Machine Learning Outwits Doctors in Predicting Heart Attacks

Machine learning technology can assess the health of the heart far more accurately than humans, a new study has found.

Doctors predict heart attacks with the help of ‘risk scores’ based on a handful of variables, but which can be inaccurate. However, machine learning algorithms can analyze large amounts of data, taking into account a lot more variables, and which can make its predictions more accurate than those carried out by humans.

“The algorithm progressively learns from the data and after numerous rounds of analyses, it figures out the high dimensional patterns that should be used to efficiently identify patients who have the event,” said Dr Luis Eduardo Juarez-Orozco, of the Turku PET Center, who authored the report.

“These advances are far beyond what has been done in medicine. We have the data but we are not using it to its full potential yet,” he said.

The study looked at 950 patients with heart disease and, during an average six-year follow-up, 24 of them suffered heart attacks, and 49 of them died. Doctors then entered all the data into a machine learning algorithm called LogitBoost, which analyzed them over and over again and identified which of the participants had died or suffered heart attacks.

“We found that machine learning can integrate the data and accurately predict individual risk,” Juarez-Orozco added.

“This should allow us to personalize treatment and ultimately lead to better outcomes for patients.”