Text Link

The VitaloJAK vs. Hyfe CoughMonitor Suite

August 29, 2024

VitaloJAK vs Hyfe? Hyfe vs VitaloJAK? TThis article offers a detailed comparison of their features, capabilities, accuracy, and other critical factors to help you make an informed decision.

Overview of VitaloJAK

The VitaloJAK was developed through a collaboration between Vitalograph, a medical device company, and the University Hospital of South Manchester. This system uses a lapel microphone, a contact microphone, and a recording device worn in a belt bag, designed to record audio for cough monitoring for up to 24 hours. It functions as a medical-grade sound recorder, continuously capturing all sounds for 24 hours.

See VitaloJAK worn by a person here

The system comes as a series of different components that include:

  • The VitaloJAK unit itself 
  • A battery and memory unit, that comes in a separate kit
  • Various cables required to connect to a dedicated PC

See VitaloJAK's Main Components diagram and list on pages 4 and 5 of their instructions manual

The device needs to be set up by a trained professional and the patient will wear it for up to 24 hours, during which time the device will record every sound, continuously.

At the end of the monitoring period, the device is collected by the research team, who connects it to a computer to transfer the unprocessed recording data. A proprietary program then compresses the recordings by removing silent periods and obvious non-cough sounds. Finally, human analysts listen to the remaining audio recordings and manually label the coughs by identifying and marking only the beginning of each of the explosive phases.

Here is the device shown next to a credit card for scale:​

The VitaloJAK is the original clinical cough monitoring tool, having been a dominant choice in clinical trials for many years. However, during a recent Center for Drug Evaluation and Research (CDER) advisory committee meeting in November 2023 regarding Merck’s 2019 Gefapixant briefing document, U.S. Food and Drug Administration (FDA) reviewers noted that extending monitoring beyond 24 hours could potentially be more effective. 'Ideally, cough frequency would have been captured through the end of the 52-week treatment periods​'.

How does VitaloJAK compare with Hyfe CoughMonitoring Suite (HCMS) Overview?

The Hyfe CoughMonitor Suite is a wristwatch that resembles an Apple Watch. It uses an AI-powered model that continuously identifies cough sounds. The watch can monitor coughs for extended periods - weeks, months, or even years - without recording or storing any audio files. Instead, it generates a timestamp for each cough. The watch syncs with a smartphone app via Bluetooth, which uploads the timestamp events to an insights server. This setup enables researchers and drug developers to access real-time cough dynamics dashboards, customized to their specific research and clinical trial needs and never worry about privacy, file sizes or computational resources. 

It also allows patients and participants to go about their lives and forget about monitoring their coughs.  

Hyfe has been used in nearly 50 studies and is increasingly used in drug development trials across all phases, from Proof of Concept to Phase I and Phase IIb, as well as in Real World Evidence (RWE) studies. Its slim and lightweight form factor, continuous monitoring capability, accuracy in real-world settings, and privacy-preserving features contribute to its utility as a leading research tool for continuous cough monitoring.

Presentation

VitaloJAK: What’s under the hood?

The VitaloJAK consists of a white plastic casing - the size of three iphones stacked on top of each other - featuring a simple LED screen, and three buttons. It operates on a 32-bit ARM 96MHz processor, optimized for sound recording with analog filters and amplifiers. Once activated, it continuously records all sounds. The on-off button is secured with a physical key, and there is a mute button for short-term silencing. The recorded sound is stored on an standard SD card inside the device, with no on-device computation or compression.

See VitaloJAK Menu Naviation using the Keypad on page 9

Given the device’s computational resources, after recording, the raw sound files must be manually transferred to a separate machine for initial processing (compression). Finally, the compressed recordings are then reviewed by human analysts who manually label participant coughs by marking the beginning of each explosive phase. VitaloJAK’s cough annotation procedures are not publicly available or published; questions about the accuracy of annotations prompted the FDA to reject the initial New Drug Application (NDA) for Gefapixant in the first cycle, citing “insufficient validation data to support that the VitaloJAK cough counting system… provided a reliable and accurate assessment of the primary endpoint, cough frequency”. After a recount of the data and a validation study, the FDA accepted the (revised) cough counts, but noted that “The validation of the VitaloJAK device and compression algorithm used in the gefapixant program is unique to the gefapixant pivotal trials and in no way implies validation beyond the existing 510(k) device clearance (i.e., this device and algorithm validation is not considered generalizable).”.

