Original paper
Pedometer accuracy in physical activity assessment of preschool children

https://doi.org/10.1016/j.jsams.2006.07.004Get rights and content

Summary

Yamax SW-200 pedometer validity during unstructured free play was assessed in 13 preschool children aged 4.1 ± 0.6 years using the Children's Activity Rating Scale (CARS). Pedometer validity and hip placement reliability were also assessed during straight line walking. Data were analysed using regression analyses, 95% upper and lower prediction intervals and 95% limits of agreement. The prediction intervals for pedometry were wide when using the CARS as a criterion. Limits of agreement and prediction intervals for directly observed step counts were also wide for pedometers, calling into question their acceptability for use with preschoolers. Limitations of employing the CARS direct observation approach as a criterion measure of physical activity are also discussed.

Introduction

Accurate quantification of physical activity in early childhood is recognised as being fundamentally important for a number of reasons including measuring physical activity levels and changes to these levels, establishing dose–response relationships between physical activity and health outcomes, allowing effectiveness of interventions to be ascertained, and to identify children at increased health risk because of their physical inactivity. While considered the “gold-standard” for physical activity measurement,1 direct observation can be impractical for epidemiological research as it is time consuming, expensive and is reliant on observer precision and presence which is often unfeasible. In contrast, accelerometry and pedometry allow for objective and more efficient data collection. While these motion sensors have been used widely in school-aged children to understand total daily physical activity, limited work has assessed the efficacy of these instruments in younger, preschool-aged children (<5 years).

Accelerometry is the most commonly used objective approach to monitoring free-living physical activity in young children, and a variety of accelerometer models have been well validated in this population.2, 3, 4, 5, 6, 7, 8, 9, 10, 11 Accelerometers are not, however, without limitations. Their cost potentially limits their use for larger research studies, and their units of measurement (counts) have less intuitive utility than pedometer step counts or accumulated time spent being active. Consequently, pedometers are often promoted as a preferred alternative to accelerometers for the objective measurement of physical activity, particularly in youth12 because of their affordability, reliability and validity in assessing free-living physical activity.

The validity and reliability of Yamax Digiwalker SW-200 pedometers (Yamax Inc., Tokyo, Japan) have been demonstrated within a preschool population by Louie and Chan13 and McKee et al.14 In both studies, the CARS was used concurrently to validate the pedometers during preschool free-play sessions, with significant correlations found between total pedometer counts and CARS scores (r = 0.637–0.950, p < 0.01). Louie and Chan13 also found no significant difference (p > 0.05) between pedometers worn on the left and right hips. There remain, however, a number of unanswered questions regarding pedometer use in early childhood populations. For example, no research has investigated pedometer validity in preschoolers using actual step counts as a criterion measure. Placement of pedometers at the back of the child (where it is not in their direct view) may mitigate any reactivity to pedometer wearing, however, the reliability of pedometers worn at the back has not been established in young children.

The current study was developed to gather more empirical evidence for the accuracy of pedometers use with preschool children and to answer some of the outstanding questions. Specifically, the validity of the Yamax SW-200 Digiwalker pedometer measures compared to direct observation values (using CARS) in free play was assessed. Further, the accuracy and reliability of pedometer readings compared with observed steps across three sites of placement (left hip, right hip and back) and over three pace conditions (slow walk, normal walk and run) in these preschool children were investigated.

Section snippets

Participants

A convenience sample of thirteen 3–5-year-old children attending the Auckland University of Technology Akoranga Crèche (childcare centre for children aged 0–5 years) was recruited for this study. The sample size required for this study was based on the report that a sample of 8–10 children was adequate to detect significant differences between motion sensor readings,7 and allowed for 25% attrition. The sole exclusion criterion was an inability to walk. Parental consent and participant assent

Results

Thirteen children (7 boys and 6 girls) aged 3.0–4.8 years (median 4.2) participated in the study.

  • 1.

    Validity in unstructured free play. Fig. 1 shows the linear prediction and 95% prediction intervals for the pedometer data against CARS physical activity levels for the 35 min free-play session. As can be seen in Fig. 1, the 95% prediction interval associated with the pedometer data is wide. For example, at a pedometer count of 1000 steps, the 95% prediction interval ranges from a CARS activity level

Discussion

The first stage of this study involved assessing the validity of pedometers using direct observation as the criterion measure. Correlations observed in the current study were similar to those found in the larger study of Louie and Chan13 (r = 0.59, compared with r = 0.64). McKee et al.14 found stronger correlations than in the present study (r = 0.64–0.95), although pedometer data were collected in 3 min intervals as opposed to steps accumulated for the entire observation period. This methodology was

Conclusion

Considerable variability was found in the ability of pedometers to accurately predict physical activity level as assessed with the CARS. Limitations of the CARS direct observation system as a criterion measure for free-living activity have been identified. A fundamental issue is then raised as to how best to assess the accuracy of physical activity measurement tools? Given that the variable of interest is physical activity rather than energy expenditure, it may be that direct observation is the

Practical implications

  • Pedometers may not be sufficiently accurate for research purposes with children aged between 3 and 5 years, and instead may be better suited for situations where a general idea of preschoolers’ accumulated physical activity is required.

