Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 36 of the 37 references it cites were found in a registry or an index; 19 of the 50 sources its claims cite were checked for support.

Investigating sources of inaccuracy in wearable optical heart rate sensors

Open access

Abstract

As wearable technologies are being increasingly used for clinical research and healthcare, it is critical to understand their accuracy and determine how measurement errors may affect research conclusions and impact healthcare decision-making. Accuracy of wearable technologies has been a hotly debated topic in both the research and popular science literature. Currently, wearable technology companies are responsible for assessing and reporting the accuracy of their products, but little information about the evaluation method is made publicly available. Heart rate measurements from wearables are derived from photoplethysmography (PPG), an optical method for measuring changes in blood volume under the skin. Potential inaccuracies in PPG stem from three major areas, includes (1) diverse skin types, (2) motion artifacts, and (3) signal crossover. To date, no study has systematically explored the accuracy of wearables across the full range of skin tones. Here, we explored heart rate and PPG data from consumer- and research-grade wearables under multiple circumstances to test whether and to what extent these inaccuracies exist. We saw no statistically significant difference in accuracy across skin tones, but we saw significant differences between devices, and between activity types, notably, that absolute error during activity was, on average, 30% higher than during rest. Our conclusions indicate that different wearables are all reasonably accurate at resting and prolonged elevated heart rate, but that differences exist between devices in responding to changes in activity. This has implications for researchers, clinicians, and consumers in drawing study conclusions, combining study results, and making health-related decisions using these devices.

Overview

CoChat Summary

## TL;DR

This study tested six wearable optical heart-rate devices in 53 participants across the full Fitzpatrick skin-tone scale, using ECG as the reference standard. The main drivers of error were device type and activity condition, not skin tone.

## Key findings

  • Across all devices, there was no statistically significant difference in heart-rate accuracy across skin tones; the authors also report no differences in HRV accuracy between skin-tone groups.
  • At rest, consumer-grade wearables were more accurate than research-grade wearables: MAE was 7.2 ± 5.4 bpm vs 13.9 ± 7.8 bpm (p < 0.0125).
  • During physical activity, error increased for both categories: MAE was 10.2 ± 7.5 bpm for consumer devices vs 15.9 ± 8.1 bpm for research devices (p < 0.0125); the abstract also states absolute error during activity was 30% higher than during rest.
  • Device performance varied substantially: at rest, the Apple Watch 4 had the lowest MAE among consumer devices (4.4 bpm) and the Empatica E4 the lowest among research devices (11.3 bpm); during activity, the Apple Watch 4 remained lowest (4.6 bpm) while the Biovotion Everion was highest among research devices (19.8 bpm).

## Methods

  • Design/sample: Observational validation study of 56 recruited participants, with 53 completing the protocol; participants were 18–54 years old, with an approximately equal distribution across six Fitzpatrick skin-tone categories.
  • Devices/protocol: Six wearables were tested—4 consumer-grade (Apple Watch 4, Fitbit Charge 2, Garmin Vivosmart 3, Xiaomi Miband 3) and 2 research-grade (Empatica E4, Biovotion Everion)—during rest, deep breathing, walking, washout, and typing, with ECG (Bittium Faros 180) as the reference standard.
  • Analysis: The paper compared wearable HR to ECG using MAE, MDE, missingness, and HRV analyses, with mixed-effects models for skin tone, condition, device, and interactions, plus t tests and Bonferroni corrections.

## Limitations

  • The authors state that the study had a limited scope of devices, so they could not tease apart the separate effects of cost, market size, and release year.
  • For HRV, analysis was limited to only the Empatica E4, because most wearables did not provide the raw data needed for HRV calculations.
  • Evident from the text: the protocol was short and structured (about 1 hour of seated rest, deep breathing, walking, and typing), so the findings may not generalize to other activities or longer-term real-world use.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41746-020-0226-6

