CoChat
Sign Up
Cited by 9Non-Invasive Vital Sign MonitoringHeart Rate Variability and Autonomic ControlECG Monitoring and Analysis

Photoplethysmography in Diverse Skin Tones: Evaluating Bias in Smartwatch Health Monitoring

Open access
Asif, Salwa, AlSaafeen, Abdallah, Nadar, Soumiya, Nambiar, Shristhi, Dannawi, Jihad, Korrapati, Harika, Wilkhoo, Harsahaj S·2025·Cureus
DOI 10.7759/cureus.94074PMC PMC12592569
Open this paper in CoChat — free accountOpen PDF

Abstract

Heart disease remains one of the most pressing global health burdens, and the growth of wearable technology has introduced new possibilities for prevention and monitoring. Smartwatches commonly use photoplethysmography (PPG) to estimate heart rate by detecting changes in light reflected from the skin. However, melanin's absorption of green light raises concerns about measurement accuracy in individuals with darker skin tones. This narrative review, conducted in accordance with the PRISMA extension for narrative reviews, examined studies published between May 2017 and May 2025. Searches were performed in PubMed, Google Scholar, and the American College of Cardiology (ACC) databases, yielding 50 records, of which 23 met the inclusion criteria. Only English-language studies were included, and duplicates were removed through automated and manual screening. Findings demonstrated significant variability across devices. One validation study of 60 participants reported that a major smartwatch brand showed mean heart rate differences of less than 5 bpm across skin tones (95% CI: -3.2 to +4.1). By contrast, other brands underestimated heart rate by 10-15 bpm at rest and by more than 20% during vigorous activity in darker-skinned users (n = 75, p < 0.01). Devices using alternate operating systems displayed inconsistent calibration, with error rates reaching up to 12%. Large-scale studies involving over 400,000 participants reported greater than 95% sensitivity for atrial fibrillation detection, though stratification by skin tone was not conducted. Skin-tone-related bias in smartwatch PPG sensors poses risks that extend beyond fitness tracking to arrhythmia detection, hypertension monitoring, and telehealth reliability. To improve equity, standardized validation protocols stratified by Fitzpatrick skin type and activity level are essential, alongside regulatory mandates for inclusive testing. Future research should include longitudinal studies across diverse cohorts to ensure wearable health technologies become reliable tools for all populations.

AI snapshot

Written from the paper’s full text.

TL;DR

This narrative review synthesizes 23 studies (May 2017–May 2025) on photoplethysmography smartwatch accuracy across skin tones, finding device-dependent bias that is worse in darker skin during activity but minimal for some Apple Watch measurements. It calls for Fitzpatrick-stratified validation, regulatory oversight, and inclusive algorithms to prevent clinical inequities.

Key findings

  • Across 23 included studies, accuracy varied by device and context: Apple Watch showed mean heart rate differences of <5 bpm across skin tones (95% CI: -3.2 to +4.1), while other brands underestimated heart rate by 10–15 bpm at rest and by >20% during vigorous activity in darker-skinned users (n = 75, p < 0.01); devices using alternate operating systems had error rates up to 12%.
  • Activity context mattered: Bent et al. reported >20% error during cycling and resistance exercise in individuals with higher pigmentation, compared with <10% error at rest; resting heart rate accuracy was generally preserved, while exercise-based monitoring was most affected by skin-tone bias.
  • Large-scale atrial fibrillation detection involved over 400,000 participants and reported >95% sensitivity, but stratification by skin tone was not conducted; the Apple Heart Study’s irregular pulse notification remained robust with sensitivity for atrial fibrillation unaffected by skin tone.
  • Proposed mitigations include red and infrared wavelengths, multimodal sensors, and AI/ML trained on diverse datasets; fiber Bragg grating pulse sensors achieved <2% error across diverse skin types, and Koerber et al. concluded no single brand was consistently superior across all skin tones and activities.

Methods

  • Narrative review conducted in accordance with the PRISMA extension for narrative reviews; searches were performed in PubMed, Google Scholar, and the American College of Cardiology (ACC) databases for May 2017 to May 2025, limited to English-language publications.
  • Fifty records were initially identified; after screening, 43 reports were sought for retrieval, 34 were assessed for eligibility, and 23 studies were included. Exclusions included non-English articles, gray literature, conference abstracts, duplicates, studies without skin-tone stratification, and case reports.
  • Data were narratively synthesized and thematically grouped into sensor accuracy, algorithmic calibration, clinical implications, and equity considerations.

Limitations

  • Stated: the long-term impacts of repeated underestimation in darker-skinned populations remain largely unexplored; longitudinal studies across diverse cohorts are needed. Limited transparency in algorithm design prevents external validation, and regulatory authorities such as the FDA and CE currently lack explicit requirements for validating wearable performance across diverse skin tones.
  • Stated: some included studies (Henriksen et al., Dias et al., Castaneda et al.) were not stratified by skin tone and were interpreted cautiously in relation to pigmentation bias; evidence was heterogeneous, with no single brand consistently superior across all skin tones and activities.
  • Evident from the text: as a narrative review, it does not pool effect sizes or perform meta-analysis, and findings rely on heterogeneous study designs and activity contexts; the abstract and body also describe somewhat different patterns for underestimation (e.g., 10–15 bpm at rest vs during moderate-to-vigorous exercise), so the magnitude of bias should be read cautiously.

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

Integrity signals

Registry

Crossref · 10.7759/cureus.94074

Cites

34 of 34 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

47 of 58 cited sources checked · 2 supported · 2 flagged · accuracy not yet measured

  • Findingcitation

    Reduced accuracy in darker skin tones has been linked to missed detection of atrial fibrillation when irregular pulse notifications fail to capture abnormal rhythms.

