Validation and Calibration of Physical Activity Monitors in Children
Children's Nutrition Research Center, Baylor College of Medicine, 1100 Bates Street, Houston, TX 77030. E-mail: email@example.com
Objective: This study was designed to validate accelerometer-based activity monitors against energy expenditure (EE) in children; to compare monitor placement sites; to field-test the monitors; and to establish sedentary, light, moderate, and vigorous threshold counts.
Research Methods and Procedures: Computer Science and Applications Actigraph (CSA) and Mini-Mitter Actiwatch (MM) monitors, on the hip or lower leg, were validated and calibrated against 6-hour EE measurements by room respiration calorimetry, activity by microwave detector, and heart rate by telemetry in 26 children, 6 to 16 years old. During the 6 hours, the children performed structured activities, including resting metabolic rate (RMR), Nintendo, arts and crafts, aerobic warm-up, Tae Bo, treadmill walking and running, and games. Activity energy expenditure (AEE) computed as EE − RMR was regressed against counts to derive threshold counts.
Results: The mean correlations between EE or AEE and counts were slightly higher for MM-hip (r = 0.78 ± 0.06) and MM-leg (r = 0.80 ± 0.05) than CSA-hip (r = 0.66 ± 0.08) and CSA-leg (r = 0.73 ± 0.07). CSA and MM performed similarly on the hip (inter-instrument r = 0.88) and on the lower leg (inter-instrument r = 0.89). Threshold counts for the CSA-hip were <800, <3200, <8200, and ≥8200 for sedentary, light, moderate, and vigorous categories, respectively. For the MM-hip, the threshold counts were <100, <900, <2200, and ≥2200, respectively.
Discussion: The validation of the CSA and MM monitors against AEE and their calibration for sedentary, light, moderate, and vigorous thresholds certify these monitors as valid, useful devices for the assessment of physical activity in children.
The increased prevalence of obesity among children and adolescents in the United States has focused attention on the need to understand its etiology, consequences, treatment, and prevention (1). Integral to understanding its etiology is the determination of the relative contributions of overeating and inactivity. Dramatic increases in the energy intake of American children have not been found in recent National Health and Nutrition Examination Surveys II and III, and, therefore, do not seem to account for the increased prevalence of obesity (1). Epidemiological evidence points to a decrease in physical activity among American youth. To study this further, there is a need for valid and reliable quantitative measures of physical activity in children (2).
Children's patterns of physical activity have been historically assessed by direct observation, questionnaires, and heart-rate monitoring (2). Activity monitors were developed in response to the lack of reliability of self-report measures, the intrusiveness of direct observation, and the complexity of heart rate monitoring. With the advent of small accelerometer-based activity monitors, the ability to monitor children's activity has improved greatly. Recent advances in integrated circuitry and memory capacity have produced sensitive, unobtrusive, accelerometer-based devices that measure continuously the intensity, frequency, and duration of movement for extended periods.
As accelerometer-based monitors evolved, investigators validated and calibrated them in populations of interest. Most validation studies have been performed in adults (3, 4, 5, 6, 7). These activity monitors have been validated against indirect calorimetry and calibrated in terms of resting metabolism equivalents (METs). The MET thresholds used to quantify the time spent in light (<3 METs), moderate (3 to 6 METs), hard (6 to 9 METs), and very hard (>9 METs) activities have been defined in adults only (4). The usual practice of defining 1 MET as 3.5 mL O2/kg per minute is incorrect in children, because their resting metabolic rate (RMR) declines from ∼6 mL O2/kg per minute at 5 years of age to 3.5 mL O2/kg per minute at 18 years of age (8). Thus, adult-derived thresholds are not applicable to children and can lead to erroneous conclusions regarding physical activity levels in children.
The lack of appropriate thresholds for children is unfortunate, because there is tremendous interest in determining the level of activity of children. The assessment and evaluation of interventions aimed at increasing physical activity would benefit from a monitoring device that classifies sedentary to vigorous levels of activity in children. Activity monitors have been calibrated to detect moderate to high physical activity levels and, therefore, do not discriminate sedentary activities such as watching television and playing video and computer games. The accuracy of the time spent in sedentary activities by subject recall is low and could be improved with the use of activity monitors calibrated with a sedentary threshold (9).
In response to these issues, the performance of two activity monitors, chosen for their small size, was evaluated in children in this study. The Computer Science and Applications Actigraph (CSA) monitor is a unidirectional accelerometer that has been tested in children (10, 11). The Mini-Mitter Actiwatch (MM) is an omnidirectional accelerometer that has been evaluated only against direct observation in pre-schoolchildren (12). The MM not only has the advantage of being omnidirectional, but it is also smaller and waterproof.