The VitaloJAK cough monitoring process involves three stages:

  1. 24 Hours of Continuous Recording: Using two microphones - a contact microphone and a lapel microphone used for validation - the device records all sounds, storing them on an SD card.
  2. Data Transfer and Initial Processing: The recording files are then downloaded to a separate machine, where a specialized program eliminates quiet periods and non-cough sounds.
  3. Manual Review: Human operators listen to the remaining recordings to identify and label explosive phases of cough sounds.

How does that compare with Hyfe CMS?

The Hyfe CMS uses a wrist-worn wearable that operates continuously. The system includes a device that is a consumer-grade smartwatch that complies with multiple international certifications, ensuring its safety and quality across various markets. Among others, the device has obtained UKCA certification for the UK, TELEC for Japan, and both FCC and IC certifications for the U.S. and Canada, respectively. Additionally, it meets the CE marking requirements for the European Union and adheres to ROHS and REACH directives, which regulate hazardous substances and chemicals. The device is also certified for safe air and sea transportation, ensuring compliance with global shipping standards.

The Device includes sensors and computational power typical of modern premium smart wearables. The cough detection model, powered by AI, runs using state-of-the-art edge computing, enabling real-time, continuous detection of coughs without recording any audio. The device does not record sound; instead, it generates timestamps corresponding to detected coughs. These timestamps are processed by Hyfe’s insights engines, which provide investigators with cough analytics in real time.

The Hyfe CoughMonitor Suite process involves the following stages:

  1. Continuous Passive Cough Monitoring: The smartwatch continuously monitors coughs passively.
  2. Data Synchronization: Timestamps of coughs, battery levels, and wear time are synchronized from the smartwatch to the smartphone-based companion app via Bluetooth. The smartphone can be provisioned, or study participants can use their own devices for the app to function. The companion app has two key functions:some text
    1. Syncing data from the smartwatch.
    2. Collecting Patient Reported Outcomes (PROs) with investigator-customized surveys, along with set notifications and alerts when necessary.
  3. Cloud Data Synchronization: Data from the companion app is synchronized to the cloud via any internet connection on the smartphone. Cough dynamics, battery levels, wearing time, and PRO inputs become instantly available to investigators, sponsors, and partners on Hyfe’s web-based dashboard.

Wearing the smartwatch during the daytime and keeping it nearby at night (for example, on the bedside table) is the only requirement. The device is discrete and familiar. It can be worn continuously - even while showering or sleeping, and its battery lasts over 80 hours. While overnight charging is recommended, the device continues to function for 3 days, if charging is missed. Researchers can access real-time cough and battery data through customizable web dashboards, regardless of their or their study participants’ locations.

Accuracy and usability

The VitaloJAK tool has a long history of being used in clinical research and has been for many years the standard in cough counting in clinical trials. There has been one published validation of the software algorithm involving10 individuals. The system is FDA cleared to capture sound, but the cough-detection component (including the compression algorithm and human annotation process) is not FDA cleared.

The VitaloJAK’s 24-hour recording limit presents a challenge, as it may not capture a comprehensive view of a patient’s cough profile over multiple days. Cough frequency can vary significantly, and a single 24-hour period does not accurately reflect overall cough patterns.

Additionally, the physical setup of VitaloJAK - requiring participants to wear a large device with connected wires - may trigger the Hawthorne effect. Though no data exists on this topic in cough, this behavioral response occurs when individuals alter their behavior due to the awareness of being observed. In plain terms, people unconsciously change how they behave and cough because they know they are being monitored

Finally, the use of VitaloJAK, along with its reliance on human annotation, introduces significant friction and delays in data access for trial sites and investigators. Since the process is not automated, it can take weeks or more for the data to become available. At scale, in large trials, this becomes a significant source of friction. 

For further insights on the limitations of 24-hour monitoring, refer to this paper. Below is a visual representation highlighting the challenges associated with a 24-hour recording period:

A patient’s cough was monitored continuously over several months, showing two key trends (indicated by red arrows from left to right): first, an initial decrease in cough frequency, followed by a subsequent increase.​

However, if you select three random 24-hour snapshots of cough frequency monitoring (indicated by red circles, e.g., using a limited 24-hour cough monitoring device), the overall trend is likely to be missed. You can learn more about day-to-day cough variability in persistent coughers by reading the recent publication in LUNG.​

How does that compare with Hyfe CMS?

