scholarly journals Comprehensive comparison of Apple Watch and Fitbit monitors in a free-living setting

PLoS ONE ◽  
2021 ◽  
Vol 16 (5) ◽  
pp. e0251975
Author(s):  
Yang Bai ◽  
Connie Tompkins ◽  
Nancy Gell ◽  
Dakota Dione ◽  
Tao Zhang ◽  
...  
Keyword(s):  
Heart Rate ◽  
Pearson Correlation ◽  
Living Condition ◽  
Equivalence Testing ◽  
Free Living ◽  
Activity Monitors ◽  
Step Counts ◽  
Comprehensive Comparison ◽  
Rate Correlation ◽  
Absolute Percent Error

Objectives The aim of this study was to evaluate the accuracy of three consumer-based activity monitors, Fitbit Charge 2, Fitbit Alta, and the Apple Watch 2, all worn on the wrist, in estimating step counts, moderate-to-vigorous minutes (MVPA), and heart rate in a free-living setting. Methods Forty-eight participants (31 females, 17 males; ages 18–59) were asked to wear the three consumer-based monitors mentioned above on the wrist, concurrently with a Yamax pedometer as the criterion for step count, an ActiGraph GT3X+ (ActiGraph) for MVPA, and a Polar H7 chest strap for heart rate. Participants wore the monitors for a 24-hour free-living condition without changing their usual active routine. MVPA was calculated in bouts of ≥10 minutes. Pearson correlation, mean absolute percent error (MAPE), and equivalence testing were used to evaluate the measurement agreement. Results The average step counts recorded for each device were as follows: 11,734 (Charge2), 11,922 (Alta), 11,550 (Apple2), and 10,906 (Yamax). The correlations in steps for the above monitors ranged from 0.84 to 0.95 and MAPE ranged from 17.1% to 35.5%. For MVPA minutes, the average were 76.3 (Charge2), 63.3 (Alta), 49.5 (Apple2), and 47.8 (ActiGraph) minutes accumulated in bouts of 10 or greater minutes. The correlation from MVPA estimation for above monitors were 0.77, 0.91, and 0.66. MAPE from MVPA estimation ranged from 44.7% to 55.4% compared to ActiGraph. For heart rate, correlation for Charge2 and Apple2 was higher for sedentary behavior and lower for MVPA. The MAPE ranged from 4% to 16%. Conclusion All three consumer monitors estimated step counts fairly accurately, and both the Charge2 and Apple2 reported reasonable heart rate estimation. However, all monitors substantially underestimated MVPA in free-living settings.

Heart Rate Equivalency of the Fitbit Charge HR During Continuous Aerobic Exercise

2018 ◽  
Vol 1 (3) ◽  
pp. 122-129
Author(s):  
Joseph M. Stock ◽  
Ryan T. Pohlig ◽  
Matthew J. Botieri ◽  
David G. Edwards ◽  
Gregory M. Dominick
Keyword(s):  
Heart Rate ◽  
Aerobic Exercise ◽  
Exercise Bout ◽  
Total Activity ◽  
Equivalence Testing ◽  
Warm Up ◽  
Activity Monitors ◽  
Continuous Exercise ◽  
Exercise Onset ◽  
Absolute Percent Error

Purpose: Consumer-grade wrist-worn activity monitors frequently include photoplethysmography (PPG) sensors for estimating heart rate (HR). The Fitbit Charge HR is marketed specifically for tracking fitness; therefore, HR accuracy is critical, especially during exercise. This study examined HR equivalency of the Fitbit Charge HR during continuous aerobic exercise. Method: Participants (N = 19) concurrently wore a Polar H1 and Fitbit Charge HR during a measurement visit that included seated rest (5 minutes), warm-up (5 minutes), continuous treadmill exercise (30 minutes), and cool-down (5–10 minutes). Mean HR differences were examined by protocol phase, total activity (i.e., warm-up, exercise, and cool-down combined), and the first, middle, and last 5 minutes of continuous exercise. Mean absolute percent error (MAPE), Bland-Altman plots, and 95% equivalence testing explained overall and individual HR agreement between devices. Results: The Fitbit Charge HR significantly underestimated HR for all measurement phases (all p ≤ .01) except cool-down (p > .33). HR agreement was notably weaker during warm-up (r = 0.66, d = 0.57) and differences were greatest for the first 5 minutes compared to the middle and end of exercise (6.94±2.16 beats per minute [bpm] vs. 1.76±0.59 bpm, and 1.74±0.58 bpm), F = 4.87, p = .04). Mean exercise HRs were equivalent between devices (±2.69 bpm, 95% CI: 1.41–3.97 bpm); MAPE was 1.96%. Conclusion: The Fitbit Charge HR is relatively accurate for measuring HR during continuous aerobic exercise. Whereas the accuracy of PPG-based HR appears limited at exercise onset, agreement improves throughout the exercise bout and HR differences are negligible.