The specific aims of our study were to validate activity monitor counts against energy expenditure (EE) measured by room respiration calorimetry, activity counts by a microwave detector, and heart rate by telemetry; to compare performance of the monitors placed above the hip or on the lower leg; to test the monitors for range of response and sensitivity under field conditions; and to establish threshold counts for sedentary, light, moderate, and vigorous levels of physical activity in children.
Research Methods and Procedures
The activity monitors were tested in 26 children under controlled laboratory and field settings. The activity monitors were validated and calibrated against continuous 6-hour measurements of EE by respiration calorimetry, activity by microwave detector, and heart rate by telemetry in a room calorimeter. While they were in the calorimeter, the children adhered to a structured protocol of physical activities. Afterward, the children were taken to a nearby track and monitored throughout a series of outdoor measurements to test the performance of the activity monitors under field conditions. The protocol was approved by the Institutional Review Board for Human Subject Research for Baylor College of Medicine and Affiliated Hospitals. All subjects and their parents gave written informed consent.
Subjects were boys (n = 14) and girls (n = 12), with mean (±SD, range) ages 10.7 ± 2.9(6 to 16) years and 11.1 ± 2.9 (7 to 16) years, respectively. The ethnically diverse sample consisted of 16 white, 2 African American, 4 Hispanic, and 4 Asian children. The inclusion criteria required the children to be healthy and free from any medical condition that would limit participation in physical activity or exercise. The subjects were recruited from the Children's Nutrition Research Center community-based database of volunteers to achieve a balance across age and sex. The boys were 146 ± 17 (125 to 178) cm tall and weighed 41 ± 12 (22 to 70) kg. The girls were 145 ± 15 (122 to 169) cm tall and weighed 37 ± 10 (20 to 56) kg. Mean body mass index(BMI) was 18.7 ± 2.0 (13.7 to 22.2) kg/m2 for boys and 17.3 ± 2.1 (13.7 to 19.7) kg/m2 for girls, corresponding to Centers for Disease Control and Prevention (CDC) BMI Z scores of 0.42 ± 0.83 (−1.7 to 1.3) for boys and −0.32 ± 0.73(−1.5 to 0.8) for girls (13). None of the children were classified as overweight (≥95th BMI percentile) by the CDC growth charts.
Physical Activity Monitors
CSA Actigraph monitor (model 7164; Computer Science and Applications, Shalimar, FL) and the MM Actiwatch (model AW16; Mini-Mitter, Bend, OR) were evaluated. The CSA Actigraph uses a unidirectional accelerometer that measures accelerations in the vertical plane. The MM Actiwatch includes an omnidirectional accelerometer built from a cantilevered rectangular piezoelectric bimorph plate and seismic mass, which is sensitive to movement in all directions, but most sensitive in the direction parallel with the longest dimension of the case. The monitors were affixed above the iliac crest of the right hip with an elastic belt and adjustable buckle. Another set of monitors was affixed to the lateral compartment of the lower right leg at the head of the fibula with a small elastic band. The monitors were always oriented in the same direction. Counts were cumulated for 1 minute and the total was saved in the memory.
Room Respiration Calorimetry
Oxygen consumption (V̇o2) and carbon dioxide production(V̇co2) were measured continuously in a 30-m3 room calorimeter for 6 hours. The performance of the respiration calorimeters has been described in detail previously (14). Errors from 24-hour infusions of nitrogen and CO2 were −0.34 ± 1.24% for V̇o2 and 0.11 ± 0.98% for V̇cO2 (14). EE was computed using the Weir equation (15). Heart rate was recorded by telemetry (DS-3000; Fukuda Denshi, Tokyo, Japan) and physical activity was monitored by a Doppler microwave sensor (D9/50; Microwave Sensors, Ann Arbor, MI).V̇o2,V̇co2, EE, and heart rate were averaged at 1-minute intervals. METs were calculated in terms of the child's measured RMR. Threshold levels were defined not in terms of METs, but in terms of activity energy expenditure (AEE) computed as EE − RMR. Sedentary activities were defined as minimal body movements in the sitting or reclining position. Light activities reflected a low level of exertion in the standing position. Moderate activities involved medium exertion in the standing position. Vigorous activities were at a high level of exertion in the standing position. The specific activities in each category are described below.
