Table 2 Characteristics of studies included in the scoping review
From: Smartwatches in healthcare medicine: assistance and monitoring; a scoping review
Author/century | Study design | Participant/Subject | Aim/Objective | Result/Outcome |
|---|---|---|---|---|
Stress monitoring | ||||
 Muhammad Ali Fauzi's et al. (2022) [1] / Norway | Experimental | 15 participants | Compared three learning strategies for stress detection tasks | The experiment involved using smartwatches to collect physiological data, create classifiers, and monitor the stress levels of hospital staff in real time |
 Morales A et al. (2022) [2] / Brazil | Systematic Review | 38 articles | The aim of the study is to develop a stress surveillance system that can help individuals manage their stress levels and improve their quality of life | The study identified several physicochemical parameters that can be used to monitor stress levels, including heart rate variation, cortisol analysis, skin conductance, body temperature, and blood volume at the wrist. The article concludes that developing a wrist wearable for stress identification using physiological and chemical sensors is challenging but possible and applicable |
 Muhammad Ali Fauzi's et al. (2021) [3] / Norway | Experimental | Seven machine learning methods | This study proposed a method for detecting continuous stress using single classifiers and classifier ensembles | It was found that single classifiers had the best accuracy, LR had the best precision, and NN had the best recall for stress detection |
 Vila G et al. (2019) [4] / France | Experimental | 1604 workers | The study focuses on developing a system that monitors passengers' stress levels during air travel and provides real-time feedback to cabin crew | The study's outcome is a proof-of-concept for a real-time stress monitoring system that can improve passengers' experience during air travel |
 Han HJ et al. (2020) [5] / United State America (USA) | Experimental | 17 participants | The study focuses on developing a system that can monitor physiological and behavioural data, including heart rate, skin conductance, and facial expressions, to estimate stress levels | The outcome of the study is a proof-of-concept for a stress monitoring system that can be used in everyday settings to improve the user's well-being |
 Gjoreski M et al. (2020) [6] / Slovenia | Experimental | 23 participants |  |  |
 Francisco de Arriba Pérez et al. (2018) [7]/ Spain | Experimental | 12 Students | This study proposes a protocol to evaluate solutions based on commercial-off-the-shelf (COTS) wrist wearables to estimate student stress | The protocol was carried out in two phases: an initial laboratory-based stress induction test and a monitoring stage in the classroom while the student performs academic activities |
Movement disorders | ||||
 Syed Mustafa Ali et al. (2021) [8] / United Kingdom (UK) | Longitudinal observational study | Fifty-three individuals | Examined the longitudinal engagement of users of a smartwatch app in people living with MLTC-M, stratifying engagement patterns by age, gender, number of disease domains and question type | This study suggests that people living with MLTC-M can use smartwatches to report multiple symptoms per day and that this data could be integrated into electronic health records to support clinical care |
 Amiri et al. (2017) [9] / USA | Feasibility study | 12 healthy subjects aged 23 to 33 years (165 samples were collected) And two subjects aged 15 and 16 | developed an Internet-of-Things (IoT) framework called WearSense that uses the sensory capabilities of modern smartwatches to detect stereotypic behaviours in children with autism | This study used a smartwatch to record motion data and develop a system to detect and monitor autistic behavioural activities. The system had an accuracy of 96.7% in detecting three autistic actions |
 Juan C. Torrado's et a. (2017)[10] / Spain | Experimental | Two people with autism disorder | This study focused on using smartwatches to help people with autism spectrum disorders with emotion regulation problems | They found that the smartwatch can help individuals with alexithymia and emotional dysregulation control their stress episodes triggered by various stimuli, except for the learning phase of the experiment |
 Stephen A et al. (2019) [11] / USA | Experimental; accelerometer data | They recruited 20 older adults | They tested whether a smartphone or smartwatch could detect whether an older adult was walking with or without an assistive device | They found that smartwatches provided much higher quality data for detecting walkers and canes compared to smartphones and that a second sensor on the hip was required for the user-generated classifiers to make the most accurate predictions |
 Tchuente, Franck et al. (2020) [12]/ Canada | Experimental | They recruited 30 able-bodied people (15 male, 15 female) | In this study, smartwatches are used to classify aggressive movements, which could benefit care providers in settings where people suffer from movement disorders | The kNN and ReliefF combination demonstrated that this smartwatch-based approach is a viable solution for identifying aggressive behavior and could be used by care providers in settings where people suffer from dementia or mental health disorders, where random aggressive behaviors often occur |