    “Reduced accuracy in darker skin tones has been linked to missed detection of atrial fibrillation when irregular pulse notifications fail to capture abnormal rhythms.”
    Quote in the paperSubstantial claim3 of 3 sources checked
    • cites [7] Sañudo B, De Hoyo M, Muñoz-López A, Perry J, Abt G. Pilot study assessing the influence of skin type on the heart rate measurements obtained by photoplethysmography with the Apple Watch. J Med Syst. 2019, 43:195. 10.1007/s10916-019-1325-2 source

      The Results and Discussion (pages unavailable) report no significant Apple Watch heart-rate difference for the darkest skin type studied, type IV; the observed differences were in types II and III. The paper assesses heart-rate measurement during cycling, not atrial-fibrillation detection or irregular-pulse notifications. (LLM, conf=0.84)

      Checked by a reasoning model — may be wrong.

      Experimental — accuracy not yet measured
    • cites [8] Henriksen A, Haugen Mikalsen M, Woldaregay AZ, Muzny M, Hartvigsen G, Hopstock LA, Grimsgaard S. Using fitness trackers and smartwatches to measure physical activity in research: analysis of consumer wrist-worn wearables. J Med Internet Res. 2018, 20:e110. 10.2196/jmir.9157 source

      The Introduction’s Sensors section describes how wearable optical heart-rate measurements can be affected by movement, ambient light, and tissue compression. Our copy does not address reduced accuracy in darker skin tones or missed atrial-fibrillation detection through irregular-pulse notifications (searched for those terms and for skin pigmentation). (LLM, conf=0.76)

      Experimental — accuracy not yet measured
    • cites [9] Koerber D, Khan S, Shamsheri T, Kirubarajan A, Mehta S. Accuracy of heart rate measurement with wrist-worn wearable devices in various skin tones: a systematic review. J Racial Ethn Health Disparities. 2023, 10:2676-84. 10.1007/s40615-022-01446-9 source

      The Results and Conclusion report that some studies found less accurate wearable heart-rate measurements in people with darker skin tones, although the evidence was mixed. The Discussion says it remains unclear whether skin tone affects arrhythmia detection; it does not link reduced accuracy to missed atrial fibrillation or failed irregular-pulse notifications. Page numbers were not available in our copy. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
  • Findingcitation

    Apple Watch devices showed less than 5 bpm variation across skin tones.

    “In contrast, Apple Watch devices were more consistent, with <5 bpm variation across skin tones, findings supported by Sañudo et al. (2019) and Shcherbina et al. (2017) [7,10].”
    • Also stated in: § Abstract · plain
      “One validation study of 60 participants reported that a major smartwatch brand showed mean heart rate differences of less than 5 bpm across skin tones (95% CI: -3.2 to +4.1).”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites [7] Sañudo B, De Hoyo M, Muñoz-López A, Perry J, Abt G. Pilot study assessing the influence of skin type on the heart rate measurements obtained by photoplethysmography with the Apple Watch. J Med Syst. 2019, 43:195. 10.1007/s10916-019-1325-2 source

      The Results report mean Apple Watch–Polar biases of −2, −2, and −1 bpm for skin types II, III, and IV during exercise, supporting a small average difference across the tested groups. But the study tested only those three skin types during rest and cycling, and its 95% limits of agreement extend beyond 5 bpm; it does not establish less than 5 bpm variation for individual measurements across skin tones generally. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites [10] Shcherbina A, Mattsson CM, Waggott D, et al.. Accuracy in wrist-worn, sensor-based measurements of heart rate and energy expenditure in a diverse cohort. J Pers Med. 2017, 7:3. 10.3390/jpm7020003 source

      The Results (§3.1) report that the Apple Watch had the lowest overall heart-rate error among the tested devices, and the Discussion (§4) notes that participants varied in skin tone. But the paper reports error as a percentage, not a less-than-5-bpm variation across skin tones; it also notes that darker skin tone correlated with increased error across multiple devices. (LLM, conf=0.87)

      Experimental — accuracy not yet measured
  • Premisecitation

    Melanin absorption of green light is a primary factor interfering with optical signal detection in PPG-based measurements.

    “In PPG-based measurements, melanin absorption of green light was identified as a primary factor interfering with optical signal detection [9].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [9] Koerber D, Khan S, Shamsheri T, Kirubarajan A, Mehta S. Accuracy of heart rate measurement with wrist-worn wearable devices in various skin tones: a systematic review. J Racial Ethn Health Disparities. 2023, 10:2676-84. 10.1007/s40615-022-01446-9 source

      The Introduction says melanin absorbs light in the epidermis and is theorized to interfere with PPG signal estimation. It does not identify absorption of green light as a primary interfering factor. The Discussion instead reports that green wavelengths showed greater heart-rate measurement accuracy across skin types than other wavelengths. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
  • Findingcitation

    Arrhythmia detection algorithms such as the irregular pulse notification in the Apple Heart Study remained robust, with sensitivity for atrial fibrillation unaffected by skin tone across a large-scale population.

    “Importantly, arrhythmia detection algorithms such as the irregular pulse notification in the Apple Heart Study (Perez et al., 2019) remained robust, with sensitivity for atrial fibrillation unaffected by skin tone across a large-scale population [15].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [15] Perez MV, Mahaffey KW, Hedlin H, et al.. Large-scale assessment of a smartwatch to identify atrial fibrillation. N Engl J Med. 2019, 381:1909-17. 10.1056/NEJMoa1901183 source

      The Discussion reports that the Apple Heart Study enrolled 419,297 participants and that 84% of subsequent notifications among participants who returned an ECG patch were concordant with atrial fibrillation. But it explicitly says the study was not designed to measure sensitivity, and our copy reports no analysis showing sensitivity was unaffected by skin tone. (LLM, conf=0.90)

      Experimental — accuracy not yet measured
  • Findingcitation

    No single brand was consistently superior across all skin tones and activities.

