A new machine-learning tool may be available that predicts the health of Covid patients and helps doctors make better treatment decisions.
A German research team from Charity-University Medicine in Berlin – one of the country’s largest university hospitals – developed an Artificial Intelligence tool that can estimate how well an infected person will fare based off of a blood sample.
A person can determine whether they will live or die from the virus by the levels of 14 proteins in their blood. Researchers have developed a tool to accurately assess this risk.
This tool helps determine who needs the most help in times of crisis and who can fight the virus.
German researchers discovered 14 proteins in blood from Covid victims can determine whether or not a person will survive the disease. A machine-learning tool was created that predicted accurately the outcomes of 23 out of 24 Covid patients.
It is easy to use, and doctors can no longer make difficult decisions with the tool. Instead it will be used by an AI system that makes more precise and accurate predictions. It is hoped that it will be able to assist healthcare systems during times of crisis. (file photo)
In a press release, Dr Florian Kurth from Charity-University, co-author of this study, stated that “our study showed that a combination to markers can fairly accurately predict whether an individual patient will survive or die.”
For the first time, researchers gathered data from 50 Covid patients, both from Germany and Austria.
The researchers took blood samples of the patients to determine if they had any biomarkers.
In the end, fifteen of 50 patients died. The researchers looked at trends in protein levels between those who survived, and those who didn’t.
Based on the data from 50 of their patients, they created an artificial intelligence system to predict whether someone will die from Covid.
This system was trialled in 24 Covid patients from the real world who were receiving treatment at their hospital.
The study found that 19 of the patients survived their disease, while five died. Machine learning software correctly predicted that all five deaths would be caused by the illness and identified 18 out of 19 survivors.
Kurth explained that 14 proteins showed opposite changes over time in patients who live in intensive care compared with patients who don’t.
“Interestingly, all those proteins were found in plasma altered by COVID-19.” [and]This makes it possible for us to be confident about our conclusions.
Researchers point out that the study had a small sample, as machine learning was limited to 50 samples and 24 participants.
The early results look promising and the team hopes to have the chance to try it out with a wider population in order to establish if their system is truly useful for future treatment decisions.
The pandemic saw overcrowded hospitals, with overworked personnel and insufficient resources. This was a common scene.
Every new wave of pandemics brings with it a flood of Covid patients.
Even though the Omicron wave is a milder strain than the previous version of the virus there have been reports that hospitals across the country had difficulty handling an influx of patients.
The situation has even forced some hospitals into having to ration care with doctors having to make tough decisions when there just are not enough resources to go around.
A system like that developed by the German team may help doctors make better use of their resources and reduce the emotional impact of making these decisions.