Accurate assessment of physical activity is necessary to determine children's physical activity level, to improve our understanding of the dose-response relationship between physical activity and health, and to evaluate intervention programs designed to increase physical activity. Physical activity has traditionally been measured with self-reports. Self-reports are easily administered, low-cost measurements. However, they do not capture the sporadic short-burst nature of children's physical activity very well (2). Self-reports tend to overestimate time spent in vigorous physical activities and to underestimate time spent in unstructured daily physical activities (1). At present, motion sensors are being used with increasing regularity to assess physical activity. Motion sensors are lightweight, unobtrusive, and relatively inexpensive compared to other objective methods, such as indirect calorimetry or doubly labeled water (DLW). In the past decades, motion sensors have evolved from simple mechanical devices to three-axial accelerometers that can be used to assess physical activity or to estimate energy expenditure (EE).
As motion sensor-based research evolved, researchers have validated and calibrated them in diverse populations, including youth. In 2005, Freedson et al. (11) reviewed the validity of the ActiGraph, Actiwatch, Actical, and RT3 accelerometers against oxygen consumption or EE in children. They concluded that none of the four accelerometers could be pointed out as superior to the other ones. Furthermore, there was no evidence to suggest that three-axial or omnidirectional accelerometers were better than uniaxial accelerometers in capturing children's physical activity. Similar conclusions can be drawn from the reviews of Trost et al. (44) and Rowlands (33). However, all three reviews focused on accelerometers. They did not include pedometers. Pedometers may be a cheap alternative to accelerometers for the objective assessment of children's physical activity. Furthermore, the three reviews did not take the quality of the studies into account. The reproducibility and validity of a motion sensor may depend on the sample and the setting in which the motion sensor is tested (38). In 2004, de Vries et al. (9) reviewed 35 studies on the reproducibility, validity, and feasibility of two pedometers (Digi-Walker and Pedoboy) and seven accelerometers (LSI, Caltrac, ActiGraph, Actiwatch, Tritrac-R3D, RT3, and Tracmor2) in youth (2-18 yr). The outcomes of the studies were weighted for their quality using a checklist. It was concluded that there was strong evidence for good reproducibility of the Caltrac in adolescents (12-18 yr), poor reproducibility of the Digi-Walker in children (8-12 yr), good validity of the ActiGraph in children and adolescents (8-18 yr), and good validity of the Tritrac-R3D in children (8-12 yr). In 2004, several pedometers and accelerometers of which the clinimetric properties in certain age groups were unknown were still there. For example, evidence for the reproducibility of three-axial accelerometers in assessing physical activity in youth was missing, and there was no information about the reproducibility of motion sensors in preschool children (2-4 yr).
Meanwhile, as the number of studies in this area has increased dramatically in the last few years, the remaining questions might have been addressed. Furthermore, the technology of existing motion sensors has advanced (e.g., improved battery life, increased memory size, USB connection, waterproof case, smaller size, improved software packages to analyze the data) and new motion sensors have been developed. Therefore, this review will provide an update of the published evidence for the reproducibility, validity, and feasibility of pedometers and accelerometers used to assess physical activity in youth. Important methodological issues are discussed, and priorities for future research are identified.
METHODS
Literature search.
In October 2007, relevant English articles were searched in PubMed (Medline), EMBASE, and PsycINFO using similar MeSH terms and text words as in the previous review in October 2004 (9). Table 1 presents the search strategy (including the number of total hits per database). Studies (written as full reports) were included in the review if: 1) their main purpose was the clinimetric evaluation of a motion sensor for assessing physical activity in healthy children and adolescents (2-18 yr); and 2) they were published between October 2004 and October 2007. Because there is a certain time delay between publication of a study and its recording in a bibliographic database, all studies that were published between January 2004 and October 2004 and were not described in the previous review (9) were also included. References of retrieved articles were screened for additional relevant studies.
;)
TABLE 1:
Literature search.
Clinimetric evaluation.
All studies were evaluated independently by two authors (S.d.V. and H.v.H.) using a 20-item checklist. The checklist has been described in more detail in the previous review in 2004 (9). It contained four items on study design, six items on reproducibility, six items on validity, and four items on feasibility of the motion sensor under study (see also Appendix) (9). A clear description of the study design includes characteristics of the sample, measurements, protocol, and statistical analyses. All quality-related items were scored 0, 0.5, or 1 point, and summed per study (see Appendix). Reproducibility of a motion sensor concerns two concepts: intrainstrument reliability (i.e., test-retest reliability) and interinstrument reliability (e.g., between a motion sensor worn on the left and the right hip). To determine whether a motion sensor measures what it is intended to measure requires evidence of criterion validity or construct validity. Criterion validity is the most powerful type of validity and requires testing of a motion sensor against the criterion standard. Direct observation and DLW are considered criterion standards for assessing physical activity and EE, respectively. Construct validity is defined as the extent to which measurements made with the motion sensor under study are associated with those made with other methods that are used to measure the same or similar aspects. Feasibility of a motion sensor is defined as the costs (including software), required expertise, and acceptability of the device. Indicators of acceptability are the level of comfort, tolerance, refusal, and the amount of missing or lost data due to malfunctioning of the motion sensor. Adequate outcome measures of reproducibility, validity, and feasibility were rated as poor (−), moderate (±), or good (+) (see Appendix). Calculation of the intraclass correlation coefficient (ICC) was considered an adequate method to quantify intra or interinstrument reliability. An ICC ≥ 0.70 was considered to reflect good reproducibility. Criterion validity was rated as good if the correlation coefficient was 0.75 or higher and construct validity was rated as good if the correlation coefficient was 0.60 or higher (9). The two authors agreed on all ratings. Ideally, a motion sensor is reproducible, valid, sensitive, feasible, and cheap.
RESULTS
The literature search identified 185 articles, of which 21 were selected. Reference tracking resulted in 11 additional articles. Thus, this review included 32 articles published between January 2004 and October 2007. The articles describe the clinimetric properties of 12 motion sensors: three pedometers (Digi-Walker, Walk4Life, and Sun TrekLINQ), one uniaxial accelerometer (ActiGraph), two two-axial accelerometers (BioTrainer and StepWatch Activity Monitor), two omnidirectional accelerometers (Actiwatch and Actical), and four three-axial accelerometers (Tritrac-R3D, RT3, ActivTracer, and Mini-Motionlogger). The general characteristics of these motion sensors are presented in Table 2 and summarized below followed by an evaluation and comparison of their clinimetric properties.
TABLE 2:
General characteristics of 12 motion sensors.
General Characteristics
Pedometers.
Three pedometers were reviewed: the Digi-Walker (SW-series; Yamax Co., Yasama Corp., Tokyo, Japan), Walk4Life (Plainfield, IL), and Sun TrekLINQ (Arvada, CO). In general, pedometers measure the total number of steps during the entire assessment period. They use a horizontal spring-suspended lever arm that moves up and down in response to the vertical displacement of the hip. With each deflection detected, an accumulated step count is displayed digitally on a screen. Most pedometers cannot store data and do not provide information about a child's physical activity pattern in terms of frequency, intensity, duration, and type of activities at a certain point in time. The three pedometers presented in Table 2 range in price between $17 and $25.
Accelerometers.
In contrast to pedometers, accelerometers are more complex electronic devices that measure accelerations produced by a body segment. With each movement, body segments accelerate and decelerate. Electric transducers and microprocessors convert the accelerations into a digital signal. Most accelerometers use a horizontal cantilever beam with a weight on the end that compresses a piezoelectric crystal when subjected to movement. Each movement generates a voltage proportional to the acceleration. Accelerometers can store data during a specified period. Recent improvements in battery life and memory size have increased data storage capacities. Nowadays, there are accelerometers that can measure up to 356 d at 1-min epochs. Most accelerometers record activity counts, time of day, and EE. In this review, nine accelerometers were evaluated (Table 2). Prices range between $300 and $4700.
The ActiGraph (ActiGraph, Pensacola, FL), formerly known as the CSA, MTI, or WAM activity monitor, is the only uniaxial accelerometer. It measures acceleration in the vertical plane. At present, this is the most widely used motion sensor for physical activity research.
The BioTrainer (IM Systems, Baltimore, MD) and StepWatch Activity Monitor (CymaTech, Mountlake Terrace, WA) are two-axial accelerometers. The BioTrainer contains one accelerometer that is positioned at an angle of 45°. It measures acceleration in both the vertical and the horizontal plane (12). The StepWatch uses a two-plane sensor (5) and is worn on the ankle.
The Actiwatch (Mini Mitter Co., Inc., Bend, OR) and Actical (Mini Mitter Co., Inc., Bend, OR) are omni-directional accelerometers. These accelerometers are most sensitive in the vertical plane but are also sensitive to movement in other directions. The output is a composite of the different signals.
The Tritrac-R3D (Hemokinetics, Inc./Professional Products, Division of Reining International, Ltd., Madison, WI), RT3 (Stayhealthy, Inc., Monrovia, CA), ActivTracer (GMS, Tokyo, Japan), and Mini-Motionlogger (Ambulatory Monitoring Inc., Ardsley, NY) are three-axial accelerometers. They measure acceleration in three planes (anterior-posterior, vertical, and mediolateral) and provide information about activity counts for each of the three dimensions, a three-axial vector representation of these counts (vector magnitude), and EE. Stayhealthy, Inc. has purchased the technology and rights of the Tritrac-R3D and repackaged it into a smaller device called the RT3. The Tritrac-R3D and RT3 differ in the number of accelerometers incorporated. The Tritrac-R3D uses three uniaxial accelerometers placed side-to-side, whereas the RT3 uses one integrated three-axial accelerometer.
Reproducibility
Of the 32 selected studies, 7 examined the intrainstrument reliability and 7 examined the interinstrument reliability of the motion sensors. Important outcomes are discussed per type of motion sensor (Table 3).
TABLE 3:
Reproducibility of motion sensors in 11 studies.
Pedometers.
Three studies examined the intrainstrument reliability and three the interinstrument reliability of the Digi-Walker. As can be seen in Table 3, the intrainstrument reliability of the Digi-Walker was high in all three studies (16,32,39). It increased with an increasing number of monitoring days: ICCs ranged from 0.69 for 2 d to 0.83 for 6 d of monitoring (32,39). Jago et al. (16) showed that the number of pedometer counts during two occasions of walking, fast walking, and running revealed no significant differences. ICCs ranged from 0.51 to 0.92 across all activities; however, the correlations were weaker for running (ICC = 0.51-0.77) than for walking (ICC = 0.75-0.89) and fast walking (ICC = 0.61-0.92). The interinstrument reliability of the Digi-Walker was also high (3,4,16). ICCs ranged between 0.73 and 0.96 for Digi-Walkers placed on the left hip, right hip, and center of the waist (3,4,16). Comparable results were found for the interinstrument reliability of the Walk4Life and Sun TrekLINQ pedometers. The correlation between left and right hip counts was ICC = 0.92 and ICC = 0.84, respectively (4).
Accelerometers.
The intrainstrument reliability has only been studied for one accelerometer, i.e., the ActiGraph (Table 3). In common with the Digi-Walker, the intrainstrument reliability of the ActiGraph increased with an increasing number of monitoring days. ICCs ranged from 0.45 for 1 d to 0.90 for 8 d of monitoring (13,20). Mattocks et al. (20) examined the seasonal and intraindividual variation of ActiGraph counts among 11- to 12-yr-old children. The children wore an ActiGraph for 7 d four times throughout the course of a year. Overall, ICC was moderate (ICC = 0.54). After adjustment for month of measurement, it increased from 0.49 to 0.53, indicating a small effect of month. The children tended to be less active in winter months. Toschke et al. (41) found low to moderate intrainstrument reliability of the ActiGraph among 3- to 5-yr old children (r = 0.31-0.51).
Garcia et al. (12) investigated the interinstrument reliability of the BioTrainer, ActiGraph, and Mini-Motionlogger among 6- to 10-yr-old children. The children wore two of each type of motion sensor on the hips during walking, running, and jumping on a force plate and during walking, jogging, and running on a treadmill. The interinstrument reliability of the BioTrainer was moderate (ICC = 0.64; ICC = 0.69). Correlations between left and right hip counts were much higher for the Mini-Motionlogger (ICC = 0.98) (12) and ActiGraph (ICC = 0.77-1.0) (12,37,42).
None of the recently published studies reported on the reproducibility of the StepWatch, Actiwatch, Actical, Tritrac-R3D, RT3, and ActivTracer in youth. However, the reproducibility of the Actiwatch and Tritrac-R3D has been described in earlier studies included in the review in 2004 (9).
Validity
All 12 motion sensors have been validated in youth in one or more studies. The criterion validity was examined in 12 studies and the construct validity in 23. The results of these studies are summarized per type of motion sensor (Table 4).
TABLE 4:
Validity of motion sensors in 28 studies.
Pedometers.
The validity of the Digi-Walker has been examined in 10 studies (Table 4). Five studies reported the criterion validity and six reported the construct validity. A high correlation was found between Digi-Walker step counts and activity EE (AEE) derived from DLW (r = 0.67) (30). Correlation between observed steps and Digi-Walker step counts was moderate to high (r = 0.59-0.90) in laboratory and field studies among 3- to 11-yr-old children (4,14,22,25). Agreement between observed steps and Walk4Life step counts was also high (ICC = 0.83) (4). This was somewhat lower (ICC = 0.64) for the Sun TrekLINQ pedometer (4). Digi-Walker step counts were highly correlated with uniaxial and three-axial accelerometer counts (r = 0.60-0.88) (7,16,30,34). Correlation with self-report measures of physical activity was low to moderate (r = 0.04-0.39) (8,39).
Accelerometers.
The validity of the ActiGraph has been examined in 10 studies involving children of all age groups (2-18 yr) (Table 4). It has been validated by direct observation (14,17,37), DLW (24,31), indirect calorimetry (12,26,35,42,46), other accelerometers (12,17), and ground reaction force (12) as reference methods. Results on the criterion validity of the ActiGraph are contradictory. Correlation between ActiGraph counts and observed activity was moderate to high (r = 0.52-0.77) (14,17,37). On the other hand, ActiGraph outcomes were poorly correlated (r = 0.22-0.33) with EE derived from DLW in young children (3-7 yr) (24). In another study using DLW, it was concluded that the ActiGraph cannot be used to estimate EE in young children (3-6 yr) (31). A similar conclusion can be drawn from the study of Trost et al. (46) in which three ActiGraph EE equations were evaluated against indirect calorimetry in 10- to 18-yr-old children. There was more agreement on the construct validity of the ActiGraph. All correlations were moderate to high. Garcia et al. (12) found a moderate correlation (r = 0.51) between ActiGraph counts and vertical ground reaction force during four ambulatory and jumping tasks on a force plate. Correlations ranged from r = 0.66 to r = 0.85 using indirect calorimetry as reference method (12,26,35,42,46) and from r = 0.36 to r = 0.92 using multiaxial accelerometers (12,17).
The validity of the BioTrainer was examined in one study (12). The outcomes of the BioTrainer were moderately to highly correlated (r = 0.52-0.74) with oxygen consumption, ground reaction force, ActiGraph outcomes, and Mini-Motionlogger outcomes in school-age children. The StepWatch has been validated in a sample of 6- to 20-yr-old children (21). During a 10-min walk, this motion sensor had an agreement of almost 100% with observed steps. The StepWatch correlated moderately (r = 0.49) with heart rate in a field setting (21).
Three studies investigated the validity of the Actiwatch. These studies showed mixed results. In a field study among 4- to 6-yr-old children, Actiwatch activity counts were poorly correlated (r = 0.27) with total EE assessed with DLW (18). Correlations with EE were much higher (r = 0.79-0.85) in a study among 7- to 18-yr olds in a respiratory room calorimeter (29). Actiwatch activity counts were also highly correlated with heart rate (r = 0.63) and Actical counts (r = 0.93) in this study (29). The correlation with ActiGraph counts was much lower (r = 0.36) in preschool children (17). The Actiwatch was also poorly correlated (r= 0.16) with direct observation of physical activity in this study (17). The validity of the Actical has also been investigated in three studies. Puyau et al. (29) found a high correlation between Actical counts and EE (r = 0.83-0.87) and heart rate (r = 0.60) among children and adolescents (7-18 yr). Similar conclusions can be drawn from the study of Pfeiffer et al. (27) in preschool children. They found a high correlation (r = 0.89) between Actical counts and oxygen consumption. Heil (15) concluded that the Actical can be used to predict AEE for groups of children.
The validity of the Tritrac-R3D has been evaluated in two studies involving children and adolescents (7-18 yr). High correlations were found between Tritrac-R3D outcomes and AEE derived from DLW (r = 0.81/0.85) (30). Tritrac-R3D outcomes were also highly correlated with oxygen consumption (r = 0.73) (23) and Digi-Walker step counts (r = 0.88) (30). The validity of the RT3 has been examined in a physical education lesson. There was a high correlation (r= 0.77-0.78) between RT3 outcomes and observed moderate to vigorous physical activity in school-age children (36). In a laboratory study among 5- to 6-yr-old children, high correlations (r = 0.86-0.93) were found between ActivTracer counts and EE for nine activities (40). Garcia et al. (12) examined the validity of the Mini-Motionlogger. Mini-Motionlogger counts were moderately to highly correlated (r = 0.46-0.92) with oxygen consumption, ground reaction force, ActiGraph counts, and BioTrainer counts in school-age children (12).
Feasibility
About half (n = 17) of the selected studies reported on the feasibility of the motion sensors in terms of the amount of missing or lost data or refusal rate. With two exceptions (30,37), the percent of missing or lost data or refusal rate was acceptable (<15%) in all studies (7,8,12,13,17,18,20,21,24,26,32,34,35,39,41). Cardon and de Bourdeauhuij (7,8) asked 4- to 5-yr-old children and 6- to 12-yr-old children how they felt about wearing a pedometer and/or accelerometer. In both age groups, more than 80% of the children found it very pleasant or pleasant to wear a pedometer and/or accelerometer. None of the children found it very unpleasant.
Comparative Ratings
To put the results of the 32 studies on the reproducibility, validity, and feasibility of the 12 motion sensors in perspective, all studies were rated using a 20-item checklist, i.e., 14 items concerning the quality of the study design and 6 items concerning the outcomes of the studies (9). All quality-related items were scored 0, 0.5, or 1 point and were summed per study. Adequate outcome measures of reproducibility, validity, and feasibility were rated as poor (−), moderate (±), or good (+). In general, the quality of the studies and, therefore, the level of evidence for the reported outcomes, was modest (mean = 6.5 ± 1.4 of 14 points, range = 4-9.5). One of the 32 studies was of poor quality (<4.5 points) (14), and three studies were of good quality (≥9 points) (17,21,37).
The outcomes of the 32 studies were weighted for their quality and summarized per type of motion sensor for each age group (Table 5). Studies of poor quality provide less evidence for the reported outcomes than studies of good quality. For example, the "high-quality" study of Sirard et al. (37) provided strong evidence for good interinstrument reliability (ICC = 0.84) (+++) and moderate criterion validity (r = 0.58) (±±±) of the ActiGraph, whereas the "low-quality" study of Hands et al. (14) provided limited evidence for good criterion validity (r = 0.90) (+) of the Digi-Walker. Table 5 shows that the ActiGraph is the only motion sensor in which the reproducibility and validity have been examined in all age groups (2-18 yr). There is strong evidence for good intra- and interinstrument reliability of the ActiGraph in preschool children (2-4 yr), moderate criterion validity in preschool children (2-4 yr) and young children (4-8 yr), and good construct validity in preschool children (2-4 yr). Furthermore, there is strong evidence for moderate construct validity of the StepWatch in children and adolescents (4-18 yr). There are currently no (new) data on the reproducibility of the StepWatch, Actiwatch, Actical, Tritrac-R3D, RT3, and ActivTracer in assessing physical activity in healthy children or adolescents. It is further shown that there is not much evidence for the reproducibility of motion sensors in preschool children (2-4 yr) and adolescents (12-18 yr) and for the criterion validity of motion sensors in adolescents (12-18 yr).
TABLE 5:
Level of evidence for the reproducibility and validity of motion sensors per age group.
DISCUSSION
This study reviewed 32 recently published studies on the reproducibility, validity, and feasibility of three pedometers and nine accelerometers that are used to assess physical activity in youth. Various reference methods have been used to validate the 12 motion sensors, ranging from self-reports to DLW and direct observation. In general, motion sensors seem to be a valid method to assess physical activity in youth. Care should be taken when using motion sensors to predict EE because significant underestimation and overestimation of EE have been reported compared to DLW (18,24,31,46). In the 32 studies, there was less evidence on the reproducibility and feasibility of the 12 motion sensors, but most motion sensors seem to be reliable and feasible in assessing physical activity in youth. The Digi-Walker is the most studied pedometer in youth. The ActiGraph is the most studied accelerometer. For both motion sensors, there is extensive evidence for moderate to good reproducibility and validity in almost all age groups (2-18 yr).
