Research progress on continuous physiological monitoring via smartwatches in the workplace for occupational health risk prevention and control: a narrative review

Introduction

Background

Occupational health in the workplace is a vital component of health protection for the working population; however, it remains relatively underdeveloped in terms of continuous, dynamic, and personalized monitoring. Traditional management approaches, which primarily rely on annual physical examinations and periodic questionnaires, struggle to promptly detect dynamic risk signals—such as fatigue, sleep disturbances, increased psychological stress, and cardiovascular abnormalities—that arise under varying shift schedules, tasks, and environmental exposure conditions. With the advancement of wearable technology, smartwatches—due to their capabilities for continuous monitoring, portability, and high user acceptance—are increasingly being applied to the long-term monitoring of indicators such as heart rate, heart rate variability (HRV), physical activity, sleep, and circadian rhythms among the working population, thereby providing new technical pathways for occupational health risk identification and management (1). From a broader research perspective, affordable wearable devices and related novel self-powered monitoring technologies have been widely adopted in health research and medical settings (2,3). In industrial and workplace contexts, relevant reviews further highlight their potential applications in safety monitoring, support, training, and tracking (4).

Rationale and knowledge gap

Existing reviews provide an important foundation for the present study. For example, reviews in the field of industrial safety have mainly focused on the application of wearable devices in hazard identification, safety training, and on-site monitoring (4). Reviews related to occupational health among nurses have placed greater emphasis on the value of wearable data for shift work, fatigue, and recovery management (5). Studies on burnout detection among healthcare workers have primarily focused on mental health risk identification and predictive modeling (6). Broader reviews of workplace wearables have addressed employee acceptance, ethical concerns, and organizational implementation conditions (7,8). However, most existing reviews have concentrated on specific occupations, single risk outcomes, or general workplace technology applications, and have not sufficiently integrated the measurement validity of smartwatch-derived indicators, their interpretive boundaries in occupational contexts, risk-warning workflows, organizational implementation conditions, and data governance requirements. In other words, the key issue in workplace applications is not only what smartwatches can measure, but also whether these measurements remain robust in real-world work environments, whether they can be interpreted in relation to shift schedules and task contexts, and whether they can be translated into a closed-loop process of verification, feedback, and intervention.

Objective

Accordingly, this review synthesizes the available evidence across physiological health monitoring, identification of mental health-related risks, workplace adaptability, implementation feasibility, closed-loop management, and data governance. The aim is to clarify which smartwatch-derived indicators may be more suitable for priority validation in occupational settings, which applications still require stronger evidence, and which implementation conditions are necessary for responsible use in workplace health protection. We present this article in accordance with the Narrative Review reporting checklist (available at https://jphe.amegroups.com/article/view/10.21037/jphe-2026-0028/rc).

Methods

This narrative review was conducted between January 2014 and March 2026. English- and Chinese-language literature was identified through targeted searches of PubMed, Web of Science, CNKI, supplemented by manual screening of reference lists of key publications. The search focused on smartwatches or wrist-worn wearable devices, occupational health or workplace settings, and outcomes including heart rate, heart rate variability, arrhythmia, sleep, circadian rhythm, physical activity, workload, stress, fatigue, implementation, and data governance. A total of 201 records were initially identified. After title and abstract screening, 76 full-text articles were assessed, and 49 publications were finally included in the narrative synthesis. In addition, two regulatory documents were cited to support the discussion of data governance.

Two authors independently screened the titles and abstracts and then assessed potentially relevant full texts. Disagreements were resolved through discussion. Priority was given to reviews, validation studies, observational studies, randomized trials, and implementation studies that were directly relevant to occupational settings or provided transferable evidence for workplace risk monitoring and management. Studies were considered directly relevant if they examined smartwatch or wrist-worn wearable applications in occupational or workplace populations. Evidence from non-occupational settings was included as transferable evidence when it addressed measurement validity, clinical verification pathways, implementation feasibility, or governance issues applicable to workplace monitoring. Articles were excluded if they focused only on non-wrist devices, purely technical algorithm development without interpretable occupational-health applications, conference abstracts, editorials, or topics with weak relevance to workplace risk prevention and control. Given the narrative design of this review, the search and selection process was reported to improve transparency, and no formal risk-of-bias assessment was conducted. The search strategy summary is presented in Table 1.

The detailed search strategy for PubMed is provided in Appendix 1.

Current applications of smartwatches in monitoring the physiological health of working adults

Monitoring of cardiovascular and autonomic nervous system function

Heart rate and HRV are among the most commonly used physiological indicators in continuous workplace monitoring (9). However, their application in occupational health requires a clear distinction between two levels of evidence: first, the agreement between heart rate or HRV metrics derived from smartwatches or wrist-worn devices and reference standards such as electrocardiography (ECG), namely measurement validity; and second, the extent to which these metrics can be interpreted as proxies for stress, fatigue, or insufficient recovery, namely interpretive validity. Validation studies have shown that, under resting or low-activity conditions, smartwatch-derived mean heart rate shows good agreement with ECG measurements (10,11). Evidence regarding HRV also suggests that some long-window HRV indices can achieve acceptable measurement accuracy and may be useful for describing within-person changes in workload and recovery across the working day (12,13). Related reviews further indicate that wrist-worn devices have considerable potential for remote monitoring of heart rate and HRV, although the selection of metrics, measurement conditions, and boundaries of interpretation need to be specified for particular contexts (14).

