An artificial intelligence(人工智能) system has shown it can save lives by warning doctors to check on patients whose heart test results show a high risk of dying. In a test with almost 16,000patients, the AI helped reduce the total deaths by 31% among people who were at high risk.
"This is actually quite amazing,"says Eric Topol at the Scripps Research TranslationalInstitute in California."It is quite uncommon for any drug to result in a 31% reduction in the number of deaths, and then even more unusual for a non-drug—this is just monitoring people withAI."
Chin Lin at Texas Medical Center and his colleagues trained their AI on more than 450,000 electrocardiogram(ECG) (心电图) tests, which measure the heart's electrical activity, along with the survival data of the ECG subjects.The AI learned to produce a percentile (百分位) score representing each patient's risk of death, whose score in the 95th percentile or higher would be considered high risk.
The researchers then tested the AI warning system with 39 doctors at two different hospitals.As nurses transmitted each new patient's ECG tests to a computer server, the AI system would analyse the results and warn doctors if it identified someone as being high risk. In addition to reducing the risk of death from all causes, the AI system also reduced the death rate from heart issues in high-risk patients by more than 90%.
Lin and his colleagues say the AI warnings may help focus attention on the highest risk group.During the research, doctors usually followed up warnings with more tests and treatment. The system is already being used across 14 hospitals in Texas."This can be put into effect in every hospital around the world," says Topol."It should be the standard of care when you see such a significant benefit."