Children entered the room calorimeter at 9:00 am and were monitored by staff at all times. The 6-hour calorimeter protocol included free time in which the children were allowed to do whatever they wished (i.e., watch television or videotapes, participate in arts and crafts, play games, or eat lunch), and the following scheduled measurements:
After a 20-minute equilibration period in a reclining position, the children were asked to remain still but awake. Television viewing was permitted. The children were monitored both visually and by the motion sensor to confirm that they were lying still (<50 counts/min) for the entire measurement. The children were in the fasted state for the RMR measurement.
Children played Nintendo for 20 minutes in a sitting position.
Arts and Crafts.
Children worked on art projects (painting, coloring, or beadwork) for 20 minutes, sitting on the floor.
Children played with a variety of toys involving small movements(cards, jacks, puzzles, Legos, miniature cars, and figurines) on a floor mat for 20 minutes.
Children performed warm-up exercises, as demonstrated on a videotape, for 10 minutes.
Children walked on a treadmill at 2.5 mph for 10 minutes.
Tae Bo Exercises.
Children performed Tae Bo martial arts exercises, as demonstrated on a videotape, for 10 minutes.
Children performed a variety of activities (basketball, paddleball, hula hoop, bouncing ball, tossing ball, darts, and jumping jacks) in a standing position for 20 minutes.
Treadmill speed was set at 3.5 mph for 6- to 7- year-old children, and at 4 mph for 8- to 16-year-old children for 10 minutes.
Treadmill speed was set at 4.5 mph for 6- to 7-year-old children; at 5 mph for 8- to 10-year-old children; and at 6 mph for 11- to 16-year-old children for 10 minutes. The treadmill speed was age-adjusted for the capability and safety of the children.
At the conclusion of the 6-hour calorimeter study, the children were escorted to a nearby track and monitored throughout the following field activities. Due to the summer heat, the time outside was restricted; however, steady-state counts were attained after 3 minutes.
Children were asked to jump rope for 3 minutes.
The children walked around the track at their own speed for 5 minutes.
The children skipped around the track at their own speed for 3 minutes.
The children jogged around the track at their own speed for 3 minutes.
The children kicked a soccer ball around cones for 3 minutes.
Data are summarized as means ± SD. Descriptive statistics, Pearson correlations, and multiple regression analyses were performed using Minitab (Release 13; Minitab, State College, PA). The sample size of 26 was based on detecting as statistically significant a correlation of 0.50 between counts and EE, and counts provided by the CSA and MM monitors with a power of 0.80 at a significance level of p ≤ 0.05.
Mean EE, AEE, MET, heart rate, and activity counts during the RMR and physical activities are summarized in Table 1. EE and AEE for a given activity are presented in terms of kcal/kg per minute. METs are calculated as EE for a given activity divided by the child's measured RMR.
Table 1. Energy expenditure, heart rate, and activity counts during physical activities
|RMR||0.026 ± 0.006||0||1||78 ± 15||23 ± 17||14 ± 17||6 ± 7||14 ± 17||15 ± 18|
|Nintendo||0.030 ± 0.006||0.005 ± 0.002||1.20 ± 0.11||82 ± 15||42 ± 33||61 ± 67||14 ± 12||92 ± 95||36 ± 33|
|Arts and crafts||0.037 ± 0.007||0.011 ± 0.004||1.44 ± 0.14||94 ± 14||174 ± 80||152 ± 171||43 ± 28||188 ± 191||75 ± 64|
|Play 1||0.040 ± 0.008||0.014 ± 0.005||1.57 ± 0.18||103 ± 12||265 ± 110||315 ± 208||69 ± 45||415 ± 452||152 ± 142|
|Aerobic warm-up||0.054 ± 0.010||0.028 ± 0.007||2.12 ± 0.32||106 ± 12||506 ± 152||518 ± 133||251 ± 110||954 ± 682||473 ± 268|
|Walk 1||0.071 ± 0.011||0.046 ± 0.008||2.90 ± 0.40||113 ± 16||1337 ± 75||2471 ± 608||940 ± 414||7272 ± 1317||2445 ± 655|
|Tae Bo||0.085 ± 0.018||0.059 ± 0.016||3.40 ± 0.86||129 ± 15||913 ± 150||1875 ± 859||867 ± 311||6398 ± 2146||2014 ± 585|
|Play 2||0.092 ± 0.018||0.066 ± 0.016||3.67 ± 0.73||132 ± 22||1149 ± 129||5886 ± 229||1287 ± 473||5249 ± 1998||1933 ± 622|
|Walk 2||0.096 ± 0.011||0.071 ± 0.010||4.04 ± 0.84||132 ± 19||1322 ± 108||4920 ± 1729||1667 ± 510||9282 ± 1574||3433 ± 1194|
|Jogging||0.154 ± 0.027||0.130 ± 0.027||6.72 ± 1.93||158 ± 39||1380 ± 108||9952 ± 5213||2647 ± 853||18404 ± 4970||4464 ± 1539|
The correlations among minute-to-minute EE or AEE, heart rate, microwave counts, and the counts given by CSA and MM over the 6-hour period were examined for each child in the calorimeter. Individual correlations were averaged for all children and are presented in Table 2. The mean correlations between EE or AEE and counts were slightly higher for the MM-hip (r = 0.78 ± 0.06) and MM-leg (r = 0.80 ± 0.05) than for the CSA-hip (r = 0.66 ± 0.08) and CSA-leg (r = 0.73 ± 0.07). EE or AEE was correlated highly with heart rate (r = 0.80 ± 0.10) and microwave activity counts in the calorimeter (r = 0.82 ± 0.04).
