Stay up to date with cutting-edge cough science

Cough Science News, March 1, 2024

March 1, 2024
Lola Jover

The Research Roundup

New cough science publications vetted and collected in one place

1. Can AI cough analysis optimize the use of chest x-rays in low-resource settings?

Key Takeaway: Based on 137 patients in Bihar, India, acoustic AI models of solicited coughs showed a signal in predicting, to some extent, whether a chest radiographic examination would have normal or abnormal results. The logistic regression model performed best, with an area under the receiver operating characteristic curves ranging from 0.7 to 0.78.

Why It Matters: Chest radiography is a valuable tool for diagnosis, but it is expensive and time-consuming, particularly in resource-constrained settings. Analyzing solicited coughs using AI-enabled algorithms could be a quick, inexpensive and widely available triage tool to determine which patients are likely to have radiographic abnormalities.

Read publication.

2. Which attributes of chronic cough and potential treatments are most important to coughers?

Key Takeaway: In a discrete choice experiment, frequency of cough attacks was the most important to participants, followed by taste change, then nighttime coughing, then daytime coughing. The experiment focused on two constructed treatment options characterized by varying attribute levels.

Why It Matters: Frequency of cough attacks appears to have the most damaging effect on chronic coughers’ quality of life, suggesting this measure should feature more prominently in the development of future therapies and efforts to measure their effectiveness.

Read publication.

3. Audio-based AI classifiers can predict Covid infection status, but are they better than simple symptom checkers?

Key Takeaway: Acoustic AI classifiers predict Covid status with high accuracy, but after matching on measured confounders, such as self-reported symptoms, performance is much weaker. In practical settings, the utility of audio AI classifiers for Covid status is outperformed by predictions on the basis of user-reported symptoms.

Why It Matters: Study design to reduce bias within the data, and the treatment of confounders are extremely important in AI-enabled diagnostics. This paper suggests limits on the transformation of diagnostics based on audio classifiers.

Read publication.


  • An exploration of the relationship between subjective and objective cough frequency data in patients with pulmonary TB (more)
  • The potential of longitudinal cough-monitoring for pulmonary sarcoidosis (more, and see this month’s Researcher Deep Dive below)

Deep Dive with Researchers

Exploring the Possibilities of Longitudinal Cough Data


Q&A with Dr Marc Judson

Chief of the Division of Pulmonary and Critical Care Medicine at Albany Medical College, NY, USA.

“...cough monitoring in sarcoidosis, will allow for an earlier reduction in toxic medications as well as earlier evaluation of pulmonary sarcoidosis exacerbations”

Dr Judson spoke with us about his research into the rate of improvement of cough symptoms for patients being treated for active pulmonary exacerbations of sarcoidosis.  The surprising speed of improvement suggests there is much more to discover about the mechanism of cough in pulmonary sarcoidosis.

Read the full Q&A here.

Cough Science Events Ahead


Barcelona International Cough Conference

Our CEO, Joe Brew, will be attending the conference with Carlos Chaccour, Clinical Lead, who will be presenting the session 'There be Dragons: Going Beyond 24 Hours in Cough Monitoring' on March 15th at 11:40 am CET.

ERS Lung Science Conference

Our COO Tamsin Chislett will be attending the ERS Lung Science Conference in Portugal 14-17 March. She's looking forward to hearing the latest in novel lung research and meeting with respiratory researchers.

If you're attending either of these conferences, drop us a line or book a call. We’d love to demo Hyfe, and hear about your research interests.

For more information and upcoming dates, visit:

Share post: