Wrist-worn Device Research Articles

Validity and Reliability of Wearable Devices during Self-Paced Walking, Jogging and Overground Skipping

Wearable technology can track unusual exercise, providing data for improving fitness. The aim of the study was to determine validity and reliability during walking, jogging, and skipping. Eighteen volunteers completed 5 min self-paced activities interspersed with 5 min rest. Variables and devices were step count (Garmin Instinct), estimated energy expenditure (Garmin Instinct, Polar Vantage M2), and heart rate (Garmin Instinct, Polar Vantage M2, Polar OH1, Polar Verity Sense). Validity measures were mean absolute percent error (MAPE) and Lin’s Concordance (CCC), and reliability were coefficient of variation (CV), and intraclass correlation (ICC). Thresholds were MAPE ≤5%, CCC≥0.90, CV≤10%, ICC≥0.70. Garmin Instinct step count during skipping was not considered valid (MAPE=90.2%, CCC=0.008) or reliable (CV=6%, ICC CI=0.4). Energy expenditure during skipping was not valid or reliable in the Garmin Instinct (MAPE=28%, CCC=0.27; CV=19%, ICC=0.61) or the Polar Vantage M2 (MAPE=19%, CCC=0.57; CV=13%). While the Polar Vantage M2 was reliable for estimated energy expenditure during walking and jogging activities, wrist-worn devices (Garmin Instinct, Polar Vantage M2) were neither valid nor reliable in returning estimated energy expenditure during overground skipping. From a wider perspective, wearable device algorithms for estimating energy expenditure should continue to be refined until they return the same level of accuracy as what is currently observed for heart rate, and to a lesser extent step count. Skipping may be an excellent unusual activity for testing wearable devices.

Vibrotactile Thresholds on the Wrist for Vibrations Normal to the Skin.

In order to facilitate the development of wrist-worn vibrotactile devices, detailed knowledge about how vibrations are perceived by the users is needed. In particular, perceptual thresholds in amplitude are really important. Thresholds have been measured in the literature for other areas of the body, but given the variability reported between areas (shape of the threshold curve, position of maximum sensitivity), thresholds on the wrist can not be inferred from previous measurements and must be measured. The amplitude thresholds for vibrations normal to the skin surface were evaluated on 28 participants, with a three interval forced choice method. They were measured for 7 frequencies that are classical in the literature about vibrotactile perception (25, 40, 80, 160, 250, 320, and 640 Hz). The classical U-shape of the amplitude-threshold curve is observed, with a maximum sensitivity at around 160 Hz, which differs from other body areas, but confirms recent results obtained for vibrations parallel to the skin surface of the same body area. The sensitivity thresholds of vibrotactile signals appear to be in the micrometer range.

Addressing the Chronic Pain-Early Cognitive Decline Comorbidity Among Older Adults: Protocol for the Active Brains Remote Efficacy Trial.