Table 2. Correlations between EE, AEE, heart rate, and microwave activity and counts provided by CSA and MM monitors
|EE or AEE (kcal/kg/min)||0.66 ± 0.08||0.78 ± 0.06||0.73 ± 0.07||0.80 ± 0.05|
|Heart rate (bpm)||0.57 ± 0.13||0.66 ± 0.13||0.63 ± 0.09||0.67 ± 0.12|
|Microwave activity (counts)||0.61 ± 0.07||0.76 ± 0.07||0.72 ± 0.06||0.83 ± 0.04|
The CSA and MM also were compared by correlation analysis (Table 3). The CSA and MM performed similarly on the hip (inter-instrument r = 0.88) and on the lower leg (inter-instrument r = 0.89). Agreement between the MM-hip and MM-leg placement (r = 0.93) was closer than between the CSA-hip and CSA-leg placement (r = 0.77).
Table 3. Correlations between CSA and MM counts at the hip and upper leg
|MM-hip (counts)||0.88 ± 0.05|| || |
|CSA-leg (counts)||0.77 ± 0.11||0.85 ± 0.07|| |
|MM-leg (counts)||0.82 ± 0.06||0.93 ± 0.04||0.89 ± 0.07|
EE for specific activities was regressed against counts given by the CSA and MM for all children. EE was significantly related to counts and age but not to sex. Age increased the r2(adj) for the prediction of EE from counts by 2% to 3%.
AEE for specific activities was regressed against counts given by the CSA and MM for all children. Age did not significantly alter the prediction of AEE from counts. The regression of AEE on counts was independent of sex and age; therefore, the regressions simplified to the following equations.
Predicting AEE from the combination of the counts from the hip and leg increased the r2(adj) to 86% for the CSA and 84% for the MM.
The linear regression of AEE on activity counts was used to define the threshold counts for the sedentary, light, moderate, and vigorous levels of physical activity (Table 4). The sedentary category was set at <0.015 kcal/kg per minute. The light category was set at ≥0.015 but <0.05 kcal/kg per minute. The moderate category was set at ≥0.05 but <0.10 kcal/kg per minute. The vigorous category was set at ≥0.10 kcal/kg per minute. The mean heart rates, which corresponded with 0.015, 0.05, and 0.10 kcal/kg per minute, were 90, 130, and 160 bpm.
Table 4. Threshold counts for CSA and MM monitors for sedentary, light, moderate, and vigorous levels of physical activity
|Sedentary (<0.015 kcal/kg/min)||<800||<100||<800||<200|
|Light (≥0.015 and <0.05 kcal/kg/min)||<3200||<900||<5100||<1800|
|Moderate (≥0.05 and <0.10 kcal/kg/min)||<8200||<2200||<12000||<4300|
|Vigorous (≥0.10 kcal/kg/min)||≥8200||≥2200||≥12000||≥4300|
The mean activity counts given by the CSA and MM monitors during field activities are displayed in Table 5. Jump roping, walking, skipping, and jogging would be classified similarly by CSA and MM, whether on the hip or leg; soccer would be classified as moderate by CSA-hip, and as vigorous by CSA-leg, MM-hip, and MM-leg.