The measurement validity of HRV does not automatically imply its interpretive validity. Theurl et al. mainly addressed the agreement between smartwatch- or photoplethysmography (PPG)-derived HRV and ECG-derived HRV, showing that the stability and consistency of different HRV parameters vary substantially. Ultra-short-term or nonlinear features, such as the root mean square of successive differences between normal heartbeats [root mean square of successive differences (rMSSD)] and detrended fluctuation analysis alpha 1 (DFA-alpha1), are more susceptible to sampling frequency, motion artifacts, wearing conditions, and algorithmic processing (15). By contrast, interpreting HRV as an indicator of stress, fatigue, or insufficient recovery requires consideration of task demands, sleep, emotion, physical activity, individual baselines, and the broader work context. Martinez et al. also suggested that HRV, when used as a proxy for stress or recovery, should be interpreted in relation to contextual information and supported by multiple sources of data (16). Therefore, in workplace applications, a more prudent approach is to use HRV for within-person trend monitoring, identifying sustained deviations from an individual baseline that may suggest inadequate recovery or cumulative workload, rather than treating HRV as a direct real-time readout of stress level or fatigue level.

In early warning practice, arrhythmia screening, particularly atrial fibrillation (AF) screening, represents one of the smartwatch applications closest to a closed-loop management model. Validation studies using synchronous Holter recordings have shown that smartwatch-derived PPG data are feasible for AF detection (17,18). Artificial intelligence (AI)-assisted ECG analysis has also shown promise in AF identification, risk prediction, and the optimization of subsequent verification processes (19). The Fitbit Heart Study further demonstrated the screening potential of continuous, multi-window signal monitoring combined with a verification pathway (18). In addition, wrist-worn PPG autocorrelation methods and reviews of smartphone- or PPG-based screening support the feasibility of early AF screening under certain conditions (20,21). Nevertheless, continuous PPG-based automatic monitoring remains vulnerable to physical activity, wearing conditions, and reduced signal quality (22,23). Therefore, smartwatch alerts generally still require subsequent ECG confirmation and clinical interpretation (24,25).

However, lessons from AF screening should be translated cautiously into occupational health monitoring. AF is a relatively well-defined, paroxysmal arrhythmic event that can be treated as a binary outcome and is supported by established clinical verification and management pathways, such as confirmatory ECG, risk assessment, and anticoagulation decision-making. In contrast, occupational fatigue, stress, and insufficient recovery are not discrete clinical events. They are continuous, context-dependent, multifactorial states with substantial inter-individual variability, and they currently lack standardized confirmation criteria and mature management pathways comparable to those used in AF screening. Thus, the value of AF screening does not lie in its direct transferability as a model for occupational fatigue or stress alerts. Rather, it illustrates that smartwatch-based warnings can become actionable only when embedded in a complete workflow that includes data acquisition, anomaly detection, professional interpretation, human review, and subsequent intervention. For occupational health management, smartwatch-derived metrics should therefore be regarded primarily as tools for risk indication and trend observation. Their interpretive validity, intervention thresholds, and management pathways in specific occupational contexts, such as highway traffic policing, require further validation.

Monitoring sleep and circadian rhythms

Sleep and circadian rhythm are important links between work schedules and health outcomes. This is particularly evident during the early transition from regular daytime work to shift work, when sleep patterns and mental health may be more vulnerable to adverse changes, highlighting the importance of continuous management of sleep, recovery, and circadian rhythm (26,27). Existing research generally agrees that smartwatches offer high practical value for tracking trend indicators such as sleep duration, bedtime, wake-up time, and sleep-wake patterns, and can be used to identify insufficient recovery, circadian misalignment, and circadian rhythm disruption (28). From an occupational health perspective, the value of these indicators lies not in providing detailed diagnoses, but in revealing whether work schedules are disrupting the recovery process. From a metabolic health perspective, persistent sleep and circadian rhythm disruption may also lead to adverse metabolic consequences (29).

However, sleep monitoring among shift workers involves greater methodological complexity. Unlike the relatively regular and nocturnally consolidated sleep patterns observed in the general population, shift workers often experience fragmented sleep, irregular sleep timing, sleep occurring at atypical circadian phases, and daytime recovery sleep or daytime anchor sleep. Most consumer-grade smartwatch sleep algorithms are developed based on nocturnal sleep patterns in the general population, and their validity remains insufficiently tested under conditions of shift work, daytime sleep, split sleep, and circadian misalignment. Harris et al. showed that the early transition to shift work is a period during which changes in sleep and mental health are particularly evident (26). Yet it is precisely in this population that the interpretive boundaries of wearable-derived sleep metrics require careful consideration. Thus, although shift workers are one of the most relevant occupational target groups for sleep monitoring, the evidence regarding the measurement and interpretive validity of wrist-worn devices in this population remains limited.

Sleep staging remains a weak point for smartwatches. Laboratory studies suggest that while wearable devices perform adequately in sleep/wake binary classification, their accuracy in identifying multiple sleep stages—such as light sleep, deep sleep, and rapid eye movement (REM) sleep—remains limited (28,30). Studies using polysomnography (PSG) as a reference have yielded similar results, showing limited accuracy in wakefulness detection and fine-grained sleep staging (28,30,31). In contrast, electroencephalography (EEG)-based wearable devices demonstrate higher accuracy in sleep staging, but comfort, cost, and suitability for specific scenarios still require careful consideration (32). Therefore, in workplace applications, sleep staging results should not be used directly as the basis for diagnosis, performance evaluation, or punitive management; rather, they are better suited as supplementary information for long-term trend management, shift optimization, and recovery support.

Among shift workers, sleep-rhythm data becomes more explanatory when combined with scheduling systems. A review on nurses’ occupational health suggests that wearable data can help compare recovery differences across different shifts (5). Related sleep and wearable reviews further indicate that, when linked with scheduling systems and recovery-oriented management, such data can provide objective support for break scheduling, shift optimization, and fatigue management (26,28). These studies suggest that sleep monitoring should not be limited to a descriptive assessment of how well one sleeps, but should be further understood in conjunction with system design and job management.

Monitoring physical activity and workload

Physical activity metrics serve as a direct link between work tasks and physiological workload. A real-life validation study by Ravanelli et al. showed that wrist-worn devices generally exhibit better consistency in step count and duration of activity than heart rate and HRV (33). Related reviews also indicate that there is substantial evidence supporting the use of activity tracking to promote physical activity, construct workload profiles, and identify behavioral changes (34,35). In occupational settings, such metrics are particularly useful for identifying consecutive days of high workload, sudden drops in activity, and workload patterns that deviate from an individual’s baseline. However, existing evidence on activity tracking primarily supports its value in behavioral change and workload characterization; it does not imply that it can directly replace the assessment of occupational fatigue. For workplace management, these metrics are better suited as part of a multi-indicator fusion analysis to provide clues for identifying workload changes; their managerial applications still require validation in conjunction with specific job contexts, review processes, and organizational implementation conditions.

However, the interpretation of activity metrics is equally dependent on the work context. For manual laborers, high step counts and high activity intensity may indicate overexertion rather than good health; for office workers, a sudden drop in activity may signal fatigue, pain, emotional issues, or impaired work performance. Reviews on occupational fatigue suggest that single activity metrics are insufficient for accurately capturing occupational fatigue (36). A relatively more robust approach involves conducting a combined analysis of activity, heart rate, sleep, work shifts, and environmental exposures to generate a more interpretive load-recovery profile (37).

In hot environments and physically demanding work scenarios, activity monitoring requires even greater caution (38). Heat strain and physical workload are not determined solely by individual activity levels; they are also closely related to ambient temperature, humidity, radiant heat, air movement, metabolic heat production, work intensity, hydration status, heat acclimatization, protective equipment, and work-rest arrangements. Step count, activity duration, and heart rate changes recorded by wrist-worn devices may reflect part of the external workload and physiological response, but they are insufficient on their own to determine heat-stress risk, dehydration status, or fatigue level. In particular, under conditions of heat, high humidity, direct sunlight, or protective equipment that limits heat dissipation, the same amount of activity may correspond to very different physiological burdens. Therefore, relying solely on a smartwatch to generate simple high-risk or low-risk labels may lead to misclassification.

This section does not aim to provide a systematic review of occupational heat physiology. Rather, it discusses the auxiliary value and limitations of smartwatches in hot and physically demanding occupational settings from the perspective of their boundaries in occupational health monitoring. Current evidence more strongly supports the integration of wrist-worn heart rate and activity data with environmental parameters, task rhythm, job type, individual baseline status, and work-rest schedules for platform-level analysis and stratified response recommendations. For example, when sustained high activity is combined with heat exposure, persistently elevated heart rate, and insufficient rest, the system may prompt increased hydration, task rotation, additional rest breaks, or human review. Such data should not be used by the device alone to make occupational risk judgments. For occupational groups such as highway traffic police officers, firefighters, outdoor inspection workers, sanitation workers, and construction workers, future studies should further validate the accuracy, interpretability, and managerial feasibility of multisource data fusion models in real-world work settings.

