Digital health technologies can make cough monitoring as common as sleep tracking – providing brand new insights for health management.
Pranavsingh Dhunnoo
6 min | 28 May 2024
Key Takeaways
Cough is a common symptom that has traditionally not been easy to track as a health metric.
New digital health technologies, including specialised hardware and artificial intelligence-based software, can provide more details into one’s coughing habit for self-evaluation as well as clinical assessments.
This AI-powered push in cough monitoring promises to empower patients and improve clinical assessments; provided that privacy concerns are addressed during development.
Let’s try a quick guessing game (without searching for the answer online): how often does a healthy person cough in a day? If you’ve guessed 4.6 coughs, then you’ve hit the bull’s eye as this is what the most recent study indicates. We usually only pay attention to our coughing if it becomes more frequent and noticeable. In fact, cough accounts for one in five reasons we seek assistance from a healthcare provider.
While it is a common symptom of an underlying health condition, ranging from reflux to lung cancer, cough has traditionally not been an easy metric to track.
We didn’t need technologies to be able to differentiate between patients based on coughs, as primary care physicians have been doing that for centuries. But one’s cough can provide insights for self-monitoring, and digital health advances have started to facilitate the cough-tracking process. These could even make cough monitoring as common as step tracking to inform individual patients and provide doctors with deeper health insights.
In this article, we will consider the digital health tools that are facilitating cough monitoring. These can essentially be categorized into two groups: smartphone apps and specialized hardware. However, in each group, artificial intelligence (AI) is playing a central role in shaping the cough monitoring landscape to better aid patients in understanding their health.
The power of cough information
One question that will inevitably arise regarding cough monitoring is: why do it at all? If we’ve been managing without such a metric so far, then is there any reason to invest further in this? It turns out that there are plenty of reasons to do so as cough is a pretty potent biomarker.
Having even basic information about a person’s daily cough count and pattern can assist them in making more informed decisions. For example, those with conditions like COPD and asthma could glean insights into whether they are about to experience a flare-up. This could prevent exacerbations which would otherwise need intensive and costly treatment. In other cases, an increased incidence of cough in a region could be indicative of allergens or an infection. Public health authorities could use this information to take adequate measures.
“There is a tremendous value in simply empowering patients and doctors with this information,” says Dr. Peter Small, chief medical officer at Hyfe AI, a company that develops cough monitoring tools. “It’s going to transform the whole clinical approach for this common and chronic symptom. Patients will come in, have the data on how much they are coughing, and the physician can suggest a treatment based on that information to see if it makes the coughs better.”
This future that Dr. Small envisions might not be too far off as there are digital cough monitoring tools that are already available and more that are coming to the market.
Tracking apps for coughs
As the average person uses 9 smartphone apps daily, having an app for cough monitoring might be the most familiar way to capture this metric. Currently, available apps such as CoughTracker and CoughPro identify and record coughs from the user, analyse them through AI models, and provide contextual summaries.
Source: https://www.technologyreview.com/
Other similar AI-powered cough-tracking apps are being developed by Swasaa and Raisonance. Swasaa’s AI models function by combining cough recordings with other health metrics such as temperature and oxygen saturation to assess a patient’s lung performance. Through such an approach, the company aims to provide an alternative to spirometry, which is more affordable and accessible for pulmonary healthcare.
In the case of Raisonance, the company began by developing the AudibleHealth Dx app to identify COVID infections from cough recordings. Once a recording is taken, the app connects to the cloud for processing and analysis with the assistance of AI. They are now expanding to influenza and tuberculosis diagnosis with the same technology.
Research insights with dedicated, cough-monitoring hardware
Aimed primarily for research purposes, dedicated cough-monitoring hardware can provide further insights. Currently, the only such hardware available on the market is the clinically validated VitaloJAK. It functions as a wearable that tracks up to 24 hours of cough recordings. The companion software’s algorithm analyses the data, separates non-cough recordings, and identifies awake and sleep periods. The device has been used in Phase 2 and 3 studies as a cost- and time-effective solution to monitor patients’ cough patterns.
Other hardware options are also in production, although not currently available at scale. One example is the RESP Biosensor from Strados, which is also aimed at clinical trials. It is a wearable patch that allows participants to collect cough and lung sounds at home for subsequent analysis by researchers. A patient-oriented option in production is the wheezo system from Respiri. It detects abnormal breath sounds such as wheeze and allows users to generate logs and potential triggers. The recording can then be played back to a healthcare professional for review.
The AI push in cough monitoring
Over time, we will likely see more cough-tracking apps and hardware become available on the market. Some might target specific conditions, while others might aim at providing some fitness insights to healthy people, akin to fitness trackers. While they will have differences in their individual aims, cough monitoring tools appear to increasingly have one common feature: AI.
All of the examples listed above rely on an underlying algorithm that analyses the cough recording sample. Applying machine learning algorithms to such ends is being termed as acoustic AI.
“When trained on sufficiently large and diverse collections of annotated sounds, these approaches can recognise and classify sounds such as a heart murmur, a wheeze or a cough,” explains Hyfe’s Dr. Peter Small. “And, when incorporated into digital stethoscopes, they can improve the diagnostic capabilities of clinicians. By tracking their frequency, they can assist with medical management decisions. And ultimately, when incorporated in consumer products, they will empower patients to have a better understanding of and be better able to cope with their chronic medical conditions.”
In addition to empowering patients with respiratory conditions or simply monitoring their health with a new metric, cough tracking can improve clinical assessments. However, the development of efficient AI tools for such ends will rely on quality training data. Developers of such software will have to address privacy concerns and obtain explicit consent from users to use their cough recordings.
Already tech giants like Google are showing interest in providing such tools to the general public, which might heighten privacy anxieties. As the field of cough monitoring is relatively nascent and evolving, it would benefit from taking these concerns into account at an early stage.
Written by Dr. Bertalan Meskó & Dr. Pranavsingh Dhunnoo
Leave a Reply