Can a video visit quickly detect if my breathing is labored before I feel unwell?

Respiratory rate is the vital sign clinicians trust most for spotting trouble early, and it is also the one most often missed. A patient whose breathing has quietly climbed from 16 to 24 breaths per minute may feel only a vague fatigue, yet that shift frequently precedes a fall in oxygen saturation by hours. For health systems building preventative virtual care programs, the question is whether a standard video encounter can surface that signal before a patient reports feeling unwell. Video visit breathing detection, which uses the same camera already running on a telehealth call to estimate respiratory rate and effort, is moving that question from speculation toward operational reality.

A hospital-based trial of 963 patients found 96.0 percent agreement between remote photoplethysmography and reference methods for respiratory rate measurement, supporting camera-based capture as a clinically meaningful modality (van Gastel and colleagues, published in PMC, 2022).

How video visit breathing detection works

Video visit breathing detection draws on two complementary camera-based signal sources. The first is remote photoplethysmography (rPPG), which reads the subtle color changes in facial skin caused by the cardiac cycle. Respiration modulates that pulse signal through amplitude, frequency, and baseline variations, allowing an algorithm to extract respiratory rate from the same facial video used for heart rate. The second source is motion analysis: small periodic movements of the chest, shoulders, and even the head track the mechanics of breathing directly. Combining a pulse-derived estimate with optical-flow movement tracking tends to produce a more robust respiratory rate than either method alone.

Crucially, respiratory rate is only part of the clinical picture. Labored breathing is defined as much by effort and pattern as by count. Research groups have shown that analyzing the relative phase of chest and abdominal wall movement can flag paradoxical breathing and chest indrawing, both markers of increased work of breathing. That distinction matters for early intervention, because a patient with a normal rate but rising effort can still be deteriorating.

For a health system CIO, the appeal is structural. No cuff, no chest strap, no patient-owned wearable, and no shipping logistics. The signal is captured passively during a visit that is already happening, then routed into the EHR alongside heart rate and other camera-derived vitals.

Camera-based respiration versus traditional methods

The methods available to a virtual care program differ sharply in cost, friction, and what they actually measure. The table below compares the common approaches to assessing breathing during or around a remote encounter.

| Method | Captures rate | Captures effort/pattern | Patient hardware | Fits a video visit | Logistics burden | |---|---|---|---|---|---| | Manual visual count over video | Approximate | Subjective | None | Yes | High clinician time | | Wearable chest strap / RIP band | Yes | Yes | Dedicated device | No | High (shipping, setup) | | Pulse oximeter (proxy) | No (oxygen only) | No | Fingertip device | Partial | Moderate | | Camera-based breathing detection (rPPG + motion) | Yes | Emerging | None | Yes | Low | | In-clinic capnography | Yes | Yes | Clinical equipment | No | Not remote |

A few points stand out for program design:

• Manual counting during a video call is possible but unreliable, and it consumes scarce clinician attention that could go to the conversation.
• Wearables capture rich respiratory data but reintroduce the device-distribution problem that virtual care is meant to avoid.
• Pulse oximetry tells you about oxygenation after compensation has begun, often later than a rising respiratory rate would.
• Camera-based detection is the only option that captures rate passively, scales without hardware, and rides on infrastructure the visit already uses.

Industry applications for health systems

Chronic respiratory and heart failure monitoring

Patients with COPD, asthma, and heart failure cycle through periods of stability and decompensation. A respiratory rate trend captured at each scheduled video follow-up creates a longitudinal record that a single in-person snapshot cannot. A reading drifting upward across three visits is an actionable signal for medication review or escalation before an emergency department visit becomes necessary.

Post-acute and post-discharge surveillance

The days after discharge carry elevated readmission risk, and respiratory deterioration is a leading driver. Embedding breathing detection into routine post-discharge video check-ins gives care teams a low-friction way to catch early decline in the window when intervention is cheapest and most effective.

Nurse triage and virtual urgent care

When a patient calls in feeling generally unwell, a triage nurse currently relies on description and visual judgment. An objective respiratory rate captured during that same video call adds a quantifiable input to disposition decisions, helping distinguish who needs an in-person evaluation now from who can be managed remotely.

