Published on
Updated
Reading 2 mins.
An app on your smartphone, using artificial intelligence, could accurately detect infection more easily than a nasopharyngeal test, a study has found.
What if the simple act of talking or breathing in front of your phone could detect a possible Covid-19 infection? In any case, this is what a study by the Institute of Data Sciences at the University of Maastricht, in the Netherlands, revealed this Monday, September 5 at the International Congress of the European Respiratory Society in Barcelona. An app using fine-tuned AI algorithms would even be better and faster at detecting infection than an antigen test or PCR test.
Cough, talk… The app gives you the answer
Since Covid-19 infection mainly affects the upper respiratory tract and the vocal cords, the research team from Maastricht University wanted to check whether voice analysis by artificial intelligence (AI) was relevant to distinguish disease. For this purpose, they used data from the University of Cambridge’s COVID-19 Sounds app containing 893 audio samples from 4,352 healthy and unhealthy participants, 308 of whom had tested positive for COVID-19. Participants, after answering a quick questionnaire about their history and smoking status, were asked to cough three times, breathe deeply through their mouths, and read a short sentence on the screen in front of their app.
The analysis of this data by a spectrogram and the construction of an artificial intelligence model resulted in an interesting figure: 89% of positive cases were detected correctly as well as 83% of negative cases. A higher accuracy rate than in nasopharyngeal tests.
Consult a GP online
AI, a medical tool that simplifies screening
“These promising results suggest that simple voice recordings and fine-tuned AI algorithms can potentially achieve high accuracy in determining which patients are infected with COVID-19.” concluded Wafaa Aljbawi, lead author of the research. Voice tests would also have certain advantages:
- They could be provided free of charge and therefore reach more people;
- They are simple to interpret;
- They are less invasive;
- They allow remote testing;
- They have a turnaround time of less than 1 minute.
The study also recalls that this same type of AI, studied by the University of Bristol, has also already been used since 2016 in the case of chronic obstructive pulmonary disease (COPD) to predict when patients could suffer from a flare-up. of their illness.
Digital health using AI models seems to be firmly settling in for impact on future healthcare.