Step-count outcomes of 13 different activity trackers: Results from laboratory and free-living experiments

Detection of sedentary time and bouts using consumer-grade wrist-worn devices: a hidden semi-Markov model

BACKGROUND

Wrist-worn data from commercially available devices has potential to characterize sedentary time for research and for clinical and public health applications. We propose a model that utilizes heart rate in addition to step count data to estimate the proportion of time spent being sedentary and the usual length of sedentary bouts.

METHODS

We developed and trained two Hidden semi-Markov models, STEPHEN (STEP and Heart ENcoder) and STEPCODE (STEP enCODEr; a steps-only based model) using consumer-grade Fitbit device data from participants under free living conditions, and validated model performance using two external datasets. We used the median absolute percentage error (MDAPE) to measure the accuracy of the proposed models against research-grade activPAL device data as the referent. Bland-Altman plots summarized the individual-level agreement with activPAL.

RESULTS

In OPTIMISE cohort, STEPHEN's estimates of the proportion of time spent sedentary had significantly (p < 0.001) better accuracy (MDAPE [IQR] = 0.15 [0.06-0.25] vs. 0.23 [0.13-0.53)]) and agreement (Bias Mean [SD]=-0.03[0.11] vs. 0.14 [0.11]) than the proprietary software, estimated the usual sedentary bout duration more accurately (MDAPE[IQR] = 0.11[0.06-0.26] vs. 0.42[0.32-0.48]), and had better agreement (Bias Mean [SD] = 3.91[5.67] minutes vs. -11.93[5.07] minutes). With the ALLO-Active dataset, STEPHEN and STEPCODE did not improve the estimation of proportion of time spent sedentary, but STEPHEN estimated usual sedentary bout duration more accurately than the proprietary software (MDAPE[IQR] = 0.19[0.03-0.25] vs. 0.36[0.15-0.48]) and had smaller bias (Bias Mean[SD] = 0.70[8.89] minutes vs. -11.35[9.17] minutes).

CONCLUSIONS

STEPHEN can characterize the proportion of time spent being sedentary and usual sedentary bout length. The methodology is available as an open access R package available from https://github.com/limfuxing/stephen/ . The package includes trained models, but users have the flexibility to train their own models.

Comparing Cadence vs. Machine Learning Based Physical Activity Intensity Classifications: Variations in the Associations of Physical Activity With Mortality

Step cadence-based and machine-learning (ML) methods have been used to classify physical activity (PA) intensity in health-related research. This study examined the association of intensity-specific PA duration with all-cause (ACM) and CVD mortality using the cadence-based and ML methods in 68 561 UK Biobank participants wearing wrist-worn accelerometers. The two-stage-ML method categorized activity type and then intensity. The one-level-cadence-method (1LC) derived intensity-specific duration using all detected steps (including standing utilitarian steps) and cadence thresholds of ≥100 steps/min (moderate intensity) and ≥130 steps/min (vigorous intensity). The two-level-cadence-method (2LC) detected ambulatory steps (i.e., walking and running) and then applied the same cadence thresholds. The 2LC exhibited the most pronounced association at the lower end of duration spectrum. For example, the 2LC showed the smallest minimum moderate-to-vigorous-PA (MVPA) duration (amount associated with 50% of optimal risk reduction) with similar corresponding ACM hazard ratio (HR) to other methods (2LC: 2.8 min/day [95% CI: 2.6, 2.8], HR: 0.83 [95% CI: 0.78, 0.88]; 1LC, 11.1[10.8, 11.4], 0.80 [0.76, 0.85]; ML, 14.9 [14.6, 15.2], 0.82 [0.76, 0.87]). The ML elicited the greatest mortality risk reduction. For example, the medians and corresponding HR in VPA-ACM association: 2LC, 2.0 min/day [95% CI: 2.0, 2.0], HR, 0.69 [95% CI: 0.61, 0.79]; 1LC, 6.9 [6.9, 7.0], 0.68 [0.60, 0.77]; ML, 3.2 [3.2, 3.2], 0.53 [0.44, 0.64]. After standardizing durations, the ML exhibited the most pronounced associations. For example, the standardized minimum durations in MPA-CVD mortality association were: 2LC, -0.77; 1LC, -0.85; ML, -0.94; with corresponding HR of 0.82 [0.72, 0.92], 0.79 [0.69, 0.90], and 0.77 [0.69, 0.85], respectively. The 2LC exhibited the most pronounced association with all-cause and CVD mortality at the lower end of the duration spectrum. The ML method provided the most pronounced association with all-cause and CVD mortality, thus might be appropriate for estimating health benefits of moderate and vigorous intensity PA in observational studies.

Comparison of step-count outcomes across seven different activity trackers: a free-living experiment with young and older adults

BACKGROUND

There are now many different types of activity trackers, including pedometers and accelerometers, to estimate step counts per day. Previous research has extensively examined step-count measurements using activity trackers across various settings while simultaneously wearing different devices.; however, older adults frequently display distinct walking patterns and gait speeds compared to younger adults. This study aimed to compare the step-count between older and younger adults by having them simultaneously wear seven different activity trackers in free-living experiments.

METHODS

This study included 35 younger adults (21-43 yrs) and 57 physically independent older adults (65-91 yrs). All participants simultaneously wore one pedometer and six activity trackers: ActiGraph GT3X + Wrist and Hip, Omron Active Style Pro HJA-350IT, Panasonic Actimarker, TANITA EZ-064, Yamasa TH-300, and Yamasa AS-200 for seven days. A regression equation was also used to assess inter-device compatibility.

RESULTS

When comparing wrist-worn ActiGraph to the six hip-worn activity trackers, the wrist-worn ActiGraph consistently recorded step counts over 4,000 steps higher than hip-worn activity trackers in both groups (range, 3000-5000 steps). Moreover, when comparing the ActiGraph worn on the wrist to that worn on the hip, the proportion was higher among older adults compared to younger ones (younger: 131%, older: 180%). The Actimarker recorded the highest average step counts among six hip-worn devices, with 8,569 ± 4,881 overall, 9,624 ± 5,177 for younger adults, and 7,890 ± 4,562 for older adults. The difference between the hip-worn ActiGraph and Active Style Pro was just about 70 steps/day overall. The correlation among all devices demonstrated a very high consistency, except for the wrist-worn ActiGraph (r = 0.874-0.978).

CONCLUSIONS

Step counts recorded from seven selected consumer-based and research-grade activity trackers and one pedometer, except for the wrist-worn ActiGraph. showed a variation of approximately 1700 steps (range, 1265-2275 steps) steps for both groups, yet maintained a high correlation with each other. These findings will be valuable for researchers and clinicians as they compare step counts across different studies or representative surveys conducted globally.

Objective Assessment of Physical Activity at Home Using a Novel Floor-Vibration Monitoring System: Validation and Comparison With Wearable Activity Trackers and Indirect Calorimetry Measurements

BACKGROUND

The self-monitoring of physical activity is an effective strategy for promoting active lifestyles. However, accurately assessing physical activity remains challenging in certain situations. This study evaluates a novel floor-vibration monitoring system to quantify housework-related physical activity.

OBJECTIVE

This study aims to assess the validity of step-count and physical behavior intensity predictions of a novel floor-vibration monitoring system in comparison with the actual number of steps and indirect calorimetry measurements. The accuracy of the predictions is also compared with that of research-grade devices (ActiGraph GT9X).

METHODS

The Ocha-House, located in Tokyo, serves as an independent experimental facility equipped with high-sensitivity accelerometers installed on the floor to monitor vibrations. Dedicated data processing software was developed to analyze floor-vibration signals and calculate 3 quantitative indices: floor-vibration quantity, step count, and moving distance. In total, 10 participants performed 4 different housework-related activities, wearing ActiGraph GT9X monitors on both the waist and wrist for 6 minutes each. Concurrently, floor-vibration data were collected, and the energy expenditure was measured using the Douglas bag method to determine the actual intensity of activities.

RESULTS

Significant correlations (P<.001) were found between the quantity of floor vibrations, the estimated step count, the estimated moving distance, and the actual activity intensities. The step-count parameter extracted from the floor-vibration signal emerged as the most robust predictor (r2=0.82; P<.001). Multiple regression models incorporating several floor-vibration-extracted parameters showed a strong association with actual activity intensities (r2=0.88; P<.001). Both the step-count and intensity predictions made by the floor-vibration monitoring system exhibited greater accuracy than those of the ActiGraph monitor.

CONCLUSIONS

Floor-vibration monitoring systems seem able to produce valid quantitative assessments of physical activity for selected housework-related activities. In the future, connected smart home systems that integrate this type of technology could be used to perform continuous and accurate evaluations of physical behaviors throughout the day.

Questionnaire-based scoring system for screening moderate-to-vigorous physical activity in middle-aged Japanese workers

OBJECTIVES

Currently available questionnaires have limited ability to measure physical activity (PA) using accelerometers as a gold standard. This study aimed to develop a PA questionnaire for middle-aged Japanese workers and propose a PA scoring system for predicting low moderate-to-vigorous PA (MVPA).

METHODS

A total of 428 participants (median age 49 years; 75.8% men) participated in a 7-day PA measurement using an accelerometer and a questionnaire. The association between questionnaire responses and low MVPA (<150 min/wk) was assessed by logistic regression analysis. A score was assigned to each response based on the correlation coefficients of the multivariate model. The ability of the sum score to predict low MVPA was assessed using the area under the receiver operating characteristic curve (AUC).

RESULTS

Five questionnaire items were used for measuring PA scores (range: 0-50; higher scores indicated a higher probability of low MVPA). The AUC was 0.741 (95% CI, 0.689-0.792), and the sensitivity and specificity at the optimal cut-off value were 66.7% and 68.2%, respectively. This predictive ability was slightly increased by body mass index (AUC 0.745 [95% CI, 0.693-0.796]; sensitivity 69.9%; specificity 66.9%). These predictive values were greater than those of conventional questionnaires used in health checkups in Japan (P < .05).

CONCLUSIONS

This questionnaire-based PA scoring system showed moderate accuracy in predicting low MVPA. It is useful for screening physically inactive workers and promoting PA.

Digital measurement of mobility in pulmonary arterial hypertension: A structured review of an emerging area

This review examined literature that has examined mobility in pulmonary arterial hypertension (PAH) using digital technology. Specifically, the review focussed on: (a) digital mobility measurement in PAH; (b) commonly reported mobility outcomes in PAH; (c) PAH specific impact on mobility outcomes; and (d) recommendations concerning protocols for mobility measurement in PAH. PubMed, Scopus, and Medline databases were searched. Two independent reviewers screened articles that described objective measurement of mobility in PAH using digital technology. Twenty-one articles were screened, and 16 articles met the inclusion/exclusion criteria and were reviewed. Current methodologies for mobility measurement in PAH with digital technologies are discussed. In brief, the reviewed evidence demonstrated that there is a lack of standardisation across studies for instrumentation, outcomes, and interpretation in PAH. The validity and reliability of digital approaches were insufficiently reported in all studies. Future research is required to standardise digital mobility measurement and characterise mobility impairments in PAH across clinical and real-world settings. The reviewed evidence suggests that digital mobility outcomes may be useful clinical measures and may be impaired in PAH, but further research is required to accurately and robustly establish findings. Recommendations are provided for future studies that encompass comprehensive reporting, validation, and measurement.

Wearable Activity Trackers Objectively Measure Incidental Physical Activity in Older Adults Undergoing Aortic Valve Replacement

BACKGROUND

For older adults with severe aortic stenosis (AS) undergoing aortic valve replacement (AVR), recovery of physical function is important, yet few studies objectively measure it in real-world environments. This exploratory study explored the acceptability and feasibility of using wearable trackers to measure incidental physical activity (PA) in AS patients before and after AVR.

METHODS

Fifteen adults with severe AS wore an activity tracker at baseline, and ten at one month follow-up. Functional capacity (six-minute walk test, 6MWT) and HRQoL (SF 12) were also assessed.

RESULTS

At baseline, AS participants (n = 15, 53.3% female, mean age 82.3 ± 7.0 years) wore the tracker for four consecutive days more than 85% of the total prescribed time, this improved at follow-up. Before AVR, participants demonstrated a wide range of incidental PA (step count median 3437 per day), and functional capacity (6MWT median 272 m). Post-AVR, participants with the lowest incidental PA, functional capacity, and HRQoL at baseline had the greatest improvements within each measure; however, improvements in one measure did not translate to improvements in another.

CONCLUSION

The majority of older AS participants wore the activity trackers for the required time period before and after AVR, and the data attained were useful for understanding AS patients' physical function.

Dose-response relationship between daily step count and prevalence of sarcopenia: A cross-sectional study

OBJECTIVES

Daily step counts are an easy-to-understand indicator of physical activity; however, there is limited evidence regarding the optimal daily step count to prevent sarcopenia. This study examined the dose-response relationship between daily step count and the prevalence of sarcopenia and explored the optimal dose.

DESIGN

Cross-sectional study.

SETTING AND PARTICIPANTS

The study included 7949 community-dwelling middle-aged and older adults (aged 45-74 years) from Japan.

MEASUREMENTS

Skeletal muscle mass (SMM) was assessed using bioelectrical impedance spectroscopy, and muscle strength was quantified through handgrip strength (HGS) measurement. Participants who exhibited both low HGS (men: <28 kg, women: <18 kg) and low SMM (lowest quartile in each sex-specific category) were defined as having sarcopenia. Daily step counts were measured for 10 days using a waist-mounted accelerometer. To examine the association between daily step count and sarcopenia, a multivariate logistic regression analysis was performed, adjusting for potential confounding factors such as age, sex, body mass index, smoking status, alcohol consumption, protein intake, and medical history. The odds ratios (ORs) and confidence intervals (CIs) were calculated based on the daily step counts categorized into quartiles (Q1-Q4). Finally, a restricted cubic spline curve was fitted to further investigate the dose-response relationship between daily step count and sarcopenia.

RESULTS

The prevalence of sarcopenia in the overall participants was 3.3 % (259/7949 participants), with a mean daily step count of 7292 ± 2966 steps. Expressed in quartiles, the mean daily step counts were 3873 ± 935 steps in Q1, 6025 ± 503 steps in Q2, 7942 ± 624 steps in Q3, and 11,328 ± 1912 steps in Q4. The prevalence of sarcopenia in each quartile of daily step count was 4.7 % (93/1987 participants) in Q1, 3.4 % (68/1987 participants) in Q2, 2.7 % (53/1988 participants) in Q3, and 2.3 % (45/1987 participants) in Q4. The ORs and 95 % CIs adjusted for covariates demonstrated a statistically significant inverse association between daily step count and sarcopenia prevalence (P for trend <0.01), as follows: Q1, reference; Q2, 0.79 (95 % CI: 0.55-1.11); Q3, 0.71 (95 % CI: 0.49-1.03); Q4, 0.61 (95 % CI: 0.41-0.90). The restricted cubic spline curve indicated that the ORs leveled off at approximately 8000 steps per day, and no statistically significant decrease in ORs was observed for daily step counts above this threshold.

CONCLUSIONS

The study found a significant inverse association between daily step count and the prevalence of sarcopenia, with the association plateauing when the daily step count exceeded approximately 8000 steps. These findings suggest that 8000 steps per day may be the optimal dose to prevent sarcopenia. Further intervention and longitudinal studies are needed to validate the results.

Acceptability of wearable devices for measuring mobility remotely: Observations from the Mobilise-D technical validation study

Background

This study aimed to explore the acceptability of a wearable device for remotely measuring mobility in the Mobilise-D technical validation study (TVS), and to explore the acceptability of using digital tools to monitor health.

Methods

Participants (N = 106) in the TVS wore a waist-worn device (McRoberts Dynaport MM + ) for one week. Following this, acceptability of the device was measured using two questionnaires: The Comfort Rating Scale (CRS) and a previously validated questionnaire. A subset of participants (n = 36) also completed semi-structured interviews to further determine device acceptability and to explore their opinions of the use of digital tools to monitor their health. Questionnaire results were analysed descriptively and interviews using a content analysis.

Results

The device was considered both comfortable (median CRS (IQR; min-max) = 0.0 (0.0; 0-20) on a scale from 0-20 where lower scores signify better comfort) and acceptable (5.0 (0.5; 3.0-5.0) on a scale from 1-5 where higher scores signify better acceptability). Interviews showed it was easy to use, did not interfere with daily activities, and was comfortable. The following themes emerged from participants' as being important to digital technology: altered expectations for themselves, the use of technology, trust, and communication with healthcare professionals.

Conclusions

Digital tools may bridge existing communication gaps between patients and clinicians and participants are open to this. This work indicates that waist-worn devices are supported, but further work with patient advisors should be undertaken to understand some of the key issues highlighted. This will form part of the ongoing work of the Mobilise-D consortium.