Physical Activity Patterns of Free Living Dogs Diagnosed with Osteoarthritis

2021 ◽  
Author(s):  
Anne H Lee ◽  
Katelyn B Detweiler ◽  
Tisha A Harper ◽  
Kim E Knap ◽  
Maria R C de Godoy ◽  
...  
Keyword(s):  
Physical Activity ◽  
Pain Intensity ◽  
Repeated Measures ◽  
Pearson Correlation ◽  
Clinical Signs ◽  
Total Activity ◽  
Activity Count ◽  
Free Living ◽  
Activity Monitors ◽  
Average Activity

Abstract Osteoarthritis (OA) affects about 90% of dogs > 5 yr of age in the US, resulting in reduced range of motion, difficulty climbing and jumping, reduced physical activity, and lower quality of life. Our objective was to use activity monitors to measure physical activity and identify how activity counts correlate with age, body weight (BW), body condition score (BCS), serum inflammatory markers, veterinarian pain assessment, and owner perception of pain in free-living dogs with OA. The University of Illinois Institutional Animal Care and Use Committee approved the study and owner consent was received prior to experimentation. Fifty-six client-owned dogs (mean age = 7.8 yr; mean BCS = 6.1) with clinical signs and veterinary diagnosis of OA wore HeyRex activity collars continuously over a 49-d period. Blood samples were collected on d 0 and 49, and dog owners completed canine brief pain inventory (CBPI) and Liverpool osteoarthritis in dogs (LOAD) surveys on d 0, 21, 35, and 49. All data were analyzed using SAS 9.3 using repeated measures and R Studio 1.0.136 was used to generate Pearson correlation coefficients between data outcomes. Average activity throughout the study demonstrated greater activity levels on weekends. It also showed that 24-h activity spiked twice daily, once in the morning and another in the afternoon. Serum C-reactive protein concentration was lower (P < 0.01) at d 49 compared to d 0. Survey data indicated lower (P < 0.05) overall pain intensity and severity score on d 21, 35 and 49 compared to d 0. BW was correlated with average activity counts (p=0.02; r=-0.12) and run activity (p=0.10; r=-0.24). Weekend average activity counts were correlated with owner pain intensity scores (p=0.0813; r=-0.2311), but weekday average activity count was not. Age was not correlated with total activity count, sleep activity, or run activity, but it was correlated with scratch (p=0.03; r=-0.10), alert (p=0.03; r=-0.13) and walk (p=0.09; r=-0.23) activities. Total activity counts and activity type (sleep, scratch, alert, walk, run) were not correlated with pain scored by veterinarians, pain intensity or severity scored by owners, or baseline BCS. Even though the lack of controls and/or information on the individual living conditions of dogs resulted in a high level of variability in this study, our data suggest that the use of activity monitors have the potential to aid in the management of OA and other conditions affecting activity (e.g., allergy; anxiety).

Validity of the SC-StepMX pedometer during treadmill walking and running

2013 ◽  
Vol 38 (5) ◽  
pp. 520-524 ◽  
Author(s):  
Rachel C. Colley ◽  
Joel D. Barnes ◽  
Allana G. Leblanc ◽  
Michael Borghese ◽  
Charles Boyer ◽  
...  
Keyword(s):  
Confidence Interval ◽  
Walking Speed ◽  
The Elderly ◽  
Measurement Bias ◽  
Convenience Sample ◽  
Activity Monitors ◽  
Step Counts ◽  
Mean Absolute Percent Error ◽  
Walking Speeds ◽  
Absolute Percent Error

The purpose of this study was to examine the validity of the SC-StepMX pedometer for measuring step counts. A convenience sample of 40 participants wore 4 SC-StepMX pedometers, 2 Yamax DigiWalker pedometers, and 2 Actical accelerometers around their waist on a treadmill at 4 speeds based on each participant's self-paced walking speed (50%, 100%, 180%, and 250%; range: 1.4–14.1 km·h–1). The SC-StepMX demonstrated lower mean absolute percent error (–0.2%) compared with the Yamax DigiWalker (–20.5%) and the Actical (–26.1%). Mean measurement bias was lower for the SC-StepMX (0.1 ± 9.1; 95% confidence interval = –17.8 to 18.0 steps·min–1) when compared with both the Yamax DigiWalker (–15.9 ± 23.3; 95% confidence interval = –61.6 to 29.7 steps·min–1) and the Actical (–22.0 ± 36.3; 95% CI = –93.1 to 49.1 steps·min–1). This study demonstrates that the SC-StepMX pedometer is a valid tool for the measurement of step counts. The SC-StepMX accurately measures step counts at slower walking speeds when compared with 2 other commercially available activity monitors. This makes the SC-StepMX useful in measuring step counts in populations that are active at lower intensities (e.g., sedentary individuals, the elderly).