The Hyfe CMS has been used in 47 studies to date. A paper on a validation trial demonstrates its real-world sensitivity above 90%, with a false positive rate of 1 false positive cough per hour.

In the Hyfe CMS validation trial, 23 adults with problematic coughs from various causes wore two smartwatches: one running Hyfe’s AI-powered, fully automated, passive cough monitoring system, and the other recording audio as an MP3 recorder for 24 hours. Participants wore the two watches while going about their daily activities. Since human-annotated cough sounds are the gold standard for evaluating automated cough monitors, two trained, blinded annotators reviewed the full 24-hour MP3 recordings for each participant, following Hyfe’s validated SOP for cough annotation. Any discrepancies were adjudicated by a third experienced expert annotator. Finally, human annotations were compared to Hyfe’s AI-powered cough counts.

Hyfe’s CMS can monitor cough for long periods of time - even years - without any interruption. The models run passively and continuously and all participants have to do is wear a small and light wristwatch - a familiar and non-obtrusive format. Due to the combination of long-term monitoring and convenience, the Hawthorne Effect is suspected to be minimal or non-existent. Researchers can monitor the cough frequency in real time on dashboards that have been customized for the specific requirements of their study.

The fact that Hyfe’s CMS does not record sounds is a significant benefit to privacy-minded individuals as well as to researchers and sponsors who want to operate at scale and do not need to worry about computational resources. However, it might arguably also be a drawback, as the lack of recordings means that backward validation/verification of data against the recordings is not possible. This means that researchers have to rely on the performance (proven, validated) of the AI models themselves. In some studies, Hyfe’s CMS has been used alongside a separate validation device - a sound recorder similar to the VitaloJAK in performance - and such approaches are always possible.     

Convenience/ Participant's comfort

The VitaloJAK device poses challenges in terms of convenience and participant comfort. The recording device is worn in a pouch attached to the belt or shoulder (see the visual below), with wires connecting it to a contact microphone on the chest and a lapel microphone. Participants are constantly aware that their surroundings are being recorded and analyzed. The device may also draw attention from bystanders, family, and friends, potentially reminding participants that their cough is being evaluated for that particular day (24 hours). On page 14, the “Vitalograph VitaloJAK 7100 Instructions For Use Manual” of the instructions, advises: “The subject should try to avoid noisy environments while wearing the VitaloJAK device. For example, the cinema, driving with the windows down, the gym, or using the hair dryer”

Participants using the VitaloJAK may need to significantly adjust their routines, potentially impacting their coughing behavior, especially if their cough is triggered by specific activities or environments.

The continuous audio recording and human review of all captured sounds may be perceived as privacy-intrusive by some, potentially affecting their comfort and willingness to engage with the device. There is also the risk that a participant may lose or misplace the device - containing hours of continuous recording - personal, identifiable data - on a standard SD card. 

How does that compare with Hyfe CMS?

The Hyfe CMS comes in a familiar form factor - a discrete, premium smartwatch (see the visual below). A participant wears the slim and light device on their wrist and can carry on with their daily activities, including showering, exercising, going to the gym or cinema, without interference or worry. The device is charged overnight, similar to other smartwatches, and if occasionally forgotten, its battery lasts up to 80 hours. Cough monitoring can be continued for weeks, months, or even years.

Hyfe’s cough monitoring device operates without recording audio. Instead, it detects coughs in real time and generates timestamps, which are synchronized with Hyfe’s Insights server. Investigators and partners can access the data through customizable real-time dashboards tailored to their specific study requirements.