    “Koerber et al. (2023), in a systematic review, emphasized this heterogeneity, concluding that no single brand was consistently superior across all skin tones and activities [9].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [9] Koerber D, Khan S, Shamsheri T, Kirubarajan A, Mehta S. Accuracy of heart rate measurement with wrist-worn wearable devices in various skin tones: a systematic review. J Racial Ethn Health Disparities. 2023, 10:2676-84. 10.1007/s40615-022-01446-9 source

      The Results (page unavailable) report mixed findings across skin tones and some differences between devices, supporting the claim of heterogeneity. But they also say that all five studies comparing manufacturers found the Apple Watch more accurate than the other devices; the review does not establish a brand ranking across all skin tones and activities. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
  • May combine two claims. 28
  • Classified as probably not a substantial claim. 10

  • May combine two claims.

    Findingcitation

    Apple and Fitbit's proprietary algorithms outperform smaller-brand or open-source counterparts, largely because of superior noise filtering and adaptive calibration.

    “Comparative evaluations indicate that Apple and Fitbit’s proprietary algorithms outperform smaller-brand or open-source counterparts, largely because of superior noise filtering and adaptive calibration [9,11].”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites [9] Koerber D, Khan S, Shamsheri T, Kirubarajan A, Mehta S. Accuracy of heart rate measurement with wrist-worn wearable devices in various skin tones: a systematic review. J Racial Ethn Health Disparities. 2023, 10:2676-84. 10.1007/s40615-022-01446-9 source

      The Results (page not available in our copy) report that all five studies comparing manufacturers found the Apple Watch more accurate than other commercial devices. The Introduction says proprietary algorithms convert optical measurements into heart-rate data, but the paper does not establish that Fitbit also outperforms other brands, compare either brand with open-source algorithms, or attribute an advantage to noise filtering or adaptive calibration. (LLM, conf=0.88)

      Experimental — accuracy not yet measured
    • cites [11] Reddy RK, Pooni R, Zaharieva DP, et al.. Accuracy of wrist-worn activity monitors during common daily physical activities and types of structured exercise: evaluation study. JMIR Mhealth Uhealth. 2018, 6:e10338. 10.2196/10338 source

      The Results (Heart Rate Accuracy; page unavailable) report no accuracy difference between Fitbit and Garmin when activity mode was indicated. The Results (Energy Expenditure Accuracy; page unavailable) report less overall bias for Garmin and significantly better Garmin accuracy in some activities, contrary to the claimed general superiority of Fitbit. The study did not compare Apple or open-source algorithms, or establish noise filtering and adaptive calibration as explanations. (LLM, conf=0.88)

      Checked by a reasoning model — may be wrong.

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingcitation

    Apple devices tended to perform better during exercise but still slightly underestimated peak values, a discrepancy deemed clinically negligible.

    “By contrast, Apple devices tended to perform better during exercise but still slightly underestimated peak values, a discrepancy deemed clinically negligible [9].”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [9] Koerber D, Khan S, Shamsheri T, Kirubarajan A, Mehta S. Accuracy of heart rate measurement with wrist-worn wearable devices in various skin tones: a systematic review. J Racial Ethn Health Disparities. 2023, 10:2676-84. 10.1007/s40615-022-01446-9 source

      The Results (page not available in our copy) report that all five studies comparing manufacturers found the Apple Watch more accurate than other commercial devices. The review does not say that this advantage was specific to exercise, or that Apple devices slightly underestimated peak heart rates or that such a discrepancy was clinically negligible. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingcitation

    Exercise-based heart rate monitoring was most affected by skin-tone bias, while resting heart rate accuracy was generally preserved.

    “When stratified by outcomes, exercise-based heart rate monitoring was most affected by skin-tone bias, while resting heart rate accuracy was generally preserved [12-14].”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites [12] Ware OR, Dawson JE, Shinohara MM, Taylor SC. Racial limitations of Fitzpatrick skin type. Cutis. 2020, 105:77-80.
    • cites [13] Xiao K, Yates JM, Zardawi F, et al.. Characterising the variations in ethnic skin colours: a new calibrated data base for human skin. Skin Res Technol. 2017, 23:21-9. 10.1111/srt.12295 source

      The Results and Samples and protocol sections describe spectrophotometer measurements of skin colour, not heart-rate monitoring. Searches for exercise, resting heart rate, wearable monitoring and photoplethysmography found no passage addressing the claimed difference in accuracy. (LLM, conf=0.78)

      Experimental — accuracy not yet measured
    • cites [14] Zhang Y, Weaver RG, Armstrong B, Burkart S, Zhang S, Beets MW. Validity of wrist-worn photoplethysmography devices to measure heart rate: a systematic review and meta-analysis. J Sports Sci. 2020, 38:2021-34. 10.1080/02640414.2020.1767348 source

  • May combine two claims.

    Findingcitation

    PPG devices showed error rates exceeding 20% during cycling and resistance exercise in individuals with higher pigmentation, while maintaining less than 10% error at rest.

    “Bent et al. (2020) reported that PPG devices showed error rates exceeding 20% during cycling and resistance exercise in individuals with higher pigmentation, while maintaining <10% error at rest, underscoring the importance of activity context [5].”
    • Also stated in: § Review · plain
      “For example, Bent et al. (2020) quantified error >20% during cycling in darker-skinned individuals compared with <10% at rest, showing how context alters reliability [5].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [5] Bent B, Goldstein BA, Kibbe WA, Dunn JP. Investigating sources of inaccuracy in wearable optical heart rate sensors. NPJ Digit Med. 2020, 3:18. 10.1038/s41746-020-0226-6 source

      The Results sections on accuracy at rest and during physical activity and on rhythmic activity report greater error during activity, supporting the importance of activity context. But the study protocol used walking, not cycling or resistance exercise, and the Results section on skin tones found no significant overall correlation between pigmentation and heart-rate error. The cited passages do not establish the claimed >20% activity error or <10% resting error for higher-pigmentation participants; page numbers are unavailable in our copy. (LLM, conf=0.90)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingcitation

    Heart rate variability applications showed promise but remain underexplored in terms of pigmentation differences.