Current applications of smartwatches in identifying and intervening in mental health risks among working populations

Monitoring work-related pressure and stress

The monitoring of psychological stress is currently the most actively researched area, yet it is also the most prone to overinterpretation. Relevant systematic reviews indicate that while some studies achieve high classification accuracy in specific task scenarios, there are significant differences across studies in terms of stress label definitions, data collection duration, algorithmic features, and evaluation methods, resulting in limited comparability (39). Some studies have attempted to use wearable digital biomarkers to predict stress, anxiety, and positive and negative emotions, but model performance remains highly dependent on the time window and specific data context (40). From the perspective of occupational health management, the critical issue is not whether the model can distinguish experimental stress tasks, but whether it can consistently reflect workers’ subjective experiences and functional status in real-world work environments.

Martinez et al. conducted an 8-week ecological momentary assessment study involving 657 information workers and found that when using only wearable HRV features to predict subjective stress, the model’s explanatory power was approximately 1%; even when restricted to the 8:00 a.m.–6:00 p.m. work hours, the explanatory power was only about 2% (16). This finding offers an important insight: the fact that autonomic nervous system changes can be detected does not equate to the ability to accurately determine the level of stress an individual is experiencing. Therefore, the value of smartwatches in stress assessment is better positioned as a supportive indicator rather than a standalone tool for determining high stress levels.

A more prudent approach is to ground stress interpretation in the integration of multimodal physiological signals and contextual information: by combining HRV fluctuations, sleep disturbances, and activity abnormalities with shift schedules, task load, subjective self-assessments, and organizational environment, we can identify the risk phenotype of persistently elevated load and insufficient recovery (41). Only under these conditions, and after completing scenario validation, incorporating review procedures, and establishing implementation processes, can stress monitoring transition from algorithmic identification to management support.

Monitoring anxiety, work function limitations, and inadequate recovery

In addition to stress, anxiety tendencies, work function limitations, and chronic inadequate recovery are also outcomes worthy of attention in occupational health management. Okawara et al. conducted a two-month longitudinal follow-up study among office workers, asking participants to complete a 30-second HRV measurement every morning upon waking. The study used high trait anxiety and work function limitations as endpoints. The results showed that both the mean HRV and intraday variability possessed a certain degree of discriminatory power and were associated with multiple dimensions of the Work Function Limitation Scale (42). This finding supports the view that longitudinal variability information is more valuable than single-point absolute values.

However, the practical implications of such findings should still be interpreted with caution. First, the specificity and stability of the relevant models are not yet sufficient to support human resources management or job screening functions; second, mental health outcomes are influenced by a variety of factors, including job demands, social support, individual personality, and life events, and a single wearable indicator can only account for a portion of the variation (43). Therefore, smartwatches are more suitable for early support and risk alerts, rather than as a basis for labeling workers.

From a broader occupational health perspective, anxiety, impaired work functioning, and inadequate recovery are often interrelated. By integrating HRV trends, sleep disturbances, and activity abnormalities with questionnaire data, shift schedules, and subjective assessments of work performance, smartwatches can provide clues for the early identification of chronic overwork and highlight their potential value in providing ongoing support (42).

Burnout, resilience, and behavioral interventions

The value of mental health monitoring depends not only on detection, but also on whether it can support meaningful feedback, intervention, and outcome improvement. In a randomized clinical trial among physicians, Dyrbye et al. found that the intervention group, who wore a smartwatch and received feedback on sleep, steps, and heart rate, showed a lower risk of burnout and higher resilience scores at 6-month follow-up (44). These findings suggest that smartwatches may serve as auxiliary tools to support self-awareness and behavioral adjustment. However, their effectiveness depends on whether monitoring data can be translated into feedback that individuals can understand, trust, and use in practice, and whether such feedback can be linked to psychological support, organizational support, and follow-up services. For high-stress occupational groups such as police officers and emergency responders, wearable devices should not be regarded as stand-alone mental health interventions. Rather, they should be embedded within a broader occupational mental health support system, with further research needed to clarify their independent contribution, applicable conditions, and long-term effects.

At the same time, reviews on the relationship between burnout and well-being suggest that the use of wearable devices is increasing in high-stress professions such as healthcare; however, there are significant differences among studies in terms of study populations, intervention methods, follow-up durations, and outcome measures, and their long-term effects and generalizability remain to be verified (6). There may be an interrelationship among sleep deprivation, declining perceived health, and the risk of clinical errors among working populations, suggesting that monitoring and feedback designs should also focus on actual work outcomes (45). Excessive reminders, synchronization burden, and unclear explanations of data significance may all undermine the intended positive effects. The evidence cited above primarily comes from physicians and healthcare-related settings; its applicability to populations in other workplaces requires careful evaluation.

Therefore, one area for further exploration is to integrate smartwatches with real-time adaptive interventions—such as providing rest recommendations, breathing exercises, activity adjustments, or advice to seek medical attention following insufficient sleep, reduced recovery, or prolonged high-intensity exertion—while enhancing intervention adherence through concise, low-intrusiveness feedback mechanisms (46). Compared to simply pursuing detection accuracy, this approach aligns more closely with the needs of occupational health management.

