7 Vital Signs You Can Capture During a Video Visit

A video visit has historically carried a clinical blind spot. The conversation transmits cleanly, the history is captured, and the assessment proceeds, but the objective measurements that anchor an in-person encounter have stayed behind in the clinic. That gap is closing. The vital signs captured during a video visit now extend well past what most clinical informatics teams assume is possible from a standard webcam, and the underlying science has matured from laboratory novelty to a body of peer-reviewed evidence. Remote photoplethysmography (rPPG), the technique of reading tiny color changes in facial skin caused by each cardiac cycle, sits at the center of this shift. For health system leaders evaluating whether televisit vital signs can carry real clinical weight, the relevant question is no longer whether a camera can measure physiology, but which signals it can measure reliably and what each one tells a clinician.

A 2023 clinical validation study of rPPG-enabled contactless pulse rate monitoring in cardiovascular disease patients reported a mean absolute error of just 1.061 beats per minute against reference measurement, demonstrating that camera-derived heart rate can approach contact-sensor agreement in a patient population that matters.

The vital signs captured during a video visit

The seven measurements below represent the practical scope of what a camera-based system can extract from a patient sitting in front of a standard device during a routine encounter. Not every signal carries the same maturity. Heart rate and respiratory rate are well established in the literature, while blood pressure and oxygen saturation remain active research frontiers with promising but still-developing validation. Understanding this gradient is essential for any informatics team building governance around the data.

The seven vital signs are:

• Heart rate (pulse rate)
• Respiratory rate (breathing rate)
• Heart rate variability (HRV)
• Blood oxygen saturation (SpO2)
• Blood pressure (estimated)
• Stress and autonomic balance indices
• Pulse rhythm and perfusion signals

Each of these derives from the same raw input, a video stream of the patient's face, but the processing and the clinical confidence attached to each differ substantially.

| Vital sign | Capture method | Reported research accuracy | Clinical maturity | |---|---|---|---| | Heart rate | rPPG facial blood-volume signal | MAE ~1.06 bpm in CVD patients (2023) | Established | | Respiratory rate | Motion and rPPG-derived modulation | RMSE ~2.13 breaths/min (2024) | Established | | Heart rate variability | Beat-to-beat interval analysis | Strong agreement in controlled settings | Emerging | | Blood oxygen (SpO2) | Dual-wavelength RGB camera analysis | MAE ~1.27%, RMSE ~1.71% in deep-learning study | Emerging | | Blood pressure | Transdermal optical imaging / pulse features | Within ~5 plus or minus 8 mmHg in normotensive adults | Research stage | | Stress / autonomic index | Derived from HRV and pulse morphology | Correlates with validated stress measures | Emerging | | Pulse rhythm / perfusion | Waveform pattern analysis | Qualitative signal, active research | Research stage |

Why each signal matters clinically

Heart rate and breathing rate

Heart rate and breathing rate from video form the foundation of any serious televisit vitals list. Both are core components of standard early-warning scores, and both shift early when a patient deteriorates. Work by Lieke Dorine van Putten, Kate Emily Bamford, Ivan Veleslavov, and Simon Wegerif on personal-device cameras has shown that pulse rate can be extracted from ordinary video, and a 2024 evaluation of photoplethysmography-derived respiration reported a root mean square error around 2.13 breaths per minute even under challenging conditions. For triage nurses and primary care providers, a reliable heart rate breathing rate video capture turns a subjective impression of how a patient looks into a documented number.

Heart rate variability and stress

HRV, the variation in time between successive heartbeats, reflects autonomic nervous system balance and has applications spanning cardiology, behavioral health, and chronic disease management. Because rPPG recovers a beat-to-beat pulse waveform, it can derive HRV metrics and downstream stress indices without any wearable. This matters most in behavioral health and longitudinal monitoring, where an objective marker of physiological stress complements a clinical interview that is otherwise entirely verbal.

Oxygen saturation and blood pressure

SpO2 and blood pressure are the two signals that draw the most interest and the most appropriate caution. Research led by groups including Mohamed Moustafa, Joseph Lemley, and Peter Corcoran has pushed camera-based SpO2 estimation toward clinically interesting territory, with one deep-learning study reporting a mean absolute error near 1.27 percent, inside the 4 percent threshold expected of approved pulse oximeters. For blood pressure, the transdermal optical imaging approach described by Luo and colleagues in Circulation (2019) predicted values within roughly 5 plus or minus 8 mmHg in normotensive adults. Both remain research-stage for general populations, and informatics teams should treat them as trend and screening signals pending broader validation across skin tones and clinical conditions.