Table 5. Counts given by CSA and MM monitors during field activities
|Jump rope||15415 ± 8051 Vigorous||2745 ± 1092 Vigorous||21353 ± 6744 Vigorous||4662 ± 1386 Vigorous|
|Walk||3465 ± 766 Moderate||1179 ± 349 Moderate||7364 ± 1581 Moderate||2959 ± 794 Moderate|
|Skip||15244 ± 3760 Vigorous||3225 ± 1221 Vigorous||25210 ± 6607 Vigorous||5782 ± 1724 Vigorous|
|Jogging||8804 ± 2057 Vigorous||3318 ± 1027 Vigorous||26562 ± 5540 Vigorous||5644 ± 1698 Vigorous|
|Soccer||7735 ± 2966 Moderate||2922 ± 904 Vigorous||20608 ± 6090 Vigorous||4578 ± 1415 Vigorous|
Accelerometer-based activity monitors proved to be valid and useful devices for the assessment of children's physical activity, defined as body movement produced by skeletal muscles resulting in EE. The high correlations between the activity counts and AEE and heart rate demonstrate that the CSA and MM monitors strongly reflect energy expended in activity. The activity monitors may be used to classify the time spent in sedentary, light, moderate, and vigorous levels of activity. Given the large standard error of the estimate (SEE) of the regression of EE or AEE on activity counts, the prediction of EE or AEE from CSA or MM activity counts, however, is inappropriate for individuals.
Validation studies of the CSA monitor have been performed against EE using portable indirect calorimeters in adults (3, 4, 5, 6, 7) and children (11). Janz (10) evaluated an earlier CSA model against heart rate monitoring in children, and Finn and Specker (12) evaluated the MM monitor against direct observation in pre-schoolchildren. Janz (10) validated an earlier accelerometer (CSA model 5032) against heart rate monitoring in 31 children 7 to 15 years of age for 6 days. The correlation between heart rate and CSA activity counts was 0.57. Trost et al. (11) reported high correlations (r = 0.86) between activity counts from two CSA monitors worn on each hip and EE measured by calorimetry while children walked at different speeds on a treadmill (r = 0.87). The SEE of the regression (0.97 kcal/min), however, was large. Finn and Specker (12) evaluated the MM in 40 children 3 to 4 years of age against 6 hours of direct observation using the Children's Activity Rating Scale. The within-child correlations between 3-minute MM counts and Children's Activity Rating Scale scores averaged 0.74.
In general, the validation studies have shown strong correlations between activity counts and EE under laboratory conditions, and weaker correlations under field conditions. The importance of evaluating monitor response to a variety of activities is now well-appreciated. Most studies indicate poor predictive validity of activity monitors for point estimates of EE for individuals (3, 11). No single regression equation accurately predicts EE for all activities. The relationship of counts with measured EE depends on the type of activity. Vertical acceleration of the body can be measured accurately with unidirectional accelerometers, but they do not accurately reflect external work performed in such activities as pushing or lifting objects, stair climbing, cycling, rowing, or resistance training. Accelerometer-based monitors tend to overestimate EE for activities with a small force:displacement ratio (jumping and running) and underestimate EE for activities with a large force:displacement ratio(stair climbing and knee bends). In our study, the omnidirectional design of the MM may account for its slightly better performance.
We validated the CSA and MM monitors under controlled laboratory conditions in a room calorimeter. Although confined, the children performed specific activities such as walking and jogging on a treadmill as well as other playtime activities in the sitting and standing position. The children also were allowed free time to participate in a variety of activities; thus, the calorimeter protocol was reflective of the broad range of gross and fine motor movements as well as the spontaneous, sporadic movements characteristic of children. We also observed the range of response of the activity monitors under field conditions in which the children performed more vigorous activities.
In general, the performances of the CSA and MM were comparable; the counts provided by the CSA and MM were highly correlated, on the order of 0.90. The individual 6-hour correlations averaged 0.78 and 0.80 for MM-hip and MM-leg, and 0.66 and 0.73 for CSA-hip and CSA-leg, respectively. Activity counts also correlated significantly with heart rate and microwave counts, independent indicators of physical activity.
Accelerometer output is influenced by the place of attachment on the body. AEE is a function of body acceleration and the mass of the body displaced; therefore, attachment to the center of the body mass seems to be an optimal choice. Melanson and Freedson (3) compared the correlations between EE on the treadmill and counts obtained with the CSA monitor on the ankle (r = 0.66), hip (r = 0.80), and wrist (r = 0.81). We found that placement of the activity monitor on the hip or the leg gave similar results for MM, and somewhat different results for the CSA. Subjectively, the children found the leg placement less comfortable and more likely to slip than the hip placement. Dual placement sites only marginally improved the predictability of AEE in this study and are not worth the increased cost, time, and effort.
