Most cardiovascular diseases are preceded by symptoms. Symptoms that, in some cases, occur even months or years before that disease is triggered. An Artificial Intelligence Predicts Heart Attacks Better Than a Doctor
Doctors use these symptoms to predict the possibility of a patient developing a particular cardiovascular disease, including a heart attack.
Every year, 20 million people worldwide die from a heart-related illness. Preventing it before it develops is key to avoiding thousands of deaths.
Researchers at the University of Nottingham have developed an artificial intelligence algorithm that predicts the occurrence of cardiovascular disease in patients, with a success rate of 5% higher than doctors themselves.
The American Heart Association and the American College of Cardiology (ACC / AHA) have created a guide to assess the risk of cardiovascular disease based on eight factors such as age, cholesterol, or blood pressure. US doctors use this guide to predicting whether patients at risk will suffer a heart attack or any other heart disease. His hit rate is a scary 72.8%. That is, three out of four people who are at high risk for a heart condition, end up suffering.
Researcher Stephen Weng and his team decided to develop an artificial intelligence algorithm that would improve this brand. To do this the software studied the medical history of 378,256 British patients. With this information, the AI created its own predictive guide. And applied it to more than 80,000 patients whose data on heart disease were known (to know if the program was successful or not).
Artificial Intelligence Predicts Heart Attacks
This artificial intelligence algorithm that predicts heart attacks hit between 74.5 and 76.4% of the time. Exceeding by more than 5% success rate doctors, as well as reducing by 1.6% false alarms.
If the medical authorities had used the algorithm in those 80,000 patients instead of the doctors’ own conclusions. An additional 355 lives could have been saved. In addition AI was able to predict diseases that were not in the initial parameters, such as mental illness and consumption of corticosteroids.
One more example of artificial intelligence promises to revolutionize all sciences in the coming decades. How far will it take us?