Brinnae Bent · npj Digital Medicine 2020 · cross-sectional validation study · n=?

Investigating sources of inaccuracy in wearable optical heart rate sensors

Cited 670 times in the scientific literature.

Level 4 - case-series / case-control

Cross-sectional validation study by design analogy.

OpenAlex W3005348902 · doi:10.1038/s41746-020-0226-6 · record verified 2026-08-29

What was done

Evaluated heart rate and photoplethysmography (PPG) data from consumer- and research-grade wearable devices across various conditions to assess the impact of diverse skin tones, motion artifacts, and device types on measurement accuracy.

What was found

No statistically significant difference in accuracy was observed across skin tones. Significant differences were observed between devices and between activity types, with absolute error during activity being on average 30% higher than during rest. All tested wearables were reasonably accurate at resting and during prolonged elevated heart rate, but varied in response to changes in activity. No exact participant numbers, baseline error values, or confidence intervals were reported in the abstract.

Why it matters

It demonstrates that physical activity and device-specific signal processing, rather than skin tone variations, are the primary contributors to optical heart rate sensor inaccuracies under the tested conditions.

Limits

The abstract does not state the sample size, participant demographics, specific device models evaluated, skin tone classification method, or the reference standard used for comparison.

Cited by