Annals of Human Biology, JulyAugust 2009; 36(4): 363378
ORIGINAL ARTICLE
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
Seasonal changes in children’s physical activity: An examination of group changes, intra-individual variability and consistency in activity pattern across season
ANN V. ROWLANDS, EMMA L. PILGRIM & ROGER G. ESTON School of Sport and Health Sciences, University of Exeter, UK (Received 21 November 2008; revised 27 January 2009; accepted 16 February 2009)
Abstract Background: Components of activity that are more variable over time may be more susceptible to manipulation in activity interventions. Aim: The present study examined variability and consistency of components of children’s activity across season. Subjects and methods: Sixty-four 911-year-old children wore an accelerometer for 6 days during winter and summer. Activity bouts (]4 s) greater than light (]LIGHT), moderate (]MOD) and vigorous (]VIG) intensity were recorded. Results: Intra-individual variability of the activity components across season was greater for bout frequency (CV: ]LIGHT6.69.9%, ]MOD10.716.1%, ]VIG 17.026.8%) than bout intensity or duration (CV: ]LIGHT 3.47.4%, ]MOD 3.67.8%, ]VIG 4.210.0%, pB0.05) and for the frequency of ]VIG bouts compared to the frequency of ]LIGHT and ]MOD bouts (p B0.05). All components of the activity pattern tended to track consistently when assessing ]LIGHT and ]MOD bouts (intra-class correlations (ICC) 0.4783, pB0.05), ]VIG bouts in boys (ICC0.690.77, pB0.01) and frequency of ]VIG bouts in girls (ICC0.82, pB0.01). Conclusions: Bout frequency was the most variable component of activity across season. However, children tended to maintain their rank for bout frequency. It would be of interest to investigate whether bout frequency can be manipulated in an activity intervention. Keywords: Intensity, vigorous, accelerometer, frequency, activitystat
Introduction There is considerable evidence concerning the benefits of physical activity for health in adults (Kesaniemi et al. 2001) and a growing body of evidence demonstrating childhood
Correspondence: Dr Ann V. Rowlands, School of Sport and Health Sciences, University of Exeter, St Luke’s Campus, Heavitree, Exeter EX1 2LU, UK. Email:
[email protected] ISSN 0301-4460 print/ISSN 1464-5033 online # 2009 Informa UK Ltd. DOI: 10.1080/03014460902824220
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
364
A.V. Rowlands et al.
activity also positively impacts on health (Strong et al. 2005; Dencker and Andersen 2008). Against this backdrop there is indirect evidence that children’s activity levels are declining. For example, the prevalence of overweight and obesity is increasing (Stamatakis et al. 2005), with research suggesting that this increase cannot be explained by changes in dietary intake alone (Dehghan et al. 2005). Studies assessing activity levels objectively have shown the activity levels of contemporary children are lower then that of those living a traditional agrarian lifestyle in Old Order Amish and Old Older Mennonite communities (Tremblay et al. 2005; Bassett et al. 2007). To address this, interventions have been designed to increase children’s physical activity. The nature of children’s activity interventions varies hugely, but one thing they appear to have in common is limited efficacy (van Sluijs et al. 2008). Other studies have shown that activity and inactivity track to a low to moderate degree over time (Janz et al. 2005) with greatest tracking being shown over periods of a few years and weaker tracking evident over longer periods (Telama et al. 2005). Research suggests that physical activity is hard to manipulate, at least in the long term. It is acknowledged that there is a biological basis for activity (Rowland 1998; Thorburn and Proietto 2000; Bouchard and Rankinen 2006) with some individual’s having a higher ‘drive’ for activity than others. The rationale and possible mechanisms responsible for this biological drive have been discussed elsewhere (Rowland 1998; Thorburn and Proietto 2000; Bouchard and Rankinen 2006; Eisenmann and Wickel 2009; Rowlands 2009). Regardless of the cause of any biological basis for activity, of interest is how flexible activity may be in the presence of this control. If biological control of activity exists, compensation for imposed or restricted activity would be expected (Rowland 1998; Wilkin et al. 2006). However, the evidence for compensation is inconsistent (Shephard et al. 1980; Blaak et al. 1992; Dale et al. 2000; Wickel and Eisenmann 2007). It is possible that the way in which activity is manipulated may impact on any drive for compensation, e.g. manipulation of frequency, intensity or duration of activity bouts may lead to relatively stronger or weaker drives for compensation. In order to address this question, a first step is to examine how aspects of the activity pattern track over time, both in the short and long term in differing environmental circumstances. The decline in activity levels in the population indicates that the environment can influence activity levels (Rowlands 2009). Children’s activity levels tend to be lower in the winter than the summer (Rifas-Shiman et al. 2001; Rowlands and