Primary care and annual virtual visits

Adding a respiratory measurement to every video encounter, even routine ones, builds baseline data for each patient. Deviation from an individual's own baseline is often more informative than comparison to a population norm.

Current research and evidence

The evidence base for camera-derived respiration has matured considerably. The 2022 hospital-based trial referenced above (van Gastel and colleagues, PMC) reported 96.0 percent agreement between rPPG and reference respiratory rate across 963 patients, a sample size large enough to take seriously for adult populations. Reviews of camera-based respiration quantification published in PMC catalogue multiple validated approaches spanning rPPG, motion magnification, and depth sensing, with some controlled-setting studies reporting mean absolute errors below 0.5 breaths per minute.

Work on effort, not just rate, is also progressing. A study on video-based non-contact monitoring of respiratory rate and chest indrawing in children with pneumonia (published via PubMed) demonstrated that camera analysis of regional chest-abdominal movement can detect indrawing, a recognized sign of respiratory distress. Separately, researchers have validated contactless respiratory rate and breathing-absence detection from head movements alone using an RGB camera (PubMed), broadening the conditions under which a signal can be recovered.

The clinical value of the underlying signal is well established independent of the capture method. A pilot prospective cohort study of continuous ward respiratory rate monitoring in COVID-19 patients (published in Frontiers) found that continuous respiratory rate tracking enabled earlier detection of deterioration than intermittent observation. That finding reinforces why bringing any reliable respiratory measurement into more frequent virtual touchpoints carries preventative value.

The honest limitations are documented too. Accuracy degrades with patient motion, poor or uneven lighting, and irregular breathing. Signal quality from camera methods is generally lower than from contact sensors, and pediatric algorithms, especially for children under 12, need further refinement. These are engineering and validation challenges, not conceptual barriers, and they define the agenda for the next several years of deployment.

The future of video visit breathing detection

Three trajectories are worth watching for informatics teams. First, multimodal fusion: combining rPPG pulse-derived respiration with chest and shoulder motion and head-movement signals to maintain accuracy when any single source is degraded. Second, the shift from rate to effort, where pattern and work-of-breathing markers move from research settings into routine reporting, closing the gap between what a camera measures and what a clinician means by "labored." Third, integration depth, where respiratory readings flow automatically into EHR flowsheets, feed early-warning scores, and trigger care-team alerts without manual entry.

For governance, the maturing of this technology means health systems should plan validation, documentation standards, and escalation protocols now. The clinical question is no longer whether a camera can estimate breathing, but how to operationalize that capability responsibly across diverse patient populations and connection conditions.

Frequently asked questions

Can a video visit really detect labored breathing before a patient feels symptoms? It can detect the objective signals that often precede symptoms. A rising respiratory rate frequently appears before a patient reports distress, and camera-based methods can surface that trend across visits. Emerging effort and pattern analysis adds detail about work of breathing, though effort detection is less mature than rate measurement.

Does breathing detection require the patient to wear or hold anything? No. The measurement uses the existing video stream, analyzing facial pulse signals and small body movements. There is no cuff, strap, or wearable, which removes the hardware-distribution burden that limits many remote monitoring programs.

How accurate is camera-based respiratory rate compared to standard methods? A hospital trial of 963 patients reported 96.0 percent agreement with reference methods for adults, and controlled studies have shown errors below 0.5 breaths per minute. Accuracy can drop with patient motion, poor lighting, and in young children, so validation in your own population and workflow conditions is essential.

Where does the respiratory data go after a visit? In a well-integrated program, readings flow into the EHR alongside other captured vitals, where they can populate flowsheets, contribute to early-warning scores, and inform triage and escalation decisions rather than living in a separate application.

Circadify is building toward exactly this preventative use case: clinical-grade vital signs, including respiratory rate, captured in every virtual visit with no patient wearables and direct EHR integration. To see the clinical workflows and request a health system demonstration, visit circadify.com/solutions/telehealth.