scholarly journals Feasibility Study of an mHealth-based Team Challenge to Promote Stepping and Stairs Climbing Activity (Preprint)

2018 ◽  
Author(s):  
Seaw Jia Liew ◽  
Alex Wilhelm Gorny ◽  
Chuen Seng Tan ◽  
Falk Muller-Riemenschneider
Keyword(s):  
Real Time ◽  
Intervention Group ◽  
Living Condition ◽  
Data Management System ◽  
Multiple Linear Regressions ◽  
Free Living ◽  
Entire Study ◽  
Step Counts ◽  
Steps Per Day ◽  
Linear Regressions

BACKGROUND mHealth approaches are gaining popularity to address low levels of physical activity (PA). OBJECTIVE This study aimed to: (1) develop an mHealth suite, combining PA wearables and an interactive smartphone application (App) supported by a web-based data management system, (2) determine the validity of the wearables in measuring steps per day and floor-count, and (3) assess feasibility and effects of a 6-week team challenge intervention. METHODS Staff and students from a public university were recruited between 2015 and 2016. In Phase 1, every participant was requested to wear a Fitbit tracker (Charge™ or Charge HR™) and an ActiGraph™ for 7 days to measure daily step counts under free-living condition. They were also asked to climb 4 bouts of floors in an indoor stairswell to measure floor-counts. Steps per day and floor-counts estimated by Fitbit™ were compared against ActiGraph and direct observation, respectively. In Phase 2, participants were allocated to control or intervention group and received a Fitbit tracker synced to the Fitbit App. Further, the intervention participants were randomized to 4 teams and used the developed mHealth suite. Teams competed in 6 weekly (Monday - Friday) real-time challenges. A valid day was defined as having accumulated ≥1,500 steps per day. Outcomes were: (i) adherence to wearing Fitbit (i.e. number of days in which all participants in each group was classified as valid users aggregated across entire study period), (ii) mean proportion of valid participants over the study period, and (iii) the effects of intervention on steps and floor-counts determined using multiple linear regressions models and generalized estimating equations (GEE) for longitudinal data analysis. RESULTS In Phase one, 32/40 (steps) and 40/40 (floors) participants provided valid data. The Fitbit trackers demonstrated a high to very high correlation (steps: Spearman Rho=0.89, P < .001, floors: Spearman Rho=0.98, P < .001), respectively. The trackers over-estimated step-counts in free-living condition (median absolute error: 17%) but accurately estimated floor-counts. In Phase two, 20 participants each were allocated to intervention and control. 24 completers (i.e. provided complete covariates and valid PA data) were included in the analyses. Multiple linear regressions revealed 15.9% higher average steps/day (95% CI: -8.9, 47.6, P= .21) and 39.4% higher average floors/day (95% CI: 2.4, 89.7, P= .04) in the intervention group during the final two intervention weeks. GEE results indicated no significant interaction effects between groups and intervention week for weekly step counts, whereas a significant effect (P< .001) was observed for weekly floor counts. CONCLUSIONS The consumer wearables integrated in our mHealth suite provided acceptable validity in estimating stepping and stairs climbing activities. The mHealth suite was feasible for implementing real-time team-challenge interventions. Compared to the controls, the intervention participants performed more stairs climbing which could be introduced as an additional PA promotion target in the context of mHealth strategies. Methodologically rigorous studies with larger sample-size and long-term follow-up are warranted to strengthen the evidence for the proposed mHealth strategy.

Validity of Pedometers to Measure Step Counts During Dance

2015 ◽  
Vol 12 (10) ◽  
pp. 1430-1435
Author(s):  
Tiago V. Barreira ◽  
John P. Bennett ◽  
Minsoo Kang
Keyword(s):  
College Students ◽  
Pearson Correlation ◽  
Validity Evidence ◽  
Individual Values ◽  
Dance Class ◽  
Percent Error ◽  
Step Counts ◽  
Mean Absolute Percent Error ◽  
Study Participants ◽  
Absolute Percent Error

Purpose:To obtain validity evidence for the measurement of step counts by spring-levered and piezoelectric pedometers during dance.Methods:Thirty-five adults in a college dance class participated in this study. Participants completed trials of 3- and 5-min of different styles of dance wearing Walk4life MVP and Omron HJ-303 pedometers, while their steps were visually counted. Pearson correlation, paired t-test, mean absolute percent error (MAPE), and mean bias were calculated between actual step and pedometer step counts for the 3- and 5-min dances separately.Results:For the Walk4life trials the correlations were .92 and .77 for the 3- and 5-min dances. No significant differences were shown by t-test for the 3- (P = .16) and 5-min dances (P = .60). However, MAPE was high, 17.7 ± 17.7% and 19.4 ± 18.3% for the 2 dance durations, respectively. For the Omron, the correlations were .44 and .58 for the 3- and 5-min dances, respectively. No significant differences were shown by t-test for the 3-min (P = .38) and for the 5-min (P = .88) dances. However, MAPE was high, 19.3 ± 16.4% and 26.6 ± 15.2% for the 2 dance durations, respectively.Conclusions:This study demonstrated that pedometers can be used to estimate the number of steps taken by a group of college students while dancing, however caution is necessary with individual values.

Step count accuracy of StepWatch and FitBit One™ among individuals with a unilateral transtibial amputation

2018 ◽  
Vol 42 (5) ◽  
pp. 518-526 ◽  
Author(s):  
Elisa S Arch ◽  
Jaclyn M Sions ◽  
John Horne ◽  
Barry A Bodt
Keyword(s):  
Intraclass Correlation ◽  
Correlation Coefficients ◽  
Living Environment ◽  
Activity Level ◽  
Step Count ◽  
Free Living ◽  
Transtibial Amputation ◽  
Activity Monitors ◽  
Step Counts ◽  
Intraclass Correlation Coefficients

Background: Step counts, obtained via activity monitors, provide insight into activity level in the free-living environment. Accuracy assessments of activity monitors are limited among individuals with lower-limb amputations. Objectives: (1) To evaluate the step count accuracy of both monitors during forward-linear and complex walking and (2) compare monitor step counts in the free-living environment. Study design: Cross-sectional study. Methods: Adult prosthetic users with a unilateral transtibial amputation were equipped with StepWatch and FitBit One™. Participants completed an in-clinic evaluation to evaluate each monitor’s step count accuracy during forward linear and complex walking followed by a 7-day step count evaluation in the free-living environment. Results: Both monitors showed excellent accuracy during forward, linear walking (intraclass correlation coefficients = 0.97–0.99, 95% confidence interval = 0.93–0.99; percentage error = 4.3%–6.2%). During complex walking, percentage errors were higher (13.0%–15.5%), intraclass correlation coefficients were 0.88–0.90, and 95% confidence intervals were 0.69–0.96. In the free-living environment, the absolute percentage difference between monitor counts was 25.4%, but the counts had a nearly perfect linear relationship. Conclusion: Both monitors accurately counted steps during forward linear walking. StepWatch appears to be more accurate than FitBit during complex walking but a larger sample size may confirm these findings. FitBit consistently counted fewer steps than StepWatch during free-living walking. Clinical relevance The StepWatch and FitBit are acceptable tools for assessing forward, linear walking for individuals with transtibial amputation. Given the results’ consistenty in the free-living enviorment, both tools may ultimiately be able to be used to count steps in the real world, but more research is needed to confirm these findings.

scholarly journals Inter-Device Agreement between Fitbit Flex 1 and 2 for Assessing Sedentary Behavior and Physical Activity

2021 ◽  
Vol 18 (5) ◽  
pp. 2716
Author(s):  
Sunku Kwon ◽  
Ryan D. Burns ◽  
Youngwon Kim ◽  
Yang Bai ◽  
Wonwoo Byun
Keyword(s):  
Physical Activity ◽  
Sedentary Behavior ◽  
Living Condition ◽  
Systematic Bias ◽  
Equivalence Testing ◽  
Strong Correlations ◽  
Time Data ◽  
Free Living ◽  
Systematic Biases ◽  
Model Agreement

This study examined the inter-model agreement between the Fitbit Flex (FF) and FF2 in estimating sedentary behavior (SED) and physical activity (PA) during a free-living condition. 33 healthy adults wore the FF and FF2 on non-dominant wrist for 14 consecutive days. After excluding sleep and non-wear time, data from the FF and FF2 was converted to the time spent (min/day) in SED and PA using a proprietary algorithm. Pearson’s correlation was used to evaluate the association between the estimates from FF and FF2. Mean absolute percent errors (MAPE) were used to examine differences and measurement agreement in SED and PA estimates between FF and FF2. Bland-Altman (BA) plots were used to examine systematic bias between two devices. Equivalence testing was conducted to examine the equivalence between the FF and FF2. The FF2 had strong correlations with the FF in estimating SED and PA times. Compared to the FF, the FF2 yielded similar SED and PA estimates along with relatively low measurement discords and did not have significant systematic biases for SED and Moderate-to-vigorous PA estimates. Our findings suggest that researchers may choose FF2 as a measurement of SED and PA when FF is not available in the market during the longitudinal PA research.