Chronic pain and early cognitive decline, which are costly to treat and highly prevalent among older adults, commonly co-occur, exacerbate one another over time, and can accelerate the development and progression of Alzheimer disease and related dementias. We developed the first mind-body activity program (Active Brains [AB]) tailored to the needs of older adults with chronic pain and early cognitive decline. Results from our previous study strongly supported the feasibility of conducting AB remotely and provided evidence for improvements in outcomes. We are conducting a single-blinded, National Institutes of Health stage-2, randomized clinical trial to establish the efficacy of AB versus a time-matched and dose-matched education control (Health Enhancement Program [HEP]) in improving self-reported and objective outcomes of physical, cognitive, and emotional functions in 260 participants. The methodology described in this paper was informed by the lessons learned from the first year of the trial. Participants are identified and recruited through multidisciplinary clinician-referred individuals (eg, pain psychologists and geriatricians), the Rally Research platform, social media, and community partnerships. Interested participants complete eligibility screening and electronic informed consent. Baseline assessments include self-report, performance-based measures (eg, 6-min walk test) and objective measures (eg, Repeatable Battery for the Assessment of Neuropsychological Status). Participants are mailed a wrist-worn ActiGraph device (ActiGraph LLC) to passively monitor objective function (eg, steps) during the week between the baseline assessment and the beginning of the programs, which they continue to wear throughout the programs. After baseline assessments, participants are randomized to either AB or HEP and complete 8 weekly, remote, group sessions with a Massachusetts General Hospital psychologist. The AB group receives a Fitbit (Fitbit Inc) to help reinforce increased activity. Assessments are repeated after the intervention and at the 6-month follow-up. Coprimary outcomes include multimodal physical function (self-report, performance based, and objective). Secondary outcomes are cognitive function (self-report and objective), emotional function, and pain. We began recruitment in July 2022 and recruited 37 participants across 4 cohorts. Of them, all (n=37, 100%) have completed the baseline assessment, 26 (70%) have completed the posttest assessment, and 9 (24%) are actively enrolled in the intervention (total dropout: n=2, 5%). In the three cohorts (26/37, 70%) that have completed the AB or HEP, 26 (100%) participants completed all 8 group sessions (including minimal makeups), and watch adherence (1937/2072, 93.48%, average across ActiGraph and Fitbit devices) has been excellent. The fourth cohort is ongoing (9/37, 24%), and we plan to complete enrollment by March 2026. We aim to establish the efficacy of the AB program over a time-matched and dose-matched control in a live video-based trial and test the mechanisms through theoretically driven mediators and moderators. Findings will inform the development of a future multisite effectiveness-implementation trial. ClinicalTrials.gov NCT05373745; https://classic.clinicaltrials.gov/ct2/show/NCT05373745. DERR1-10.2196/47319.

Screening for hypertension in adults - the use of tonometric blood pressure monitoring.

Arterial hypertension increases the risk of developing cardiovascular disease. Reliable screening tools for diagnosing hypertension are important to ensure correct risk stratification of subjects. In this study, we aimed to analyse if a wrist-worn device using a tonometric technique for measuring of 24-hour blood pressure could be used to diagnose hypertension and non-dipping. A conventional device using oscillometric measurements was used as golden standard. Secondary aim was to compare the degree of discomfort related to monitoring with the two devices. In 89 subjects with a history of normal blood pressure and naive to ambulatory BP monitoring (ABPM), 24-hour ABPM was measured simultaneously with A&D TM2430 (oscillometric technique) and BPro (tonometric technique). When comparing measurements from the two devices, we found that the tonometric device misclassified 46% of hypertensive subjects and 69% of non-dippers. The tonometric device measured significantly lower systolic 24-hour and daytime blood pressure. The subjects reported less discomfort related to the tonometric than the oscillometric device. Despite less discomfort related to usage of the tonometric device for 24-hour blood pressure monitoring compared to an oscillometric device, misclassification of hypertension and non-dipping makes the tonometric device inappropriate as a screening instrument.

Overnight Cardiac Chronotropy Assessed By Wrist-worn Device On Markers Of Allostatic Load In Marines

Overnight Cardiac Chronotropy Assessed By Wrist-worn Device On Markers Of Allostatic Load In Marines

Prediction of injurious falls in older adults using digital gait biomarkers extracted from large-scale wrist sensor data.

To determine whether digital gait biomarkers captured by a wrist-worn device can predict injurious falls in older people and to develop a multivariable injurious fall prediction model. Population-based longitudinal cohort study. Community-dwelling participants of the UK Biobank study aged 65 and older (n = 32,619) in the United Kingdom. Participants were assessed at baseline on daily-life walking speed, quality, quantity and distribution using wrist-worn accelerometers for up to 7 days. Univariable and multivariable Cox proportional hazard regression models were used to analyse the associations between these parameters and injurious falls for up to 9 years. Five percent of the participants (n = 1,627) experienced at least one fall requiring medical attention over a mean of 7.0 ± 1.1years. Daily-life walking speed, gait quality, quantity of walking and distribution of daily walking were all significantly associated with the incidence of injurious falls (P < 0.05). After adjusting for sociodemographics, lifestyle factors, comorbidities, handgrip strength and reaction time; running duration, total step counts and usual walking speed were identified as independent and significant predictors of falls (P < 0.01). These associations were consistent in those without a history of previous fall injuries. In contrast, step regularity was the only risk factor for those with a previous fall history after adjusting for covariates. Daily-life gait speed, quantity and quality, derived from wrist-worn sensors, are significant predictors of injurious falls in older people. These digital gait biomarkers could potentially be used to identify fall risk in screening programs and integrated into fall prevention strategies.

