A team of researchers from Stanford University in the USA. UU He has developed a computer program that can predict when a hospital patient will die with a baffling level of accuracy.
According to the team, your program is correct 90% of the time.
Will be used to provide an appropriate end-of-life car to hospital patients with a terminal illness.
The program uses artificial intelligence and was trained by analyzing 160,000 files of adult patients and children from Stanford and Lucile Packard. Children's hospital.
Things were observed in the health records, such as the diagnosis, the procedures performed and what medicine a patient was taking.
When the algorithm was applied to 40,000 active patients (it was asked to predict what would die in the next three to twelve months) it was correct in 90% of the cases.
"The scale of available data allowed us to build a prediction model of all-cause mortality, rather than being disease-specific or demographic," said Anand Avati, a member of the IA laboratory at Stanford University. IBTimes.
The researchers plan to continue refining the system with more data before it is implemented in hospitals and medical personnel.
Kenneth Jung, research scientist at Stanford University said: "We believe that keeping a doctor in the know and thinking about this as 'machine learning plus the doctor' is the way to go instead of doing blindly medical interventions based on algorithms … that puts us on a firmer ground both ethically and safety "