Compared to our review in 2004 (9), there is considerably more information on the reproducibility and validity of the Digi-Walker in youth. Evidence for the clinimetric quality of the ActiGraph further increased since 2004. Evidence for the reproducibility of multiaxial accelerometers is still limited and not many studies have focused on the reproducibility of motion sensors in preschool children (2-4 yr) or adolescents (12-18 yr) or on the criterion validity of motion sensors in adolescents (12-18 yr). In youth, clinimetric properties obtained in a certain age group cannot be generalized to other age groups, because of the differences in their physical activity pattern and anthropometrics (e.g., leg length) (38). Although there are still several pedometers and accelerometers of which the clinimetric properties in certain age groups are unknown, most motion sensors seem to work reasonably well in youth. Our results are in line with previous reviews of Bjornson and Belza (6), Freedson et al. (11), Trost et al. (44), and Rowlands (33). They concluded that although there is some evidence that certain motion sensors perform better than others under certain conditions, the reported differences are not consistent or sufficiently compelling to single out one brand or type of motion sensor as being superior to the others. The choice of an appropriate motion sensor therefore remains an issue of cost per unit, monitor size, battery life, memory size, technical support, software packages to download and analyze the data, and comparability of findings with other studies. Low cost and ease of use are the primary advantages of pedometers over accelerometers. However, they only provide a total score of activity or EE during the entire assessment period and do not provide information about children's physical activity pattern.
Besides selection of a motion sensor, other considerations include the site of placement of the motion sensor on the body, the sampling interval, the number of monitoring days, instruction, and blinding. According to Trost et al. (44) accelerometers are best placed on the hip or lower back. It is recommended to use a short sampling interval (<1 min) to capture children's physical activity pattern (2,33,44). At least a 7-d period is required to reliably assess children's habitual physical activity (45). Rowlands (33) emphasizes the need to assess both weekend days, because in Western countries, there are differences in children's physical activity pattern between Saturday and Sunday. Compliance with the monitoring protocol can be increased by reminder phone calls or text messages, flyers on refrigerators, activity logs, lists of frequently asked questions, examples of accelerometer output, and incentives (44). Wearing a motion sensor day and night might also increase compliance (33). Face-to-face distribution and collection of motion sensors is the best option for field-based research, but delivery and return by mail have also been successful (e.g., in the National Health and Nutrition Survey) (43).
To date, there is no standardized method for cleaning, analyzing, and reporting accelerometer data. This hampers comparison between studies. Data processing decisions, such as the definition of spurious data, the definition of a valid day, the choice of count cutoffs for moderate to vigorous physical activity, the choice of EE equations, and the manner in which missing data is dealt with, all influence the outcomes (10,19). In most studies, the raw acceleration signal is converted into activity counts. Total or mean activity counts per day and minutes per day spent above a certain intensity threshold are reported. This does not value the richness of accelerometer data. More sophisticated approaches to data processing, such as pattern recognition-based approaches (e.g., quadratic discriminant analysis, hidden Markov model), can improve the accuracy of accelerometer measurements (11,28). With these methods, different types of activities can be distinguished.
Our study has some limitations. The study was hampered by the use of various samples, test protocols, reference methods, and outcome measures. Although this diversity helps building the evidence for the clinimetric properties of motion sensors, it complicates comparison between studies and between motion sensors. Furthermore, there are no standardized criteria to evaluate and compare the quality of clinimetric studies and their outcomes. The criteria we used may be disputed. However, our intention was not to develop a standardized evaluation checklist but to provide information on the clinimetric quality of different motion sensors in a standardized manner. Taking these limitations into account, from this update, it can be concluded that:
- Although there are still several pedometers and accelerometers of which the clinimetric properties in certain age groups are unknown, most motion sensors seem to be reproducible, valid, and feasible in assessing physical activity in youth (2-18 yr).
- There is no information on the criterion validity of motion sensors in adolescents (12-18 yr).
- There is limited information on the reproducibility of multiaxial accelerometers in youth (2-18 yr).
- There is limited information on the reproducibility of motion sensors in preschool children (2-4 yr) and adolescents (12-18 yr).
Because the technology of motion sensors is still improving, we can expect them to continue to change. Researchers are encouraged to report the clinimetric properties of (new) devices, although not without improving the quality of the reported information. Freedson et al. (11) formulated seven guidelines to improve the quality of calibration studies. Furthermore, consensus about collecting, cleaning, analyzing, and reporting motion sensor data is urgently needed.
The study was funded by the Dutch Ministry of Health, Welfare, and Sport.
The results of the present study do not constitute endorsement by ACSM.
Conflict of interest: There are no conflicts of interest.
REFERENCES
1. Armstrong N, Welsman JR. The physical activity patterns of European youth with reference to methods of assessment.
Sports Med. 2006;36(12):1067-86.
2. Baquet G, Stratton G, van Praagh E, Berthoin S. Improving physical activity assessment in prepubertal children with high-frequency accelerometry monitoring: a methodological issue.