Challenges and solutions for smartwatch-based occupational health monitoring and risk management in the workplace

Measurement errors and adaptation of device algorithms to workplace conditions

A key difference between workplace applications and general health management scenarios lies in the complexity of workplace conditions and the abundance of interfering factors. Factors such as physical activity, vibration, perspiration, changes in device fit, skin condition, and ambient temperature and humidity can all affect the quality of PPG signals, making high-activity segments a primary source of errors and false positives (23). This is also a key reason why many metrics that perform well in laboratory settings experience a decline in performance once applied to real-world workplace environments. At the same time, issues related to the alignment of device selection, sampling rate, and data quality with the specific usage context can also significantly impact the interpretation of results (47).

In terms of countermeasures, first, data quality control rules and wearing guidelines should be established for the target positions, clearly defining which metrics can be used under what operating conditions; second, activity gating, confidence evaluation, and anomaly segment removal mechanisms should be introduced at the algorithmic level; third, in metric selection, trend-based and long-window metrics that are more robust for PPG should be prioritized, with interpretation primarily based on deviations from individual baselines rather than directly applying uniform thresholds (48). For high-risk positions, consideration should also be given to incorporating environmental sensors and task information into the model to collectively enhance interpretability.

Long-term usage rates, worker acceptance, and organizational feasibility

Whether smartwatches can truly deliver value depends on whether they can be used consistently over the long term. Existing research and reviews have repeatedly highlighted that long-term adherence can be affected by usability and implementation issues such as charging demands, synchronization burden, wearing comfort, workflow fit, and notification burden (1,47). In certain job roles, workers may even voluntarily stop using the devices out of concern that they will interfere with their work or create an additional burden.

Therefore, when implementing these systems at scale, evaluation should not focus solely on algorithm accuracy; it should also incorporate factors such as wear time, data completeness, loss-to-follow-up rates, user acceptance, and usage costs into the evaluation framework. In terms of design, the principle of minimizing burden should be upheld by reducing charging and synchronization frequency as much as possible, minimizing manual data entry, and limiting the number of reminders. Additionally, participatory design involving workers should be employed to enhance the system’s adaptability and acceptability (49). From an implementation perspective, successful deployment also depends on clear objectives, workflow integration, and stakeholder collaboration (49). Only when the monitoring system is sufficiently lightweight and beneficial to workers will there be a realistic basis for sustained use.

From risk alerts to a closed-loop management process

The key to occupational health risk management lies not merely in identifying anomalies, but in establishing a closed-loop process of “alert—verification—intervention—evaluation”. Studies on AF screening have clearly demonstrated that the effectiveness of early warning systems is highly dependent on the design of trigger rules, verification pathways, and subsequent medical or management processes (17,18,22,25). This logic offers valuable insights for addressing issues such as fatigue, insufficient recovery, abnormal physiological fluctuations, and sleep deterioration in the workplace; however, its specific applicability across different work settings requires further validation.

Therefore, in the workplace, smartwatches should not be viewed as automatic detectors, but rather as risk indicators. Once an alert is triggered, verification and response must be conducted by considering the job context, self-assessment scales, further testing when necessary, and organizational support pathways. Subsequent evaluations should focus on metrics such as lead time, false alarm costs, alert fatigue, compliance with interventions, and improvements in outcomes (48). Without these components, even the most advanced technology will struggle to translate into a sustainable occupational health management program.

Data governance, boundaries of use, and ethical standards

Health data in occupational settings is highly sensitive. If the boundaries of data use are unclear, workers may easily interpret continuous monitoring as surveillance rather than support, leading to resistance, avoidance of wearing devices, or even behavioral distortion (43). This form of monitoring alienation may undermine the trust required for technological implementation and threaten the long-term adoption of workplace wearable programs (8,43).

European regulatory frameworks provide a more concrete reference point for data governance in occupational wearable programs. The General Data Protection Regulation (GDPR) classifies health data as a special category of personal data and requires its processing to comply with principles such as lawfulness, transparency, purpose limitation, data minimization, and security protection (50). In employment relationships, because workers’ consent may be affected by power imbalances, continuous monitoring programs should pay particular attention to data protection impact assessments, opt-out mechanisms, and strict limitations on data use. The EU Artificial Intelligence Act further classifies certain AI systems used in employment and worker management as high-risk systems, particularly those involving task allocation, performance assessment, or the monitoring and evaluation of worker behavior (51). Therefore, once smartwatch-derived data are used for shift scheduling, job reassignment, performance appraisal, or managerial decision-making, the associated legal and governance requirements differ substantially from those of voluntary health promotion programs. The sustainable implementation of occupational wearable systems must therefore be grounded in clearly defined boundaries of use, transparent communication, human oversight, audit trails, and the protection of workers’ rights.

Future initiatives should prioritize data governance on par with algorithmic performance: clearly define the boundaries of use to focus on health promotion rather than punitive management; adhere to the principles of data minimization and tiered authorization; establish audit trails, informed consent, and opt-out mechanisms; and avoid using psychological stress, sleep, or HRV results directly for performance evaluations or job screening (43). From a longer-term perspective, the value of smartwatches lies not in creating new tools of control, but in providing foundational support for the development of a digital occupational health support system tailored to frontline workers.

Strengths and limitations of this review

This review integrates evidence across physiological monitoring, mental-health-related risk identification, implementation feasibility, and governance issues, thereby linking smartwatch performance with the practical requirements of occupational health risk management. Its main strength lies in the thematic synthesis of representative evidence relevant to workplace monitoring, implementation, and governance.

This review has several limitations. As a narrative review, it did not aim to exhaustively include all relevant studies, nor did it conduct a formal risk-of-bias assessment. More importantly, a substantial proportion of the available evidence comes from general populations, consumer health settings, clinical screening studies, or controlled laboratory environments, whereas studies directly based on real occupational cohorts and complex working conditions remain limited. Therefore, our conclusions regarding indicators such as heart rate trends, activity levels, sleep duration, and sleep rhythm should be interpreted as cautious inferences about their potential occupational applicability, rather than direct evidence that they already have generalizable warning value. These conclusions still require further validation in specific occupational populations, shift arrangements, task demands, and environmental exposure conditions.

Conclusions

Overall, smartwatches show potential for occupational health monitoring and workplace risk management. At present, however, they are more appropriately regarded as auxiliary tools for monitoring and risk-signal identification rather than mature automated early-warning systems. Existing evidence suggests that indicators such as heart rate trends, activity levels, sleep duration, and sleep rhythm are relatively suitable for continuous data collection and long-term trend observation. Nevertheless, much of this evidence comes from general populations, consumer health settings, clinical validation studies, or studies in selected occupational groups, and its applicability across different occupational cohorts requires further validation.

The indicators identified in this review as being closer to occupational application should therefore be viewed as candidate indicators for priority validation in workplace settings, rather than as mature tools ready for general deployment. By contrast, stress detection, interpretation of ultra-short-term HRV, multi-parameter identification of occupational fatigue risk, and device-generated risk labels still require more rigorous measurement validation, context-specific interpretation, and support from prospective occupational cohort studies. In complex work environments, such as high workload, shift work, heat exposure, open-road duty, or emergency response settings, smartwatch-derived data must be interpreted in conjunction with shift schedules, task demands, environmental exposures, individual baselines, and organizational response processes.

Future research could focus on four key directions: First, conduct metrological validation in real-world job settings, actual working conditions, and across different devices to establish a system for signal quality control and individual baseline modeling; Second, promote the multimodal fusion of sleep, activity, heart rate/HRV, and contextual information such as shifts, tasks, and environmental exposures to enhance interpretability and actionable insights; Third, establish an evaluation system that balances lead time, false-positive costs, and compliance, guided by outcomes such as fatigue, work function limitations, errors, accidents, and medical visits; fourth, concurrently refine the framework for scope of use, worker participation, privacy protection, and platform governance at the organizational implementation level. Only in this way will smartwatches be more likely to evolve from measuring devices into a vital component of digital support systems for occupational health promotion and risk management.

Acknowledgments

The authors thank all colleagues who provided assistance during the development and revision of this manuscript.

Footnote

Reporting Checklist: The authors have completed the Narrative Review reporting checklist. Available at https://jphe.amegroups.com/article/view/10.21037/jphe-2026-0028/rc

Peer Review File: Available at https://jphe.amegroups.com/article/view/10.21037/jphe-2026-0028/prf