scholarly journals Agreement between the Apple Series 1, LifeTrak Core C200, and Fitbit Charge HR with Indirect Calorimetry for Assessing Treadmill Energy Expenditure

2019 ◽  
Vol 16 (20) ◽  
pp. 3812
Author(s):  
Peng Zhang ◽  
Ryan Donald Burns ◽  
You Fu ◽  
Steven Godin ◽  
Wonwoo Byun
Keyword(s):  
Young Adults ◽  
Energy Expenditure ◽  
Indirect Calorimetry ◽  
Mean Difference ◽  
Treadmill Running ◽  
Equivalence Testing ◽  
Activity Monitors ◽  
Individual Level ◽  
Statistical Equivalence ◽  
Absolute Percent Error

The purpose of this study was to examine agreement in energy expenditure between the Apple Series 1 Watch, LifeTrak Core C200, and Fitbit Charge HR with indirect calorimetry during various treadmill speeds in young adults. Participants were a sample of college-aged students (mean age = 20.1 (1.7) years; 13 females, 17 males). Participants completed six structured 10-minute exercise sessions on a treadmill with speeds ranging from 53.6 m·min−1 to 187.7 m·min−1. Indirect calorimetry was used as the criterion. Participants wore the Apple Watch, LifeTrak, and Fitbit activity monitors on their wrists. Group-level agreement was examined using equivalence testing, relative agreement was examined using Spearman’s rho, and individual-level agreement was examined using Mean Absolute Percent Error (MAPE) and Bland-Altman Plots. Activity monitor agreement with indirect calorimetry was supported using the Apple Watch at 160.9 m·min−1 (Mean difference = −2.7 kcals, 90% C.I.: −8.3 kcals, 2.8 kcals; MAPE = 11.9%; rs = 0.64) and 187.7 m·min−1 (Mean difference = 3.7 kcals, 90% C.I.: −2.2 kcals, 9.7 kcals; MAPE = 10.7%; rs = 0.72) and the Fitbit at 187.7 m·min−1 (Mean difference = −0.2 kcals, 90% C.I.: −8.8 kcals, 8.5 kcals; MAPE = 20.1%; rs = 0.44). No evidence for statistical equivalence was seen for the LifeTrak at any speed. Bland-Altman Plot Limits of Agreement were narrower for the Apple Series 1 Watch compared to other monitors, especially at slower treadmill speeds. The results support the utility of the Apple Series 1 Watch and Fitbit Charge HR for assessing energy expenditure during specific treadmill running speeds in young adults.

scholarly journals Can We Trust Inertial and Heart Rate Sensor Data from an APPLE Watch Device?

Proceedings ◽  
2020 ◽  
Vol 49 (1) ◽  
pp. 128
Author(s):  
Hugo G. Espinosa ◽  
David V. Thiel ◽  
Matthew Sorell ◽  
David Rowlands
Keyword(s):  
Heart Rate ◽  
Inertial Sensor ◽  
Pearson Correlation ◽  
Smart Phone ◽  
Human Movement ◽  
Stair Climbing ◽  
Sensor Data ◽  
Activity Monitors ◽  
Step Counting ◽  
Rate Sensor

The use of wearable technologies for the monitoring of human movement has increased considerably in the past few years, with applications to sports and other physical activities. Energy expenditure, walking and running distance, step count, and heart rate are some of the metrics provided by such devices via smart phone applications. Most of the research studies have involved validating the accuracy and reliability of the activity monitors by using the post-processed data from the device. The aim of this preliminary study was to determine if we can trust sensor data obtained from an Apple watch. This study evaluated the pre-processed data from the watch through step counting and heart rate measurements, and compared it with known validated devices (in-house 9DOF inertial sensor and Polar H10TM). Repeated activities (walking, jogging, and stair climbing) of varying duration and intensity were conducted by participants of varying age and body mass index (BMI). Pearson correlation (r > 0.95) and Bland–Altman statistical analyses were applied to the data to determine the level of agreement between the validated devices and the watch. The sensors from the Apple watch counted steps and measured heart rate with a minimum error and performed as expected.

Sign in / Sign up

Export Citation Format

Share Document