Sjaan R. Gomersall

Accuracy of Heart Rate Watches: Implications for Weight Management

Background Wrist-worn monitors claim to provide accurate measures of heart rate and energy expenditure. People wishing to lose weight use these devices to monitor energy balance, however the accuracy of these devices to measure such parameters has not been established. Aim To determine the accuracy of four wrist-worn devices (Apple Watch, Fitbit Charge HR, Samsung Gear S and Mio Alpha) to measure heart rate and energy expenditure at rest and during exercise. Methods Twenty-two healthy volunteers (50% female; aged 24 ± 5.6 years) completed ~1-hr protocols involving supine and seated rest, walking and running on a treadmill and cycling on an ergometer. Data from the devices collected during the protocol were compared with reference methods: electrocardiography (heart rate) and indirect calorimetry (energy expenditure). Results None of the devices performed significantly better overall, however heart rate was consistently more accurate than energy expenditure across all four devices. Correlations between the devices and reference methods were moderate to strong for heart rate (0.67–0.95 [0.35 to 0.98]) and weak to strong for energy expenditure (0.16–0.86 [-0.25 to 0.95]). All devices underestimated both outcomes compared to reference methods. The percentage error for heart rate was small across the devices (range: 1–9%) but greater for energy expenditure (9–43%). Similarly, limits of agreement were considerably narrower for heart rate (ranging from -27.3 to 13.1 bpm) than energy expenditure (ranging from -266.7 to 65.7 kcals) across devices. Conclusion These devices accurately measure heart rate. However, estimates of energy expenditure are poor and would have implications for people using these devices for weight loss.

Assessing Sedentary Behavior with the GENEActiv: Introducing the Sedentary Sphere.

BACKGROUND The Sedentary Sphere is a method for the analysis, identification, and visual presentation of sedentary behaviors from a wrist-worn triaxial accelerometer. PURPOSE This study aimed to introduce the concept of the Sedentary Sphere and to determine the accuracy of posture classification from wrist accelerometer data. METHODS Three samples were used: 1) free living (n = 13, ages 20-60 yr); 2) laboratory based (n = 25, ages 30-65 yr); and 3) hospital inpatients (n = 10, ages 60-90 yr). All participants wore a GENEActiv on their wrist and activPAL on their thigh. The free-living sample wore an additional GENEActiv on the thigh and completed the Multimedia Activity Recall for Children and Adults. The laboratory-based sample wore the monitors while seated at a desk for 7 h, punctuated by 2 min of walking every 20 min. The free-living and inpatient samples wore the monitors for 24 h. Posture was classified from wrist-worn accelerometry using the Sedentary Sphere concept. RESULTS Sitting time did not differ between the wrist GENEActiv and the activPAL in the free-living sample and was correlated in the three samples combined (rho = 0.9, P < 0.001), free-living and inpatient samples (r ≃ 0.8, P < 0.01). Mean intraindividual agreement was 85% ± 7%. In the laboratory-based and inpatient samples, sitting time was underestimated by the wrist GENEActiv by 30 min and 2 h relative to the activPAL, respectively (P < 0.05). Posture classification disagreed during reading while standing, cooking while standing, and brief periods during driving. Posture allocation validity was excellent when the GENEActiv was worn on the thigh, evidenced by the near-perfect agreement with the activPAL (96% ± 3%). CONCLUSIONS The Sedentary Sphere enables determination of the most likely posture from the wrist-worn GENEActiv. Visualizing behaviors on the sphere displays the pattern of wrist movement and positions within that behavior.

