New test with AI could help doctors diagnose heart attacks faster – research | tech news

A new test developed using artificial intelligence could help doctors diagnose heart attacks more quickly and accurately, according to a new study.

Researchers developing a computer algorithm hope it can reduce unnecessary admissions to busy A&E units — and stop the clinical bias that currently causes some women to miss life-saving treatment.

A trial of 10,286 patients with chest pain found that a diagnostic tool called CoDE-ACS was able to rule out heart attacks in twice as many patients as current tests, with 99.6 percent accuracy.

A clinical trial is currently underway in Scotland, backed by Wellcome Leap, to assess whether the tool can reduce the pressure on overcrowded emergency rooms.

Professor Nicholas Mills, Professor of Cardiology at the University of Edinburgh’s Center for Cardiovascular Sciences, who led the study, said: “For patients with acute chest pain from a heart attack, early diagnosis and treatment can save lives.

Please use Chrome for a more accessible video player

“We don’t understand how AI works”

“Unfortunately, many conditions can cause these common symptoms, and diagnosis is not always straightforward.

“Using data and artificial intelligence to support clinical decision-making has enormous potential to improve patient care and increase efficiency in our busy emergency departments.”

More about artificial intelligence

The current gold standard for diagnosing heart disease is measuring levels of the protein troponin in the blood.

But the same threshold is used for every patient — though levels can be affected by age, gender and other health conditions.

Previous research has shown that women are 50 percent more likely to be initially wrongly diagnosed. People who were initially wrongly diagnosed had a 70 percent increased risk of dying after 30 days.

But according to the British Heart Foundation, which funded the work, the new algorithm could prevent that from happening.

CoDE-ACS works well regardless of patient characteristics, according to research published in the journal Nature Medicine.

It was developed using artificial intelligence based on data from over 10,000 patients in Scotland.

It uses information including age, gender, EKG test results, medical history, and troponin levels to predict someone’s likelihood of having a heart attack.

Professor Sir Nilesh Samani, Medical Director of the British Heart Foundation, said: “The CoDE-ACS has the potential to rule out or rule out a heart attack more accurately than current methods.

read more:
Geoffrey Hinton: Who is the “Godfather of AI”?
AI market faces probe from UK competition watchdog

“This could be transformative for emergency departments, reducing the time it takes to make a diagnosis and better for patients.”

Professor Steve Goodaker, professor of emergency medicine at the University of Sheffield, called the study “interesting”, adding that it showed “how artificial intelligence can use complex analysis, rather than simple rules, to improve diagnosis”.

No matter where you get the podcast, you can click to subscribe to Sky News Daily

“It’s not [yet] Showing that we can replace doctors with computers,” he added. “Experienced clinicians know that diagnosis is a complex business.

“Indeed, the ‘ground truth’ used to judge whether an AI algorithm is accurate is a judgment made by clinicians.”

Source link