the Heart Lung Foundation
Heart rhythm disorders, collectively known as arrhythmias, encompass a wide spectrum of conditions. While occasional extra heartbeats are often harmless, more severe disturbances can drastically affect a person’s quality of life and even lead to cardiac arrest. Atrial fibrillation, in particular, is so prevalent that it has been labeled a “folk disease” in some regions, with a significant portion of the aging population affected.
This study is a promising step forward, and the next phase involves testing the AI algorithm in a larger clinical trial. Such research is essential to confirm its benefits in everyday healthcare settings and to explore new diagnostic and treatment methods for rhythm disorders.
Source: My Newsdesk
