Q&A: The New Standard for Measuring Cough with Tamsin Chislett

Let’s start with the basics: what is the CoughMonitor Suite?

CoughMonitor Suite is Hyfe’s research tool designed specifically for objectively measuring cough in clinical studies. It is used by pharmaceutical sponsors and academic researchers who want to understand cough in different diseases and how it changes over time.

Unlike traditional tools that only capture 24-hour snapshots, our system monitors cough continuously, 24 hours a day, seven days a week. That opens up the ability to observe long-term patterns across a study or treatment period for the first time.

Watch the full interview here or scroll down to continue reading

‍

Can you walk us through how the system works?

CoughMonitor Suite has three main components:

• The Watch: A lightweight device preloaded with Hyfe’s cough detection software. It continuously detects and timestamps coughs, without recording audio.
• The Companion App: Available on iOS and Android, the app collects patient-reported outcomes (PROs) and transfers cough data from the watch to the cloud.
• The Dashboard: This is where study teams and sponsors can manage participants, track adherence in real time, and view cough data as needed.

The core technology works like a step counter but for coughs. Our software uses two AI models: one to detect explosive sounds and another to classify whether those sounds are actually coughs. The result is a precise log of when each cough occurred, with no need for human review or manual counting.

How does CMS handle privacy?

Privacy was a top priority in our design. We knew that participants would not want to wear a device that records or transmits audio from their daily lives. So we engineered the system so that all audio processing happens directly on the watch, and only cough timestamps ever leave the device.

This on-device approach means no audio is recorded, stored, or transmitted at any point. It has been a key factor in building trust with participants and ensuring they are comfortable wearing the device for long periods.

Can you share a bit more about the validation?

Yes. We recently completed a 30-person validation study. Participants wore the Hyfe device on one wrist and an audio recorder on the other for 24 hours. The audio was labeled by trained professionals, and we compared that against what the Hyfe device detected.

We found a Lin’s concordance correlation coefficient of 0.975, median sensitivity over 90 percent, and just 1.56 false positives per hour. All of this was achieved while participants went about their normal daily activities, including driving, using tools, and listening to music. We are proud of how accurate the results are in real-world conditions.

Can you tell us a bit more about Hyfe’s experience in providing CMS to researchers?

We have supported over 50 studies globally across a wide range of conditions and research phases, from proof of concept through to Phase 2b, and now into Phase 3 trials. We are also supporting large real-world evidence studies running in multiple countries.

Over that time, we have built out comprehensive services that make CMS easy to use, including logistics, training materials, helpdesk support, and dashboard customization. Whether researchers need full-service support or just the core technology, we can adapt to their needs.

How does Hyfe support high adherence to the technology?

Adherence is essential. If participants are not consistently wearing or charging the device, the data will not be usable. So we made a number of design choices to support this.

We chose a familiar, low-profile watch form factor that is comfortable and discreet. The battery lasts over three days, but we encourage nightly charging so we can monitor that process. Our system includes automated alerts for study teams if a participant is not syncing, wearing, or charging the device.

Importantly, syncing is fully automatic. Participants do not need to take any action. And because all processing happens on the watch, privacy is never a barrier to wearing it.

Why is this kind of tool needed in clinical research?

Historically, researchers had two imperfect options. One was to ask patients how much they think they are coughing, which is subjective and often unreliable. The other was to use audio recorders that require manual labelling. Those are expensive, intrusive, and only cover short timeframes.

Cough is also highly variable. Measuring it for just 24 hours could capture a good day or a bad day, but not the full picture. CMS makes it possible to observe how cough patterns change over time, which is essential for understanding treatment effects.

What can researchers learn from using CMS?

Because we timestamp every cough, researchers get granular, high-resolution data that opens up a wide range of analysis.

You can study bouts, which are clusters of coughs in quick succession. These often have the greatest impact on quality of life and can be linked to other symptoms. You can also analyse silences, or periods when the patient is cough-free, which are often highly valued. And you can detect change points, where cough patterns shift, to understand when medication starts working or when triggers arise.

It is not just about counting coughs. It is about discovering patterns that were previously invisible.

‍