Prev Med. 2007;44(2):143-7.
3. Barfield JP, Rowe DA, Michael TJ. Interinstrument consistency of the Yamax Digiwalker pedometer in elementary school-aged children.
Meas Phys Educ Exerc Sci. 2004;8(2):109-16.
4. Beets MW, Patton MM, Edwards S. The accuracy of pedometer steps and time during walking in children.
Med Sci Sports Exerc. 2005;37(3):513-20.
5. Bjornson KF. Physical activity monitoring in children and youths.
Pediatr Phys Ther. 2005;17(1):37-45.
6. Bjornson KF, Belza B. Ambulatory activity monitoring in youth: state of the science.
Pediatr Phys Ther. 2004;16(2):82-9.
7. Cardon G, de Bourdeauhuij I. Comparison of pedometer and accelerometer measures of physical activity in preschool children.
Pediatr Exerc Sci. 2007;19(2):205-14.
8. Cardon G, de Bourdeauhuij I. A pilot study comparing pedometer counts with reported physical activity in elementary school children.
Pediatr Exerc Sci. 2004;16:355-67.
9. de Vries SI, Bakker I, Hopman-Rock M, Hirasing RA, van Mechelen W. Clinimetric review of motion sensors in children and adolescents.
J Clin Epidemiol. 2006;59(7):670-80.
10. Esliger DW, Copeland JL, Barnes JD, Tremblay MS. Standardizing and optimizing the use of accelerometer data for free-living physical activity monitoring.
J Phys Act Health. 2005;(3):366-83.
11. Freedson P, Pober D, Janz KF. Calibration of accelerometer output for children.
Med Sci Sports Exerc. 2005;37(11 suppl):S523-30.
12. Garcia AW, Langenthal CR, Angulo-Barroso RM, Gross MM. A comparison of accelerometers for predicting energy expenditure and vertical ground reaction force in school-age children.
Meas Phys Educ Exerc Sci. 2004;8(3):119-44.
13. Garnier D, Benefice E. Reliable method to estimate characteristics of sleep and physical inactivity in free-living conditions using accelerometry.
Ann Epidemiol. 2006;16(5):364-9.
14. Hands B, Parker H, Larkin D. Physical activity measurement methods for young children: a comparative study.
Meas Phys Educ Exerc Sci. 2006;10(3):203-14.
15. Heil DP. Predicting activity energy expenditure using the Actical activity monitor.
Res Q Exerc Sport. 2006;77(1):64-80.
16. Jago R, Watson K, Baranowski T, et al. Pedometer reliability, validity and daily activity targets among 10- to 15-year-old boys.
J Sports Sci. 2006;24(3):241-51.
17. Kelly LA, Reilly JJ, Fairweather SC, Barrie S, Grant S, Paton JY. Comparison of two accelerometers for assessment of physical activity in preschool children.
Pediatr Exerc Sci. 2004;16:324-33.
18. Lopez-Alarcon M, Merrifield J, Fields DA, et al. Ability of the actiwatch accelerometer to predict free-living energy expenditure in young children.
Obes Res. 2004;12(11):1859-65.
19. Masse LC, Fuemmeler BF, Anderson CB, et al. Accelerometer data reduction: a comparison of four reduction algorithms on select outcome variables.
Med Sci Sports Exerc. 2005;37(11 suppl):S544-54.
20. Mattocks C, Leary S, Ness A, et al. Intraindividual variation of objectively measured physical activity in children.
Med Sci Sports Exerc. 2007;39(4):622-9.
21. McDonald CM, Widman L, Abresch RT, Walsh SA, Walsh DD. Utility of a step activity monitor for the measurement of daily ambulatory activity in children.
Arch Phys Med Rehabil. 2005;86(4):793-801.
22. McKee DP, Boreham CAG, Murphy MH, Nevill AM. Validation of the Digiwalker pedometer for measuring physical activity in young children.
Pediatr Exerc Sci. 2005;17:345-52.
23. McMurray RG, Baggett CD, Harrell JS, Pennell ML, Bangdiwala SI. Feasibility of the Tritrac R3D accelerometer to estimate energy expenditure in youth.
Pediatr Exerc Sci. 2004;16:219-30.
24. Montgomery C, Reilly JJ, Jackson DM, et al. Relation between physical activity and energy expenditure in a representative sample of young children.
