A tool has been developed to help healthcare professionals identify hospitalized patients at the highest risk of death from COVID-19 using artificial intelligence (AI).
The algorithm can help doctors direct critical care resources to those most in need, which the developers of the artificial intelligence tool say could be especially valuable for countries with limited resources.
And as the end of the coronavirus pandemic is not in sight, with new options leading to new waves of disease and hospitalization, scientists behind the tool say there is a need for generalized tools like this that can be easily introduced.
To develop the tool, the researchers used biochemical data from routine blood samples taken from nearly 30,000 patients hospitalized in more than 150 hospitals in Spain, the United States, Honduras, Bolivia and Argentina between March 2020 and February 2022.
Taking blood from so many patients meant that the team was able to capture data from people with different immune status – vaccinated, unvaccinated and naturally immune – and from people infected with each variant of COVID-19.
With so much data, they were able to train an artificial intelligence program to predict the signs of a bad prognosis, regardless of different immune statuses or variants.
In addition, they tested whether the time of blood collection affected the operation of the instrument by comparing data from different time points of blood collection before patients recovered or died.
They found that the algorithm predicts with high accuracy the survival or death of hospitalized patients up to nine days before each outcome.
The resulting algorithm, called COVID-19 Disease Outcome Predictor (CODOP), uses measurements of 12 blood molecules that are usually collected during hospital admission, which means that the instrument can be easily integrated into any hospital.
The peer-reviewed findings are published in the journal eLife.
“More hospitalizations are likely”
“The emergence of new variants of SARS-CoV-2 that reduce immune protection and alleviate mitigation measures means that we are likely to continue to see an influx of infections and hospitalizations,” said David Gomez-Varela, head of the international project and senior author. , former Head of the Max Planck Group and current Senior Research Fellow in the Department of Pharmacology and Toxicology at the University of Vienna in Austria.
“There is a need for clinically valuable and generalizing triage tools to help allocate hospital resources for COVID-19, especially in areas where resources are scarce. But these tools must be able to cope with the ever-changing global pandemic scenario and must be easy to implement.
Gomez-Varela added that the team is now working on a follow-up dual model, “in line with the current pandemic scenario for increased infections and cumulative immune protection, which will predict the need for 24-hour hospitalization for primary care and intensive care patients.” within 48 hours for those who have already been hospitalized. “
Scientists from a number of institutions have co-developed the tool, including the Max Planck Institute for Experimental Medicine, Turku University in Finland, the Spanish Society of Internal Medicine, the Argentine Medical Society and the International Forum on Internal Medicine.
Add Comment