Assessing Volume of Accelerometry Data for Reliability in Preschool Children

1Interdisciplinary Educational Research Institute, University of Wollongong, Wollongong, New South Wales, Australia 2Centre for Physical Activity and Nutrition Research, Deakin University, Burwood, Victoria, Australia.
Medicine and science in sports and exercise (Impact Factor: 3.98). 07/2012; 44(12). DOI: 10.1249/MSS.0b013e3182661478
Source: PubMed


This study examines what volume of accelerometry data (h·d) is required to reliably estimate preschool children's physical activity and whether it is necessary to include weekday and weekend data.

Accelerometry data from 493 to 799 (depending on wear time) preschool children from the Melbourne-based Healthy Active Preschool Years study were used. The percentage of wear time each child spent in total (light-vigorous) physical activity was the main outcome. Hourly increments of daily data were analyzed. t-tests, controlling for age and clustering by center of recruitment, assessed the differences between weekday and weekend physical activity. Intraclass correlation coefficients estimated reliability for an individual day. Spearman-Brown prophecy formula estimated the number of days required to reach reliability estimates of 0.7, 0.8, and 0.9.

The children spent a significantly greater percentage of time being physically active on weekend compared with weekdays regardless of the minimum number of hours included (t = 12.49-16.76, P < 0.001 for all). The number of days required to reach each of the predetermined reliability estimates increased as the number of hours of data per day decreased. For instance, 2.7-2.8 d of data were required to reach a reliability estimate of 0.7 with 10 or more hours of data per day; 3.3-3.4 d were required to meet the same reliability estimate for days with 7 h of data.

Future studies should ensure they include the minimum amount of data (hours per day and number of days) as identified in this study to meet at least a 0.7 reliability level and should report the level of reliability for their study. In addition to weekdays, at least one weekend day should be included in analyses to reliably estimate physical activity levels for preschool children.