    “On the other hand, heart rate variability (HRV) applications showed promise but remain underexplored in terms of pigmentation differences [16].”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [16] Haxha S, Nwibor C, Ali M, Sakel M, Saunders K, Dyo V, Nabakooza S. Effect of skin pigmentation and finger choice on accuracy of oxygen saturation measurement in an IoT-based pulse oximeter. Sensors (Basel). 2024, 24:3301. 10.3390/s24113301 source

      The Abstract describes measurements of oxygen saturation (SpO2) and heart rate, while Results and Discussion §5.2 compares SpO2 readings across skin colours. A search for “heart rate variability” and HRV applications found no passage assessing their promise or whether pigmentation differences remain underexplored. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingcitation

    A majority of skin tone-related algorithmic models work best when applied to diagnostic data that includes patients with lighter complexions, while generating inferior outcomes for patients with darker complexions.

    “A majority of skin tone-related algorithmic models work best when applied to diagnostic data that includes patients with lighter complexions, while generating inferior outcomes for patients with darker complexions.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [29] Pope J, Hassanuzzaman M, Sherpa M, Emara O, Joshi A, Adhikari N. Skin cancer machine learning model tone bias [PREPRINT]. arXiv. 2024, source

      The Abstract reports that models trained on both the imbalanced and balanced datasets detected malignant images significantly better for lighter than darker skin tones. It does not establish that a majority of skin tone-related models have this pattern; the study evaluates models trained on one subset of skin images. (LLM, conf=0.88)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Skin pigmentation affects light absorption and reflection such that darker skin tones reduce sensor signal detection capabilities and increase misread errors.

    “Skin pigmentation affects light absorption and reflection rates such that darker skin tones reduce sensor signal detection capabilities while increasing misread errors.”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [20] Castaneda D, Esparza A, Ghamari M, Soltanpur C, Nazeran H. A review on wearable photoplethysmography sensors and their potential future applications in health care. Int J Biosens Bioelectron. 2018, 4:195-202. 10.15406/ijbsbe.2018.04.00125 source

      The Introduction’s “PPG sensors” section describes how sensors detect reflected light, and “Factors affecting PPG sensor recordings” discusses other sources of signal variation. Neither section addresses skin pigmentation, darker skin tones, or increased misread errors; searches for those terms found no supporting passage in our copy. No page numbers were available. (LLM, conf=0.72)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Red and infrared light wavelengths, which are less absorbed by melanin, have potential to reduce error rates and enhance inclusivity.

    “Additionally, Dias et al. (2018) and Castaneda et al. (2018) highlighted the potential of red and infrared light wavelengths, which are less absorbed by melanin, to reduce error rates and enhance inclusivity [16,17].”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites [16] Haxha S, Nwibor C, Ali M, Sakel M, Saunders K, Dyo V, Nabakooza S. Effect of skin pigmentation and finger choice on accuracy of oxygen saturation measurement in an IoT-based pulse oximeter. Sensors (Basel). 2024, 24:3301. 10.3390/s24113301 source

      Section 3.1 says red and near-infrared light penetrate biological tissue more deeply than several other wavelengths and are used in pulse oximeters. It does not say they are less absorbed by melanin or that using them reduces pigmentation-related errors or enhances inclusivity; Section 5.2 reports no significant difference between the skin-colour groups tested, not a wavelength-based improvement. (LLM, conf=0.82)

      Experimental — accuracy not yet measured
    • cites [17] Li K, Cardoso C, Moctezuma-Ramirez A, Elgalad A, Perin E. Heart rate variability measurement through a smart wearable device: another breakthrough for personal health monitoring?. Int J Environ Res Public Health. 2023, 20:7146. 10.3390/ijerph20247146 source

      Section 5.2 (PPG; page not provided) says skin color can affect photoplethysmography accuracy. It does not say that red or infrared light is less absorbed by melanin, or that those wavelengths could reduce errors and improve inclusivity. (LLM, conf=0.85)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingcitation

    Technological limitations are primarily sensor-related (LED wavelength, light intensity, placement), while algorithmic limitations stem from biased datasets, poor calibration, and underrepresentation of darker skin in training cohorts.

    “Technological limitations are primarily sensor-related - such as LED wavelength, light intensity, and placement - while algorithmic limitations stem from biased datasets, poor calibration, and underrepresentation of darker skin in training cohorts.”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites [27] Chow HW, Yang CC. Accuracy of optical heart rate sensing technology in wearable fitness trackers for young and older adults: validation and comparison study. JMIR Mhealth Uhealth. 2020, 8:e14707. 10.2196/14707 source

      The Introduction (Accuracy of Optical Heart Rate Monitoring; Research Gaps, page unavailable) identifies sensor placement and skin structure as factors in optical heart-rate accuracy. The Discussion (Limitations, page unavailable) says devices may use different PPG modules or algorithms, but the paper does not identify LED wavelength or light intensity as primary limitations, or attribute algorithmic limitations to biased training datasets, poor calibration, or underrepresentation of darker skin in training cohorts. Its observation that earlier validation studies enrolled mostly white participants concerns study samples, not algorithm training cohorts. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
    • cites [28] Fine J, Branan KL, Rodriguez AJ, et al.. Sources of inaccuracy in photoplethysmography for continuous cardiovascular monitoring. Biosensors (Basel). 2021, 11:126. 10.3390/bios11040126 source

      Section 2.1 (Skin Tone; page unavailable) links PPG accuracy to LED wavelength and source–detector placement, and reports greater error for darker skin tones. The Introduction also acknowledges hardware and software sources of noise. However, the cited passages do not establish that technological limitations are primarily sensor-related or that algorithmic limitations stem from biased training datasets, poor calibration, or underrepresentation of darker skin in training cohorts; the study with fewer dark-skinned participants is described as a device validation, not a training cohort. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Interpretationcitation

    The problem arises primarily from model biases as well as from issues with feature selection and calibration because better settings for lighter skin types create unequal outcomes for people with darker skin.

