Coronavirus: Algorithm that identifies 'Covid-19 cough' could boost detection rates

The programme developed by researchers in the United States was able to pick up all asymptomatic infections and nearly all symptomatic cases

• 

  A volunteer checks a list of people suspected of having Covid-19 as they arrive in an ambulance at a quarantine centre in Yangon, Myanmar. AFP

• 

  Pupils wait for the opening of their primary school on the first day of the new school year, in Ouagadougou, Burkina Faso. AFP

• 

  A health worker holds a sample taken for testing at clinic in Moscow, Russia. Reuters

• 

  Snack bar attendants wait for customers at a Kinoplex movie theater in Rio de Janeiro, Brazil. Bloomberg

• 

  People walk in the tourist area surrounding Houhai Lake during Chinese National Day holidays in Beijing, China. Reuters

• 

  Ali Ewoldt performs during Radial Park's production of "The Phantom of the Opera" in New York, which was presented to the audience via a movie screen. AP Photo

• 

  A horse is hosed down after racing at Warwick Racecourse in central England. Getty Images

• 

  Two pigeons pass through an outdoor dining area sealed off due to the coronavirus pandemic in Huntington Park, California, USA. AP Photo

• 

  White House social media director Dan Scavino, left, wears a face mask aboard Marine One as they depart at Andrews Air Force Base, Maryland. AP Photo

• 

  American Airlines employees work in a check-in area at Los Angeles International Airport in Los Angeles, California. AFP

• 

  Honduran migrants trying to reach the US hitchhike on a truck after bursting through a border checkpoint to enter Guatemala illegally, in Entre Rios, Guatemala. Reuters

• 

  New Zealand Prime Minister Jacinda Ardern meets shoppers in Newmarket in Auckland. Getty Images

• 

  A woman walks past a mural promoting awareness of the coronavirus disease outbreak in Jakarta, Indonesia. Rueters

An algorithm developed by researchers in the United States has identified people with the coronavirus by the sound of their coughs.

The programme, which analyses sound patterns, achieved a 98.5 per cent success rate in tests involving people with an official Covid-19 diagnosis.

In asymptomatic patients, the figure rose to 100 per cent.

The researchers from Massachusetts Institute of Technology said the virus changes the way a person produces sound, but the difference is not detectable by human ears.

They collected about 70,000 audio samples, each of which contained a number of coughs.

Of those, 2,500 were from people with positive coronavirus tests.

MIT scientist Brian Subirana co-authored the paper, published in the IEEE Open Journal of Engineering in Medicine and Biology.

“The way you produce sound changes when you have Covid, even if you’re asymptomatic,” he told the BBC.

Researchers said the algorithm could be used for the daily screening of students, staff and members of the public at schools, in workplaces and on public transport, or for pool testing to quickly flag outbreaks among groups.

The team is now working on using the findings to create an easy-to-use app which could provide a free and widely available method of identifying infections.

A user could log in each day, cough into their phone and instantly be told whether they might be infected and should follow up with a standard coronavirus test.

Such an app would require approval from the US Food and Drugs Administration, MIT said.

A similar project reported an 80 per cent success rate in identifying a combination of breath and cough sounds in July.

The PCR swab remains the gold standard for Covid-19 testing, but it is not the only screening tool at governments’ disposal.

Others include diffractive phase interferometry tests, which are in wide use across the UAE. They scan a person’s blood cells for signs of infection.

Sniffer dogs are also said to be highly accurate at picking out sufferers from a crowd. They have been stationed at airports across the country.