Artificial intelligence is used for the first time to predict the oxygen demand of new crown patients

Artificial intelligence is used for the first time to predict the oxygen demand of new crown patients

International war "epidemic" operation

Science and Technology Daily, Beijing, September 16 (Reporter Liu Xia) According to a report on the University of Cambridge website on the 15th, scientists from the university, together with 20 hospitals and pharmaceutical companies from all over the world, used artificial intelligence technology (AI) and machine learning technology for the first time , Predicting the oxygen demand of patients with new coronary pneumonia on a global scale, the accuracy rate is as high as 95%
.
This new method provides new ideas for scientists and medical staff to deal with future epidemics and other health challenges

The researchers explained that this study, called EXAM, is one of the largest and most diverse clinical joint learning studies to date.
It aims to use data from four continents to build an artificial intelligence tool to predict the hospitalization of patients with new coronary pneumonia.
How much oxygen might be needed for the first few days of treatment
.

In the study, a technology called joint learning used an algorithm to analyze chest x-rays and electronic health data of hospitalized patients with new coronary pneumonia
.
To protect patient privacy, patient data is completely anonymous, and researchers send algorithms to each hospital, so no data will be shared or left behind

To verify the accuracy of this tool, researchers conducted tests in multiple hospitals on five continents and analyzed the results of approximately 10,000 COVID-19 patients from all over the world, including 250 patients who were hospitalized between March and April 2020 Data
.
The results show that this tool can predict the required oxygen within 24 hours after the patient arrives in the emergency room, with an accuracy of 95%, and it only takes two weeks of artificial intelligence "learning" time to achieve high-quality predictions

Professor Fiona Gilbert, the head of the latest research, said: “Our ongoing research on EXAM shows that such global cooperation is reproducible and more efficient, which can satisfy clinicians in dealing with complex health challenges and future epidemics.

Dr.