Funding: This study was supported by the Natural Science Foundation Project of Hunan Province (grant No. 2025JJ80776) and the Natural Science Foundation Project of Changsha City (grant No. kq2502280).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jphe.amegroups.com/article/view/10.21037/jphe-2026-0028/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Triantafyllidis A, Kondylakis H, Katehakis D, et al. Smartwatch interventions in healthcare: A systematic review of the literature. Int J Med Inform 2024;190:105560. [Crossref] [PubMed]
2. Huhn S, Axt M, Gunga HC, et al. The Impact of Wearable Technologies in Health Research: Scoping Review. JMIR Mhealth Uhealth 2022;10:e34384. [Crossref] [PubMed]
3. Zhang LF, Gao GW. Application and research progress of TENG in wearable human body health status monitoring. Transducer and Microsystem Technologies 2024;43:1-5.
4. Svertoka E, Saafi S, Rusu-Casandra A, et al. Wearables for Industrial Work Safety: A Survey. Sensors (Basel) 2021;21:3844. [Crossref] [PubMed]
5. Chen XW, Huang HL, Hong XP, et al. Advances in the application of wearable devices in occupational health monitoring for nurses. Chinese Journal of Nursing 2025;60:1277-81.
6. Barac M, Scaletty S, Hassett LC, et al. Wearable Technologies for Detecting Burnout and Well-Being in Health Care Professionals: Scoping Review. J Med Internet Res 2024;26:e50253. [Crossref] [PubMed]
7. Maltseva K. Wearables in the workplace: the brave new world of employee engagement. Bus Horiz 2020;63:493-505.
8. Patel V, Chesmore A, Legner CM, et al. Trends in workplace wearable technologies and connected-worker solutions for next-generation occupational safety, health, and productivity. Adv Intell Syst 2022;4:2100099.
9. Nelson BW, Low CA, Jacobson N, et al. Guidelines for wrist-worn consumer wearable assessment of heart rate in biobehavioral research. NPJ Digit Med 2020;3:90. [Crossref] [PubMed]
10. Fuller D, Colwell E, Low J, et al. Reliability and Validity of Commercially Available Wearable Devices for Measuring Steps, Energy Expenditure, and Heart Rate: Systematic Review. JMIR Mhealth Uhealth 2020;8:e18694. [Crossref] [PubMed]
11. Miller DJ, Sargent C, Roach GD. A Validation of Six Wearable Devices for Estimating Sleep, Heart Rate and Heart Rate Variability in Healthy Adults. Sensors (Basel) 2022;22:6317. [Crossref] [PubMed]
12. Li K, Cardoso C, Moctezuma-Ramirez A, et al. Heart Rate Variability Measurement through a Smart Wearable Device: Another Breakthrough for Personal Health Monitoring?. Int J Environ Res Public Health 2023;20:7146. [Crossref] [PubMed]
13. O'Grady B, Lambe R, Baldwin M, et al. The Validity of Apple Watch Series 9 and Ultra 2 for Serial Measurements of Heart Rate Variability and Resting Heart Rate. Sensors (Basel) 2024;24:6220. [Crossref] [PubMed]
14. Alugubelli N, Abuissa H, Roka A. Wearable Devices for Remote Monitoring of Heart Rate and Heart Rate Variability-What We Know and What Is Coming. Sensors (Basel) 2022;22:8903. [Crossref] [PubMed]
15. Theurl F, Schreinlechner M, Sappler N, et al. Smartwatch-derived heart rate variability: a head-to-head comparison with the gold standard in cardiovascular disease. Eur Heart J Digit Health 2023;4:155-64. [Crossref] [PubMed]
16. Martinez GJ, Grover T, Mattingly SM, et al. Alignment Between Heart Rate Variability From Fitness Trackers and Perceived Stress: Perspectives From a Large-Scale In Situ Longitudinal Study of Information Workers. JMIR Hum Factors 2022;9:e33754. [Crossref] [PubMed]
17. Chang PC, Wen MS, Chou CC, et al. Atrial fibrillation detection using ambulatory smartwatch photoplethysmography and validation with simultaneous holter recording. Am Heart J 2022;247:55-62. [Crossref] [PubMed]
18. Lubitz SA, Faranesh AZ, Selvaggi C, et al. Detection of Atrial Fibrillation in a Large Population Using Wearable Devices: The Fitbit Heart Study. Circulation 2022;146:1415-24. [Crossref] [PubMed]
19. Tao YR, Liu W. Recent advances in artificial intelligence-based ECG analysis for atrial fibrillation. Chinese Journal of Cardiac Pacing and Electrophysiology 2026;40:13-7.
20. Väliaho ES, Kuoppa P, Lipponen JA, et al. Wrist Band Photoplethysmography Autocorrelation Analysis Enables Detection of Atrial Fibrillation Without Pulse Detection. Front Physiol 2021;12:654555. [Crossref] [PubMed]
21. Gill S, Bunting KV, Sartini C, et al. Smartphone detection of atrial fibrillation using photoplethysmography: a systematic review and meta-analysis. Heart 2022;108:1600-7. [Crossref] [PubMed]
22. Antiperovitch P, Mortara D, Barrios J, et al. Continuous Atrial Fibrillation Monitoring From Photoplethysmography: Comparison Between Supervised Deep Learning and Heuristic Signal Processing. JACC Clin Electrophysiol 2024;10:334-45. [Crossref] [PubMed]
23. Bent B, Goldstein BA, Kibbe WA, et al. Investigating sources of inaccuracy in wearable optical heart rate sensors. NPJ Digit Med 2020;3:18. [Crossref] [PubMed]
24. Fiorina L, Chemaly P, Cellier J, et al. Artificial intelligence-based electrocardiogram analysis improves atrial arrhythmia detection from a smartwatch electrocardiogram. Eur Heart J Digit Health 2024;5:535-41. [Crossref] [PubMed]