Am J Clin Nutr. 2004;80(3):591-6.
25. Oliver M, Schofield GM, Kolt GS, Schluter PJ. Pedometer accuracy in physical activity assessment of preschool children.
J Sci Med Sport. 2007;10(5):303-10.
26. Pate RR, Almeida MJ, McIver KL, Pfeiffer KA, Dowda M. Validation and calibration of an accelerometer in preschool children.
Obesity (Silver Spring). 2006;14(11):2000-6.
27. Pfeiffer KA, McIver KL, Dowda M, Almeida MJ, Pate RR. Validation and calibration of the Actical accelerometer in preschool children.
Med Sci Sports Exerc. 2006;38(1):152-7.
28. Pober DM, Staudenmayer J, Raphael C, Freedson PS. Development of novel techniques to classify physical activity mode using accelerometers.
Med Sci Sports Exerc. 2006;38(9):1626-34.
29. Puyau MR, Adolph AL, Vohra FA, Zakeri I, Butte NF. Prediction of activity energy expenditure using accelerometers in children.
Med Sci Sports Exerc. 2004;36(9):1625-31.
30. Ramirez-Marrero FA, Smith BA, Sherman WM, Kirby TE. Comparison of methods to estimate physical activity and energy expenditure in African American children.
Int J Sports Med. 2005;26(5):363-71.
31. Reilly JJ, Kelly LA, Montgomery C, et al. Validation of Actigraph accelerometer estimates of total energy expenditure in young children.
Int J Pediatr Obes. 2006;1(3):161-7.
32. Rowe DA, Mahar MT, Raedeke TD, Lore J. Measuring physical activity in children with pedometers: reliability, reactivity and replacement of missing data.
Pediatr Exerc Sci. 2004;16:343-54.
33. Rowlands AV. Accelerometer assessment of physical activity in children: an update.
Pediatr Exerc Sci. 2007;19(3):252-66.
34. Rowlands AV, Eston RG. Comparison of accelerometer and pedometer measures of physical activity in boys and girls, ages 8-10 years.
Res Q Exerc Sport. 2005;76(3):251-7.
35. Schmitz KH, Treuth M, Hannan P, et al. Predicting energy expenditure from accelerometry counts in adolescent girls.
Med Sci Sports Exerc. 2005;37(1):155-61.
36. Scruggs PW, Beveridge SK, Clocksin BD. Tri-axial accelerometry and heart rate telemetry: relation and agreement with behavioral observation in elementary physical education.
Meas Phys Educ Exerc Sci. 2005;9(4):203-18.
37. Sirard JR, Trost SG, Pfeiffer KA, Dowda M, Pate RR. Calibration and evaluation of an objective measure of physical activity in preschool children.
J Phys Act Health. 2005;3:324-36.
38. Stone MR, Esliger DW, Tremblay MS. Comparative validity assessment of five activity monitors: does being a child matter?
Pediatr Exerc Sci. 2007;19(3):291-309.
39. Strycker LA, Duncan SC, Chaumeton NR, Duncan TE, Toobert DJ. Reliability of pedometer data in samples of youth and older women.
Int J Behav Nutr Phys Act. 2007;4:4.
40. Tanaka C, Tanaka S, Kawahara J, Midorikawa T. Triaxial accelerometry for assessment of physical activity in young children.
Obesity (Silver Spring). 2007;15(5):1233-41.
41. Toschke JA, von KR, Rosenfeld E, Toschke AM. Reliability of physical activity measures from accelerometry among preschoolers in free-living conditions.
Clin Nutr. 2007;26(4):416-20.
42. Treuth MS, Schmitz K, Catellier DJ, et al. Defining accelerometer thresholds for activity intensities in adolescent girls.
Med Sci Sports Exerc. 2004;36(7):1259-66.
43. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer.
Med Sci Sports Exerc. 2008;40(1):181-8.
44. Trost SG, McIver KL, Pate RR. Conducting accelerometer-based activity assessments in field-based research.
Med Sci Sports Exerc. 2005;37(11 suppl):S531-43.
45. Trost SG, Pate RR, Freedson PS, Sallis JF, Taylor WC. Using objective physical activity measures with youth: how many days of monitoring are needed?
Med Sci Sports Exerc. 2000;32(2):426-31.
46. Trost SG, Way R, Okely AD. Predictive validity of three ActiGraph energy expenditure equations for children.
Med Sci Sports Exerc. 2006;38(2):380-7.
