A team of scientists have repurposed an artificial intelligence/machine learning-powered cough analysis mobile app for assessing respiratory health and associated COVID-19 risks.

The mobile app called kAs has been developed by Zensark Technologies (Hyderabad, Telangana, India), in collaboration with Salcit Technologies (Hyderabad, Telangana, India), to assess respiratory health of the patients by leveraging machine learning and disease specific cough signatures. It brings the functionality of Spirometry using one's smartphone to carry out on demand assessments without the need to step out of the house. Patients can use their mobile phones to track the sound of their cough to assess and track their health.

The app has been developed based on the guidelines laid out by the WHO and asks 15 questions to the subject. The subject then cough into the mobile app which sends the cough sounds to the company’s AI platform Swaasa, where it undergoes an audiometric analysis and produces a composite COVID Risk Score, based on the answers to the questionnaire and the coughing sound. The app generates a rating on a scale of 1 to 10, with 10 being the highest risk level. The Risk Score of any individual can be tracked across multiple days or weeks. The subjects can be asked to submit cough samples as many times a day for as many days as necessary. Based on the samples and symptoms, the app tracks the risk level of each individual over time.

Anytime a subject is found to be breaching a threshold risk score, it is highlighted to public health officials on their dashboards. The dashboards are designed to show the population level statistics and can drill down to the detailed reports of an individual subject. The app also allows doctors, with an external microphone and wireless stethoscope to track various breathing and risk parameters.

http://m.allmycbs.com/en/index.asp