Smartwatches in the Clinic: Should Doctors Trust the Data or the Patient?

At a French sports medicine congress, cardiologist Stéphane Doutreleau laid out when wrist sensors mislead, when a heart rate spike deserves an ECG, and why heart rate variability only works as a trend. His message for clinics: ask which device, ask under what conditions, and treat the watch as a starting point, never a diagnosis.

PetaTech24 Technology Desk, October 16, 2026

Patients now walk into clinics carrying weeks of heart data on their wrists, and many expect their doctor to make sense of it. The question dividing physicians is blunt: should they trust the data, or the patient sitting in front of them?

At the 19th National Congress on Sports Medicine and Traumatology in France, Stéphane Doutreleau, MD, PhD, of Grenoble Alpes University Hospital in La Tronche, offered a practical answer. Wearable data cannot replace a clinical examination, he said, but it can serve as a useful starting point, provided the doctor knows exactly which device produced it and under what conditions.

That condition matters more than most people think. Optical wrist sensors can underestimate heart rate by dozens of beats per minute during exercise, yet the same watches can also record genuine arrhythmias that deserve a proper ECG. Sorting signal from artifact is becoming part of the clinical job.

The Short Version

• Chest strap ECG remains the reference. Among consumer devices, the chest strap’s electrical signal is the most reliable way to measure heart rate, Doutreleau said.
• Wrist PPG sensors are conditional. Optical watches and bands measure blood flow with light, and they are sensitive to movement, sensor position, sweat, and exercise type.
• Exercise is where watches fail most. At rest or while walking, optical readings are often satisfactory; during running or cycling, the gap with a chest strap can reach dozens of beats per minute, usually an underestimation.
• Spikes are not always artifacts. Simultaneous recordings with Holter monitors show a significant share of heart rate anomalies are genuine arrhythmias, not interference.
• HRV only works as a trend. A single heart rate variability reading has little value; it must be measured at rest, under standardized conditions, with the same device.
• The clinical rule. Ask which device and under what conditions, verify consistency, look for symptoms, compare with the patient’s usual values, and confirm extremes with a real medical examination.

Why It Matters

Patients armed with weeks of unverified biodata are now a routine part of clinical life, and doctors are split on what to do with the numbers. A 2026 survey of 2,222 physicians across the US, Canada, France, Germany, Spain, and the UK, conducted by the AMA’s Center for Digital Health and AI with Medscape, found that 97% of physicians had reviewed at least one type of wearable data, and most saw at least some clinical advantage for patient care.

Yet routine use remains rare. According to reporting on the survey, no more than 6% of physicians in any surveyed country said wearable data was fully integrated into clinical workflows. The barriers named were confidence in data quality, false-positive alerts, liability, workflow pressure, and reimbursement.

That gap is exactly what Doutreleau’s talk addresses. The technology is already sitting in the exam room. The framework for judging it is not.

When the Watch Misleads

Optical sensors work by photoplethysmography: light shines into the skin and the sensor reads blood volume changes with each pulse. It is comfortable and practical, but the method has weak points. Movement shakes the signal, a loose fit lets light leak, sweat changes the optical path, and some exercises swing the wrist constantly.

The result is predictable. At rest or during a walk, the numbers are often fine. During a run or a ride, the watch can lag the true heart rate by dozens of beats per minute, and the error runs in one direction: underestimation. Performance also varies by brand and model, which is why Doutreleau told the audience that the first clinical question should be which device produced the data.

There is a second trap. Watches can automatically recalculate maximum heart rate after an isolated spike. If that spike was an artifact, or an arrhythmia, the watch bakes the error into every training metric that follows. The fix is simple: do not accept an auto-adjusted maximum heart rate without verification.

Do Not Dismiss the Spike

Here the talk took an unexpected turn. When a heart rate graph shows an irregular spike, the reflex is to blame interference. Doutreleau warned against that reflex. Premature beats and certain arrhythmias briefly change the intervals between heartbeats, and the device records that as a sharp, short rise in heart rate.

Data collected simultaneously by a heart rate monitor and a Holter monitor showed that a significant proportion of these anomalies were not interference at all. In practice, the talk says, an unusual waveform has to be judged with the symptoms, the circumstances, and the sensor type in mind.

