An algorithm capable of identifying people with COVID-19 by only the sound of their coughs has been developed in the United States.
The artificial intelligence (AI) algorithm, which was built at the Massachusetts Institute of Technology (MIT), achieved a success rate of 98.5% among people who had officially tested positive for coronavirus. It was also able to perfectly detect COVID-19 in those who had no other symptoms, according to the BBC.
This is an important innovation as asymptomatic people with COVID-19 display no physical symptoms of the disease and are thus less likely to seek testing and could inadvertently spread the virus to others, says MIT News.
The researchers say that they would need regulatory approval before they can develop the algorithm into an app.
The evidence shows that the AI algorithm can discern differences in coughing that humans cannot and if this detection system can be incorporated into smartphones, or similar devices, the researchers believe that it could become a valuable tool for early COVID screening, according to Science Alert.
Brian Subirana, a co-author of the paper, said: “AI techniques can produce a free, non-invasive, real-time, any-time, instantly distributable, large-scale COVID-19 asymptomatic screening tool to augment current approaches in containing the spread of COVID-19.
“Practical use cases could be for daily screening of students, workers, and public as schools, jobs, and transport reopen, or for pool testing to quickly alert of outbreaks in groups.”
The main value of the algorithm is its ability to differentiate between healthy and unhealthy coughs in asymptomatic people, according to the researchers. It is intended to be an early warning system and not a way of diagnosing COVID-19, a proper test would be required for that.
“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,” said Subirana.
The next step for the researchers is to test the system on a more diverse set of data to see if there are other factors at play that could explain such a remarkably high detection rate.