As of November of 2020, there have been more than 225,000 deaths due to COVID-19 in the United States alone. One of biggest barriers during this pandemic is the inability to test for COVID-19 at a global scale. However, this does not mean that researchers have not been working around the clock to develop a screening tool. One research team is currently developing and testing an Artificial Intelligence (AI)-powered screening solution for COVID-19 called A14COVID-19, which is deployable via a smartphone app. The app records and sends three 3-second cough sounds to an AI engine operating in the cloud and returns results within two minutes.
Coughing is a symptom of over thirty related medical conditions, which makes the diagnosis of COVID-19 by cough alone very challenging. However, the research team investigated the distinctness of pathomorphological alteration in the respiratory system induced by COVID-19 infection when compared to other respiratory infections. They found that their app can distinguish among COVID-19 coughs and several types of non-COVID-19 coughs.
The accuracy of the study was promising enough to encourage a large-scale collection of labeled cough data to examine the generalization capability of the app. While the app is not a clinical grade testing tool, it is a screening tool that can be deployed anytime, anywhere, by anyone. It can also be used to assist clinical decisions to channel clinical-testing and treatment to those who need it most.
References
Imran, A., Posokhova, I., Qureshi, H. N., Masood, U., Riaz, S., Ali, K., John, C., Hussain, I., & Nabeel, M. (2020). AI4COVID-19: AI enabled preliminary diagnosis for COVID-19 from cough samples via an app. Informatics in Medicine Unlocked, 20(1).
James Bender, MA
WKPIC Doctoral Intern