A major study confirms that artificial intelligence surpasses human capabilities for analyzing extensive heart rhythm monitors.
Researchers note in their Nature Medicine study the extensive amount of time required for human technicians to read electrocardiograms (ECGs) that document daily to weekly heart beat patterns because the heart beats around 120,000 times per day.
A team analyzed recording data from 14,606 patients with average ECG monitoring times of 14 days to begin with data obtained from standard human technician analyses.
The scientists performed a new analysis on the data through DeepRhythm AI which Medical Logrithmics from Poland developed to fulfil this purpose.
AI analysis showed superior results in detecting severe arrythmias with a rate of 0.3% compared to human technicians who failed to identify such problems in 4.4% of patients.
During a 14-day ECG recording session the AI model achieved absolute confidence at 99.9% in identifying no severe arrhythmias.
The shortage of qualified personnel who perform walking ECG analysis creates a global healthcare processing backlog according to Linda Johnson from Lund University who led the research team in Sweden. Patients would obtain better outcomes when electrocardiograms are recorded across extended walking periods instead of limited durations.
The researcher insisted artificial intelligence offers a suitable solution to this problem.
