Remote trials, real challenges: The quest for accurate cough monitoring

By Liza Laws

- Last updated on GMT

© Getty Images
© Getty Images

Related tags cough monitoring Hyfe AI

Within the rapidly evolving landscape of decentralized clinical trials (DCTs), one critical challenge persists: the accurate monitoring of cough symptoms.

As the healthcare industry continues to embrace remote trial methods, the need for reliable cough monitoring solutions becomes increasingly apparent, says Dr Peter Small, of Hyfe AI. OSP put forward a number of questions to him to delve into the complexities surrounding cough monitoring and explore the potential implications for clinical research and patient care.

He believes we’re in a perfect storm of wearable sensors, powerful AI models, reliable monitoring platforms and digital therapeutics, which can change the process and enhance the accuracy of monitoring, ultimately leading to new treatments and a radical change in respiratory care. OSP: Why has the widespread use of cough counting tools, to date, been limited by a reliance on human input to determine cough frequency?

PS:​ It's not been for a lack of effort. Folks have been trying to automatically monitor coughs for over 25 years. The initial efforts were based on a variety of sensors such as stretch receptors on vests or EMG recordings from straps. A couple of decades ago a major advance came from using contact and ambient microphones linked to a belt with a sound recorder and data compression algorithms. It was only by having people listen to these recordings that coughing could be quantified.

A variety of rules-based approaches were attempted for identifying cough sounds from such recordings, but the performance of these approaches was limited and insufficient for use in clinical trials. The game changer has been the application of acoustic AI to massive cough databases to create cough recognition algorithms. With sufficient quantities of well annotated diverse data to train on these algorithms can automatically recognize cough from other explosive sounds in near real time with amazing accuracy. These algorithms can be rapidly scaled given the ubiquity of high-quality microphones and chips sets in smart devices.

Over the last two decades, advances in digital technology and audio capture have reduced this dependence, could you give your thoughts on this?

The major impediment to innovation in cough monitoring is regulatory uncertainty. As per Dr. Rachel Bean, Medical Officer at the FDA, 'the VitaloJAK device holds an FDA 510(k) clearance as an audio recording device only. This does not include compression or cough counting​'. It has been for many years the best option and one that is well known to sponsors and regulators. However, it was clear from the November 2023 CDER advisory committee meeting about Merck's 2019 Gefipixant briefing document that FDA reviewers consider that monitoring for longer than 24 hours would be better 'Ideally, cough frequency would have been captured through the end of the 52-week treatment periods​'.

Moving to continuous, passive, privacy-preserving monitoring is a scientific and statistical no-brainer but has not to my knowledge been proposed to the FDA by a sponsor. There has been an explosion in cough science in the past couple of years - providing new insights into how to best identify research subjects and determine if their cough has really changed with interventions. For example, most coughers naturally have good and bad days so that just monitoring 24 hours can be misleading; if you enroll a subject on a good day and re-measure on a bad day you will miss any therapeutic impact. And if you enroll on a bad day and re-measure on a good day, you’ll overestimate or misattribute therapeutic effect. It’s my view that the FDA will be convinced of the value of continual monitoring when confronted with rigorous science-based endpoints. 

OSP: Cough frequency is increasingly recognized as a measurable parameter of respiratory disease, what difference is this making?

PS: ​Cough is a big deal for patients and health systems. Chronic cough is as prevalent as asthma and can have devastating social and medical consequences for sufferers. Furthermore, cough is a harbinger of worsening for many conditions such as COPD exacerbations and heart failure decompensations. It's embarrassing that in this era of precision health, where everything gets measured, one of the most common reasons people seek care is simply not measured! Scalable, low-cost and accurate cough monitors will do for cough what the thermometer did for fever 300 years ago. Imagine trying to treat hypertension without knowing how a patient’s blood pressure is varying over time? I predict that in a few years anyone with a cough-related condition will have a cough monitor,like they now have a BP cuff for hypertension, and that soon thereafter the ability to monitor coughs will be ubiquitous on medical and consumer devices and patients will pay attention to those counts when sick or managing a chronic condition.