Ease of use

The VitaloJAK device is custom-built and relatively large, roughly the size of three stacked iPhones 14. It comes with a pouch and the manufacturers recommend the use of this pouch (which comes with a sling over the shoulder).   VitaloJAK includes wires connecting a contact microphone, which must be adhered to the participant’s chest, as well as a lapel microphone. Once activated, the device is “locked,” and the user can interact with it by tapping one of three buttons: one to mute the device (e.g., during a sensitive call) and the other two to mark timestamps for specific events. The device is worn in the standard issue pouch for 24 hours, after which it is returned for processing. According to the device’s manual, participants cannot shower, go to the gym, drive, or get the device wet. They also are recommended to sleep with the device attached or to put the device on the bedside and connect the microphone to their night clothes or pillowcase. The Contact microphone must stay connected at all time:

 “If required for comfort overnight, the pouch may be placed on a bedside locker, with the audio microphone clipped to the lapel of the user's nightclothes or edge of the pillowcase. The chest sensor must remain in place for the entire recording period”.

Data processing involves a delay, as recordings must uploaded, compressed, and then manually reviewed by human annotators.

In terms of ease of use for researchers, in order to commence and complete monitoring, specially trained staff need to set up the device and then collect it 24 hours later. Specially trained staff need to attach the contact microphone and detach it at completion. The devices need to be connected to a dedicated computer running a “web portal” for data transfer and the first phase of processing (compression). Finally, the compressed audio needs to be labeled by specially trained human analysts.   

How does that compare with Hyfe CMS?

The Hyfe CMS is worn on the wrist and is indistinguishable from any modern premium smartwatch. It is waterproof, requires no special attention and functions anywhere in the real world. Participants can charge it overnight on their nightstand, even while it continues monitoring, with charging required at least every other day. The device charges in full in a few hours.

Cough detection occurs directly on the device without the need to record or store any sound. Participants can go about their lives with the peace of mind that no sound is recorded and there are no privacy risks involved. The data, consisting of timestamps, synchronizes via Bluetooth with a companion app on the participant’s smartphone (or a provisioned phone for the study). 

Researchers do not need to be physically present on site. Participants receive the device (at home/a central location) and they are ready to go. Researchers can then access real-time cough data through web-based dashboards the minute the device is on.

Essentially, the researchers get:

  • Individual, millisecond-granularity cough timestamps in raw, analysis-readable tabular format
  • Real-time dashboard access for monitoring participant enrollment, usage, and cough rates
  • Detailed adherence usage, including minute-to-minute data on when the device was worn, charging, etc.
  • Cough-derived aggregated analysis endpoints including bouts, epochs, oases, etc.

Additionally, the system offers the ability to automatically generate QR codes for subject enrollment, as needed.​

Available Cough Data and Analytical Insights

After the VitaloJAK records 24 hours of raw audio, trained staff connect the device to a computer and download the audio file via USB. A proprietary compression program is then run on the computer, eliminating silence and obvious non-cough sounds. The remaining, smaller audio files are passed to human annotators who listen and manually label cough explosive phases. The output is a series of human-generated, manually inputted timestamps representing the participant’s cough frequency (by counting the number of explosive phases) during the 24-hour recording, excluding any periods when the participant may have muted the device. No additional cough insights are provided.

How does that compare with Hyfe CMS?

The Hyfe CMS monitors cough continuously for weeks, months, or even years, marking each cough automatically with microsecond-precise timestamps. These timestamps sync with Hyfe’s Insights API, where dashboards are generated in real time, tailored to the analytics most relevant to investigators. Beyond cough counts, the dashboards can provide additional insights such as cough bouts,periods of relief from cough,  cough timing, and metadata like patient-reported outcomes entered via the companion app. The researchers also get insights on detailed adherence, including minute-to-minute data on when the device was worn, charging, etc.

Additional data streams can be easily integrated as needed by the study design (activity, movement heart rate, etc).

Industry Recognition

The VitaloJAK device was a pioneering tool and, for many years, the only automated solution for cough monitoring in clinical trials. It has been widely used in numerous studies worldwide and has become a default standard in clinical research. However, the FDA advisory committee’s emphasis on monitoring cough beyond 24-hour snapshots has led to a swift industry shift toward alternative cough monitoring solutions, focused on continuous longitudinal monitoring.

How does that compare with Hyfe CMS?

Hyfe’s CoughMonitor Suite is currently being used by drug developers seeking FDA approval, including publicly announced partnerships with Nocion and SensoryCloud. Multiple other developers are actively discussing Hyfe with the FDA’s Center for Drug Evaluation and Research (CDER). Furthermore, Hyfe is increasingly being explored as an endpoint in indications beyond refractory chronic cough / unexplained chronic cough (RCC/UCC), particularly in Idiopathic pulmonary fibrosis (IPF), chronic bronchitis and bronchiectasis for Phase 2b and Phase 3 trials (confidential).