Detection of Medication Taking Using a Wrist-Worn Commercially Available Wearable Device.

Medication nonadherence is a persistent and costly problem across health care. Measures of medication adherence are ineffective. Methods such as self-report, prescription claims data, or smart pill bottles have been used to monitor medication adherence, but these are subject to recall bias, lack real-time feedback, and are often expensive. We proposed a method for monitoring medication adherence using a commercially available wearable device. Passively collected motion data were analyzed on the basis of the Movelet algorithm, a dictionary learning framework that builds person-specific chapters of movements from short frames of elemental activities within the movements. We adapted and extended the Movelet method to construct a within-patient prediction model that identifies medication-taking behaviors. Using 15 activity features recorded from wrist-worn wearable devices of 10 patients with breast cancer on endocrine therapy, we demonstrated that medication-taking behavior can be predicted in a controlled clinical environment with a median accuracy of 85%. These results in a patient-specific population are exemplar of the potential to measure real-time medication adherence using a wrist-worn commercially available wearable device.

Nurses’ Sleep Characteristics by Shift Type in a Tertiary Hospital With Flexible Working Arrangements

Objectives: To improve the adaptation of shift work, various flexible shift patterns for nurses have introduced in addition to the existing eight-hour-three-shift system. This study aimed to compare the differences in rotating nurses’ sleep characteristics according to shift patterns. Methods: A total of 62 rotating nurses (all females, mean age 29.79±4.30 years) participated in sleep monitoring for consecutive two weeks. Objective sleep was obtained using Readiband (Fatigue Science, Inc., Vancouver, BC, Canada), a wrist-worn device that uses an accelerometer to distinguish sleep and wakefulness. Subjective sleep characteristics were measured using the Morningness-Eveningness questionnaire (MEQ), Bedtime Procrastination Scale (BPS), Epworth Sleepiness Scale (ESS), and Insomnia Severity Index (ISI). Results: Mean BPS was 30.42±5.10, ESS was 10.42±4.27, and ISI was 11.58±4.16, and there were no statistical differences in sleep parameters except for total sleep time (TST). Although significant differences were not found in sleep parameters according to shift patterns, 27.4% of rotating nurses report subjective sleep problems. More than half (59.7%) suffer from excessive daytime sleepiness. ISI was negatively correlated with age (rho=-0.275, &lt;i&gt;p&lt;/i&gt;=0.031) and shift work period (rho=-0.278, &lt;i&gt;p&lt;/i&gt;=0.028), but it was positively correlated with ESS (rho=0.306, &lt;i&gt;p&lt;/i&gt;=0.015). Furthermore, BPS was negatively correlated with MEQ (rho=-0.351, &lt;i&gt;p&lt;/i&gt;=0.005) and TST (rho=-0.307, &lt;i&gt;p&lt;/i&gt;=0.016). Conclusions: To improve the sleep of rotating nurses, more active interventions, such as sleep education to reduce bedtime procrastination and short naps during night shifts, are needed.

Automated, ecologic assessment of frailty using a wrist-worn device

The COVID-19 pandemic has considerably shifted the focus of scientific research, speeding up the process of digitizing medical monitoring. Wearable technology is already widely used in medical research, as it has the potential to monitor the user’s physical activity in daily life. This study aims to explore in-home collected wearable-derived signals for frailty status assessment. A sample of 35 subjects aged 70+, autonomous in basic activities of daily living and cognitively intact, was collected. After being clinically assessed for frailty according to Fried’s phenotype, participants wore a wrist device equipped with inertial motion sensors for 24 h, during which they led their usual life in their homes. Signal-derived traces were split into 10-s segments and labeled classified as gaits, other motor activities, or rests. Gait and other motor activity segments were used to calculate the Subject Activity Level (SAL), an index to quantify how users were active throughout the day. The SAL index was then combined with gait-derived features to design a novel frailty status assessment algorithm. In particular, subjects were classified as robust or non-robust, a category that includes both Fried’s frail and pre-frail phenotypes. For some users, activity levels alone enabled accurate frailty assessment, whereas, for others, a Gaussian Naive Bayes classifier based on the gait-derived features was required to assess frailty status. Overall, the proposed method showed extremely promising results, allowing discrimination of robust and non-robust subjects with an overall 91% accuracy, stemming from 95% sensitivity and 88% specificity. This study demonstrates the potential of unobtrusive, wearable devices in objectively assessing frailty through unsupervised monitoring in real-world settings.

Stress Detection Using Context-Aware Sensor Fusion From Wearable Devices

Wearable medical technology has become increasingly popular in recent years. One function of wearable health devices is stress detection, which relies on sensor inputs to determine a patient’s mental state. This continuous, real-time monitoring can provide healthcare professionals with vital physiological data and enhance the quality of patient care. Current methods of stress detection lack: (i) robustness—wearable health sensors contain high levels of measurement noise that degrades performance, and (ii) adaptation—static architectures fail to adapt to changing contexts in sensing conditions. We propose to address these deficiencies with SELF-CARE, a generalized selective sensor fusion method of stress detection that employs novel techniques of context identification and ensemble machine learning. SELF-CARE uses a learning-based classifier to process sensor features and model the environmental variations in sensing conditions known as the noise context. SELF-CARE uses noise context to selectively fuse different sensor combinations across an ensemble of models to perform robust stress classification. Our findings suggest that for wrist-worn devices, sensors that measure motion are most suitable to understand noise context, while for chest-worn devices, the most suitable sensors are those that detect muscle contraction. We demonstrate SELF-CARE’s state-of-the-art performance on the WESAD dataset. Using wrist-based sensors, SELF-CARE achieves 86.34% and 94.12% accuracy for the 3-class and 2-class stress classification problems, respectively. For chest-based wearable sensors, SELF-CARE achieves 86.19% (3-class) and 93.68% (2-class) classification accuracy. This work demonstrates the benefits of utilizing selective, context-aware sensor fusion in mobile health sensing that can be applied broadly to Internet of Things applications.

Accuracy and Reliability of a Suite of Digital Measures of Walking Generated Using a Wrist-Worn Sensor in Healthy Individuals: Performance Characterization Study.

Mobility is a meaningful aspect of an individual's health whose quantification can provide clinical insights. Wearable sensor technology can quantify walking behaviors (a key aspect of mobility) through continuous passive monitoring. Our objective was to characterize the analytical performance (accuracy and reliability) of a suite of digital measures of walking behaviors as critical aspects in the practical implementation of digital measures into clinical studies. We collected data from a wrist-worn device (the Verily Study Watch) worn for multiple days by a cohort of volunteer participants without a history of gait or walking impairment in a real-world setting. On the basis of step measurements computed in 10-second epochs from sensor data, we generated individual daily aggregates (participant-days) to derive a suite of measures of walking: step count, walking bout duration, number of total walking bouts, number of long walking bouts, number of short walking bouts, peak 30-minute walking cadence, and peak 30-minute walking pace. To characterize the accuracy of the measures, we examined agreement with truth labels generated by a concurrent, ankle-worn, reference device (Modus StepWatch 4) with known low error, calculating the following metrics: intraclass correlation coefficient (ICC), Pearson r coefficient, mean error, and mean absolute error. To characterize the reliability, we developed a novel approach to identify the time to reach a reliable readout (time to reliability) for each measure. This was accomplished by computing mean values over aggregation scopes ranging from 1 to 30 days and analyzing test-retest reliability based on ICCs between adjacent (nonoverlapping) time windows for each measure. In the accuracy characterization, we collected data for a total of 162 participant-days from a testing cohort (n=35 participants; median observation time 5 days). Agreement with the reference device-based readouts in the testing subcohort (n=35) for the 8 measurements under evaluation, as reflected by ICCs, ranged between 0.7 and 0.9; Pearson r values were all greater than 0.75, and all reached statistical significance (P<.001). For the time-to-reliability characterization, we collected data for a total of 15,120 participant-days (overall cohort N=234; median observation time 119 days). All digital measures achieved an ICC between adjacent readouts of >0.75 by 16 days of wear time. We characterized the accuracy and reliability of a suite of digital measures that provides comprehensive information about walking behaviors in real-world settings. These results, which report the level of agreement with high-accuracy reference labels and the time duration required to establish reliable measure readouts, can guide the practical implementation of these measures into clinical studies. Well-characterized tools to quantify walking behaviors in research contexts can provide valuable clinical information about general population cohorts and patients with specific conditions.

