US Pacific Fleet/Flickr
An artificial intelligence (AI) system has proven capable of saving lives by prompting doctors to check on patients whose heart test results indicated an elevated risk of death.
In a clinical trial with around 16,000 patients in two hospitals in Taiwan, a AI predicted and reduced by 31% the total number of deaths among high-risk patients.
The study that made it possible to reach the aforementioned number was carried out by researchers from the National Defense Medical Center in Taiwan and published this Monday, in Nature Medicine.
Across 450,000 electrocardiogram tests (ECG), the AI learned to produce the percentage that represents the risk of death for each patient – those who are at least in the percentile 95 considered to be high risk.
As ECG tests were performed, the AI system analyzed the results and quickly alerted the 39 doctors involved in the investigation, if the patient was identified as being at high risk.
As New Scientist reveals, in addition to reducing the risk of death from all causes by 31%, this AI reduced deaths by more than 90% of high-risk patients due to heart problems. The same magazine states that this system is already being used in 14 military hospitals in Taiwan.
“This is extraordinary (…) It is very rare for a medicine to produce a 31% reduction in mortality and even rarer than a non-medicine – this is just monitoring people with AI”, he praises Eric Topolfrom the Scripps Research Translational Institute, in California (USA), cited by New Scientist.
“This can be implemented in all hospitals around the world – it shouldn’t be expensive (…) It should be the standard of care when you see this magnitude of benefits”, added the specialist.
