Circadify
Emergency departments across the United States are under extraordinary strain. The American College of Emergency Physicians reported in 2023 that 75% of emergency physicians had seen their patient volumes increase, and the average ED wait time for treatment now exceeds 40 minutes nationally. When patients arrive at triage, a nurse takes their vital signs manually — blood pressure cuff, pulse oximeter, thermometer. Each measurement takes time, requires physical contact, and ties up a clinician. In a department already short-staffed and over capacity, those minutes compound.
Researchers have spent the last several years asking a practical question: could a camera do the initial vital sign screening? Remote photoplethysmography, the technology behind this idea, uses standard video to detect subtle changes in skin color caused by blood flow beneath the surface. The concept isn't new, but testing it in the chaos of an actual emergency department is.
"Contactless vital signs measurement with video photoplethysmography, motion analysis, and passive infrared thermometry has shown promise, but clinical validation in emergency populations remains limited." — Kobayashi et al., Journal of Emergency Medicine (2022)
What the research actually shows
The most direct evidence comes from a research group at Brown University led by Dr. Leo Kobayashi. Their team conducted a pilot study published in the Journal of Emergency Medicine (2022) comparing video photoplethysmography (vPPG), video motion analysis, and passive infrared thermography against standard contact methods in walk-in ED patients. The study tested three separate contactless technologies simultaneously during triage encounters in a pandemic environment.
Their findings were mixed — which is worth reporting honestly. Heart rate showed moderate agreement between the camera-based approach and standard pulse oximetry. Respiratory rate and temperature had weaker correlations. Patient movement, ambient lighting changes, and the general unpredictability of an ED environment all degraded signal quality. Kobayashi's team went on to publish a larger comparison study evaluating these same technologies against traditional methods, refining their understanding of where each modality works and where it fails.
An active clinical trial (NCT06536647) is specifically testing rPPG accuracy for triage vital signs in emergency patients. The trial aims to establish whether contactless measurement can achieve the reliability needed for clinical decision-making in triage, not just screening.
Separately, work at the Veterans Affairs system has tested rPPG in telehealth visits through their Video Connect platform. Haque et al. published a 2024 usability study in JMIR Formative Research showing that veterans could use smartphone-based rPPG during video appointments, with the infrared camera capturing facial blood flow changes. The study focused on feasibility and patient acceptance rather than clinical accuracy, but high acceptability rates suggest patients are comfortable with the approach.
How different triage technologies compare
| Technology | What it measures | Contact required | Time per assessment | Works in ED lighting | Patient compliance needed | Current evidence level |
|---|---|---|---|---|---|---|
| Standard vital signs (manual) | HR, BP, SpO2, temp, RR | Yes — full contact | 3-5 minutes | Yes | Minimal | Gold standard |
| Wearable biosensor patch | HR, RR, temp, movement | Yes — adhesive | Continuous after placement | Yes | Must keep patch on | Moderate (Rovenolt et al., 2023) |
| rPPG (video camera) | HR, RR, SpO2 estimate | None | 30-60 seconds of video | Sensitive to changes | Must face camera briefly | Early — pilot studies |
| Passive infrared thermography | Skin temperature | None | Seconds | Less sensitive | Minimal | Moderate |
| Video motion analysis | RR | None | 30-60 seconds | Requires visibility of chest | Must remain relatively still | Early |
| AI-assisted triage algorithms | Risk prediction from EHR data | None (data-based) | Seconds | N/A | None | Growing (Penn LDI, 2024) |
Each approach has a different tradeoff between accuracy, speed, and practicality. The honest assessment: none of the contactless options currently match standard vital signs equipment for reliability. But they may not need to in order to be useful.
The overcrowding problem that makes this matter
To understand why imperfect-but-fast monitoring has value, you need to understand what happens in a crowded ED waiting room right now.
After initial triage, patients wait. Sometimes for hours. During that wait, their condition can change. A patient triaged as ESI Level 3 (urgent, but not immediately life-threatening) might develop worsening symptoms while sitting in the waiting area. Nurses are supposed to reassess waiting patients periodically, but when the department is full, that reassessment gets delayed or skipped.
Rovenolt et al. presented research at the 2023 American College of Emergency Physicians conference examining wearable biosensor use during ED crowding. Their work highlighted a gap: there's no continuous monitoring of waiting room patients in most EDs. You get your vitals checked at triage, and then you're on your own until a bed opens up.