Prediction of Incident Depression in Middle-Aged and Older Adults using Digital Gait Biomarkers Extracted from Large-Scale Wrist Sensor Data.

To determine if digital gait biomarkers captured by a wrist-worn device can predict the incidence of depressive episodes in middle-age and older people. Longitudinal cohort study. A total of 72,359 participants recruited in the United Kingdom. Participants were assessed at baseline on gait quantity, speed, intensity, quality, walk length distribution, and walk-related arm movement proportions using wrist-worn accelerometers for up to 7days. Univariable and multivariable Cox proportional-hazard regression models were used to analyze the associations between these parameters and diagnosed incident depressive episodes for up to 9years. A total of 1332 participants (1.8%) had incident depressive episodes over a mean of 7.4 ± 1.1years. All gait variables, except some walk-related arm movement proportions, were significantly associated with the incidence of depressive episodes (P < .05). After adjusting for sociodemographic, lifestyle, and comorbidity covariates; daily running duration, steps per day, and step regularity were identified as independent and significant predictors (P < .001). These associations held consistent in subgroup analysis of older people and individuals with serious medical conditions. The study findings indicate digital gait quality and quantity biomarkers derived from wrist-worn sensors are important predictors of incident depression in middle-aged and older people. These gait biomarkers may facilitate screening programs for at-risk individuals and the early implementation of preventive measures.

An End-to-end Energy-efficient Approach for Intake Detection With Low Inference Time Using Wrist-worn Sensor.

Automated detection of intake gestures with wearable sensors has been a critical area of research for advancing our understanding and ability to intervene in people's eating behavior. Numerous algorithms have been developed and evaluated in terms of accuracy. However, ensuring the system is not only accurate in making predictions but also efficient in doing so is critical for real-world deployment. Despite the growing research on accurate detection of intake gestures using wearables, many of these algorithms are often energy inefficient, impeding on-device deployment for continuous and real-time monitoring of diet. This paper presents a template-based optimized multicenter classifier that enables accurate intake gesture detection while maintaining low-inference time and energy consumption using a wrist-worn accelerometer and gyroscope. We designed an Intake Gesture Counter smartphone application (CountING) and validated the practicality of our algorithm against seven state-of-the-art approaches on three public datasets (In-lab FIC, Clemson, and OREBA). Compared with other methods, we achieved optimal accuracy (81.60% F1 score) and very low inference time (15.97 msec per 2.20-sec data sample) on the Clemson dataset, and among the top performing algorithms, we achieve comparable accuracy (83.0% F1 score compared with 85.6% in the top performing algorithm) but superior inference time (13.8x faster, 33.14 msec per 2.20-sec data sample) on the In-lab FIC dataset and comparable accuracy (83.40% F1 score compared with 88.10% in the top-performing algorithm) but superior inference time (33.9x faster, 16.71 msec inference time per 2.20-sec data sample) on the OREBA dataset. On average, our approach achieved a 25-hour battery lifetime (44% to 52% improvement over state-of-the-art approaches) when tested on a commercial smartwatch for continuous real-time detection. Our approach demonstrates an effective and efficient method, enabling real-time intake gesture detection using wrist-worn devices in longitudinal studies.