    “The problem arises primarily from model biases as well as from issues with feature selection and calibration because better settings for lighter skin types create unequal outcomes for people with darker skin [29].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [29] Pope J, Hassanuzzaman M, Sherpa M, Emara O, Joshi A, Adhikari N. Skin cancer machine learning model tone bias [PREPRINT]. arXiv. 2024, source

      The Abstract (page unavailable) reports that the models detected malignant lesions better in lighter than darker skin tones, and the Introduction (page unavailable) says the results provide evidence that the models learn tone-related features. The paper does not establish that feature selection or calibration issues, or settings optimized for lighter skin, cause the unequal outcomes. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Darker skin, with higher melanin levels, absorbs more light, reducing signal strength and limiting light penetration to blood vessels.

    “Darker skin, with higher melanin levels, absorbs more light, reducing signal strength and limiting light penetration to blood vessels [19].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [19] Singh S, Bennett MR, Chen C, Shin S, Ghanbari H, Nelson BW. Impact of skin pigmentation on pulse oximetry blood oxygenation and wearable pulse rate accuracy: systematic review and meta-analysis. J Med Internet Res. 2024, 26:e62769. 10.2196/62769 source

      The Discussion’s Strengths and Limitations section says skin tone is determined by the amount of melanin in the epidermis. But neither that section nor the Principal Findings discussion states that darker skin absorbs more light, weakens the signal, or limits light penetration to blood vessels; the discussion only suggests that melanin levels could affect the PPG signal. No page numbers are provided in our copy. (LLM, conf=0.82)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingcitation

    Smartwatches often emit insufficient light for accurate readings on darker-skinned users, leading to inconsistent results and potential misdiagnosis.

    “Studies show that smartwatches often emit insufficient light for accurate readings on darker-skinned users, leading to inconsistent results and potential misdiagnosis.”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [21] Parak J, Korhonen I. Evaluation of wearable consumer heart rate monitors based on photopletysmography. Annu Int Conf IEEE Eng Med Biol Soc. 2014, 2014:3670-3. 10.1109/EMBC.2014.6944419 source
    • cites [22] Duncan M, Murawski B, Short CE, et al.. Activity trackers implement different behavior change techniques for activity, sleep, and sedentary behaviors. Interact J Med Res. 2017, 6:e13. 10.2196/ijmr.6685 source

      The Methods section (page unavailable) describes an assessment of behavior-change techniques in three wrist-worn activity-tracker systems. The Discussion mentions a limitation in how those systems classify sedentary behavior, but our copy does not address emitted light, skin tone, reading accuracy, or misdiagnosis (searched for darker skin, light, optical sensors, and accuracy). (LLM, conf=0.78)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingcitation

    Li et al. demonstrated reliable HRV tracking overall but did not stratify outcomes by skin tone.

    “Li et al. (2023) demonstrated reliable HRV tracking overall but did not stratify outcomes by skin tone [17].”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [17] Li K, Cardoso C, Moctezuma-Ramirez A, Elgalad A, Perin E. Heart rate variability measurement through a smart wearable device: another breakthrough for personal health monitoring?. Int J Environ Res Public Health. 2023, 20:7146. 10.3390/ijerph20247146 source

      Section 7 (page unavailable) reports that verification studies found wearable devices can provide reliable HRV readings compared with clinical ECG. But Li et al. review those studies rather than demonstrate reliability in their own trial, and Section 10 cautions that reliable long-term tracking remains difficult. Section 5.2 notes that skin color can affect PPG accuracy, but the paper does not present HRV outcomes stratified by skin tone. (LLM, conf=0.87)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Interpretationcitation

    Weaker signals and increased scattering from darker skin degrade measurement accuracy.

    “This results in weaker signals and increased scattering, which degrade measurement accuracy [19,20].”
    Quote in the paperSubstantial claim2 of 2 sources checked
    • cites [19] Singh S, Bennett MR, Chen C, Shin S, Ghanbari H, Nelson BW. Impact of skin pigmentation on pulse oximetry blood oxygenation and wearable pulse rate accuracy: systematic review and meta-analysis. J Med Internet Res. 2024, 26:e62769. 10.2196/62769 source

      The Results (PPG Accuracy and Bias by Skin Pigmentation; page not provided) report higher pooled pulse-oximetry bias for darker than lighter skin, supporting a difference in measurement accuracy. The paper does not establish weaker signals or increased scattering as the cause of that difference. (LLM, conf=0.85)

      Experimental — accuracy not yet measured
    • cites [20] Castaneda D, Esparza A, Ghamari M, Soltanpur C, Nazeran H. A review on wearable photoplethysmography sensors and their potential future applications in health care. Int J Biosens Bioelectron. 2018, 4:195-202. 10.15406/ijbsbe.2018.04.00125 source

      The Introduction’s “Factors affecting PPG sensor recordings” section says that anatomy and tissue differences can change how light propagates and alter the signal. But our copy does not link darker skin to weaker signals, increased scattering, or degraded measurement accuracy. Searches for skin pigmentation, melanin, scattering, and measurement accuracy did not find that link. (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Interpretationcitation

    Misleading heart rate readings lead to negative consequences such as substandard recovery assessments, improper exercise intensity control, and cardiovascular condition identification, resulting in deteriorated fitness results and healthcare outcomes.