Cites

36 of 36 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Graphical abstract of research. Graphical abstract of research study presented. We present a full characterization of HR accuracy across skin tones, clinical metrics of HRV accuracy across skin tones, and HR during activity, rest, deep breathing, and typing for six wearable devices representing both consumer wearables and research-grade wearables. HR metrics are compared to the clinical-grade electrocardiogram (ECG) as the standard for heart rate measurement.
Fig. 1 Graphical abstract of research. Graphical abstract of research study presented. We present a full characterization of HR accuracy across skin tones, clinical metrics of HRV accuracy across skin tones, and HR during activity, rest, deep breathing, and typing for six wearable devices representing both consumer wearables and research-grade wearables. HR metrics are compared to the clinical-grade electrocardiogram (ECG) as the standard for heart rate measurement.CC BY · PubMed Central
Fig. 2 Error in heart rate across skin tones and devices at rest and during activity. Mean error in heart rate (bpm) across skin tones and devices at a rest and b during physical activity. The green horizontal line represents no error (no difference from the true measurement of HR from ECG). Mean absolute error in heart rate (bpm) across skin tones and devices at c rest and d during physical activity. Error is calculated as the difference between the ECG and wearable reported heart rate at every simultaneous measurement. Fitzpatrick skin tones 1–6 are represented with an approximately equal number of participants in each skin tone. Error bars represent the 95% confidence interval. Mean absolute error across devices and across skin tones at rest (e) and during activity (f). Error bars represent the 95% confidence interval.
Fig. 2 Error in heart rate across skin tones and devices at rest and during activity. Mean error in heart rate (bpm) across skin tones and devices at a rest and b during physical activity. The green horizontal line represents no error (no difference from the true measurement of HR from ECG). Mean absolute error in heart rate (bpm) across skin tones and devices at c rest and d during physical activity. Error is calculated as the difference between the ECG and wearable reported heart rate at every simultaneous measurement. Fitzpatrick skin tones 1–6 are represented with an approximately equal number of participants in each skin tone. Error bars represent the 95% confidence interval. Mean absolute error across devices and across skin tones at rest (e) and during activity (f). Error bars represent the 95% confidence interval.CC BY · PubMed Central
Fig. 3 Error in heart rate across all devices and analysis of missing values across consumer devices. a Mean absolute error in heart rate (bpm) across devices during rest (teal) and activity (orange). This shows the true difference in HR from the ECG but does not show the sign of the difference. The green horizontal line represents no error (no difference from the true measurement of HR from ECG). Error bars show the 95% confidence interval. ** indicates significant difference in error between baseline and activity with a Bonferroni multiple hypothesis corrected p value of 0.0042. b Mean relative error in heart rate (bpm) across devices during rest (teal) and during activity (orange) shows the relative differences from the ECG. The green horizontal line represents no error (no difference from the true measurement of HR from ECG). Error bars show the 95% confidence interval. ** indicates significant difference in error between baseline and activity with a Bonferroni multiple hypothesis corrected p value of 0.0042. c Analysis of missing values across skin tones for rest and activity for consumer wearables. Research-grade wearables (Empatica, Biovotion) down-sample and/or interpolate to have exactly 1 Hz sampling rate and thus we could not calculate missingness values for those devices. Missingness is calculated from the expected sampling rate (reported sampling rate for Apple Watch and Garmin and study average sampling rate for Garmin and Miband, which do not report sampling rate). Missingness that is positive indicates percentage of values with missingness. Missingness that is negative indicates a greater than expected sampling rate (more values than expected).
Fig. 3 Error in heart rate across all devices and analysis of missing values across consumer devices. a Mean absolute error in heart rate (bpm) across devices during rest (teal) and activity (orange). This shows the true difference in HR from the ECG but does not show the sign of the difference. The green horizontal line represents no error (no difference from the true measurement of HR from ECG). Error bars show the 95% confidence interval. ** indicates significant difference in error between baseline and activity with a Bonferroni multiple hypothesis corrected p value of 0.0042. b Mean relative error in heart rate (bpm) across devices during rest (teal) and during activity (orange) shows the relative differences from the ECG. The green horizontal line represents no error (no difference from the true measurement of HR from ECG). Error bars show the 95% confidence interval. ** indicates significant difference in error between baseline and activity with a Bonferroni multiple hypothesis corrected p value of 0.0042. c Analysis of missing values across skin tones for rest and activity for consumer wearables. Research-grade wearables (Empatica, Biovotion) down-sample and/or interpolate to have exactly 1 Hz sampling rate and thus we could not calculate missingness values for those devices. Missingness is calculated from the expected sampling rate (reported sampling rate for Apple Watch and Garmin and study average sampling rate for Garmin and Miband, which do not report sampling rate). Missingness that is positive indicates percentage of values with missingness. Missingness that is negative indicates a greater than expected sampling rate (more values than expected).CC BY · PubMed Central

Verification

Claims the paper makes

5 Valid · 4 Not valid · 8 Not checked · 77 Own results and interpretations

41 more claims

Create a free account to see all 94 claims and to verify the 2 references still open.

See all claims — free account
  • It may state two separate claims as one. 34
  • Probably not a substantial claim. 17

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

43 more claims about this paper

Create a free account to see every claim other papers make about this one.

See all claims — free account

Citations

Cites

Show all 37 references

Research works printed but not linked: 2

  1. Adams, C. & Pinkas, D. Internet X.509 Public Key Infrastructure Time-Stamp Protocol (TSP). https://dl.acm.org/doi/book/10.17487/RFC3161. · doi:10.17487/rfc3161
  2. Hugh et al. Consumer-Grade Wrist-worn Ppg Sensors Can Be Used to Detect Differences in Heart Rate Variability among A Heterogenous Prediabetic Population (IEEE International Conference on Biomedical and Health Informatics, 2018).

Cited by

See all 673

Metadata

Authors
Bent, B., Goldstein, B. A., Kibbe, W. A., & Dunn, J. P.
Affiliations
Duke University (US)
Corresponding author
Dunn, Jessilyn P.
Venue
npj Digital Medicine
Published
2020-02-10
Volume
3
Issue
1
Pages
18-18
ISSN
2398-6352 (ISSN-L)
OpenAlex
W3005348902
Cited by
673
References
37
Citations by year
2026: 127 · 2025: 114 · 2024: 111 · 2023: 98 · 2022: 114 · 2021: 85 · 2020: 24
FWCI
25.20
Type
journal article
Language
en
Open access
Licence
Locations
  • npj Digital Medicine (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
Duke FORGE Fellowship
Topics
  • Physical Sciences › Engineering › Biomedical Engineering › Non-Invasive Vital Sign Monitoring
  • Health Sciences › Medicine › Cardiology and Cardiovascular Medicine › Heart Rate Variability and Autonomic Control
  • Health Sciences › Medicine › Physiology (medical) › Cardiovascular and exercise physiology
Keywords
wearable devices, heart rate monitoring, photoplethysmography, measurement accuracy, skin tone, motion artifacts, device comparison, physical activity

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Investigating sources of inaccuracy in wearable optical heart rate sensors. https://app.cochat.ai/papers/10.1038/s41746-020-0226-6
The paper
Bent, B., Goldstein, B. A., Kibbe, W. A., & Dunn, J. P. (2020). Investigating sources of inaccuracy in wearable optical heart rate sensors. npj Digital Medicine. https://doi.org/10.1038/s41746-020-0226-6

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.