Sleep tracking | ||||
 Chen et al. (2021) [13] / China | Experimental | 20 subjects | In this study, ApneaDetector was developed as a smartwatch-based system to detect sleep apnoea | The study found that while the ApneaDetector accurately classified normal and sleep apnoea events, it had problems with more specific categorisations such as OSA, CSA, and hypopnoea |
 Chen et al. (2021) [14] / China | Experimental | 20 patients | A pilot study was conducted to analyze the effectiveness of a smartwatch with seven sensors in screening for Obstructive Sleep Apnea (OSA). The study generated respiratory waveforms and detected sleep–wake states using PPG (photoplethysmography) signals | PPG-based smartwatches were more effective than simultaneous in-lab PSG or HSAT devices when screening suspected cases of OSA |
 Mehrabadi et al. (2020) [15] / USA | Experimental | 45 healthy individuals | This study investigated the accuracy of sleep data from the Oura ring compared to medically approved actigraphy devices in 45 healthy individuals aged between 18 and 55 | The results revealed that the Oura ring was significantly more accurate than the Samsung Gear Sport smartwatch in detecting heart rate, sleep, and activity parameters |
Blood pressures | ||||
 Falter et al. (2022) [16] / Belgium | Prospective, single-arm, cross-sectional study | 40 patients | Consecutive patients scheduled for 24-h ambulatory blood pressure monitoring were recruited from the cardiology outpatient clinic | This study has found that the Samsung Galaxy Watch Active 2 has an anchoring point when calibrating the device, resulting in an overestimation of lower blood pressures (BPs) and an underestimation of higher BPs. This indicates that the smartwatch is not ready for clinical usage due to its systematic bias toward a certain calibration point |
 Mark Tsou et al. (2021) [17] / Taiwan | Experimental | 49 subjects | This study examined the applicability of smartwatches in PM2.5 health assessment | The results indicated that the elevated PM2.5 concentration was significantly associated with G-HR for low-intensity activities and marginally associated with G-HR for moderate- to high-intensity activities |
 Yen et al. (2022) [18] / Taiwan | A randomised controlled trial (RCT) | adults aged 20–65 | This study was conducted as a single-blinded, two-arm study to test the effectiveness of a commercial smartwatch with a BP-monitoring feature | This study demonstrates that using a smartwatch with BP-monitoring features can improve blood pressure and other health parameters, increase awareness of high blood pressure, and modify related risk factors |
Heart disease | ||||
 Bumgarner et al. (2018) [19]/ USA | A prospective, nonrandomised, and adjudicator-blinded study | 100 patients | This was a non-randomized study in that the accuracy of the KB automated algorithm was evaluated by comparing its results to physician-interpreted KB rhythm strips and simultaneous ECGs | Physician interpretation of KB tracings produced by smartwatches demonstrated similar results in testing with 99% sensitivity, 83% specificity, and a K coefficient of 0.83. This suggests that the quality of KB tracings produced by smartwatches is highly reliable and accurate |
 Liao et al. (2022) [20] / Taiwan | Randomised | 116 patients | This study evaluated the accuracy of the AF detection algorithm by obtaining ECG waveforms and PPG signals from patients undergoing AF catheter ablation while considering other arrhythmias' impact on it | Results suggest that using a longer length (25-beat) for analysing PPG data leads to higher accuracy in discriminating AF from SR compared to using only 10 beats |
Covid pandemic | ||||
 Abbasi et al. (2020) [21] / USA |  | 333 participants | In this study, researchers analysed data collected from 333 participants who used the DETECT smartphone app to enter symptoms and test results | The model analysed both symptoms and sensor data, and the results showed that the distinction between positive and negative cases was more accurate |
 Niela-Vilén et al. (2021) [22]/ Finland | A longitudinal cohort study design | 38 pregnant women | This longitudinal cohort study investigated the use of an Internet of Things (IoT)-based system and smartwatch technology for monitoring pregnant women | The findings of this study showed that the pandemic-related restrictions were associated with increased heart rate variability, stress levels, decreased physical activity, and decreased sleep duration |
 Hunter et al. (2022) [23] / Great Britain or UK | Multicenter Observational Study | mean age of 57 years | In the study, Fitbit Charge 3 watches were given to each participant to follow up on their complications after the recovered disease | This study demonstrated that smartwatches can monitor physical activity remotely |