    “Misleading heart rate readings lead to several negative consequences, such as substandard recovery assessments and improper exercise intensity control, and cardiovascular condition identification, resulting in deteriorated fitness results and healthcare outcomes [27].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [27] Chow HW, Yang CC. Accuracy of optical heart rate sensing technology in wearable fitness trackers for young and older adults: validation and comparison study. JMIR Mhealth Uhealth. 2020, 8:e14707. 10.2196/14707 source

      Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationcitation

    Socioeconomic disparities in access to smartwatches may interact with skin-tone-related bias, raising additional equity concerns.

    “However, regional variations may exist, and socioeconomic disparities in access to these devices may further interact with skin-tone-related bias, raising additional equity concerns [17-23].”
    Quote in the paperNot a substantial claim5 of 7 sources checked
    • cites [17] Li K, Cardoso C, Moctezuma-Ramirez A, Elgalad A, Perin E. Heart rate variability measurement through a smart wearable device: another breakthrough for personal health monitoring?. Int J Environ Res Public Health. 2023, 20:7146. 10.3390/ijerph20247146 source

      Section 5.2 (PPG; page not provided) says skin color can affect the accuracy of wearable-device measurements, supporting the skin-tone-related concern. The passages do not address socioeconomic disparities in smartwatch access or say that access and measurement bias interact to create additional equity concerns. (LLM, conf=0.83)

      Experimental — accuracy not yet measured
    • cites [18] Shi C, Zhang H, Ni X, Wang K. An FBG-based sensor with both wearable and handheld forms for carotid arterial pulse waveform measurement. IEEE Trans Instrum Meas. 2023, 72:1-10. 10.1109/TIM.2023.3311052 source
    • cites [19] Singh S, Bennett MR, Chen C, Shin S, Ghanbari H, Nelson BW. Impact of skin pigmentation on pulse oximetry blood oxygenation and wearable pulse rate accuracy: systematic review and meta-analysis. J Med Internet Res. 2024, 26:e62769. 10.2196/62769 source

      The Introduction (page unavailable) links potential skin-pigmentation bias in wearable devices to concerns about inequity and discrimination. It does not discuss socioeconomic disparities in smartwatch access or whether those disparities interact with skin-tone-related bias. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites [20] Castaneda D, Esparza A, Ghamari M, Soltanpur C, Nazeran H. A review on wearable photoplethysmography sensors and their potential future applications in health care. Int J Biosens Bioelectron. 2018, 4:195-202. 10.15406/ijbsbe.2018.04.00125 source

      The Introduction discusses consumer interest in smartwatches, and the 'Factors affecting PPG sensor recordings' section discusses anatomical and movement-related effects; neither addresses socioeconomic access or skin-tone-related bias. Searches for disparities in smartwatch access and for skin pigmentation or skin-tone bias found no passage linking those concerns. Page numbers were not available in our copy. (LLM, conf=0.75)

      Experimental — accuracy not yet measured
    • cites [21] Parak J, Korhonen I. Evaluation of wearable consumer heart rate monitors based on photopletysmography. Annu Int Conf IEEE Eng Med Biol Soc. 2014, 2014:3670-3. 10.1109/EMBC.2014.6944419 source
    • cites [22] Duncan M, Murawski B, Short CE, et al.. Activity trackers implement different behavior change techniques for activity, sleep, and sedentary behaviors. Interact J Med Res. 2017, 6:e13. 10.2196/ijmr.6685 source

      The Introduction reports overall device ownership, and the Methods describe an assessment of behavior-change features in three wrist-worn trackers; neither addresses socioeconomic differences in access. Searches for skin-tone bias, affordability, and equity found no passage discussing their proposed interaction. Our copy therefore does not verify the claim. (LLM, conf=0.76)

      Experimental — accuracy not yet measured
    • cites [23] Nelson BW, Low CA, Jacobson N, Areán P, Torous J, Allen NB. Guidelines for wrist-worn consumer wearable assessment of heart rate in biobehavioral research. NPJ Digit Med. 2020, 3:90. 10.1038/s41746-020-0297-4 source

      The ‘Biobehavioral and demographic data’ section (page not available) discusses reduced heart-rate accuracy on darker skin tones in some studies and the resulting generalizability concerns. It does not address socioeconomic disparities in smartwatch access or whether those disparities interact with skin-tone-related bias. (LLM, conf=0.86)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Premisecitation

    Melanin levels can change PPG sensors, producing inaccurate results.

    “As darker skin absorbs and scatters lighter light, melanin levels can change PPG sensors, which measure blood flow through the skin, producing inaccurate results [3].”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites [3] Huhn S, Axt M, Gunga HC, et al.. The impact of wearable technologies in health research: scoping review. JMIR Mhealth Uhealth. 2022, 10:e34384. 10.2196/34384 source

      The Results section (Wearable Characteristics; page not specified) describes PPG as an optical technique that detects blood-volume changes in tissue, supporting the sentence’s description of what the sensors measure. The paper passages do not say that melanin or darker skin affects PPG accuracy; searches for melanin, skin pigmentation, and sensor accuracy found no such discussion. (LLM, conf=0.84)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    PPG detects heart rate by measuring light absorption and reflection occurring within the skin's microvascular bed during blood pulses.

    “Photoplethysmography (PPG) detects heart rate by measuring light absorption and reflection occurring within the skin's microvascular bed during blood pulses according to [20].”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites [20] Castaneda D, Esparza A, Ghamari M, Soltanpur C, Nazeran H. A review on wearable photoplethysmography sensors and their potential future applications in health care. Int J Biosens Bioelectron. 2018, 4:195-202. 10.15406/ijbsbe.2018.04.00125 source

  • May combine two claims.

    Findingcitation

    Errors during exercise monitoring may result in inappropriate training intensities or misleading recovery assessments.

