The Alert Before the Symptom: Triaging Consumer Wearable Notifications in UK Practice

General practice has absorbed many new streams of unsolicited information over the past decade. The latest arrives on a wrist. A patient attends not because they feel unwell but because a device told them something might be wrong: an irregular rhythm notification, a high resting heart rate, a pause, a suggestion of disturbed breathing overnight. They are frequently asymptomatic, occasionally frightened, and almost always holding a phone that contains a data file nobody in the building knows how to read.

This is a genuinely new consultation type, and it deserves a considered response rather than a reflexive one in either direction.

What these devices measure, and what they infer

Most irregular rhythm notifications rely on photoplethysmography — the optical sensor used for heart rate — analysing pulse intervals for irregularity. Some devices add a single-lead electrocardiogram taken by the wearer touching a crown or bezel. Sleep-related features infer breathing disturbances from movement, blood oxygen and heart rate variability during the night.

These are screening signals, not diagnoses. A single-lead tracing obtained at home, in unknown conditions, with unverified electrode contact, cannot replace a twelve-lead ECG interpreted by a clinician. Nor is the regulatory picture uniform: some features carry UKCA or CE marking as medical devices with a specified intended purpose, while others are marketed as general wellness tools that make no diagnostic claim at all. Checking the instructions for use is not bureaucratic pedantry; it determines how much weight the output can bear.

The atrial fibrillation notification: useful, but conditional

The clinically strongest use case is atrial fibrillation detection in older adults, where opportunistic case-finding has a real evidence base and where anticoagulation decisions follow a confirmed diagnosis. NICE guidance on atrial fibrillation is clear that a diagnosis requires an ECG demonstrating the arrhythmia, recorded by an appropriately qualified professional and interpreted with clinical context.

Where a wearable notification precipitates assessment, the useful sequence is: establish whether the patient has symptoms, obtain and review the device's exported data, repeat a resting twelve-lead ECG, and proceed to ambulatory monitoring — typically a 24-hour or longer recording — where paroxysmal AF is suspected but not captured. If AF is confirmed, assessment of stroke risk and bleeding risk follows in the usual way, alongside a search for reversible contributors such as thyroid dysfunction, alcohol excess, infection or electrolyte disturbance.

What the notification changes is the threshold for investigation. It does not shorten the diagnostic pathway.

Sleep apnoea features: a hint, not an answer

Breathing-disturbance notifications are generating a parallel stream of referrals. Obstructive sleep apnoea is common, under-recognised and consequential, and a device prompt may well be the first nudge a patient receives. But the reference standard remains overnight oximetry or limited-channel sleep study, with assessment of daytime sleepiness, comorbidities and occupational risk — particularly for drivers and safety-critical workers. A notification should prompt a conversation about symptoms and risk factors, not a self-purchased device and not a diagnosis.

The anxiety loop, and the capacity it consumes

There is a temptation to frame this as a problem of the worried well. That framing is both unkind and inaccurate. Health anxiety is a real clinical presentation, and it can be reinforced — sometimes created — by continuous physiological monitoring in people who were previously unconcerned. A device that surfaces a benign ectopic beat or a normal nocturnal dip as an alert has done harm, not surveillance.

At the same time, the volume effect is real. Each notification that converts into a same-day appointment or an emergency department attendance has an opportunity cost for a system operating within tight constraints. The proportionate response is neither to dismiss the signal nor to investigate indiscriminately, but to triage it consistently.

A checklist for clinicians

• Ask what the patient actually felt. Asymptomatic alerts and symptomatic alerts follow different pathways.
• Obtain the exported PDF or data file rather than the summary on the watch face; note date, time, duration and the patient's activity at that moment.
• Check whether the feature is a regulated medical device function or a wellness indicator.
• Perform and document a twelve-lead ECG in the practice where feasible.
• Arrange ambulatory monitoring if the alert was plausible but the resting ECG is normal.
• Red flags — syncope, exertional symptoms, chest pain, breathlessness, a family history of sudden cardiac death — bypass the algorithm and go straight to urgent assessment.
• Address the anxiety component explicitly and without condescension, including safety-netting and, where appropriate, a discussion about whether continued monitoring is helping.

A checklist for patients

• If you feel faint, breathless, or have chest pain or palpitations with symptoms, seek urgent care rather than waiting for the next notification.
• Do not adjust or stop any prescribed medication on the basis of a device alert.
• Bring the data, not just the message, to your appointment.
• Note what you were doing when the alert fired; context changes interpretation.
• Ask whether repeat recording, rather than more investigation, is the right next step.

Where population surveillance fits

Aggregated wearable data — resting heart rate shifts, sleep disruption, changes in activity — is an intriguing population signal, and interest in using it for early warning is legitimate. It is also unvalidated, and it is skewed: devices are concentrated among younger, healthier and more affluent groups, which is close to the inverse of the populations most often under-represented in health data. Any surveillance use must be built on validation and equity analysis, not on enthusiasm.

The proportionate position

A wearable alert is best understood as a prompt to have a conversation, not a finding to be managed. Handled well, it can shorten the distance between an arrhythmia and its diagnosis. Handled badly — by clinician or patient — it manufactures anxiety and consumes capacity that other patients need. The checklist, not the algorithm, is where the clinical value sits.