Scientists use AI to determine which COVID-19 patients require hospitalization
As the coronavirus (COVID-19) continues to grow across the world, hospitals are getting crowded with an influx of patients. Under this great stress, hospitals are facing difficulties to manage their developing care and hospitalization strategies & resources that can prioritize the high-risk patients.
This is the situation where advanced artificial intelligence (AI) can assist, experts at Jvion believe. The company specializes in clinical Artificial Intelligence and is now working on a data analysis system that will help doctors take a proactive approach to manage the influx of patients in the inpatient and outpatient settings.
Jvion believes machine learning algorithms can be used to understand the social risk factors that make people more likely to acquire or spread an infection that requires hospitalization.
“There are two things to consider: First, at some point hospital capacity and resources will be surpassed by the demand, and second, x-rays, vital signs, etc. are not the right indicators for a patient’s final outcome,” John Frownfelter, MD, CMIO at Jvion, told TechTalks. “Which patient could get well at home, and which patient is not likely to survive regardless of the therapies being provided is one of the more clinical challenges. However, AI can help in this space.”
“As hospitals are overwhelmed with patients, these people will be safe at home. One cannot generalize it as a practice for all people, but it can be life-saving for people at highest,” said Frownfelter.
Some of the methods artificial intelligence use to fight against coronavirus are: automatically measuring people’s temperature in the crowded areas, using chest x-ray scans to diagnose coronavirus infections, and machine learning algorithms to foresee the spread of the virus.
Over data from 2 million patients, the AI’s initial analysis determined which among hundreds of factors increase the risks of catching infections that cause end-organ damage, such as respiratory failure.
Social Risk Factors
Collaborating with other researches on COVID-19, the findings have highlighted chronic conditions and old age as major risk factors for poor outcomes. However, the machine learning algorithm was also able to search social risk factors such as shopping in person, college dorms, long commutes, and living in dense suburban areas. Understanding social risk factors can make a big difference in dealing with coronavirus.
As for now, Jvion’s efforts are currently available in the U.S. However, to help other places where individual data is not available, the AI could take a large sample size to build a model that is representative of the population.