Limited Progress in Smartwatch Accuracy for Atrial Fibrillation Detection Over Time

 The lack of a significant decline in inconclusive readings across devices suggests improvements are needed in the algorithms.

The accuracy of several popular consumer smart devices for detecting atrial fibrillation (AF) did not improve over a roughly 2-year period.

This finding comes from the single-center Basel Wearable Study, despite frequent software updates to these devices.

The proportion of single-lead ECG readings tagged as inconclusive declined numerically over time but not statistically significantly.

The five devices studied were AliveCor KardiaMobile, Apple Watch 6, Fitbit Sense, Samsung Galaxy Watch 3, and Withings ScanWatch.

This is the first study assessing the improvement of diagnostic accuracy of algorithm-based AF detection in smart devices over time.

The study included 247 patients with a median age of 66 years, 30% of whom were women.

Each participant received both a 12-lead ECG and single-lead ECGs from all five devices.

The total number of software updates ranged from six for the Apple Watch to 26 for the KardiaMobile and Withings ScanWatch.

Of 1,235 total single-lead ECG readings, 16% (202 readings) were labeled as inconclusive by the manufacturers’ algorithms.

Inconclusive rates ranged from 15% to 19% across devices.

The inconclusive rates declined numerically for all devices, but none showed statistically significant trends.

Rates fell from 14% to 6% for KardiaMobile and Apple Watch, 20% to 14% for Fitbit Sense, 18% to 10% for Samsung Galaxy, and 22% to 14% for Withings ScanWatch.

When the same single-lead ECGs were read by two cardiologists, the inconclusive rate was less than 1%.

This indicates that algorithm performance, not the ECG technology itself, needs improvement.

External algorithms applied to single-lead ECGs have been shown to outperform smart devices.

This demonstrates the potential for future improvement in smart device algorithms.

A challenge is that many smartwatch users are younger, a group at lower risk for atrial fibrillation.

Screening older populations with higher risk factors might make smart devices more clinically useful.

Smart devices could be useful for monitoring patients with palpitations, paroxysmal supraventricular tachycardia, or post-catheter ablation.

These devices could also help patients monitor AF recurrence after a stroke.

Smartwatches are not yet commonly integrated into clinical practice.

In the future, as more people adopt wearables, their integration into care is likely to increase.

Take-Home Points

• Smart devices show no significant accuracy improvements for AF detection over time.
• Algorithm refinement is needed, not hardware changes.
• Integration into clinical practice will depend on meaningful usage strategies.