Industry Applications

Primary care and chronic disease

In primary care, the value is in continuity. A hypertension or heart failure patient who previously contributed no objective data during a virtual follow-up can now generate a documented heart rate, respiratory rate, and blood pressure trend at every encounter. That converts the video visit from a check-in into a measurable touchpoint in a longitudinal record.

Behavioral Health

Behavioral health televisits gain the most from the autonomic signals. HRV and stress indices give clinicians a physiological reference point during sessions that are otherwise built entirely on self-report, supporting assessment of anxiety, treatment response, and overall arousal.

Triage and urgent virtual care

For nurse triage and on-demand virtual urgent care, the combination of heart rate, breathing rate, and oxygen saturation supports faster, better-documented escalation decisions. A camera-derived vital set helps a triage nurse distinguish a patient who can be managed remotely from one who needs in-person evaluation.

Current research and evidence

The peer-reviewed base under camera-based vitals has grown quickly. Systematic reviews of contactless vital sign monitoring now catalog dozens of studies across heart rate, respiratory rate, SpO2, and blood pressure, and meta-analytic work on consumer-grade contactless monitors has begun to quantify how these systems perform outside the laboratory. The consistent finding is a gradient of maturity: heart rate sits on the firmest ground, respiratory rate follows closely, and HRV, SpO2, and blood pressure carry more variance and more sensitivity to real-world conditions.

The recurring limitations across this literature are consistent and worth naming directly for any clinical governance framework:

• Motion artifact from a patient who shifts or gestures during capture
• Variable and uncontrolled lighting in home environments
• Performance differences across skin tones, an equity issue that demands diverse validation cohorts
• Camera quality and subject-to-camera distance
• The need for validation in larger, more representative clinical populations rather than healthy volunteers

These are engineering and study-design challenges rather than fundamental barriers, and the trajectory of reported accuracy across successive studies has been toward narrower error margins.

The future of camera-based televisit vitals

The direction of the field points toward more signals captured more reliably, with two forces driving progress. The first is methodological: deep-learning models trained on larger and more diverse datasets continue to reduce error and improve robustness to motion and lighting. The second is integration: the clinical value of a camera-derived vital is only realized when it flows automatically into the electronic health record, attaches to the encounter, and respects the same data-quality and confidence-flagging standards as any other documented measurement. Health systems that treat camera vitals as a structured data source, complete with confidence indicators and clear documentation of measurement modality, will extract far more value than those treating it as a novelty feature. The likely near-term outcome is a tiered model in which established signals such as heart rate and respiratory rate inform direct clinical decisions, while emerging signals serve as screening and trend tools that prompt confirmatory measurement when they fall outside expected ranges.

Frequently asked questions

Which vital signs can a standard video visit actually capture? A standard video visit can capture heart rate, respiratory rate, heart rate variability, estimated oxygen saturation, estimated blood pressure, stress and autonomic indices, and pulse rhythm signals. Heart rate and respiratory rate carry the strongest validation, while blood pressure and SpO2 remain active research areas best used as trend and screening signals.

How does a camera measure heart rate and breathing rate without any device on the patient? The technique is remote photoplethysmography. A camera detects subtle color changes in facial skin caused by blood volume shifts with each heartbeat, and breathing rate is derived from the modulation of that signal and small body movements. A 2023 study reported heart rate error near 1.06 bpm in cardiovascular patients.

Are camera-based blood pressure and oxygen readings reliable enough to act on? They are promising but still maturing. Research has shown SpO2 errors inside the 4 percent oximeter standard and blood pressure within roughly 5 plus or minus 8 mmHg in normotensive adults, but both need broader validation across diverse populations and conditions. Most programs treat them as screening and trend signals, not standalone diagnostic values.

What affects the accuracy of vitals captured during a video visit? Patient motion, lighting, skin tone, camera quality, and distance from the camera all influence accuracy. Strong programs pair the technology with capture guidance for patients and confidence flags that tell clinicians when a reading should be confirmed.

Circadify is building toward this clinical depth, capturing vital signs in every virtual visit without patient wearables and integrating the results directly into the EHR. Clinical informatics teams evaluating how camera-based vitals fit their virtual care architecture can review the clinical overview and workflow documentation at circadify.com/solutions/telehealth.