Objective, scalable cough measurement has pushed to a new frontier in respiratory health.
In modern medical sciences, respiratory health has the dubious distinction of not having joined the era of precision medicine. Even while ubiquitous sensors, advances in artificial intelligence (AI), and other innovations are reshaping modern medicine, respiratory health – and in particular anything to do with chronic conditions such as refractory chronic cough – has not advanced much in many decades.
It is telling that the last time the U.S. Food and Drug Administration (FDA) approved a cough medication was in 1958, when it approved dextromethorphan. There has been no new antitussive since.
The main reason for this lack of progress is the fact that respiratory health in general – and cough in particular – lacks quantification. Until very recently, there was no reliable way to measure cough. And because it could not be measured, it was very hard to generate evidence that would satisfy a modern regulator.
This presents problems for multiple diseases affecting significant portions of the population, across all ages. From childhood asthma to chronic cough, chronic obstructive pulmonary disease (COPD), idiopathic pulmonary fibrosis, and lung cancer.
Even as recently as mid-November 2023, in a meeting of the FDA’s Pulmonary-Allergy Drugs Advisory Committee discussing a new drug application for a new generation antitussive, the Committee voted 12-to-1 against approving the new drug, due to a failure to generate reliable data.
In addition to stagnating innovation, the lack of reliable ways to measure cough consistently is interfering with quality health care and making it difficult for patients and providers to even have a meaningful conversation about cough.
Millions of people suffer from chronic respiratory conditions and, in addition to a debilitating impact to their quality of life, most of them share a frustration with the quality of the medical care they are receiving.
In short, there is a critical need for modern solutions in the management of chronic respiratory health conditions.
Among the most common chronic respiratory conditions, the management of COPD is a prime example of this. COPD is prevalent, occurring in more than 5% of the population in the U.S., and debilitating. Additionally, it poses a substantial economic burden on health care systems. COPD exacerbations account for the lion’s share of the $49 billion spent annually in the U.S. on COPD-related health care costs. Acute exacerbations are particularly costly due to patients’ frequent emergency room visits and hospitalizations. If hospitalized once for an acute COPD exacerbation, the likelihood of a subsequent hospitalization in the next 30 days is 25% and in the next eight weeks is 30%.
Cough is a primary symptom of COPD and one of only three “cardinal symptoms,” reported by more than 70% of COPD patients. Persistent chronic cough in patients with COPD has been linked with more severe disease and increased health care utilization. Critically, an increase in coughing (as reported by patients) is one of the most common signs of a COPD exacerbation. Appropriate COPD management can alleviate symptoms, lower the frequency and severity of exacerbations, reduce hospitalizations and readmissions, enhance health status, improve exercise capacity and extend survival.
Cough monitoring has been shown to predict 45% of acute COPD exacerbations with an average lead time of 3.4 days, and an extremely low false positive rate (approximately one for every 100 days of person-time). At-home cough monitoring is not only more predictive of exacerbations than survey-based methods, it also has the advantage of being more amenable to long-term domiciliary contexts than methods for the monitoring of other symptoms (such as pulse oximetry or self-report) since cough monitoring can run continuously in the background without requiring any patient involvement.
Historically, clinical evaluation of cough has been exclusively subjective, meaning that doctors had to rely exclusively on a patient’s description and recollection of their cough. However, as this recent paper proves, there is a big difference between people's perception of how much they cough (subjective cough rate) and how much they actually cough (objective cough rate). This has a significant (negative) impact on both patient care and the broader landscape of medical research and science.
Cough is highly stochastic (variable), which is why it poses such challenges in its assessment. Cough fluctuates seasonally, throughout disease progression, and even within a single day. Basically, coughers have bad days and good days, seemingly randomly. The current approach, such as it is, does not reflect these realities. Even seemingly objective tools, such as continuous recorders, fail to reflect the variability of cough frequency, because they are often limited to a maximum 24-hour window. This leads to misunderstandings and dissatisfaction, with patients feeling unheard and doctors missing significant clues. The lack of objective measurement exacerbates the frustration for both patients and health care providers.
Progress in AI capabilities, combined with the widespread use of smart devices and wearables, essentially powerful computers in everyone’s pocket, offer a generational opportunity to address one of the biggest challenges in COPD management: continuous and unobtrusive cough monitoring over extended periods and the unveiling of patterns that have been impossible to see through short-term assessments or using patient-reported measures. AI models can identify patterns in cough frequency that reveal imminent risks of things like COPD exacerbations days ahead of time. They can help prevent exacerbations that cause emergency room visits, hospitalizations, and increased likelihood of further respiratory and cardiovascular deterioration. And this is not a future capability, platforms like this one can do this right now.
Improved cough monitoring facilitated by objective ways to measure cough over time, holds promise for improved management of chronic respiratory conditions such as COPD precisely because it can be done at any scale using AI continuously, inexpensively, accurately, and with little need for patient involvement. We can expect – and hope for – a change in the standards of care when it comes to the evaluation, diagnostics and management of chronic respiratory conditions. This will save time, money and lives.
Peter Small, MD, is the chief medical officer at Hyfe AI. He has had an eclectic career, with the common theme being the use of innovation to improve health care. He was chief medical resident at University of California San Francisco during the dawn of the HIV epidemic, did pioneering molecular epidemiologic research at Stanford University, and built and ran the TB program for the Bill and Melinda Gates Foundation. In 2015, he founded the Global Health Institute at Stony Brook University focused on the use of technology to deliver health care in remote Madagascar and Nepal. In 2019, he stepped in as the technical lead of a Gates-funded design-build firm which he recently left to focus on making cough count.