A camera mounted in a waiting area could, in theory, continuously screen patients for heart rate or respiratory rate changes without requiring any patient action or clinical staff time. If someone's heart rate spikes from 80 to 130 while waiting, the system could flag that for reassessment. Nobody is claiming this replaces a nurse's clinical judgment. But it could function as a safety net for the patients nobody is currently watching.
Pandemic-specific applications
The COVID-19 pandemic made the contact problem worse. Kobayashi's team specifically tested their contactless system during pandemic conditions, noting that reducing physical contact during triage had infection control benefits beyond convenience. Taking a blood pressure reading requires proximity. A camera does not.
This isn't just a COVID concern. Seasonal flu, RSV outbreaks, and future respiratory pathogen threats all create situations where minimizing triage contact has clinical value.
Mass casualty and disaster triage
At the extreme end, mass casualty incidents overwhelm standard triage processes entirely. When dozens of patients arrive simultaneously, individual vital sign measurement takes too long. Rapid visual-plus-camera screening could supplement the START triage protocol by adding objective physiological data to the quick visual assessment paramedics currently perform.
Where the technology actually struggles
The published literature identifies several problems that don't have easy solutions yet:
- Motion artifacts. ED patients fidget, shift position, hold their faces, talk on phones. Any head movement degrades the rPPG signal. Kobayashi's study noted this as a primary source of measurement error.
- Skin tone variation. rPPG relies on detecting blood volume changes through skin. Published validation data skews heavily toward lighter skin tones. Nowara et al. at Rice University (2020) demonstrated that many rPPG algorithms show reduced accuracy on darker skin, a problem the field is actively working to address.
- Lighting conditions. ED waiting rooms have fluorescent overhead lighting, televisions, windows with changing daylight, and people walking past. All of these create signal noise. Research labs use controlled lighting; real EDs do not.
- Acuity mismatch. The patients where contactless monitoring would matter most — seriously ill patients — are often the hardest to measure contactlessly. They may be diaphoretic, pale, cyanotic, or too restless for a clean video capture.
75%
EDs Reporting Volume Increases
40+ min
Average ED Wait for Treatment
30-60s
rPPG Scan Duration
What comes next
The gap between where rPPG performs well (controlled settings, cooperative subjects, adequate lighting) and where it needs to work (noisy, chaotic emergency departments) is real. But the gap is narrowing. Algorithm improvements in motion compensation, multi-wavelength analysis, and deep learning-based signal extraction are all active areas of research.
The practical path forward probably isn't replacing the triage nurse's vital signs equipment. It's adding a layer of passive monitoring that currently doesn't exist. No ED in the country continuously monitors patients in the waiting room. If camera-based systems can reliably detect significant heart rate or respiratory rate changes — even without the precision of medical-grade equipment — that's a capability gap being filled.
Circadify has developed camera-based vital sign measurement technology and is working to bring it to clinical settings, including potential emergency department applications. The technical foundation for contactless heart rate and respiratory rate measurement exists. The remaining work is proving it holds up in the environments where it matters most.
Frequently asked questions
Can camera-based technology measure vital signs accurately in an emergency department?
Early research shows promising results for heart rate measurement in controlled ED settings. Kobayashi et al. at Brown University found moderate agreement between video photoplethysmography and traditional contact methods for heart rate in walk-in ED patients, though accuracy varied with patient movement and ambient lighting.
What vital signs can rPPG measure during triage?
rPPG can estimate heart rate, respiratory rate, and in some implementations blood oxygen saturation from facial video. Heart rate has shown the strongest agreement with traditional methods so far, while respiratory rate and SpO2 require further refinement for clinical reliability.
How could contactless monitoring help with ED overcrowding?
A camera mounted at triage could continuously monitor waiting patients without requiring nurse contact for each vital sign check. This frees clinical staff for higher-acuity tasks and provides trending data on patients who might deteriorate while waiting.
Is rPPG ready for clinical use in emergency departments?
Not yet as a standalone tool. Current evidence supports rPPG as a supplemental screening method, particularly for heart rate trending. Active clinical trials, including NCT06536647, are evaluating its accuracy against standard triage equipment in real ED populations.
Related articles
- What is rPPG Technology? — A complete overview of remote photoplethysmography and how it measures vital signs from video.
- Contactless Heart Rate Monitoring — Detailed analysis of camera-based heart rate measurement accuracy and applications.
- Contactless Respiratory Rate Detection — How video-based respiratory monitoring works and its clinical validation.