Respiratory digital biomarkers
Breathing is continuous and physiologically informative, yet respiratory endpoints remain underrepresented relative to their clinical importance.
%20(1).png)
Respiratory measurement sits at an interesting point in digital health. Breathing is continuous, physiologically informative and increasingly captured by devices that people already own, yet respiratory endpoints remain underrepresented in registered trials relative to their clinical importance.
This category collects the measures that connected devices can produce: breathing rate estimated from movement or optical sensing, peripheral oxygen saturation, cough recorded and counted by automated monitors, and the lung function measures captured by connected spirometers. Together they describe respiratory status in the home rather than in a pulmonary function laboratory.
Two things distinguish this domain. The first is that some of its measures are acoustic rather than mechanical, which raises questions of privacy and consent that step counting does not. Continuous audio recording to count coughs is a substantially more sensitive data collection practice than counting steps, and study designs need to address that directly rather than treat it as an implementation detail.
The second is that respiratory measures respond quickly. Activity measures shift over weeks, but cough frequency and oxygen saturation can change within days of an exacerbation, which makes this domain well suited to detecting acute deterioration rather than only tracking slow decline.
Respiratory measures in this library
How these measures are used
The dominant research use is exacerbation detection and characterisation in chronic respiratory disease. Because these measures respond within days, they can identify the onset of an exacerbation earlier than a scheduled visit would, and they can quantify how long recovery takes, which conventional endpoints capture poorly.
Cough frequency has a distinctive position. In the public registry of digital endpoints, every registered trial that has used it has used it as a primary endpoint, which is true of almost no other digital measure. That reflects both its clinical face validity in chronic cough and the existence of dedicated monitoring hardware built for the purpose.
Across the domain, protocols normally pair the objective measure with a validated respiratory questionnaire covering symptom burden and daily impact. Oxygen saturation and breathing rate are also collected as safety parameters in trials that have nothing to do with the lungs, which is worth remembering when designing consent and data handling.
What the evidence supports today
Evidence in this domain is uneven and worth stating plainly. Connected spirometry inherits the strong evidence base of conventional spirometry, with the caveat that unsupervised home manoeuvres are more variable than supervised ones. Cough monitoring has a smaller literature but an unusually high proportion of primary endpoint use, which is a meaningful signal of how seriously the field takes it.
Peripheral oxygen saturation from consumer wearables is the weakest link. The clinical measure is well established, but wrist based estimation is considerably less accurate than fingertip pulse oximetry, and accuracy varies with skin pigmentation, perfusion and motion. Consumer implementations are generally positioned as wellness features rather than cleared medical measurements, and treating them as equivalent to clinical oximetry is a mistake.
Breathing rate derived from wearables is credible for tracking change within a person overnight, when the participant is still, and much less credible during the day.
Common questions
How accurate is oxygen saturation from a smartwatch?
Less accurate than a fingertip pulse oximeter, which is itself less accurate than arterial blood gas measurement. Wrist based estimation is affected by motion, perfusion and skin pigmentation. Most consumer implementations are positioned as wellness features rather than cleared medical devices, and clinical decisions should not rest on them.
How is cough measured objectively?
By recording ambient sound continuously with a body worn monitor and applying an algorithm that identifies cough events, usually reported as coughs per hour. This requires explicit consent covering audio capture, and study designs should specify what is stored, for how long, and whether speech is retained or discarded at the point of processing.
Can home spirometry replace laboratory testing?
It complements rather than replaces it. Home devices allow far more frequent measurement, which reveals variability that periodic laboratory testing misses. Unsupervised manoeuvres are more variable because technique drifts without a technician present, so protocols typically include coaching and quality checks on each blow.
Which respiratory measures respond fastest to change?
Cough frequency and oxygen saturation shift within days during an exacerbation, and breathing rate can move earlier still. Activity based measures respond more slowly. Studies aiming to detect acute deterioration usually rely on the fast responding measures and use activity as context.
Run a study on these measures
WeGuide captures wearable data and patient reported outcomes in one platform, from screening through to analysis.
