Mount Sinai AI predicts outcomes for COVID-19 patients in the ER: 5 details

New York City-based Mount Sinai researchers have developed an artificial intelligence algorithm that predicts outcomes for COVID-19 patients who arrive in the emergency room with mild symptoms.

The researchers published their findings in the December issue of Radiology: Artificial Intelligence. Five details:

1. The researchers used the electronic medical records of 338 patients, combined with lab tests and chest X-rays, to train an algorithm to detect whether patients are likely to become sicker and need closer observation. The patients were between the ages of 21 and 50 years old.

2. The algorithm could help clinicians more quickly triage patients and accelerate treatment for those at high risk. This is especially important for younger patients who may not have specific symptoms when they arrive at the ER but may need intervention before symptoms worsen.

3. Researchers developed a severity score based on the likelihood of intubation or death within 30 days of arrival. They tested the algorithm with other patient data in all age groups and found the algorithm has 82 percent sensitivity.

4. This algorithm uses information just from the initial patient encounter in the ER and is different from other algorithms predicting COVID-19 outcomes among admitted patients with severe symptoms.

5. Mount Sinai is incorporating the algorithm-generated sensitivity score into its clinical workflow.

More articles on artificial intelligence:
AdventHealth joins clinical biotech research initiative on age-related diseases: 3 things to know
20 key health system AI tool rollouts in 2020
Microsoft to deploy tumor detection AI tool in hospitals

 

© Copyright ASC COMMUNICATIONS 2021. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Featured Whitepapers

Featured Webinars