Telerehabilitation in patients with recent hospitalisation due to acute decompensated heart failure: protocol for the Tele-ADHF randomised controlled trial

BackgroundCardiac rehabilitation in patients with chronic heart failure (CHF) has favourable effects on exercise capacity, the risk at hospital (re-)admission and quality of life. Although cardiac rehabilitation is generally recommended it is still under-utilised in daily clinical practice, particularly in frail elderly patients after hospital admission, mainly due to low referral and patient-related barriers. Cardiac telerehabilitation (CTR) has the potential to partially solve these barriers. The purpose of this study is to evaluate the effects of CTR as compared to standard remote care after hospital admission on physical functional capacity in CHF patients.MethodsIn this randomised controlled trial, 64 CHF patients will be recruited during hospitalisation for acute decompensated heart failure, and randomised to CTR combined with remote patient management (RPM) or RPM alone (1:1). All participants will start with RPM after hospital discharge for early detection of deterioration, and will be up titrated to optimal medical therapy before being randomised. CTR will start after randomisation and consists of an 18-week multidisciplinary programme with exercise training by physical and occupational therapists, supported by a (remote) technology-assisted dietary intervention and mental health guiding by a physiologist. The training programme consists of three centre-based and two home-based video exercise training sessions followed by weekly video coaching. The mental health and dietary programme are executed using individual and group video sessions. A wrist-worn device enables remote coaching by the physical therapist. The web application is used for promoting self-management by the following modules: 1) goal setting, 2) progress tracking, 3) education, and 4) video and chat communication. The primary outcome measure is physical functional capacity evaluated by the Short Physical Performance Battery (SPPB) score. Secondary outcome measures include frailty scoring, recovery after submaximal exercise, subjective health status, compliance and acceptance to the rehabilitation programme, and readmission rate.DiscussionThe Tele-ADHF trial is the first prospective randomised controlled trial designed for evaluating the effects of a comprehensive combined RPM and CTR programme in recently hospitalised CHF patients. We hypothesize that this intervention has superior effects on physical functional capacity than RPM alone.Trial registrationNetherlands Trial Registry (NTR) NL9619, registered 21 July 2021.

Physical Activity Pattern of Adults With Metabolic Syndrome Risk Factors: Time-Series Cluster Analysis.

Physical activity plays a crucial role in maintaining a healthy lifestyle, and wrist-worn wearables, such as smartwatches and smart bands, have become popular tools for measuring activity levels in daily life. However, studies on physical activity using wearable devices have limitations; for example, these studies often rely on a single device model or use improper clustering methods to analyze the wearable data that are extracted from wearable devices. This study aimed to identify methods suitable for analyzing wearable data and determining daily physical activity patterns. This study also explored the association between these physical activity patterns and health risk factors. People aged >30 years who had metabolic syndrome risk factors and were using their own wrist-worn devices were included in this study. We collected personal health data through a web-based survey and measured physical activity levels using wrist-worn wearables over the course of 1 week. The Time-Series Anytime Density Peak (TADPole) clustering method, which is a novel time-series method proposed recently, was used to identify the physical activity patterns of study participants. Additionally, we defined physical activity pattern groups based on the similarity of physical activity patterns between weekdays and weekends. We used the χ2 or Fisher exact test for categorical variables and the 2-tailed t test for numerical variables to find significant differences between physical activity pattern groups. Logistic regression models were used to analyze the relationship between activity patterns and health risk factors. A total of 47 participants were included in the analysis, generating a total of 329 person-days of data. We identified 2 different types of physical activity patterns (early bird pattern and night owl pattern) for weekdays and weekends. The physical activity levels of early birds were less than that of night owls on both weekdays and weekends. Additionally, participants were categorized into stable and shifting groups based on the similarity of physical activity patterns between weekdays and weekends. The physical activity pattern groups showed significant differences depending on age (P=.004) and daily energy expenditure (P<.001 for weekdays; P=.003 for weekends). Logistic regression analysis revealed a significant association between older age (≥40 y) and shifting physical activity patterns (odds ratio 8.68, 95% CI 1.95-48.85; P=.007). This study overcomes the limitations of previous studies by using various models of wrist-worn wearables and a novel time-series clustering method. Our findings suggested that age significantly influenced physical activity patterns. It also suggests a potential role of the TADPole clustering method in the analysis of large and multidimensional data, such as wearable data.

