NEW AI MAY HELP IDENTIFY COVID-19 POSITIVE CASES BY SOUND OF COUGHING

I recently had a dental examination. Before allowing me to make an appointment, I was asked a series of questions: Have you recently been around anyone who has been diagnosed with COVID-19? Have you had a COVID-19 test in the past 14 days? Have you experienced any COVID-19 symptoms? Have you recently travelled abroad?

Then, when I went to the appointment two days later, I had to answer all the same questions, and then they used a digital no-touch thermometer to check my temperature. “97.3,” the volunteer announced. “You’re free to go on in.

But what if I was an asymptomatic carrier? Asymptomatic COVID-19 patients display no symptoms of the coronavirus infection. They can, however, spread the disease to healthy people, who in turn may spread it to others, whether they become symptomatic themselves or not. This has been one of the biggest issues facing the control of COVID-19 and is why social distancing regulations are as severe as they are.

Now, there may be a dramatic innovation to help solve this problem.

Thanks to a new, artificial intelligence (AI) test engineered by researchers at MIT, it may soon be a simple matter indeed to discover asymptomatic COVID-19 positive cases. The best part of this new test? It could be free to administer.

A paper recently published in the IEEE Journal of Engineering in Medicine and Biology shows that an AI model has been created that seemingly can accurately distinguish between healthy and asymptomatic but COVID-19 positive individuals, through forced-cough recordings.

Tens of thousands of people submitted forced-cough recordings – i.e., when your doctor says, “Cough for me” and listens to your chest through a stethoscope. Instead, people voluntarily submitted recordings through web browsers, cellphones and laptops. Researchers discovered that the new AI could identify 98.5% of coughs from people who were confirmed to have COVID-19. Of people who were asymptomatic but tested positive for the virus, the AI model was able to identify with 100% accuracy.

This could turn out to be a massive breakthrough in the fight against the spread of COVID-19.

Take, for example, the quantum leap in checking on elementary school students – tens of millions who have been stuck at home for months, now. Imagine if every student at their school were required to do a home check-in, before even leaving the house every morning. That check in could be a simple cough into a phone-app. No temperature reading, no series of questions, just a quick “cough, cough.”

And, should the quick test find a “hit,” instead of that pupil exposing hundreds of students and staff to potential infection, they instead would stay home and schedule a doctor’s appointment.

“The effective implementation of this group diagnostic tool could diminish the spread of the pandemic if everyone uses it before going to a classroom, a factory, or a restaurant,” Brian Subirana, a co-author of the research paper, told MIT News.

Hundreds of thousands of lives could be saved and containment of the virus could begin in earnest.

The simple, immediate check could be especially helpful in plugging the hole in America’s testing approach, to date. People who are asymptomatic do not typically get tested unless someone in their immediate proximity is diagnosed as COVID-19 positive. For one thing, testing supplies are simply too limited. Also, each test performed on someone who is diagnosed as COVID-19 negative is a test that, in theory, could have been used to test someone who, indeed, is found to be positive and could then have quarantined themselves to prevent further infections.

This is especially good news, given the New York Times reported Friday that the United States has set an all-new daily record for number of coronavirus cases, with two dozen states reporting their worst-ever weeks for increased cases and no states reporting improvements.

Dr. Deborah Birx, response coordinator to the virus for the White House, told governors Friday that a full one-third of the country, over 1,200 counties, are considered coronavirus hot-spots.

Using AI technology to diagnose respiratory ailments isn’t exactly new. An article posted to docwirenews.com on June 7, 2019, reported that AI technology developed by Curtin University and The University of Queensland, Australia, was used to accurately diagnose pneumonia, croup, asthma, and other lung ailments in patients.

The MIT team responsible for the research paper is currently working on creating a user-friendly app, which will need FDA approval. Once that is achieved, we could all be breathing a little easier in regards not only to the coronavirus pandemic, but future outbreaks as well.