According to a new study published in the journal, artificial intelligence can help identify asymptomatic COVID-19 patients by simply coughing “forcefully”. IEEE Journal of Engineering in Medicine and Biology. But some experts expressed doubts about this technology.
For this study, MIT researchers set up a website where people can record a number of different forced coughs from their cell phones. The researchers gathered data from more than 5,300 people. According to an MIT press release, about 2,500 recordings were sent by people with a confirmed COVID-19 diagnosis, including those who tested positive for the virus but had no symptoms, ” along with 2,500 other recordings they randomly selected from the collection to balance the dataset. “
The researchers reported that the AI model had a 98% more accuracy rate in determining the coughs of people confirmed to have COVID-19. They also found that the technology accurately differentiated 100% of forced cough cases of asymptomatic people from healthy people.
“MIT researchers have now found that asymptomatic people may differ from healthy people in the way they cough,” according to the press release. (Yahoo Life contacted the study’s lead author but received no response.)
In the statement, one of the study’s co-authors Brian Subirana, director of MIT’s Automated ID Lab, said: “We think this shows how you make the sound change as you have. COVID, even if you don’t have symptoms. ”
While both Dr. Jonathan Chen, assistant professor of medicine at Stanford University Medical Center and Dr. Megan Conroy, assistant professor of clinical medicine in the lung disease division, critical care and sleep medicine sleeping at Ohio State University’s Wexner Medical Center, told Yahoo Life the results were “interesting,” both skeptical.
Chen, who specializes in biomedical informatics, calls the driving force behind the research “amazing” and “smart”, but questions the plausibility of such a high success rate out of the words “Self-reporting”. “Even one [COVID-19] PCR test is not accurate, he said. Chen added that the technology “could still be useful, but it is very likely that it is much less accurate than what they are finding.”
Conroy shares that she has some “serious concerns,” saying: “First, this is not an established way to screen for any other respiratory disorder or cause of cough; The study did not include any other respiratory tract infections, which raised some concerns about the ability to clearly differentiate between pathological and normal. Furthermore, they claim that the technology can identify a cough of an asymptomatic person as infected with COVID-19, but it’s important to note that if you have recently had a cough, you will not be considered. no symptoms. “
She also called the idea of differentiating diseases by the sound of a cough “questionable”. Chen added: “What if I have a cough, but is it a common cold?”
According to the release, the AI model is not designed to diagnose people with symptoms, “whether their symptoms are caused by COVID-19 or other illnesses like the flu or asthma”, adding: ” of the tool lies in its ability to distinguish asymptomatic coughs from healing ones. ”
Researchers are currently trying to create an application that uses the technology as a “free, convenient, non-invasive pre-screening tool,” according to the press release. “A user can log in daily, cough on their phone and immediately receive information about whether they might be infected and should therefore confirm with an official check.”
However, Conroy warns “strongly” about using “unproven technology to guide behaviors”. Instead, she encouraged “to continue to develop field trials for the SARS-CoV2 virus, in addition to the many measures we know to help reduce the spread of the virus: wearing a universal mask. Stay at home when possible, avoid crowds and keep society away whenever you’re around people outside of your home. “
For Latest news and updates about coronavirus, follow along at https://news.yahoo.com/coronavirus. According to experts, people over the age of 60 and those with the immunosuppression continue to be at greatest risk. If you have any questions, please consult CDC‘sand Who resource guide.
Read more from Yahoo Life:
Want lifestyle and health news delivered to your inbox? Sign up here for the Yahoo Life newsletter.