Researchers from the Maastricht University of the Netherlands have developed an app that can detect COVID-19 from the sound of your voice. The app claims to detect infections “more accurately than lateral flow tests”.
It is called the COVID-19 Sounds App, and it is already available on Google Play Store as well as Apple’s App Store.
One of the main symptoms of COVID-19 infections is inflammation in the respiratory tract and vocal cords, which usually leads to a change in the patient’s voice. The scientists investigated whether they could use these symptoms to accurately diagnose the disease.
The team of scientists made use of artificial intelligence (AI) with an accuracy rate of more than 89%, making it more reliable than rapid antigen tests. The AI model was trained using data from over 893 audio samples gathered from 4,352 healthy and non-healthy participants. 308 of these patients turned out to be positive for COVID-19.
One of the researchers at the Institute of Data Science, Maastricht University, Wafaa Aljbawi said:
These promising results suggest that simple voice recordings and fine-tuned AI algorithms can potentially achieve high precision in determining which patients have Covid-19 infection. Such tests can be provided at no cost and are simple to interpret. Moreover, they enable remote, virtual testing and have a turnaround time of less than a minute.
The research was presented to the European Respiratory Society International Congress in Barcelona on Monday.
How The App Works
The app asks for some basic information such as demographics, medical history, smoking status, and then asks patients to record a sample of respiratory sounds. This includes coughing three times, breathing deeply through their mouth three to five times, and reading a short sentence on the screen three times.
The app uses a voice analysis technique called Mel-spectrogram analysis that can diagnose using voice features such as loudness, power, and variation over time.
As mentioned earlier, the app’s accuracy rate for detecting positive cases is 89%, but its ability to correctly identify negative cases is around 83%.
These results show a significant improvement in the accuracy of diagnosing Covid-19 compared to state-of-the-art tests such as the lateral flow test. In other words, with the AI LSTM model, we could miss 11 out 100 cases who would go on to spread the infection, while the lateral flow test would miss 44 out of 100 cases.