Cough frequency is now apparently the gold standard primary endpoint for trials of new treatments, what sort of trials and how can this assist with treatments?

The ability to passively quantify coughing will have impacts across the product development cycle. Natural history studies will provide real world evidence about the magnitude of the burden in populations and monitor-based patient registries will accelerate subject recruitment while improved understanding of cough statistics will focus selection on subjects so as to reduce placebo effects and minimize sample sizes. In studies, cough will first be used as an exploratory endpoint then soon as a primary efficacy endpoint. Proper frequency measurement will also bring rigor to cough as an adverse drug reaction such as with ACE inhibitors and inhaled medications. Finally, they will play a big role in post-marketing surveillance, patient engagement, and medication adherence.

I believe cough monitoring will have an even more immediate and profound impact on patient’s lives. It turns out that people don't really know how much they cough and so it's hard for them to know if an empiric therapy is working or they are having a disease exacerbation. Monitoring cough will quickly become a standard part of remote patient monitoring, hospital at home care, and post-hospital discharge.          

OSP: ​Can you talk us through automated cough detection and the current landscape of the technology's use in clinical research?

PS: ​The application of acoustic AI to cough science is a very hot topic. Many groups are testing the hypothesis that the intrinsic sound of a cough can be used to diagnose a condition. These are methodologically complex studies trying to do this rigorously. And yet, the overall direction of this work suggests that while cough might be a good screening tool, it will never reach the performance of in vitro diagnostics.

I work for Hyfe AI, a company focused on continuous cough monitoring that utilizes longitudinal data streams to detect cough dynamics, patterns, and signals over time, leveraging AI to identify even the most subtle variations in cough frequency. Our algorithms are built around a powerful cough / non-cough classifier that runs continuously and therefore doesn’t require any action on the part of the patient. Cough / non-cough classifiers are a much easier problem to solve than disease-specific classifiers because they do not rely on potentially weak indication-specific signals within individual coughs. They also don't require a controlled acoustic environment or special hardware or software. Our wellness apps can be downloaded in the usual places and the wearable is undergoing evaluation by the FDA for use in disease management. Our technology is being used in over 45 investigator-initiated and Pharma-sponsored trials globally.

OSP: ​What are your predictions on the future directions in the field based on recent developments?

PS: ​Looking backwards, the FDA has not approved a new antitussive for over 60 years and thus the drugs on the market are largely ineffective. But it’s not hyperbole to say we are entering the golden era of cough. Covid has focused the world's attention on cough and further stigmatized those of us with non-infectious chronic cough. Basic science has shown that cough is not only a symptom of common diseases - but cough also itself is a neural hypersensitivity disease. This neural pathway is now understood, and pharma has very promising anti-tussive in their pipelines. Automated continual cough monitoring running on medical and consumer devices is the glue that holds all of this together and promises to greatly improve cough care and the lives of those of us who suffer from it. 

Related news

Show more

Related products

show more

Saama accelerates data review processes

Saama accelerates data review processes

Content provided by Saama | 25-Mar-2024 | Infographic

In this new infographic, learn how Saama accelerates data review processes. Only Saama has AI/ML models trained for life sciences on over 300 million data...

More Data, More Insights, More Progress

More Data, More Insights, More Progress

Content provided by Saama | 04-Mar-2024 | Case Study

The sponsor’s clinical development team needed a flexible solution to quickly visualize patient and site data in a single location

Using Define-XML to build more efficient studies

Using Define-XML to build more efficient studies

Content provided by Formedix | 14-Nov-2023 | White Paper

It is commonly thought that Define-XML is simply a dataset descriptor: a way to document what datasets look like, including the names and labels of datasets...

Why should you use clinical trial technology?

Why should you use clinical trial technology?

Content provided by Formedix | 01-Nov-2023 | White Paper

New, innovative clinical trial technology is helping to revolutionize the research landscape. COVID-19 demonstrated that clinical trials can be run much...

Related suppliers

Follow us


View more