Development and evaluation of an adult use-of-time instrument with an energy expenditure focus

Measurement in behavioural epidemiology depends on high resolution and precise and accurate measures of the behaviour of interest. Few questionnaires in the adult population are able to simultaneously collect the multidimensional information that is emerging as being important in the relationship between behaviour and health. This project had two objectives: (1) to develop an adult version of the computer-delivered Multimedia Activity Recall for Children and Adolescents (MARCA), a 24-h activity recall instrument that can measure use-of-time and estimate energy expenditure and (2) to determine the test-retest reliability and convergent validity of the developed adult MARCA. Thirty-eight healthy subjects (mean±SD, 31.7±12.1 yr) completed two recalls of the adult MARCA within 24-h and accelerometer counts were measured on 30 of the subjects. Bland-Altman analysis and intraclass correlation coefficients (ICC) were used to quantify the test-retest reliability of the adult MARCA. Spearman rank correlation coefficients (rho) were used to quantify convergent validity of the adult MARCA compared to accelerometer counts. The test-retest reliability coefficients of the adult MARCA were high with intra-class coefficients ranging from 0.99 to 1.00. Moderate to strong validity was observed for physical activity level (PAL) (MET.min score of accelerometer wear time) and accelerometer counts per minute (rho=0.72). The adult MARCA is a valid and reliable self-report measure of use-of-time and energy expenditure, capable of a wide variety of flexible use-of-time analyses related to a wide range of behaviours.

The validity of the GENEActiv wrist-worn accelerometer for measuring adult sedentary time in free living

OBJECTIVES Based on self-reported measures, sedentary time has been associated with chronic disease and mortality. This study examined the validity of the wrist-worn GENEactiv accelerometer for measuring sedentary time (i.e. sitting and lying) by posture classification, during waking hours in free living adults. DESIGN Fifty-seven participants (age=18-55 years 52% male) were recruited using convenience sampling from a large metropolitan Australian university. METHODS Participants wore a GENEActiv accelerometer on their non-dominant wrist and an activPAL device attached to their right thigh for 24-h (00:00 to 23:59:59). Pearsons Correlation Coefficient was used to examine the convergent validity of the GENEActiv and the activPAL for estimating total sedentary time during waking hours. Agreement was illustrated using Bland and Altman plots, and intra-individual agreement for posture was assessed with the Kappa statistic. RESULTS Estimates of average total sedentary time over 24-h were 623 (SD 103) min/day from the GENEActiv, and 626 (SD 123) min/day from the activPAL, with an Intraclass Correlation Coefficient of 0.80 (95% confidence intervals 0.68-0.88). Bland and Altman plots showed slight underestimation of mean total sedentary time for GENEActiv relative to activPAL (mean difference: -3.44min/day), with moderate limits of agreement (-144 to 137min/day). Mean Kappa for posture was 0.53 (SD 0.12), indicating moderate agreement for this sample at the individual level. CONCLUSIONS The estimation of sedentary time by posture classification of the wrist-worn GENEActiv accelerometer was comparable to the activPAL. The GENEActiv may provide an alternative, easy to wear device based measure for descriptive estimates of sedentary time in population samples.

Estimating physical activity and sedentary behavior in a free-living context: a pragmatic comparison of consumer-based activity trackers and ActiGraph accelerometry

