As the number of COVID-19 cases rise once again throughout New York, state and local leaders are keeping a close eye on the number of available hospital and ICU beds.

But a research team at Rensselaer Polytechnic Institute recently published a discovery that could help the state foresee how many ICU beds it will need, before they are needed.

Pingkun Yan, an assistant professor of Biomedical Engineering at Rensselaer Polytechnic Institute, led a team that created an algorithm using artificial intelligence that has so far been successful in predicting the outcome of COVID-19 patients.

“So we tried to use this algorithm to predict whether a patient will develop into a severe condition,” Yan explained. “And if that is the case, we can give them more virus supportive treatment to prevent that from happening.”

Here’s how it works.

A computed tomography (CT) scan is taken of a patient’s chest to determine the severity of the lung infection.

They then take this person’s demographic information, vital signs, and more and the AI does the rest.

“We put all this information together and designed the AI algorithm to analyze all this data and predict whether the patient will need the ICU treatment,” Yan said.

Yan says they have tested this on over 300 patients and so far it has been successful 94% of the time.

It can not only help doctors decide if this patient will eventually need ICU treatment, but also what sort of treatment and medication might be the most useful.

Back in March, the number of hospitalizations doubled so quickly the state had to scramble to free up hospital beds.

Although hospitalizations are not increasing the way they were in March and April, Governor Andrew Cuomo said that available hospital beds will be the main focus of his administration moving forward into the winter months.

In total the state has 53,000 hospital beds. Out of those, 6,000 are ICU beds.

Right now, there are only 2,144 total ICU beds available.

“It’s all about hospitalizations,” Cuomo emphasized. “We’re dealing with hospitalization rate, hospitalization capacity. You start with 53,000 beds, you can end elective surgery, reduce the number of people in those beds.”

Yan and his team at Rensselaer Polytechnic Institute are partnered with Massachusetts General Hospital to continue testing this algorithm in the hopes it can be implemented soon.  

In the future, Yan says this algorithm could also help predict mortality risk of other lung and heart diseases as well.