Communications of the ACM
Computer Can Determine Risk of Dying from COVID
Machine learning may be used to identify drivers of progression to more severe cases of COVID-19.
Artificial intelligence developed by researchers at Denmark's University of Copenhagen (UCPH) can predict a patient's likelihood of dying from COVID-19 with up to 90% accuracy.
The investigators trained the algorithm on data from 3,944 Danish COVID-19 patients, in order to recognize patterns and correlations in prior illnesses and in COVID infections. The UCPH team also determined that, once someone is hospitalized with the virus, the algorithm can predict their need for a respirator with 80% accuracy.
Age and body mass index are the biggest determinants of infection severity, said UCPH's Mads Nielsen, while being male and having high blood pressure or a neurological disease also elevate mortality or respirator risk. "We are working towards a goal that we should be able to predict the need for respirators five days ahead by giving the computer access to health data on all COVID positives in the region," Nielsen said.
From University of Copenhagen
View Full Article
Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA
No entries found