The Case for Objective Cough Interpretation
What if we placed the same importance on interpreting cough as we do in the heart and lung auscultation findings or ECG results?
Coughs, as we know, are nuanced in their sounds, consistency, and frequency. Coughs are even seasonal for many, either making an appearance with a change in weather or increased respiratory allergens. Some coughs occur after taking certain medications.
Assessing A Cough
Although a cough could be a benign process in some situations, as a response to transient throat irritation, in other instances, a cough is a manifestation of an underlying disease.
If recorded, the acoustic variations of cough hold potential clinical benefits if adequately analyzed. Unfortunately, manually capturing this data is not reliable enough for clinical applications. However, cough acoustics have clinical value, as any sufficiently experienced expert can attest (e.g., a pulmonologist who can suggest diagnoses by the sound of the cough). For instance, some coughs, such as whooping cough or barking cough, have unique sounds that can help to diagnose a possible cause.
If recorded longitudinally, it becomes possible to register cough frequency. Additionally, individual symptoms associated with a cough, such as dyspnea, fatigue, or fever, can provide a broader picture and provide context for that cough.
Reliably capturing cough data can also enable inferring other dimensions of cough, such as providing a more precise picture of their productivity. For instance, a dry cough could be a hacking cough or come in paroxysms, as observed with whooping cough. A wet cough can produce mild mucus to profuse mucus observed with bacterial infections. Some instances might record the production of blood.
Digital Technology and Cough
This is essentially where technology steps in to bridge the gap.
Surprisingly, there is already a lot of technology currently available to track and understand the differences in cough. Since manual cough tracking can result in human error, technology consisting of computing devices including a microphone and specialized software algorithms can reliably provide continuous monitoring.
Besides enabling cough data capture, digital technology can also register what accompanies the cough. This complementary data could include symptoms, mucus production, medication intake, and even the feeling right after the cough. In addition, environmental triggers potentially contributing can also be registered.
The intervals between the coughs are also of vital significance. In conditions like heart failure, chest heaviness, swelling, difficulty sleeping due to fluid collection in the lungs can come together to point to a cause external to the lungs. Just mentioning a nagging cough, in this case, would not provide an accurate overall picture.
With all this data, AI technology, or machine learning algorithms, can be created and refined to help interpret potential causes. Moreover, with the possibility of gathering data around the globe (big data), the variations in both cough and its related symptoms for just one condition can help refine the algorithms. The increasingly large data sets, also known as “big data,” can allow algorithm interpretation reliability to increase substantially. Most importantly, reliability can rise to fulfill the criteria mandated by health regulatory agencies.
The extremely high scalability and remote diagnosis possibilities of AI technology allow the precious healthcare professional’s time to be optimized, directed towards confirming alerts raised by the algorithms and dedicating more time to the most important people in their practice — their patients. Ultimately, healthcare providers can save both time and expenses by properly implementing AI technology.
The Future of Cough Investigation
Today an ECG can help identify heart blocks, myocardial infarctions, pulmonary edema, and anomalies in basic heart function. This diagnostic richness derives from the changes in the electrical signals from the heart.
Similarly, when tracking a cough, the acoustic variations, the symptoms, the environment resulting in a cough can all help conceptualize a possible reason for a cough. As our current state of affairs involves dealing every day with the Covid-19 pandemic, something as simple as an increase in the coughs tracked clustered geographically could draw attention, enable finding the underlying causes, and substantially help contain an outbreak. This acoustic cough surveillance specifically, or syndromic surveillance more broadly, can significantly help curb global catastrophes wherever possible.
In many respiratory conditions, early cough tracking can help prevent the progress of hazardous infectious diseases.
Tracking coughs can also encourage fine-tuning devices to pick up smaller variations that might otherwise not be noticeable. Cough tracking can also encourage additional research into what clinicians require when making a conclusive diagnosis using cough and related symptoms.
With time, the technology of machine learning will fine-tune these needs towards more accurate diagnosis.
Having a tech device at your fingertips is the norm. A large percentage of the population is already using smart devices to track heart rate, steps, and even the number of hours they sleep. In their pocket or at their fingertips, many people have the technology to track everything from the calories we intake to pregnancy symptoms using a device.
Cough trackers aim to provide a holistic approach to disease diagnosis. As a result, the previously neglected cough is slowly but surely paving the way towards more efficient clinical diagnosis and management.
Dr. Michelle Frank is a healthcare consultant working in the FemTech space. Her work centers around building and fostering online women’s health communities. Read more about her latest work here.
