A team of researchers at the University of Pittsburgh has developed a new artificial intelligence system that can use electrocardiogram (ECG) readings to diagnose heart attacks with greater accuracy and speed.
“When a patient goes to the hospital with chest pain, we first ask whether the patient is experiencing a heart attack. The process of reading the ECG and completing the necessary medical examinations for diagnosis takes about 24 hours”, said Salah Al-Zaiti, a researcher specializing in cardiac diseases at the School of Nursing and Emergency Medicine at the University of Pittsburgh.
“The new artificial intelligence model addresses these important challenges by improving the ability to assess risks, ensuring that patients receive the healthcare they need without delay”, he added.
By analyzing the waveform patterns of the ECG, doctors can usually diagnose the type of heart attack the patient is experiencing. However, the new artificial intelligence system can detect fine details in the ECG that are sometimes difficult for doctors to identify when diagnosing the cause of chest pain in patients.
According to “Nature”, where the new study on the artificial intelligence system was published, the research team compared the new computational model with three main criteria for diagnosing heart attacks: clinical readings of ECGs, specialized algorithms for reading ECGs, and diagnostic data such as medical history, smoking, cholesterol levels, and diabetes.
The researchers found that the new system outperformed the three criteria in diagnosing heart attacks, categorizing patients with chest pain into three categories.
