Introducing AI that predicts 'COVID-19 patients who need treatment in the intensive care unit' with 96% accuracy
With the new coronavirus infection (COVID-19), the severity of symptoms varies greatly from patient to patient, depending on whether the infected person has asymptomatic or severe symptoms. Under the pandemic of the new coronavirus, where the number of infected people is exploding, the number of beds and intensive care units that can handle patients is limited, so take appropriate measures at the right time for patients. Is very important to survive the spread of the infection. Meanwhile, artificial intelligence (AI) has been developed that can identify COVID-19 patients who require treatment in the intensive care unit with 96% accuracy.
Integrative Analysis for COVID-19 Patient Outcome Prediction
AI Predicts 96% of COVID-19 Pneumonia Cases That Will Require Intensive Care
https://www.sciencealert.com/new-algorithm-works-out-which-covid-19-pneumonia-cases-require-intensive-care
The number of cases where AI is used to confront the pandemic of COVID-19 is increasing, and AI that can detect COVID-19 infection from the 'sound of cough' ...
AI that can detect new coronavirus infection from 'coughing sound' will be developed, 100% detection of asymptomatic patients --GIGAZINE
Attempts are being made to explore new ways to make COVID-19 therapeutics using AI.
Attempt to develop 'new way of making' new coronavirus infection drug with artificial intelligence --GIGAZINE
Most people will eventually recover when infected with COVID-19, but some will develop severe pneumonia at a rapid pace, with diffuse axonal injury andacute respiratory failure , and eventually. It can be fatal. AI has been developed to identify such severely ill COVID-19 patients.
AI has succeeded in predicting critically ill patients in need of intensive care unit treatment by combining chest scan images with non-image data such as demographic information, vital signs, and blood test results. AI is in the early stages of development, but in order to verify its accuracy, a test using data from 295 COVID-19 patients admitted to the United States, Iran, and Italy showed that AI was 96% accurate. He has succeeded in predicting patients who need treatment in the intensive care unit.
'As an AI developer, I believe in that power. AI analyzes large amounts of data and is less obvious to the human eye,' said
In another study, it ventilator usage and potassium level is the variable of the most important non-image revealed has become, percentage or total bilirubin lymphocytes Kueachinin, values such as albumin in important variables It turns out that there is. Yet another study found that non-imaging data such as age, fever, and respiratory abnormalities alone could predict whether COVID-19 patients would need to be treated in the intensive care unit.
Compared to the AI developed in these studies, the AI developed by Yan et al. Is one step further. First, we will quantify the range and amount of lung features, and then focus on the most important functions in predicting COVID-19-derived pneumonia. Finally, other contextual factors are used to predict which patients will develop the most serious symptoms. 'As far as we know, this study is the first to use patient-wide information, including both imaging and non-imaging data, to predict outcomes,' the treatise states.
The research team hopes to improve AI in the future and use it to screen patients with COVID-19 pneumonia who require closer monitoring and care. In this study, no tests were conducted to compare the difference between COVID-19-derived pneumonia and respiratory illness.
Related Posts:
