According to Science (AAAS), artificial intelligence taught itself how to diagnosis heart attacks and performs better than doctors. Doctors have tools. Now, they may have another.
The human body is complex. Heart attacks are a problem. Computers and doctors working together perform better than doctors alone. It has been reported that “scientists have shown that computers capable of teaching themselves can perform even better than standard medical guidelines, significantly increasing prediction rates.”
The novel method might save 1,000 to 1,000,000 of lives per annum. A vascular surgeon at Stanford University, Elsie Ross, said, “I can’t stress enough how important it is…and how much I really hope that doctors start to embrace the use of artificial intelligence to assist us in care of patients.”
More than 20,000,000 people die from cardiovascular disease every year, specifically, “heart attacks, strokes, blocked arteries, and other circulatory system malfunctions.” The issue is prediction. Doctors need better extrapolation from diagnostics to know the probabilities of heart attacks.
The American College of Cardiology/American Heart Association (ACC/AHA) describes the 8 main factors or variables in the risk of heart attacks including, “age, cholesterol level, and blood pressure—that physicians effectively add up.”
Biology is complex, as is the human body. The human body can prevent cardiovascular problems with fat at times and it depends. An epidemiologist from the University of Nottingham in the UK, Stephen Weng, said, “What computer science allows us to do is to explore those associations.”
The recent research by Wend used the ACC/AHA guidelines. They were put through the rigours. The rigours of 4 machine-learning algorithms. Each analysed human amounts of data: 378,256 patient profiles in the UK.
“First, the artificial intelligence (AI) algorithms had to train themselves. They used about 78% of the data—some 295,267 records—to search for patterns and build their own internal ‘guidelines.’”
The AI tested themselves on the records left out of the 295,267 taken out of the 378,256. “…They predicted which patients would have their first cardiovascular event over the next 10 years, and checked the guesses against the 2015 records.”
The machine-learning methods took 22 points of data to make the extrapolation such as kidney disease or ethnicity.
“The best [machine-learning algorithm]—neural networks—correctly predicted 7.6% more events than the ACC/AHA method, and it raised 1.6% fewer false alarms. In the test sample of about 83,000 records, that amounts to 355 additional patients whose lives could have been saved,” Science said.
A data scientist from the University of Manchester, Evangelos Kontopantelis, considered the work important with the possibility of leading to greater gains. “Going forward, Weng hopes to include other lifestyle and genetic factors in computer algorithms to further improve their accuracy.”