Activity monitors can be used to measure the duration and frequency of activity bouts in different intensity categories. Janz (10) calibrated an earlier accelerometer (CSA model 5032) against heart rate monitoring in children. Heart rates of 75, 130, and 150 bpm were selected to represent sedentary, moderate, and vigorous activity levels. Melanson and Freedson (3) calibrated the CSA monitor (model 5032) against energy expended (kcal/min) at three speeds on the treadmill in 34 adults. Under these laboratory conditions, CSA counts were highly correlated with EE (r = 0.80), relativeV̇o2 per kilogram (r = 0.82), and heart rate (r = 0.66); however, the range of individual differences between actual and predicted values was large, ranging from −4.17 to 2.04 kcal/min.
The intensity categories of CDC-American College of Sports Medicine for physical activity (light: <3 METs; moderate: 3 to 6 METs; vigorous:>6 METs; 1 MET = 3.5 mL O2/kg per minute) have been used to define activity thresholds in adults (4, 7, 16). Freedson et al. (4) calibrated the CSA monitor in 50 adults using three speeds on the treadmill. Cross-validation of the prediction equation in 15 subjects showed no significant differences between actual and predicted EE (SEE = 0.50 to 1.40 kcal/min). In 25 adults, Hendelman et al. (16) developed regression equations to predict METs from counts obtained from the CSA monitor during walking only and walking plus performing recreational and household activities. Correlations between counts and METS were higher during treadmill walking (r = 0.77) than all activities combined (r = 0.59). Energy costs of golf and household activities were underestimated by 30% to 60% based on equations derived from level walking. The inherent limitations of accelerometry to detect upper body movement, load carriage, or changes in terrain were illustrated in field settings. In 70 adults, Swartz et al. (7) calibrated the CSA monitor on the hip and wrist using EE during several lifestyle activities: yard work, housework, family care, occupation, recreation, and conditioning. The correlation between counts and EE was moderate for the hip (r = 0.56) and low for the wrist (r = 0.18). The addition of the data from the wrist accelerometer to the data obtained by the hip accelerometer resulted in only minor improvement in the prediction of EE and was considered not worth the extra personnel time and instrument cost (7).
The energy costs of resting metabolism (1 MET = 3.5 mL O2/kg per minute) and activity in adults are not applicable to children, and their use creates substantial bias in the estimation of EE in children (17). A further impediment is that the bias is age-dependent, which severely limits comparisons of physical activity across different age groups. Sallis et al. (17) found that the child-to-adult ratio of energy cost of submaximal work decreases from 1.37 at 5 years of age to 1.03 at 17 years of age, indicating that the energy cost of activity would be underestimated by ∼40%, 20%, and 5% in 5-, 10-, and 16-year-old children, respectively. The higher energy cost of activity at younger ages has been attributed to the higher RMR, higher stride frequency, and less efficient economy of locomotion in younger children. The use of the adult MET value of 3.5 mL O2/kg per minute to define thresholds in children whose resting EEs range from 6 to 3.5 mL O2/kg per minute would create an erroneous age-dependent bias in the prediction equation.
We defined threshold levels not in terms of METs, but in terms of the energy expended in activity (AEE), our parameter of interest. We found that the relationship between EE and activity counts was dependent on age but not on sex. The relationship between AEE and activity counts was independent of age and sex. Activity monitors can be used to assess not only moderate to vigorous physical activity, but also physical inactivity. In establishing a sedentary threshold level, we distinguished sedentary from light activities.
In conclusion, the validation of the CSA and MM monitors against AEE, and their calibration for sedentary, light, moderate, and vigorous levels of physical activity certify these monitors as valid, useful devices for the assessment of physical activity in children.
This work is a publication of the U.S. Department of Agriculture/Agricultural Research Service Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine and Texas Children's Hospital, Houston, TX. This project has been funded in part with federal funds from the U.S. Department of Agriculture/Agricultural Research Service under Cooperative Agreement 58-6250-6-001. The contents of this publication do not necessarily reflect the views or policies of the U.S. Department of Agriculture, and mention of trade names, commercial products, or organizations does not imply endorsement by the U.S. Government. We thank the children who participated in this study and Carolyn Heinz for study coordination, Sopar Seributra for nursing, Sandra Kattner for dietary support, Idelle Tapper for secretarial assistance, and Leslie Loddeke for editorial work.