1 Follower
8 Reads
  • Source
    • "By application of standard data reduction wear criteria (≥6–10 h/day and ≥3 and 5 days/week), we found reliability estimates ≥0.75 for all outcome variables, except for SED (min/day). Thus, in terms of ICC, our results were consistent with previous studies that have estimated reliability over one week of measurement in preschool-(Addy et al., 2014; Hinkley et al., 2012; Hislop et al., 2014; Penpraze et al., 2006) and older children (Basterfield et al., 2011; Chinapaw et al., 2014; Janz et al., 1995; Kang et al., 2009; Murray et al., 2004; Ojiambo et al., 2011; Rich et al., 2013; Treuth et al., 2003; Trost et al., 2000), which indicates generalizability to other study samples. Still, studies that have applied several measurement periods over the course of a year have yielded substantially lower reliability estimates in adults (Levin et al., 1999) and children (Mattocks et al., 2007; Wickel and Welk, 2010). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Objective: To determine the intra-individual agreement for objectively measured physical activity (PA) and sedentary behavior (SED) over two subsequent weeks in preschool children. Method: Ninety-one children aged 3 to 5 years (49% boys) from three preschools in Sogn og Fjordane, Norway, provided 14 consecutive days of accelerometer data (Actigraph GT3X +) during the autumn of 2014. Week-by-week reliability was assessed using intraclass correlation (ICC), Bland-Altman plots and 95% limits of agreement for different wear time criteria (≥ 6, 8 and 10 h/day and ≥ 3 and 5 days/week). Results: The week-by-week ICC was ≥ 0.75 for all variables across all wear criteria applied, except for absolute sedentary time (ICC 0.61-0.81). Using a ≥ 8 h/day and ≥ 3 days/week criterion (n = 78), limits of agreement were ± 209.5 cpm for overall PA, ± 68.6 min/day for SED, ± 43.8 min/day for light PA, ± 20.2 min/day for moderate-to-vigorous PA, and ± 55.9 min/day for light-to-vigorous PA, equaling 1.0-1.6 standard deviation units. Conclusion: Considerable week-by-week variability was found for all variables. Researchers need to be aware of substantial intra-individual variability in accelerometer-measurements and take necessary actions according to the hypothesis under study, as noise in any measurement will preclude researchers' ability to arrive at valid conclusions in epidemiology.
  • Source
    • "Especially triaxial accelerometers are valid devices with similar classification accuracy for sedentary, light, and moderate-to-vigorous levels of PA in preschoolers [19]. However, in order to obtain reliable data, a minimum amount of measured days/daily hours of measurement should be obtained, and at least one weekend day should be included in analyses to reliably estimate physical activity levels for preschool children [20]. In our study, PA was measured by accelerometers (SenseWear Pro 2, Bodymedia, SMT medical GmbH&Co Würzburg, Germany [21], [22] over a period of 7 consecutive days, thereby including one weekend. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Data on objectively measured physical activity (PA) in preschoolers are controversial. Direct accelerometry was performed in children aged 3-6 years, and differences in PA patterns over the course of the week were evaluated. Data were analyzed with gender, BMI, lifestyle, and socioeconomic parameters as covariates. PA was measured in 119 children by the SensewearPro® accelerometer and analyzed in the 92 (40 girls) that wore it for at least 4 days including one day of the weekend. Median measuring time in this group was 7 consecutive days (median/mean daily measuring time: 23.5 h/d and 21.8 h/d, respectively), corresponding to 834,000 analyzed minutes. PA questionnaires were completed by 103 parents and 87 preschool teachers to collect anthropometric, lifestyle, and socioeconomic data. Median daily PA (MET>3) was 4.3 hours (mean: 4.4 hours). Boys spent an estimated 52 min/week more being very active (MET>6) than girls (95% CI [6, 96] min/week, p = 0.02). PA was lower during the weekend (3.7 h/d) compared to weekdays (4.5 h/d), p = 3×10(-6)), where a 95% CI for the difference is [0.5, 1.0] h/d. PA levels did not differ between overweight/obese children (median 4.7 h/d) and normal-weight peers (median 4.2 h/d). Daily media consumption increased with decreasing social class on weekdays (p = 0.05) and during the weekend (p = 0.01), but was not related to the amount of daily PA. A multivariate regression with BMI-SDS as independent variable and gender, age, amount of PA>6 MET, parental BMI, media time and socioeconomic status as explanatory variables revealed that only SES had a significant contribution. The negative impact of obesity-promoting factors in older children is rather low for preschoolers, but there is evidently a gradient in PA between weekdays and weekends already in this age group. Weight status of preschoolers is already considerably influenced by SES, but not physical activity levels.
    PLoS ONE 04/2013; 8(4):e60619. DOI:10.1371/journal.pone.0060619 · 3.23 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Early childhood is a critical time for promoting physical activity. Few studies have investigated the effect of interventions in this population. The aim of this study was to investigate the effect of a school-based active play intervention on preschool children's sedentary time and physical activity. Preschool children were recruited from randomly selected preschools. Schools were randomly assigned to an intervention or comparison group. One teacher per intervention school received training from active play professionals in the delivery of a 6-week active play programme. Comparison schools continued their usual practice. Children wore a uni-axial accelerometer for 7 days at baseline, immediately after and at 6-month post-intervention. No significant intervention effects were observed for sedentary time or physical activity. However, sex and hours spent at school were significant predictors of physical activity. Children who spent fewer hours (half-day children) at school were significantly more active than their fullday counterparts. Physical activity during the intervention classes was high even though neither daily physical activity nor sedentary time changed. Notably children who spent more time at preschool were less active suggesting that preschool was not as conducive to physical activity engagement as other environments.© The Author 2013. Published by Oxford University Press. All rights reserved.
    International Journal of Behavioral Nutrition and Physical Activity 10/2012; 9(1):117. DOI:10.1186/1479-5868-9-117 · 4.11 Impact Factor
Show more