  • When using pedometers with children aged 3–5 years, pedometers should be placed at either the left or right hip, rather than at the back of the child.

  • If the Children's Activity Rating Scale is used to validate physical activity measurement tools in preschoolers,

Acknowledgements

Melody Oliver is supported by a Tertiary Education Commission Top Achiever Doctoral Scholarship. Part of this study was funded by Sport and Recreation New Zealand as a component of the Active Movement Scoping Exercise and Programme Evaluation.25 The staff and children at the Auckland University of Technology Akoranga Crèche are also thanked for their participation in this study.

References (25)

  • K. Finn et al.

    Factors associated with physical activity in preschool children

    J Pediatr

    (2002)
  • C. Tudor-Locke et al.

    BMI-referenced standards for recommended pedometer-determined steps/day in children

    Prev Med

    (2004)
  • J.R. Sirard et al.

    Physical activity assessment in children and adolescents

    Sports Med

    (2001)
  • K.F. Janz et al.

    Physical activity and bone measures in young children: the Iowa bone development study

    Pediatrics

    (2001)
  • D.M. Jackson et al.

    Objectively measured physical activity in a representative sample of 3–4-year old children

    Obes Res

    (2003)
  • J.R. Sirard et al.

    Calibration of the Computer Science and Applications Inc. physical activity monitor in preschool children

    Med Sci Sports Exerc

    (2001)
  • L.A. Kelly et al.

    Comparison of two accelerometers for assessment of physical activity in preschool children

    Pediatr Exerc Sci

    (2004)
  • S.C. Fairweather et al.

    Using the computer science and applications (CSA) activity monitor in preschool children

    Pediatr Exerc Sci

    (1999)
  • K. McIver et al.

    Validity of the ActiGraph and Actical accelerometers in 3–5-year-old children

    Pediatr Exerc Sci

    (2004)
  • K.A. Pfeiffer et al.

    Validation and calibration of the Actical accelerometer in preschool children

    Med Sci Sports Exerc

    (2006)
  • J.R. Sirard et al.

    Calibration and evaluation of an objective measure of physical activity in preschool children

    J Phys Act Health

    (2005)
  • K.J. Finn et al.

    Validation of the Actiwatch activity monitor in children

    Med Sci Sports Exerc

    (2001)
  • Cited by (40)

    • Measurement of obesity prevention in childcare settings: A systematic review of current instruments

      2017, Obesity Research and Clinical Practice
      Citation Excerpt :

      Pedometers were used in four studies set in schools or preschools, two of which calculated average daily steps as an outcome [74–76], one that calculated steps per minute during school day, and one that calculated total steps during one recess period [77]. One study found that pedometers are correlated (r = .59) with levels from the CARS but show wide prediction intervals [73]. Two studies used heart rate monitors, one in a preschool-based cross-sectional study for three consecutive days [78] and another for a longitudinal study, annually, for three years, during the afterschool (3–6 pm) period [79].

    • Validity of the Omron pedometer and the actigraph step count function in preschoolers

      2015, Journal of Science and Medicine in Sport
      Citation Excerpt :

      We can conclude that pedometer-based steps from the Omron Walking Style Pro are a good measure to assess preschoolers’ PA levels. Oliver et al.29 gave a different conclusion regarding the use of pedometers in preschool children, but this conclusion was mainly based on the Bland–Altman methodology instead of the correlation coefficient found in their study (r = 0.59). However, the latter findings were limited by a small sample size, and the approach used in the study of Oliver et al.29 is different from the approach used in the current study.

    • A cluster randomized trial of a multi-level intervention, delivered by service staff, to increase physical activity of children attending center-based childcare

      2014, Preventive Medicine
      Citation Excerpt :

      Craig et al. as part of a large nationally representative survey of pedometer-determined physical activity in youth including children, reported that one day of pedometer monitoring yielded a valid representation of steps per day relative to the whole week in terms of both reliability (ICC = 0.79) and validity (relative absolute percent error [APE] ≤ 10%) (Craig et al., 2010). Pedometers measure vertical oscillations of body movement (Louie and Chan, 2003), and provide a total count of accumulated movements over the data collection time period (McKee et al., 2005) and are suitable for assessing accumulated time spent being physically active (Oliver et al., 2007a, 2007b). Pedometry has been shown to be a reliable and valid measure of physical activity in preschool age children (age 3–5 years) (Cardon and De Bourdeaudhuij, 2007; Louie and Chan, 2003; McKee et al., 2005; Oliver et al., 2007a, 2007b) and has been used in intervention studies assessing child physical activity levels attending center based care (Boldemann et al., 2006; Eliakim et al., 2007).

    • Pedometer Efficacy for Clinical Care in Pediatric Cardiology

      2023, Journal of Physical Activity and Health
    View all citing articles on Scopus
    View full text