Photoplethysmography in Wearable Devices: A Comprehensive Review of Technological Advances, Current Challenges, and Future Directions

Photoplethysmography (PPG) is an affordable and straightforward optical technique used to detect changes in blood volume within tissue microvascular beds. PPG technology has found widespread application in commercial medical devices, enabling measurements of oxygen saturation, blood pressure, and cardiac output; the assessment of autonomic nerve function; and the diagnosis of peripheral vascular disease. Recently, the growing demand for non-invasive, portable, cost-effective technology, along with advancements in small semiconductor components, has led to the integration of PPG into various wrist-worn wearable devices. Multiple sensor structures have been proposed and, through appropriate signal processing and algorithmic application, these wearable devices can measure a range of health indicators during daily life. This paper begins by addressing the market status of wrist-worn wearable devices, followed by an explanation of the fundamental principles underlying light operation and its interaction with living tissue for PPG measurements. Moving on to technological advancements, the paper addresses the analog front end for the measurement of the PPG signal, sensor configurations with multiple light emitters and receivers, the minimum sampling rate required for low-power systems, and the measurement of stress, sleep, blood pressure, blood glucose, and activity using PPG signals. Several challenges in the field are also identified, including selecting the appropriate wavelength for the PPG sensor’s light source, developing low-power interpolation methods to extract high-resolution inter-beat intervals at a low sampling rate, and exploring the measurement of physiological phenomena using multi-wavelength PPG signals simultaneously collected at the same location. Lastly, the paper presents future research directions, which encompass the development of new, reliable parameters specific to wearable PPG devices and conducting studies in real-world scenarios, such as 24-h long-term measurements.

1823-PUB: Expanded Dataset Improves Accuracy of Wrist-Worn Noninvasive BGM Using AI

Background and Aims: To determine accuracy, safety and specificity of a novel non-invasive wrist-worn continuous BGM device which analyses resonance shifts in the microwave spectrum using AI. We present results from an ongoing study in patients with T1D and T2D from an expanded dataset (cohorts 1-3). Methods: In this open, pilot, adaptive design study, subjects (N=5/cohort) attended 4 test occasions (n=2/session), each ≤7 days apart. Devices automatically collected data every 60 secs for 500 msec over 3 hours/session, with plasma glucose measured every 5 mins. A global AI model was evaluated by MARD using venous blood glucose. Interim results from every 5 completed patients informed the next device iteration with 10 iterations possible across the study. Results: Data from each cohort was used to train a neural net algorithm to predict a new trial. Each cohort improved overall MARD prediction accuracy. All analyses followed a leave-one-trial-out methodology where the data for the omitted trial was predicted for each analysis cycle. Using the first cohort, the MARD was 21%, adding the second reduced MARD to 15% and with all 3 cohorts the predictive MARD was 13%. Improvement in the SEG plots showed data falling into low risk SEG categories. Conclusions: We have shown that by giving the neural network an expanded dataset, the MARD decreased to nearer commercially available minimally-invasive BGMs by 38% from a predictive MARD of 21% to 13%. Disclosure M.S.Chaudhry: None. G.J.Dunseath: None. J.Ryan: None. L.Barlow: None. I.C.Carrillo masso: Employee; Afon Technology Ltd. J.H.Crane: None. M.R.A.Qureshi: Employee; Afon Technology Ltd. S.C.Bain: Advisory Panel; Novo Nordisk, Sanofi, Lilly, Boehringer Ingelheim Inc., AstraZeneca. S.D.Luzio: None. C.M.Handy: None. B.Love: Consultant; Afon Technologies LTD. N.M.M.Silva: None. L.M.Ferreira: None. K.Wareham: None.

Automatic sleep-stage classification of heart rate and actigraphy data using deep and transfer learning approaches

Manual sleep-stage scoring based on full-night polysomnography data recorded in a sleep lab has been the gold standard of clinical sleep medicine. This costly and time-consuming approach is unfit for long-term studies as well as assessment of sleep on a population level. With the vast amount of physiological data becoming available from wrist-worn devices, deep learning techniques provide an opportunity for fast and reliable automatic sleep-stage classification tasks. However, training a deep neural network requires large annotated sleep databases, which are not available for long-term epidemiological studies. In this paper, we introduce an end-to-end temporal convolutional neural network able to automatically score sleep stages from raw heartbeat RR interval (RRI) and wrist actigraphy data. Moreover, a transfer learning approach enables the training of the network on a large public database (Sleep Heart Health Study, SHHS) and its subsequent application to a much smaller database recorded by a wristband device. The transfer learning significantly shortens training time and improves sleep-scoring accuracy from 68.9% to 73.8% and inter-rater reliability (Cohen’s kappa) from 0.51 to 0.59. We also found that for the SHHS database, automatic sleep-scoring accuracy using deep learning shows a logarithmic relationship with the training size. Although deep learning approaches for automatic sleep scoring are not yet comparable to the inter-rater reliability among sleep technicians, performance is expected to significantly improve in the near future when more large public databases become available. We anticipate those deep learning techniques, when combined with our transfer learning approach, will leverage automatic sleep scoring of physiological data from wearable devices and enable the investigation of sleep in large cohort studies.