Hughes 2006) and on weekdays compared to weekend days (Rowlands et al. 1999; Gavarry et al. 2003). It is hypothesized that a component of the activity pattern (i.e. frequency, intensity or duration of activity bouts) that is more variable over season or type of day may be more susceptible to manipulation and this information may inform future activity interventions. Children’s activity is sporadic with 80, 93 and 96% of activity bouts of moderate, vigorous and very high intensity, respectively, shorter than 10 s and high intensity bouts typically lasting 3 s (Bailey et al. 1995; Berman et al. 1998). Counts accumulated in vigorous and very high intensity activity have been shown to account for over a third of the children’s total physical activity, despite accounting for less than 3% of the monitored time (Baquet et al. 2007). This highlights the importance of quantifying short bouts of intense activity. Previous research has indicated that differences in weekday and weekend activity may be largely due to the intensity of the most frequent bouts of activity and frequency of the most intense bouts (Rowlands et al. 2008a). Whether these components of activity also explain seasonal differences is not clear. Therefore, the aim of this study was to examine intra-individual variability and consistency in the components of children’s activity pattern across different types of day (weekdays and weekend days) and across season (winter and summer). It was hypothesized that some components of the activity pattern would be more variable than others and
Intra-individual variability and consistency in physical activity
365
therefore may offer a window of opportunity for the design of interventions aiming to increase physical activity.
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
Methods The initial sample consisted of 84 911 year-old children (45 boys and 39 girls) recruited from two schools in the southwest of England. A total of 64 children (32 boys (age (mean9 SD): 9.990.3 years; height: 1.3890.06 m; mass: 34.196.8 kg) and 32 girls (age: 9.890.3 years; height: 1.3890.05 m; mass: 34.396.7 kg)) provided sufficient data at both time points for analysis. Height was measured to the nearest 0.1 cm and body mass was measured to the nearest 0.1 kg. The age, height and mass of these children did not differ from those with insufficient data. Procedures followed were approved by the School Ethics committee and were in accordance with the Helsinki Declaration of 1975, as revised in 1983. Parents or guardians provided written informed consent and children gave verbal assent.
Physical activity assessment Physical activity was assessed objectively over up to 6 days using uniaxial accelerometry (GT1M Actigraph, Pensacola, FL, USA) at two time points (time point 1: weekdays 3.99 0.4, weekend 1.890.4; time point 2: weekdays 3.890.7, weekend 1.690.7). The first data collection took place during term time in January/February 2007 and the second data collection took place during term time in June/July 2007. All children were assessed within a seven week period at each time point. Mean daily temperature during the winter period was lower, hours of sunlight lower and rainfall similar relative to the summer period (mean daily temperature 8.793.08C cf. 17.88 91.58C; hours of sunshine per month 60.8 cf. 154.0 and monthly rainfall 159.2 mm cf. 155.4 mm for winter and summer, respectively). Children wore the accelerometers on the hip for up to four weekdays and two weekend days from when they got up in the morning until they went to bed. The accelerometer was programmed to record activity counts in 2 s epochs. This was the lowest epoch that would allow data to be stored for the entire measurement period without download. One hour of consecutive zeros was taken to indicate the monitor had been removed (Rowlands et al. 2008b). Data were analysed between 6 am and 9 pm. For inclusion in the data analyses, a child needed a minimum of 10 h wearing time for each of three weekdays and a minimum of 8 h wearing time for at least 1 weekend day at each time point. The use of different criteria for weekdays and weekend days tackles the issue of differing sleep patterns for weekend days and/or the greater difficulty in ensuring that an accelerometer is worn from waking, use of the same criteria for both types of day can lead to very few children having data for weekend days (Rowlands 2007). Published threshold values (Trost et al. 1998) relating counts/minute to activity intensity were divided by 30 to provide thresholds for the 2 s epoch data (Table I). Additionally, a lower threshold of 10 cts 2 s 1 (300 cts min 1) was added to categorize light intensity activity (defined as whole body movement, Rowlands et al. (2008a)). Output measures included composite measures of activity and activity pattern variables. Composite variables were: total activity; time spent in moderate activity (MOD); time spent in ]vigorous activity (]VIG). Pattern variables were: frequency of bouts of at least light (]LIGHT), moderate (]MOD), vigorous (]VIG) intensity; mean duration of bouts above each
366
A.V. Rowlands et al. Table I. Classification of ActiGraph output into intensity categories.
