The Covid-19 pandemic spread primarily through the dispersion of droplets containing the virus through expulsions such as coughs or sneezes.

Testing Failed to Overtake the COVID-19 Spread

The COVID-19 pandemic showed that humanity was unprepared. Concretely, we lacked the means to test the population reliably and quickly enough to contain the spread of the virus. We knew, though, that people with Covid-19 had specific symptoms, cough being a primary symptom.

As a result, the virus SARS-CoV-2 spread worldwide, from nation to nation, from city to city. Essentially, it seemed to spread at the rate people would get COVID-19 and breath or cough the virus onto others.

To sum it up, the virus seemed several steps ahead of our capacity to test the population, identify and contain outbreaks.

The standard means of testing adopted worldwide during the pandemic has been real-time reverse transcription-polymerase chain reaction (RT-PCR).

RT-PCR was insufficient to contain the pandemic. Although it provides reliable test results, RT-PCR had three major shortcomings:

  • Costly – Requires expensive chemical reagents, devices, and expert personnel.
  • Time-consuming – The results typically take 2 to 48 hours.
  • Unsafe sample collection – social distancing has to be violated to collect samples, thus risking contagion.

AI-Powered Point-of-Care Testing for COVID-19 Through Cough

The potential payoff of implementing rapid point-of-care tests (POCT) is immense. For instance, it could reduce unnecessary contact with healthcare professionals. Most importantly humanity would possess practical tools for outbreak detection and containment. that  could be immense. In essence, it is a technology that can save lives.

Recently, larger datasets have become available. Simultaneously, reliable machine learning algorithms are a reality. And finally, the smartphone has become widely accessible.

As a result, this combination of data, code, and tech is the perfect storm to make POCT a reality. Most importantly, it can increase the power of diagnostic and disease stratification efforts.

In short, consider the two following statements:

On the one hand, doctors routinely use audio signals – such as breath sounds during auscultation – as a tool to diagnose diseases or monitor their progression. And have been doing so pretty much since the dawn of the medical field.

On the other hand, the typical symptoms of COVID-19 include cough and breathing difficulties.

As a result, combining both facts mentioned above implies there is potential to use the power of information technology to give healthcare professionals some help.

COVID-19 Cough Acoustic Epidemiology: The DiCOVA Challenge

Several research teams and industry groups worldwide are working to achieve an accurate COVID-19 diagnostic tool. Such a tool would analyze the sound of a forced cough with a specialized algorithm developed by machine learning.

For instance, an exceptionally high moment of this effort happened during early 2021 in the DiCOVA (Diagnostics of COVID-19 using Acoustics) challenge. In short, the main goals of this challenge were twofold:

Firstly, to provide a benchmark for monitoring the progress in acoustic-based diagnostics of COVID-19.

Secondly, to challenge speech and audio researchers with a problem they can solve with their particular set of skills. Moreover, a challenge that is of immediate societal relevance.

Additionally, a more ambitious goal was to achieve the requirements defined by the World Health Organization (WHO). Specifically, the WHO stated that a sensitivity of 70% (at a specificity of 95%) is necessary for a proper POCT tool.

The DiCOVA challenge in a nutshell

Firstly, teams competed to develop the best possible classification system to diagnose COVID-19 by analyzing acoustic biomarkers. Secondly, these biomarkers consisted mainly of cough recording features. Thirdly the challenge used data from a public dataset named Coswara. Finally, the team’s tracked their results on a digital leaderboard that operated in near-real-time.

In conclusion, not even the highest-scoring team reached the criteria at the end of the challenge. However, a fusion of the four top-scoring teams achieved the WHO’s benchmark by a relatively small margin, with a sensitivity of 70.73%.

This Cough-Interpreting AI vs. COVID-19 Story Is Not Over Yet

The DiCOVA challenge demonstrated an incredible concept. Namely, the idea of using machine learning algorithms to diagnose acoustic biomarkers to result in a practical POCT. If validated, such technology would indeed be a game-changer. Essentially, it can aid humanity’s fight against pandemics and respiratory diseases in general.

The researchers who participated in the DiCOVA challenge presented their works at the DiCOVA Special Session, at the Interspeech 2021 conference. Here is a page that summarizes their results.

A Question Easier to Answer: “Is my Cough COVID-19?”

What if anyone with a smartphone could quickly tell if their cough sounds like they have COVID-19? In short, could you think of a nobler use of the smartphone?

With such technology, a citizen can answer the question:

Is my cough COVID-19?

We are not there yet. However, until then, you can download a free smartphone application to track cough. With it, you can see the dynamics of your cough, and show relevant data to your doctor.


  1. Meister JA, Nguyen KA, Luo Z. Audio feature ranking for sound-based COVID-19 patient detection. arXiv:210407128 [cs, eess]. Published online April 14, 2021. Accessed August 3, 2021.
  2. Pahar M, Klopper M, Warren R, Niesler T. COVID-19 cough classification using machine learning and global smartphone recordings. Computers in Biology and Medicine. 2021;135:104572. doi:10.1016/j.compbiomed.2021.104572
  3. Sadiq F, Masood K, Muhammad A, Riaz Khan A. Screening of COVID-19 using Cough Audio Frequencies. IJATCSE. 2021;10(3):1851-1861. doi:10.30534/ijatcse/2021/511032021
  4. Sharma N, Krishnan P, Kumar R, et al. Coswara — A Database of Breathing, Cough, and Voice Sounds for COVID-19 Diagnosis. In: Interspeech 2020. ISCA; 2020:4811-4815. doi:10.21437/Interspeech.2020-2768
  5. Sharma NK, Muguli A, Krishnan P, Kumar R, Chetupalli SR, Ganapathy S. Towards sound based testing of COVID-19 — Summary of the first Diagnostics of COVID-19 using Acoustics (DiCOVA) Challenge. arXiv:210610997 [cs, eess]. Published online June 21, 2021. Accessed August 3, 2021.
  6. Topol EJ. Is my cough COVID-19? Lancet. 2020;396(10266):1874. doi:10.1016/S0140-6736(20)32589-7
  7. WHO Target Product Profiles for COVID-19 Diagnostics. World Health Organization; 2020

Leave a Reply

Your email address will not be published. Required fields are marked *

Post comment

Recent Posts