Each year, over 20 million people die from heart attacks caused by cardiovascular disease. One of the reasons this keeps happening is due to doctors not having a better way to predict when a person might have a heart attack.

As it stands right now, a doctor is the only individual capable of summarizing whether or not a patient could have a heart attack soon. However, things could change in the coming years as a team from the University of Nottingham in the UK is working on ways to make the task easier.

According to the latest report, the team established a machine-learning algorithm designed to predict when folks would have a heart attack or even a stroke. The algorithm should do a better job than any doctor could, but this is yet to be proven in the field.

Interestingly enough, the artificial intelligence in use here is no slouch since it’s capable of predicting a person’s risk of heart attack at the rate of 72.8 percent. That’s quite impressive for a computer, but it only gets better from here.

Advertisement

The leader of the team behind this AI, Stephen Weng, is not satisfied with the results, and as such, he went ahead to make it much better than before. The report claims the team created four computers designed with machine learning algorithm, then fed them data based on nearly 400,000 patients headquartered in the United Kingdom.

At first, the AI went through 295,000 patient’s data to generate the required internal predictive information for each. After that, the system used the remaining to refine its findings. This means the algorithm managed to outpace the AAA/AHA guidelines.

Increase in accuracy

The AI now has a 76.4 percent accuracy, much better than any human doctor. The report says it beats the current guideline by 7.6 percent, while false alarms are below 2 percent.

Now, when we look at what this system has done, it’s clear it could have saved an extra 355 lives. This is reason enough to refine the AI better and get it on the field as soon as possible.

We should point out how the AI is capable of predicting oral corticosteroids and severe mental illness. Bear in mind these things are not part of the existing guidelines, which means, the AI is moving into uncharted territories.

With more refinements, we expect to see a rise of artificial intelligence in the health industry. It will only mean good things for the longevity of human beings, and less reliance on doctors who are more likely to make a mistake.

[See More: Are Artificial Intelligence Systems Learning to Be Racist?]

Comments

comments

LEAVE A REPLY

Please enter your comment!
Please enter your name here