The warning signs it lists: palpitations, a very high heart rate, a feeling of discomfort, unusual exercise intolerance, or a marked break from the patient’s usual pattern. In those cases the data should be confirmed with an ECG or Holter monitoring, not waved away as a measurement error. Some portable systems can now record an ECG when symptoms occur, which can speed up the diagnostic process.

Heart Rate Variability: Trends Beat Snapshots

Heart rate variability is one of the most watched numbers on modern smartwatches, and one of the easiest to misread. The talk’s guidance was strict: measure at rest, under standardized conditions, always with the same device. A suggested protocol for athletes is a few minutes lying down, then a measurement after standing up.

The reason is the long list of confounders: fatigue, sleep, stress, temperature, breathing rate, meals, hydration, noise, light, and recent activity. A single reading, taken on a random morning, carries almost no information. What matters is the trend across several weekly measurements under comparable conditions, with the patient serving as their own control.

The patterns to watch are relative, not absolute. A persistent drop in variability, especially alongside a rising resting heart rate, unusual fatigue, or falling performance, may point to insufficient recovery or excessive training load. A lower resting heart rate with stable or rising variability may signal favorable adaptation. But the talk stressed that these markers are nonspecific: they only mean something next to symptoms, a clinical examination, and the training context.

Confirmed vs Reported vs Still Open

• Confirmed. Chest-strap ECG is the most reliable consumer heart rate method. Wrist PPG accuracy degrades with movement and can underestimate heart rate by dozens of beats per minute during exercise. Accuracy varies by brand and model. Watches can auto-adjust maximum heart rate after a single spike.
• Reported. The AMA and Medscape 2026 survey findings on physician attitudes and integration barriers, as covered by Healio and the AMA. The congress talk’s guidance, as published by Medscape from the JIM translation.
• Still open. How each manufacturer’s algorithm recalculates maximum heart rate. The exact share of spike anomalies that turn out to be genuine arrhythmias across devices. The threshold at which watch data should trigger a clinical workup.

What to Watch Next

Can a smartwatch detect atrial fibrillation?
Some watches and portable systems can record an ECG when symptoms occur, and the talk notes this can help the diagnostic process. But a watch tracing is a starting point, not a diagnosis; confirmation with medical-grade equipment remains the standard the talk describes.

Should I show my doctor my smartwatch data?
The talk’s framework says yes, with context: which device, what you were doing, whether the pattern is new, and what symptoms you felt. A doctor weighs the data against your history and examination. The watch is never read in isolation.

Why does my watch underread my heart rate when I run?
Optical sensors track blood flow with light, and arm movement, sweat, and sensor position all disturb the signal during exercise. The error typically runs toward underestimation, sometimes by dozens of beats per minute. A chest strap avoids most of this because it reads an electrical signal.

Is tracking heart rate variability worth it?
Only as a trend. Single readings are swamped by daily factors from sleep to hydration. Measured at rest, under consistent conditions, with the same device, a multi-week trend can hint at recovery status. The talk treats it as one input among many, never a verdict.

The Takeaway

The watch is a witness, not a judge. For doctors, the talk’s method is simple: ask which device and under what conditions, check consistency, weigh symptoms, and compare against the patient’s own baseline. For patients, the useful material is not a single number but patterns, and any worrying symptom deserves a clinician, not a gadget verdict.

Trust the data enough to investigate it. Trust the patient enough to examine them.

This article summarizes a medical conference discussion for a general audience and is not medical advice. The AMA’s patient guidance adds this: seek medical care if a wearable repeatedly flags abnormal readings and you feel chest pain, shortness of breath, dizziness, or fainting.

Sources / Attribution

• Medscape: “Smartwatches in Clinic: Trust the Data or Trust the Patient?”, October 14, 2026 (translated from JIM, Medscape Professional Network). https://www.medscape.com/viewarticle/smartwatches-clinic-trust-data-or-trust-patient-2026a10012jc
• Healio: “Structural barriers limit use of wearable technology in medicine,” July 29, 2026 (reporting on the AMA/Medscape 2026 International Physician Survey on Consumer Wearables, 2,222 physicians across six countries). https://www.healio.com/news/rheumatology/20260729/ama-structural-barriers-limit-uptake-of-wearable-technology-in-medicine
• AMA: “What doctors want patients to know about health wearables” and “International physician survey on consumer wearables.” https://www.ama-assn.org/practice-management/digital-health/what-doctors-want-patients-know-about-health-wearables