The COVID+ VOICES study uses an algorithm that aims to detect unique sound signatures associated with COVID-19. The software was developed in 2020 by VoiceMed. These signatures are also known as vocal biomarkers and can be present in the sound of coughing, breathing, or speech. The project builds on the initial results of one of VoiceMed’s projects carried out in Italy. In this first phase of the project, the algorithm achieved a sensitivity (ability to detect positive cases) of 60.0% and specificity (ability to detect negative cases) of 74.5% and met all the EU requirements applicable to software with a medical purpose (CE mark). « This is a truly innovative and, above all, convenient and cost-effective approach that would allow a large number of people to be tested simultaneously without the need for the unpleasant nasopharyngeal swab« , explains Arianna Arienzo, CEO of VoiceMed.
Patients are offered the possibility of participating in the study
Since March 2022, patients coming to the LNS collection centre (Centre de Prélèvement-Dudelange), with a prescription for a COVID-19 diagnosis by PCR, are offered the possibility of participating in the study. Participants log in to a web application developed for the study just using a code. Their identity remains confidential to VoiceMed. After giving their consent, they answer a short questionnaire and record different sounds of their voice (cough, breath, and vowels). In 4 months, already 235 people have participated in the study.
« If the algorithm ultimately shows sufficient sensitivity, it would offer a rapid and inexpensive alternative for mass testing. Results can be combined with the more robust PCR tests to confirm all suspected cases identified by the application, » adds Dr Guillaume Fournier, the former main researcher of the study recently replaced by Dr Amy Parrish.
Selected to accelerate a digital solution against COVID-19
The study was selected to be part of the COVID-X program under the name « VADI » and it receives European funding in order to accelerate digital solutions in the fight against COVID-19.
“The collaboration with the LNS offers two major advantages: first, there was the possibility of IT integration between the LNS and VoiceMed which maximises both efficiency and security; second, the international environment allows the algorithm to be validated in several languages at the same time”, points out Dr Ayana Martins, Chief Scientific Officer of VoiceMed.
The COVID+ VOICES study involved comprehensive expertise and knowledge from both partners. Furthermore, apart from the joint IT system, the study also required the establishment of a legal framework, before it could be approved by the National Research Ethics Committee (CNER).
Preliminary results on an updated version in October 2022
The COVID+ VOICES study will run until, at least, March 2023. However, VoiceMed will release preliminary results on an updated version of the algorithm in October this year. In the meantime, VoiceMed is also developing an application to help people with asthma self-manage their condition. « We realised that it was even more important to make the experience as easy and convenient as possible in the case of chronic diseases. We are creating an app that includes clinically validated breathing exercises for asthma. With the recording of these breathing exercises, we are first helping people to learn to breath better but also developing vocal biomarkers that can predict an asthma crisis. », says Arianna Arienzo.