A German research team has reportedly built an artificial intelligence (AI) tool which can predict how a person will react after being hospitalised based on a blood sample, reports WION News.
Reports suggest that the machine learning tool can precisely predict whether the patient will survive or die based on levels of protein present in a person's blood.
The tool is meant to assist health workers to understand whether a patient needs critical care or can fight the virus without requiring much help. The prediction model identifies 14 levels of protein in a person before coming a decision.
Researchers say they tested the tool on 50 patients at Charité University Hospital in Germany as it conducted machine learning. Reports state 15 patients died out of the 50 as researchers sought protein levels and developed the tool based on findings from data of the patients.
After the initial trial 24 COVID-19 patients who had received treatment were also tested as researchers found 19 survived and five patients died. The researchers said the AI tool accurately forecasted the five deaths.
The researchers however underlined that the sample size was small as it only consisted of 50 patients and 24 other patients during open trials.
The new findings come as Europe and America grapple with the Omicron wave with hospitals under immense stress in several areas. Several health care workers have been hit due to the wave making it difficult for hospitals to function efficiently.