    “Errors during exercise monitoring may result in inappropriate training intensities or misleading recovery assessments [7,8,9].”
    Quote in the paperSubstantial claim3 of 3 sources checked
    • cites [7] Sañudo B, De Hoyo M, Muñoz-López A, Perry J, Abt G. Pilot study assessing the influence of skin type on the heart rate measurements obtained by photoplethysmography with the Apple Watch. J Med Syst. 2019, 43:195. 10.1007/s10916-019-1325-2 source

      The Perspectives section (page unavailable) says heart-rate monitors are used to control exercise intensity and that measurement accuracy matters when prescribing activity. The Results section (page unavailable) reports small measurement differences during exercise, supporting a potential concern about training intensity. The paper does not address recovery assessments. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites [8] Henriksen A, Haugen Mikalsen M, Woldaregay AZ, Muzny M, Hartvigsen G, Hopstock LA, Grimsgaard S. Using fitness trackers and smartwatches to measure physical activity in research: analysis of consumer wrist-worn wearables. J Med Internet Res. 2018, 20:e110. 10.2196/jmir.9157 source

      The Introduction (Sensors; Wearable Usage Scenario, pages unavailable) says wearable measurements vary in accuracy and that researchers can use continuous data to adjust physical-activity interventions. It does not specifically say that monitoring errors may lead to inappropriate training intensities or misleading recovery assessments. (LLM, conf=0.78)

      Experimental — accuracy not yet measured
    • cites [9] Koerber D, Khan S, Shamsheri T, Kirubarajan A, Mehta S. Accuracy of heart rate measurement with wrist-worn wearable devices in various skin tones: a systematic review. J Racial Ethn Health Disparities. 2023, 10:2676-84. 10.1007/s40615-022-01446-9 source

      The Results (page unavailable) report heart-rate measurement errors in some wrist-worn devices, which supports the premise that exercise-monitoring readings can be inaccurate. The paper does not address whether those errors lead to inappropriate training intensities or misleading recovery assessments. (LLM, conf=0.79)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Interpretationcitation

    These discrepancies undermine trust in wearable health devices, especially when inaccurate readings affect fitness tracking or miss critical signs like arrhythmias.

    “These discrepancies undermine trust in wearable health devices, especially when inaccurate readings affect fitness tracking or miss critical signs like arrhythmias [21,22].”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites [21] Parak J, Korhonen I. Evaluation of wearable consumer heart rate monitors based on photopletysmography. Annu Int Conf IEEE Eng Med Biol Soc. 2014, 2014:3670-3. 10.1109/EMBC.2014.6944419 source
    • cites [22] Duncan M, Murawski B, Short CE, et al.. Activity trackers implement different behavior change techniques for activity, sleep, and sedentary behaviors. Interact J Med Res. 2017, 6:e13. 10.2196/ijmr.6685 source

      The Methods describe assessing behavior-change techniques in activity trackers, and the Discussion mentions technical limitations in how the trackers classify sedentary behavior. Neither section says that inaccurate readings undermine users’ trust or cause devices to miss arrhythmias (searched for accuracy, trust, and arrhythmia). (LLM, conf=0.76)

      Experimental — accuracy not yet measured

  • May combine two claims.

    Findingcitation

    Garmin vívosmart HR+ and Fitbit Charge 2 achieved reasonable heart rate accuracy during steady-state activity but produced large deviations during cycling, particularly in darker-skinned users.

    “Comparisons across brands also revealed that Garmin vívosmart HR+ (Garmin Ltd., Olathe, KS) and Fitbit Charge 2 were shown to achieve reasonable heart rate accuracy during steady-state activity but produced large deviations during cycling, particularly in darker-skinned users.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    Smartwatch PPG accuracy is influenced by skin tone, with the magnitude of the effect varying widely across devices and contexts.

    “Studies consistently demonstrated that smartwatch accuracy is influenced by skin tone, though the magnitude of this effect varied widely across devices and contexts (Table 3).”
    • Also stated in: § Review · boosted
      “The precision of smartwatches depends heavily on skin tone since accurate wearable health measurements require reliable data.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingnone given

    Large-scale studies involving over 400,000 participants reported greater than 95% sensitivity for atrial fibrillation detection, though stratification by skin tone was not conducted.

    “Large-scale studies involving over 400,000 participants reported greater than 95% sensitivity for atrial fibrillation detection, though stratification by skin tone was not conducted.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based).

    Findingnone given

    Mitigation strategies, including red and infrared light wavelengths or multimodal sensors that combine optical with accelerometer data, have shown potential in reducing pigmentation-related error.

    “Mitigation strategies, including red and infrared light wavelengths or multimodal sensors that combine optical with accelerometer data, have shown potential in reducing pigmentation-related error.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingnone given

    Other brands underestimated heart rate by 10-15 bpm at rest and by more than 20% during vigorous activity in darker-skinned users (n = 75, p < 0.01).

    “By contrast, other brands underestimated heart rate by 10-15 bpm at rest and by more than 20% during vigorous activity in darker-skinned users (n = 75, p < 0.01).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    Some manufacturers have attempted to adjust algorithms to improve accuracy, but these modifications have not consistently eliminated bias.

    “While some manufacturers have attempted to adjust algorithms to improve accuracy, evidence suggests that these modifications have not consistently eliminated bias, and error rates are still reported across different skin tones in recent studies.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Findingnone given

    Adaptive algorithms and machine learning models are beginning to be explored, but their validation remains inconsistent.

    “Although adaptive algorithms and machine learning models are beginning to be explored, their validation remains inconsistent, and multi-wavelength approaches, such as red and infrared light, face trade-offs related to cost, battery consumption, and device design.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    PPG sensors generate heart rate measurements by applying LEDs to measure skin-based light reflection through photodetectors that evaluate changes in intensity.

    “PPG sensors generate heart rate measurements by applying LEDs to measure skin-based light reflection through photodetectors that evaluate changes in intensity [21].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [21] Parak J, Korhonen I. Evaluation of wearable consumer heart rate monitors based on photopletysmography. Annu Int Conf IEEE Eng Med Biol Soc. 2014, 2014:3670-3. 10.1109/EMBC.2014.6944419 source

  • May combine two claims.