Regulatory status

VitaloJAK has FDA approval (510K) as an audio recording device. Its audio compression algorithm, removal of silent and non-cough periods, and human annotation processes are not approved by the FDA or any other regulator. This means that the VitaloJAK system is not FDA-approved for cough monitoring, only for audio recording. 

In fact, there are currently no FDA-approved devices for cough detection or cough counting.

How does that compare with Hyfe CMS?

Hyfe’s CMS is also not yet approved by the FDA. Hyfe’s AI-powered cough monitoring system has a CE Class I mark, and the company is pursuing FDA de-novo clearance for its fully automated, passive continuous CMS in adults with problematic cough of any etiology.

Conclusion

The VitaloJAK has been a long-established tool for cough monitoring in clinical studies, though it is limited to 24-hour periods. Its design reflects a traditional approach to sound recording, relying on analog technology and manual processes for data handling, human annotation, and analysis. While it has been a standard choice due to its historical use and familiarity in clinical trials, it is limited in terms of form factor, ability to monitor for long periods, and privacy elements. These factors are critical considerations for researchers, clinicians, and biopharmaceutical professionals when selecting tools for cough monitoring as an endpoint in current and future studies.

Who is VitaloJAK best for?

Given its long history and the inherently high accuracy that comes with manual labeling carried out by humans, the VitaloJAK may be the ideal choice for studies that (a) only require 24h of data collection, (b) do not require any real-time monitoring, (c) are unconcerned by the potential “Hawthorne effect” of participants changing their behavior due to the heightened awareness of monitoring from using a large device, (d) have a large budget and/or (e) have a small sample size.

More specifically, The VitaloJAK is best suited for short-term studies that prioritize detailed audio recordings and human-annotated data but can tolerate its limitations in terms of convenience, privacy, and long-term monitoring.

  1. Short-term clinical studies: It is designed for monitoring coughs over maximum 24-hour periods, making it ideal for studies that require limited-duration cough data, such as those focused on capturing detailed cough activity within a specific short time frame.
  2. Studies that value continuity with previous methods - As a well-established tool in clinical research, VitaloJAK may be preferred in studies where consistency with past data collection protocols is essential.
  3. Research environments with manual data processing capabilities: The VitaloJAK's data processing relies heavily on human annotators to review audio recordings and manually label coughs, which makes it suited for research teams equipped to handle this labor-intensive process.
  4. Studies that require audio recordings: Since VitaloJAK captures raw sound files, it is valuable in studies where investigators need the actual audio data to analyze or validate later 
  5. Controlled environments: It may be useful in settings where participants are monitored in controlled environments, as the device is more intrusive, requiring participants to wear microphones and adjust their routines to avoid interfering noises.

Who is HCMS for?

Hyfe's Cough Monitoring Suite (CMS) is best suited for studies where real-world usability, long-term data collection, adherence to protocol, privacy, and participant comfort are priorities.

Specifically, CMS is ideal for researchers, drug developers, and clinical trial investigators who require:

  1. Long-term monitoring: Hyfe CMS can monitor coughs for extended periods—weeks, months, or even years—without interruption, making it ideal for studies that need comprehensive, continuous data over time. There is evidence that short monitoring periods - such as 24h only - are insufficient due to substantial within-day and day-to-day variability  
  2. Real-time data and ease of use: The system provides instant access to cough data through customizable dashboards, reducing delays in data processing. Researchers can work with any number of participants concurrently, even across multiple locations. 
  3. Privacy concerns: Since the system does not record audio but only generates timestamps for cough events, it is beneficial for privacy-minded individuals or studies where participant privacy is a concern.
  4. Adherence due to Convenience and participant comfort: With its familiar smartwatch form factor, Hyfe CMS is discrete, comfortable, and easy to wear continuously, even during activities like showering or exercising. This is reducing the likelihood of adherence failures during studies.

Scalability and automation: The system's automated AI-powered cough detection, combined with real-time synchronization and no need for manual data review, makes it highly efficient for large-scale studies.

Ready to get started?

talk to our clinical expert