Your phone could soon detect your cough as UW doctors, engineers work on app

Debating going to work because you have a cough? University of Washington doctors and engineers are working on an app that could soon differentiate your cough between tuberculosis, asthma, COVID-19, or other respiratory illnesses.

“Researchers collected over 33,000 coughs from 149 patients with TB with the coughs from 46 patients who had other respiratory illnesses,” said a news release from UW Medicine. “They then fed audio of the coughs, collected through various microphones, into a machine-learning model called TBscreen, which the researchers had developed.”

Researchers found that the app and phone microphone were able to predict which coughs were active TB at a better rate than more expensive microphones.

Dr. Thomas Hawn, one of the researchers and co-director of the UW Medicine’s Center for Emerging and Re-emerging Infectious Diseases, pointed out that although the app is not a confirmed diagnosis, it could alert the person or provider that further testing is needed.

“Working with smartphones and machine learning allowed us to look at cough frequency rather than sound,” said lead author Manuja Sharma, a UW doctoral student in electrical and computer engineering. “It is not possible to differentiate TB coughs by just hearing with the human ear.”

UW Medicine said study participants sat in a quiet room for two hours while researchers recorded their coughs.

Researchers found that the phone recording was 82% accurate when compared with two microphones, which were around 70% accurate at predicting TB.

Normally a TB diagnosis “requires a sophisticated clinical laboratory, with steady access to supplies and reliable electricity,” said UW Medicine.

Dr. David Horne, senior author of the paper and an associate professor of medicine, said TB has passed COVID as the leading cause of infectious disease-related death worldwide.

Horne estimates that 10 million people contracted TB in 2022 and 1.4 million people died from it.

The World Health Organization said a detection method for TB needs to be 90% accurate. Horne is confident the app will meet the standard.

“This new technology could take TB testing into the field and stop its spread,” said Horne.

Next, researchers will validate the app’s effectiveness with a larger population sample.

©2024 Cox Media Group