Validity of Estimated Results from a Wearable Device for the Tests Time Up and Go and Sit to Stand in Young Adults and in People with Chronic Diseases.

Health care professionals need a valid tool to assess the physical ability of patients with chronic diseases. We aimed to assess the validity of the results of physical fitness tests estimated by a wrist wearable device in young adults and chronic disease people. Participants wore a sensor placed on their wrist and performed two physical fitness tests (sit to stand (STS) and time up and go (TUG)). We checked the concordance of sensor-estimated results using Bland-Altman analysis, root-mean-square error, and intraclass coefficient of correlation (ICC). In total, 31 young adults (groups A; median age = 25 ± 5 years) and 14 people with chronic diseases (groups B; median age = 70 ± 15 years) were included. Concordance was high for both STS (ICCA = 0.95, and ICCB = 0.90), and TUG (ICCA = 0.75, ICCB = 0.98). The best estimations were given by the sensor during STS tests in young adults (mean bias = 0.19 ± 2.69; p = 0.12) and chronic disease people (mean bias = -0.14 ± 3.09 s; p = 0.24). The sensor provided the largest estimation errors over 2 s during the TUG test in young adults. This study showed that the results provided by the sensor are consistent with those of the gold standard during STS and TUG in both healthy youth and people with chronic diseases.

Validation of photoplethysmography- and acceleration-based sleep staging in a community sample: comparison with polysomnography and Actiwatch.

Although wrist-worn consumer wearables are widely used for home sleep monitoring, few have been validated. It is unclear whether consumer wearables could be an alternative to Actiwatch. This study aimed to establish and validate an automatic sleep staging system (ASSS) utilizing photoplethysmography and acceleration data collected from a wrist-worn wearable device. Seventy-five participants from a community population underwent overnight polysomnography (PSG) while wearing a smartwatch (MT2511) and Actiwatch Spectrum Plus (Philips Respironics, Inc; Murrysville, PA, USA). Photoplethysmography and acceleration data collected from the smartwatches were utilized to build a 4-stage (wake, light sleep, deep sleep, and rapid eye movement [REM] sleep) classifier, which was validated against PSG. The performance of the sleep/wake classifier was compared with Actiwatch. All analyses were conducted separately for participants with PSG sleep efficiency (SE) ≥ 80% and SE < 80%. The 4-stage classifier and PSG showed fair overall epoch-by-epoch agreement (kappa, 0.55; 95% confidence interval, 0.52 to 0.57). The deep sleep and REM times were comparable between ASSS and PSG, while ASSS underestimated the wake time and overestimated the light sleep time among participants with SE < 80%. Moreover, ASSS underestimated sleep-onset latency and wake after sleep onset and overestimated total sleep time and SE among participants with SE < 80%, while all were comparable among participants with SE ≥ 80%. The biases were smaller for ASSS than for Actiwatch. Our photoplethysmography- and acceleration-based ASSS was reliable for participants with SE ≥ 80% and had a smaller bias than Actiwatch among those with SE < 80%. Thus, ASSS may be a promising alternative to Actiwatch. Registry: ClinicalTrials.gov; Name: Validation of Sleep Healthcare System; URL: https://clinicaltrials.gov/study/NCT04252482; Identifier: NCT04252482. Liu P-K, Ting N, Chiu H-C, etal. Validation of photoplethysmography- and acceleration-based sleep staging in a community sample: comparison with polysomnography and Actiwatch. J Clin Sleep Med. 2023;19(10):1797-1810.