New Machine Learning Tool Detects Heart Attack with 100% Accuracy

In Scotland hospitals, they experience software based on artificial intelligence, which can with high accuracy to identify a heart attack in the patient. The purpose of the project is to improve the speed and quality of medical care, as well as reduce mortality from myocardial infarction.

Cardiac attacks are difficult to diagnose, because symptoms, such as chest pain, dizziness or shortness of breath, are characteristic of many other diseases. Patients and even doctors themselves can often write off the identified symptoms for something else, which is why they miss the precious time. But if you do not have timely treatment, then the probability of an imminent death from this ailment is about 70 percent.

In the development of a tool called Code-Acs (Collaboration for the Diagnosis and Evaluation of Acute Coronary Syndrome), a serious contribution of the British Fund (BHF), which provided the necessary investments.

The developed tool was trained on data from above 10 thousand patients hospitalized in Scotland with suspicion of a heart attack. According to researchers, Code-AC is able to identify myocardial infarction with a probability of 99.6 percent.

The tool analyzes the age, gender, medical history of the patient, studies the data of electrocardiograms and uses a blood test to search for troponin – protein, which is produced when the muscles of the heart are damaged. The tool expresses its analysis in assessment from 0 to 100, and the higher it is, the higher the chance of a heart attack.

“For patients with acute chest pain due to heart attack, early diagnosis and treatment saves life,” explained Nicholas Mills, a professor of cardiology at the center of the cardiovascular science of the University of Edinburgh, who headed the study published in the journal Nature.

Researchers believe that this algorithm will allow doctors to quickly identify people who contact the departments of emergency medical care with a real cordial attack for life. This will strongly save time in treatment, and also help to weed out patients with low risks for health and life in order to devote time to those patients who cannot wait.

It is surprising to observe such innovative developments from the outside. It is possible that already in the near future many diseases dangerous for human life will be able to identify simple fitness braids or other wearable equipment, without the need for a medical examination.

Specialists from Philips, for example, have recently shown the world smart watches capable of determining the entering the human body of various infections. Surely the number of such developments will only increase over time, every day promoting the possibilities of medicine further and further.

/Reports, release notes, official announcements.