Over the past couple of years, one of the biggest things that has come to strike fear into our hearts when we hear it, is the sound of a cough. Whereas before the pandemic, we might at most throw an annoyed glance at someone in a restaurant if they cough in the direction of our food, these days, we’re likely to get up and walk away—that is, if we’d even venture into an open-dining restaurant in the first place.
Every respiratory disease has its own coughing pattern
Many illnesses have very distinct coughing patterns which may not be distinguishable by the human ear, but which could be picked up by artificial intelligence and potentially diagnosed. This is true for most respiratory diseases such as tuberculosis, pneumonia, asthma, COVID-19, bronchitis, and others.
For example, a pneumonia cough can be recognized by crackling in parts of the lungs, says Peter Small (chief medical officer of Hyfe Inc.). Asthma coughs, on the other hand, are marked by a wheezing sound. Researchers are still working to detect a noticeable pattern in COVID-19 patients’ coughs.
Although the research needed to create AI technology which can consistently and accurately identify and diagnose respiratory diseases through the sound of their cough still has a long way to go, thousands of cough recordings from U.S. and Australian patients have already been collected in a University of Queensland database, along with their medical records, with that end goal in mind.
This technology could save thousands of lives in third-world countries
Besides potentially making it possible to identify COVID-19 patients through their coughing patterns, researchers say such a breakthrough in AI technology could have a tremendous impact particularly in low-socioeconomic countries, such as South Africa.
In South Africa, half of all people who die from TB have never done any testing for the disease, even though a tuberculosis cough has distinctive acoustic patterns. If those can be picked up through a smartphone app and TB patients are caught earlier on (at a much lower cost per patient), the TB fatality rate could be drastically reduced.
The project of combining personal technology and healthcare in this new way is currently gaining traction around the world, with medical and AI researchers collaborating to see if something like this could eventually become possible.