    Premisecitation

    PPG technology typically relies on green light wavelengths, which are particularly affected by melanin's higher absorption.

    “Photoplethysmography (PPG) technology typically relies on green light wavelengths, which are particularly affected by melanin’s higher absorption [1].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [1] Ray I, Liaqat D, Gabel M, de Lara E. Skin tone, confidence, and data quality of heart rate sensing in WearOS smartwatches. 2021 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events (PerCom Workshops). Kassel, Germany; 2021. 2021-22. 10.1109/PerComWorkshops51409.2021.9431120 source

  • May combine two claims.

    Interpretationreasoning

    The clinical implications of PPG inaccuracies extend beyond fitness tracking to arrhythmia detection, hypertension monitoring, and telehealth assessments, where even modest errors can alter diagnostic and therapeutic decisions.

    “Skin-tone-related bias in smartwatch PPG sensors poses risks that extend beyond fitness tracking to arrhythmia detection, hypertension monitoring, and telehealth reliability.”
    • Also stated in: § Review · hedged
      “Beyond fitness tracking, the clinical implications are broader: smartwatch-derived atrial fibrillation notifications have been validated in large-scale cohorts [25], yet blood pressure monitoring, hypertension surveillance, and telehealth systems relying on wearable data could be jeopardized if measurement bias persists.”
    • Also stated in: § Conclusions · plain
      “The clinical implications of these inaccuracies extend beyond fitness tracking to critical domains, including arrhythmia detection, hypertension monitoring, and telehealth assessments, where even modest errors can alter diagnostic and therapeutic decisions.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Clinically meaningful error in cardiology monitoring is typically greater than 10 bpm.

  • May combine two claims.

    Interpretationreasoning

    Theoretical implications include flawed training load assessment, stress evaluation, and recovery modeling, whereas real-world outcomes documented in clinical contexts involve misclassification of exercise intensity and delayed arrhythmia detection.

    “Theoretical implications include flawed training load assessment, stress evaluation, and recovery modeling, whereas real-world outcomes documented in clinical contexts involve misclassification of exercise intensity and delayed arrhythmia detection.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    Multi-wavelength approaches, such as red and infrared light, face trade-offs related to cost, battery consumption, and device design.

    “Although adaptive algorithms and machine learning models are beginning to be explored, their validation remains inconsistent, and multi-wavelength approaches, such as red and infrared light, face trade-offs related to cost, battery consumption, and device design.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The review was conducted in accordance with PRISMA guidelines for narrative syntheses.

    “This narrative review was conducted in accordance with PRISMA guidelines for narrative syntheses and literature search with appropriate strategies, inclusion, and exclusion criteria (Figure 2).”
    • Also stated in: § Abstract · plain
      “This narrative review, conducted in accordance with the PRISMA extension for narrative reviews, examined studies published between May 2017 and May 2025.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    Integrating larger, more diverse training datasets with adaptive AI systems capable of real-time adjustment may reduce discriminatory biases and enhance equity of performance across populations.

    “Integrating larger, more diverse training datasets with adaptive AI systems capable of real-time adjustment may reduce discriminatory biases and enhance equity of performance across populations.”
    Quote in the paperNot a substantial claim
34 of 34

  • The paper relies on this without support.

    Clinically meaningful error in cardiology monitoring is typically greater than 10 bpm.

    “Clinically meaningful error is typically considered >10 bpm in cardiology monitoring, meaning even modest inaccuracies may alter clinical decision-making.”

  • The paper relies on this without support.

    Regulatory authorities such as the FDA and CE currently lack explicit requirements for validating wearable performance across diverse skin tones.

    “Despite these developments, regulatory authorities such as the FDA and CE currently lack explicit requirements for validating wearable performance across diverse skin tones, raising concerns for health equity.”
2 of 2

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Atrial fibrillation sensitivity is said to be unaffected by skin tone despite no stratification

    Major

    The abstract says the large-scale atrial fibrillation studies did not stratify results by skin tone, but the Results section concludes that sensitivity was unaffected by skin tone. An unstratified result cannot establish that comparison.

  • The 10-15 bpm underestimation is assigned to different activity conditions

    Major

    The abstract places the 10-15 bpm underestimation at rest, whereas the Results section reports that magnitude during moderate to vigorous exercise.

  • Included studies conflict with the skin-tone stratification exclusion criterion

    Minor

    The eligibility table excludes studies without skin-tone or pigmentation stratification, yet the Results section explicitly describes unstratified studies as included.

3 of 3

Cites

Flagged first, then by citations

Cited by

Still being checked.

Authors
Asif, S., AlSaafeen, A., Nadar, S., Nambiar, S., Dannawi, J., Korrapati, H., & Wilkhoo, H. S.
Affiliations
  • Tbilisi State Medical University (GE)
  • Makassed General Hospital (LB)
ORCID
Venue
Cureus
Publisher
Springer Science and Business Media LLC
Published
2025-10-07
Volume
17
Issue
10
Pages
e94074-e94074
ISSN
2168-8184 (ISSN-L)
DOI
10.7759/cureus.94074
PMC
PMC12592569
OpenAlex
W4414924044
Semantic Scholar
4be0d2a5202e672828ebe34af20aff39d1d1b3fa
Cited by
9
References
34
Citations by year
2026: 9
FWCI
2.24
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • Cureus (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Europe PMC (PubMed Central) (submittedVersion, cc-by)
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 › Cardiology and Cardiovascular Medicine › ECG Monitoring and Analysis
Keywords
photoplethysmography, smartwatches, skin tone, melanin, heart rate measurement, algorithmic bias, wearable health technology, Fitzpatrick skin type, atrial fibrillation detection, hypertension monitoring, telemedicine, health equity, device validation

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.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-09.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-09.

Cited sources are checked by gpt-6-sol.

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