 Mishra T et al. (2020) [24] / USA | Cohort | 5,262 participants | This study explored whether wearable devices could detect COVID-19 at an early, pre-symptomatic stage | It was determined that abnormal physiological events, such as elevated resting heart rate and increased heart rate relative to the number of steps, can be detected using a smartwatch at or near the time of infection |
 Quer et al. (2022) [25] / USA | Cohort | 7298 volunteers | This study hypothesised that there are digital, objective biomarkers of reactogenicity that could be identified by detecting subtle deviations from an individual's normal resting heart rate | They demonstrated that it is possible to detect physiologic manifestations of reactogenicity to COVID-19 vaccination through individual changes in RHR |
 Guan G et al. (2022) [26] / USA | Cohort | 1534 participants | This study uses the Garmin Vivosmart 4 smartwatch to measure heart rate and heart rate variability | This study showed that smartwatches are more accurate than patients' self-reports in predicting post-vaccination physiological conditions |
Safety | ||||
 Lacour P et al. (2020) [27] / Germany | Single case | 148 patients | This prospective observational study evaluated the potential for electromagnetic interference with CIEDs | This study shows that there is no risk of EMIs between the iPhone and CIEDs, but relatively frequent telemetry interferences do occur between the iPhone and the CIEDs |
 Tzeis S et al. (2021) [28] / Greece | Cohort | 171 patients | This prospective, multicenter study was conducted to investigate whether the use of the latest generation smartwatches might interfere with the proper functioning of the CIED | The emission levels of the tested smartwatches and their magnetic chargers were evaluated by measuring low-frequency magnetic fields between 110 and 400 kHz |
 Fischer et al. (2022) [29] / Switzerland | Prospective comparative | 13 different settings | This study evaluated a popular smartwatch for its ability to monitor noise levels in 13 occupational and recreational settings accurately | Smartwatch accuracy was lower in settings with rapid acoustic fluctuations but comparable to the sound level meter across different pressure levels based on SD of LAeq differences and ICC results |
Validations or evaluation | ||||
 Pope et al. (2019) [30] / USA |  | 21 healthy college students | This study examined four popular smartwatches' validity, measurement bias, and precision in assessing EE, average HR, and peak HR during active play | This study suggested that smartwatches should not be used as part of a systems medicine approach to health care |
 Shenglong et al. (2022) [31] / China |  | 20 healthy Chinese participants | This study measured individual VE, VO2 and VCO2 using the Cosmed K5 system for a wide range of metabolic rates | The findings of this study indicate that smartwatches may have moderate validity in estimating energy expenditure for outdoor walking and running |
 Sarhaddi et al. (2022) [32] / Finland | Observational study | 28 healthy individuals | This study evaluated the validity of the Samsung Gear Sport smartwatch in terms of HR and HRV parameters compared with a medical-grade chest ECG monitor in a 24-h continuous free-live setting monitoring | The smartwatch can accurately measure HR and HRV parameters during sleep and awake time and provide acceptable RMSSD, SDNN, LF, and HF |
 Brew B et al. (2022) [33] / Australia | Cross-sectional study | 22 volunteer participants | This study used a threshold-based algorithm programmed for different smartwatches to detect a fall on 22 volunteer participants automatically | This study found that an algorithm programmed in commercially available smartwatches to detect induced falls had an overall sensitivity of 77% and specificity of 99%. The fall detection performance depends on the algorithm used, and the sensitivity ranges from 70 to 100% and the specificity from 80 to 100%, depending on the type of fall. In addition, they showed that the performance of a fall detection algorithm could be strongly dependent on the smartwatch model |
 Auepanwiriyakul et al. (2020) [34] / UK | Combined methods design | 15 healthy volunteers | This study evaluated raw IMU sensor data quality, followed by a trial of the feasibility of wearable health devices in a clinical environment | In this study, wearable IMU technology, such as smartwatches and research-grade IMU Xsens, was cleaner linear acceleration signals and fewer errors than Axivity due to additional magnetometer and strap-down integration technology. This technology is viewed positively by healthcare professionals, according to this study on its acceptability among hospital patients |
 Varghese et al. (2021) [35] / Germany | Prospective study | 400 participants | This study was a prospective study from 2018 to the end of 2021, in which they recruited and measured the hand movements of 400 participants using Apple smartwatches and smartphones. The aim was to distinguish PD from other movement disorders and healthy individuals | Results showed that smartwatches had high agreement with seismological sensor validation in measuring movement subtleties or hand-tremor amplitudes and frequencies more accurately than clinical documentation or human vision |