Researchers from Geisinger Health System have trained an AI (Artificial Intelligence) to predict which patients are at a high risk or dying within the next year. The AI was fed 1.77 million electrocardiogram logs from 400,000 patients to detect patterns that can point to future cardiac problems like heart attacks and atrial fibrillation. The AI performed better than the traditional approaches taken by cardiologists.
The results are very impressive and a little scary. It even detected cardiac problems in patients who had been cleared by doctors. The cardiac risk patient’s ECGs were reviewed by three cardiologists separately and they were unable to figure out the risk patterns that the AI detected. “This is the most important finding of this study. This could completely alter the way we interpret ECGs in the future,” said Brandon Fornwalt, chair of the Department of Imaging Science and Innovation at Geisinger in Danville, Pennsylvania. Another study was done by the same group of researchers that found that AI based models can analyse ECG test results and pin point patients at high risk of developing arrhythmia.
Both studies are one of the few to use AI to predict future health issues from an ECG instead of detecting current health problems. “This is exciting and provides more evidence that we are on the verge of a revolution in medicine where computers will be working alongside physicians to improve patient care,” said Fornwalt.
However, there’s a major catch with the AI: they are struggling to explain how the AI works. Professionals are worried about making any decisions based on the AI algorithm.