Background Activity trackers are increasingly popular with both consumers and researchers for monitoring activity and for promoting positive behavior change. However, there is a lack of research investigating the performance of these devices in free-living contexts, for which findings are likely to vary from studies conducted in well-controlled laboratory settings. Objective The aim was to compare Fitbit One and Jawbone UP estimates of steps, moderate-to-vigorous physical activity (MVPA), and sedentary behavior with data from the ActiGraph GT3X+ accelerometer in a free-living context. Methods Thirty-two participants were recruited using convenience sampling; 29 provided valid data for this study (female: 90%, 26/29; age: mean 39.6, SD 11.0 years). On two occasions for 7 days each, participants wore an ActiGraph GT3X+ accelerometer on their right hip and either a hip-worn Fitbit One (n=14) or wrist-worn Jawbone UP (n=15) activity tracker. Daily estimates of steps and very active minutes were derived from the Fitbit One (n=135 days) and steps, active time, and longest idle time from the Jawbone UP (n=154 days). Daily estimates of steps, MVPA, and longest sedentary bout were derived from the corresponding days of ActiGraph data. Correlation coefficients and Bland-Altman plots with examination of systematic bias were used to assess convergent validity and agreement between the devices and the ActiGraph. Cohen’s kappa was used to assess the agreement between each device and the ActiGraph for classification of active versus inactive (≥10,000 steps per day and ≥30 min/day of MVPA) comparable with public health guidelines. Results Correlations with ActiGraph estimates of steps and MVPA ranged between .72 and .90 for Fitbit One and .56 and .75 for Jawbone UP. Compared with ActiGraph estimates, both devices overestimated daily steps by 8% (Fitbit One) and 14% (Jawbone UP). However, mean differences were larger for daily MVPA (Fitbit One: underestimated by 46%; Jawbone UP: overestimated by 50%). There was systematic bias across all outcomes for both devices. Correlations with ActiGraph data for longest idle time (Jawbone UP) ranged from .08 to .19. Agreement for classifying days as active or inactive using the ≥10,000 steps/day criterion was substantial (Fitbit One: κ=.68; Jawbone UP: κ=.52) and slight-fair using the criterion of ≥30 min/day of MVPA (Fitbit One: κ=.40; Jawbone UP: κ=.14). Conclusions There was moderate-strong agreement between the ActiGraph and both Fitbit One and Jawbone UP for the estimation of daily steps. However, due to modest accuracy and systematic bias, they are better suited for consumer-based self-monitoring (eg, for the public consumer or in behavior change interventions) rather than to evaluate research outcomes. The outcomes that relate to health-enhancing MVPA (eg, “very active minutes” for Fitbit One or “active time” for Jawbone UP) and sedentary behavior (“idle time” for Jawbone UP) should be used with caution by consumers and researchers alike.

The Elasticity of Time: Associations Between Physical Activity and Use of Time in Adolescents

The way an individual uses one’s time can greatly affect his or her health. The purpose of this article was to examine the cross-sectional cross-elasticity relationships for use of time domains in a sample of Australian adolescents. This study analyzed 24-hour recall time use data collected using the Multimedia Activity Recall for Children and Adults (N = 2,200). Using simple linear regression, the authors calculated the difference in time devoted to a reference activity (i.e., screen time, sleep, or social) given 1 hour extra in the time devoted to a criterion activity (i.e., physical activity). The most elastic activities were screen time and school-related time. Every additional hour committed to physical activity was associated with 32 minutes less screen time. This relationship was more pronounced in obese adolescents (−56 minutes screen time) compared with normal (−31 minutes) and overweight (−27 minutes) adolescents. Promising behavior patterns exist, with screen time appearing as a highly elastic behavior.

Results from Australia’s 2014 Report Card on Physical Activity for Children and Youth

BACKGROUND Like many other countries, Australia is facing an inactivity epidemic. The purpose of the Australian 2014 Physical Activity Report Card initiative was to assess the behaviors, settings, and sources of influences and strategies and investments associated with the physical activity levels of Australian children and youth. METHODS A Research Working Group (RWG) drawn from experts around Australia collaborated to determine key indicators, assess available datasets, and the metrics which should be used to inform grades for each indicator and factors to consider when weighting the data. The RWG then met to evaluate the synthesized data to assign a grade to each indicator. RESULTS Overall Physical Activity Levels were assigned a grade of D-. Other physical activity behaviors were also graded as less than average (D to D-), while Organized Sport and Physical Activity Participation was assigned a grade of B-. The nation performed better for settings and sources of influence and Government Strategies and Investments (A- to a C). Four incompletes were assigned due to a lack of representative quality data. CONCLUSIONS Evidence suggests that physical activity levels of Australian children remain very low, despite moderately supportive social, environmental and regulatory environments. There are clear gaps in the research which need to be filled and consistent data collection methods need to be put into place.

Nine year changes in sitting time in young and mid-aged Australian women: findings from the Australian Longitudinal Study for Women's Health.