25. Abdelhamid K, Reissenberger P, Piper D, et al. Fully Automated Photoplethysmography-Based Wearable Atrial Fibrillation Screening in a Hospital Setting. Diagnostics (Basel) 2025;15:1233. [Crossref] [PubMed]
26. Harris R, Kavaliotis E, Drummond SPA, et al. Sleep, mental health and physical health in new shift workers transitioning to shift work: Systematic review and meta-analysis. Sleep Med Rev 2024;75:101927. [Crossref] [PubMed]
27. Jaiswal SJ, Pawelek JB, Warshawsky S, et al. Using new technologies and wearables for characterizing sleep in population-based studies. Curr Sleep Med Rep 2024;10:82-92.
28. de Zambotti M, Goldstein C, Cook J, et al. State of the science and recommendations for using wearable technology in sleep and circadian research. Sleep 2024;47:zsad325. [Crossref] [PubMed]
29. Depner CM, Stothard ER, Wright KP Jr. Metabolic consequences of sleep and circadian disorders. Curr Diab Rep 2014;14:507. [Crossref] [PubMed]
30. Birrer V, Elgendi M, Lambercy O, et al. Evaluating reliability in wearable devices for sleep staging. NPJ Digit Med 2024;7:74. [Crossref] [PubMed]
31. Huang ZH, Yao FL, Zhang FJ, et al. Analysis of the consistency and accuracy of sleep monitoring using smartwatches. Chinese Journal of Medical Engineering 2025;33:1-7.
32. de Gans CJ, Burger P, van den Ende ES, et al. Sleep assessment using EEG-based wearables - A systematic review. Sleep Med Rev 2024;76:101951. [Crossref] [PubMed]
33. Ravanelli N, Lefebvre K, Brough A, et al. Validation of an Open-Source Smartwatch for Continuous Monitoring of Physical Activity and Heart Rate in Adults. Sensors (Basel) 2025;25:2926. [Crossref] [PubMed]
34. Brickwood KJ, Watson G, O'Brien J, et al. Consumer-Based Wearable Activity Trackers Increase Physical Activity Participation: Systematic Review and Meta-Analysis. JMIR Mhealth Uhealth 2019;7:e11819. [Crossref] [PubMed]
35. Ferguson T, Olds T, Curtis R, et al. Effectiveness of wearable activity trackers to increase physical activity and improve health: a systematic review of systematic reviews and meta-analyses. Lancet Digit Health 2022;4:e615-26. [Crossref] [PubMed]
36. Moshawrab M, Adda M, Bouzouane A, et al. Smart Wearables for the Detection of Occupational Physical Fatigue: A Literature Review. Sensors (Basel) 2022;22:7472. [Crossref] [PubMed]
37. Adão Martins NR, Annaheim S, Spengler CM, et al. Fatigue Monitoring Through Wearables: A State-of-the-Art Review. Front Physiol 2021;12:790292. [Crossref] [PubMed]
38. Moon J, Ju BK. Wearable sensors for healthcare of industrial workers: a scoping review. Electronics 2024;13:3849.
39. Morales A, Barbosa M, Morás L, et al. Occupational Stress Monitoring Using Biomarkers and Smartwatches: A Systematic Review. Sensors (Basel) 2022;22:6633. [Crossref] [PubMed]
40. Saylam B, İncel ÖD. Quantifying Digital Biomarkers for Well-Being: Stress, Anxiety, Positive and Negative Affect via Wearable Devices and Their Time-Based Predictions. Sensors (Basel) 2023;23:8987. [Crossref] [PubMed]
41. Pataca AO, Zdravevski E, Coelho PJ, et al. Use of machine learning for predicting stress episodes based on wearable sensor data: A systematic review. Comput Biol Med 2025;198:111166. [Crossref] [PubMed]
42. Okawara H, Shiraishi Y, Sato K, et al. Visually assessing work performance using a smartwatch via day-to-day fluctuations in heart rate variability. Digit Health 2024;10:20552076241239240. [Crossref] [PubMed]
43. Herold M, Simbula S, Gallucci M. Can smartphone applications and wearable technologies improve workplace well-being and help manage stress? A systematic review. Curr Psychol 2024;43:28650-73.
44. Dyrbye LN, West CP, Wilton AR, et al. Smartwatch Use and Physician Well-Being: A Randomized Clinical Trial. JAMA Netw Open 2025;8:e2527275. [Crossref] [PubMed]
45. Trockel MT, Menon NK, Rowe SG, et al. Assessment of Physician Sleep and Wellness, Burnout, and Clinically Significant Medical Errors. JAMA Netw Open 2020;3:e2028111. [Crossref] [PubMed]
46. Fiedler J, Bergmann MR, Sell S, et al. Just-in-Time Adaptive Interventions for Behavior Change in Physiological Health Outcomes and the Use Case for Knee Osteoarthritis: Systematic Review. J Med Internet Res 2024;26:e54119. [Crossref] [PubMed]
47. Roos LG, Slavich GM. Wearable technologies for health research: Opportunities, limitations, and practical and conceptual considerations. Brain Behav Immun 2023;113:444-52. [Crossref] [PubMed]
48. Bayoumy K, Gaber M, Elshafeey A, et al. Smart wearable devices in cardiovascular care: where we are and how to move forward. Nat Rev Cardiol 2021;18:581-99. [Crossref] [PubMed]
49. Smuck M, Odonkor CA, Wilt JK, et al. The emerging clinical role of wearables: factors for successful implementation in healthcare. NPJ Digit Med 2021;4:45. [Crossref] [PubMed]
50. European Parliament and Council. Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation). Official Journal of the European Union 2016;L119:1-88.
51. European Parliament and Council. Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act). Official Journal of the European Union 2024;L2024/1689.