In two studies, artificial intelligence was used with electrocardiogram (ECG) results to identify patients who are at increased risk for a potentially risky irregular heartbeat, and those more likely to die within a year, researchers say.
We have an AI technique that can forecast volcanic eruption, and an algorithm that can write believable fake news within seconds. Patients predicted to develop a-fib within one year were 45% more likely to develop the disorder over 25 years than other patients. Yes, that exists now. To that end, researchers from Pennsylvania healthcare provider, Geisinger have trained an AI to predict which patients are at risk of dying within a course of a year, reports New Scientist.
So, how does it work? An ECG is a simple test that can be used to check a patient's heart rhythm and electrical activity as a graph of voltage over time. So, patients with cardiac conditions such as heart attacks and atrial fibrillation show a noticeable pattern change.
Fornwalt and his colleagues trained two versions of the AI.
Two tests were carried out with one machine taught to only examine the raw data from the ECGs while the other was fed information about the people it related to - their sex, age.
Then, they tasked the AI models with examining 1.77 million ECG results from nearly 400,000 people. Even for ECG outcomes that cardiologists made a decision to be regular, the AI was capable of decide up on different patterns and precisely predict deadly well being dangers inside a 12 months's time. A score can range from 0 to 1, where 1 is where predictions are 100% correct, and 0.5 implies that it can't distinguish between the patients living more than a year and those who die within a year.
When one is a flawless score matching every ECG with the correct outcome, the machine looking at the raw data alone scored eight point five, achieving a clear distinction between the groups - five is no distinction at all. The AUC scoring models now used by doctors range between 0.65 and 0.8, says Fornwalt. What's more, three cardiologists who separately reviewed seemingly ordinary ECG data missed the risk patterns that the AI picked up.
Dr Fornwalt added: 'That finding suggests that the model is seeing things that humans probably can't see, or at least that we just ignore and think are normal. "This could completely alter the way we interpret ECGs in the future".