Could your phone replace your RAT?
A new smartphone app could replace Rapid Antigen Tests (RATs) by detecting whether you have COVID-19 or not based on one telling symptom - the sound of your cough.
ResApp, a digital health company based in Brisbane, announced in late March that its new screening test, which you take with your smartphone, had successfully detected the virus in 92 percent of people who were infected.
The promising results come from the company’s pilot clinical trial of its machine learning technology, which analysed coughing sounds from 741 patients, including 446 who were confirmed to be infected with Covid using RATs.
It was also found that the technology could correctly identify eighty percent of those who didn’t have the virus as being negative.
However, the technology will still need to be tested in double-blind clinical trials and successfully pass through the process for regulatory approval before it hits the shelves, per 7News.
Dr Lucy Morgan, the Chair of the Lung Foundation Australia and a professor of respiratory medicine, told 7News the technology could be a promising alternative to RATs.
“What’s so exciting about this pilot project, a relatively small project, is that this app has been able to predict that a cough is due to COVID-19 at a very, very high rate,” she said.
But, Dr Morgan stressed that it still wouldn’t replace the technology currently used to diagnose COVID-19.
“It’s not a diagnosis. It’s not a blood test, it’s not a PCR test, it’s not a RAT test. It’s coughing into your phone and predicting,” she explained.
This isn’t the first time scientists have turned to our smartphones to diagnose Covid either.
In January, a team of US researchers published a paper detailing how they developed an inexpensive testing kit. Using a similar method to PCR tests to make copies of DNA in a saliva sample, the test then uses your smartphone camera and an app to detect whether any DNA from the virus is present.
Meanwhile, researchers from the Massachusetts Institute of Technology (MIT) have been training their own Artificial Intelligence (AI) using a database of over 5000 forced coughs. They found that over 98 percent of Covid-positive samples were correctly identified by the AI, and that more than 94 percent of those without Covid were correctly identified as being Covid-negative (known as specificity).
Their findings were even more promising with asymptomatic subjects, with 100 percent of Covid-positive subjects being correctly identified (known as sensitivity).
In comparison, the trial of the new ResApp technology used significantly fewer subjects, but Dr Morgan said there was no reason the results could be replicated in a trial with a larger sample size.
However, she noted that it’s unknown whether having underlying illnesses, such as asthma or chronic obstructive pulmonary disease (COPD), could affect how sensitive the test actually is.
“How do we know the cough for a person who has background lung disease is going to be as sensitively detected to have Covid as someone who was previously well and then develops a cough?” she said.
Though this and other questions are still unanswered, the numerous tests required before the technology can be approved are sure to resolve at least some.
Image: Getty Images