OBJECTIVE To examine changes in sitting time (ST) in women over nine years and to identify associations between life events and these changes. METHODS Young (born 1973-78, n=5215) and mid-aged (born 1946-51, n=6973) women reported life events and ST in four surveys of the Australian Longitudinal Study on Womens Health between 2000 and 2010. Associations between life events and changes in ST between surveys (decreasers ≥2 h/day less, increasers ≥2 h/day more) were estimated using generalized estimating equations. RESULTS Against a background of complex changes there was an overall decrease in ST in young women (median change -0.48 h/day, interquartile range [IQR]=-2.54, 1.50) and an increase in ST in mid-aged women (median change 0.43 h/day; IQR=-1.29, 2.0) over nine years. In young women, returning to study and job loss were associated with increased ST, while having a baby, beginning work and decreased income were associated with decreased ST. In mid-aged women, changes at work were associated with increased ST, while retiring and decreased income were associated with decreased ST. CONCLUSIONS ST changed over nine years in young and mid-aged Australian women. The life events they experienced, particularly events related to work and family, were associated with these changes.

Field evaluation of a random forest activity classifier for wrist-worn accelerometer data

OBJECTIVES Wrist-worn accelerometers are convenient to wear and associated with greater wear-time compliance. Previous work has generally relied on choreographed activity trials to train and test classification models. However, validity in free-living contexts is starting to emerge. Study aims were: (1) train and test a random forest activity classifier for wrist accelerometer data; and (2) determine if models trained on laboratory data perform well under free-living conditions. DESIGN Twenty-one participants (mean age=27.6±6.2) completed seven lab-based activity trials and a 24h free-living trial (N=16). METHODS Participants wore a GENEActiv monitor on the non-dominant wrist. Classification models recognising four activity classes (sedentary, stationary+, walking, and running) were trained using time and frequency domain features extracted from 10-s non-overlapping windows. Model performance was evaluated using leave-one-out-cross-validation. Models were implemented using the randomForest package within R. Classifier accuracy during the 24h free living trial was evaluated by calculating agreement with concurrently worn activPAL monitors. RESULTS Overall classification accuracy for the random forest algorithm was 92.7%. Recognition accuracy for sedentary, stationary+, walking, and running was 80.1%, 95.7%, 91.7%, and 93.7%, respectively for the laboratory protocol. Agreement with the activPAL data (stepping vs. non-stepping) during the 24h free-living trial was excellent and, on average, exceeded 90%. The ICC for stepping time was 0.92 (95% CI=0.75-0.97). However, sensitivity and positive predictive values were modest. Mean bias was 10.3min/d (95% LOA=-46.0 to 25.4min/d). CONCLUSIONS The random forest classifier for wrist accelerometer data yielded accurate group-level predictions under controlled conditions, but was less accurate at identifying stepping verse non-stepping behaviour in free living conditions Future studies should conduct more rigorous field-based evaluations using observation as a criterion measure.

Introducing novel approaches for examining the variability of individuals’ physical activity

Abstract Tudor-Locke and colleagues previously assessed steps/day for 1 year. The aim of this study was to use this data set to introduce a novel approach for the investigation of whether individual’s physical activity exhibits periodicity fluctuating round a mean and, if so, the degree of fluctuation and whether the mean changes over time. Twenty-three participants wore a pedometer for 365 days, recorded steps/day and whether the day was a workday. Fourier transform of each participant’s daily steps data showed the physical activity had a periodicity of 7 days in half of the participants, matching the periodicity of the workday pattern. Activity level remained stable in half of the participants, decreased in ten participants and increased in two. In conclusion, the 7-day periodicity of activity in half of the participants and correspondence with the workday pattern suggest a social or environmental influence. The novel analytical approach introduced herein allows the determination of the periodicity of activity, the degree of variability in activity that is tolerated during day-to-day life and whether the activity level is stable. Results from the use of these methodologies in larger data sets may enable a more focused approach to the design of interventions that aim to increase activity.