Count threshold (cts 2 s 1) 10$ 65* 176*
Metabolic equivalent (METs)
Intensity
NA 3 6
light moderate vigorous
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
*Published threshold values (Trost et al. 1998). $Threshold to differentiate sedentary activity (seated, standing still) from whole body movement (Rowlands et al. 2008a).
intensity threshold; mean intensity of bouts above each intensity threshold as described previously (Rowlands et al. 2008a). Although the minimum intensity and duration was stated, bouts could be made up of any intensity and duration above these minimums. Reporting the mean intensity and duration of bouts of a minimum intensity and duration reflects the open ended nature of the variables. This allowed the investigation into variability of bouts of any intensity (]LIGHT) and also allowed the higher intensity bouts undertaken by some children to be accounted for within the ]VIG bouts. This gives greater scope than simply recording the frequency of bouts within constrained categories. A bout was defined as a minimum of two consecutive epochs (i.e. 4 s) above the intensity threshold (] 4 s bouts). This corresponds to the typical duration of children’s moderate and vigorous intensity activity bouts (6 and 3 s, respectively, Bailey et al. 1995) and therefore captures the children’s characteristic sporadic activity pattern. Whether these very short bouts are related to health is not clear although preliminary data indicates the interval between bouts is positively related to fatness and the intensity of bouts is positively related to aerobic fitness (Chu et al. 2006). All output variables were calculated for all days combined and separately for weekdays and weekend days.
Statistical analysis Descriptive statistics were calculated for all variables by sex. Tests for normal distribution revealed that some of the physical activity variables were skewed. Log transformation of these variables (Tabachnick and Fidell 1996), resulted in distributions that more closely approximated normal. Differences in composite measures of activity and activity pattern by season and type of day (weekday and weekend) were assessed with a series of two-factor repeated measures ANOVAs for boys and girls separately. To assess individual variability in composite measures of activity and activity pattern, the coefficient of variation (CV) across seasons for all days (mean of all days measured including weekend and weekdays) at each time point was calculated for each child. Differences in the variability of activity by sex and by component of activity assessed were investigated using a series of two-factor (sex activity component) ANOVAs. The first ANOVA assessed composite activity components, specifically total activity and time accumulated at MOD and ]VIG intensities. Three further ANOVAs addressed differences in the variability of the activity pattern component (frequency, duration and intensity of activity bouts) for ]LIGHT, ]MOD and ]VIG intensity activity, respectively. Individual consistency in composite measures of activity and activity pattern across type of day and across season was investigated using a series of intra-class correlations (ICC, mixed effects model) for boys and girls separately. This assesses the tracking of activity from one time to the next, specifically how the ranking of the individual’s activity level within the group tracks as opposed to whether the actual activity level changes.
Intra-individual variability and consistency in physical activity
367
Where appropriate, analyses were run using the log-transformed variables. As the use of log-transformed variables did not impact notably on the ANOVA results, the data from the untransformed variables is reported for ease of interpretation. Significant ANOVA results were followed-up using the Tukey’s test adapted for repeated measures (Stevens 1996). An alpha level of 0.05 was used for all statistical tests. SPSS version 15.0 for Windows (SPSS Inc., Chicago, IL, USA) was used for all statistical analyses. Results
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
Activity level by season, type of day and sex Composite physical activity data are presented in Figure 1ac (total, MOD and ]VIG, respectively) and activity pattern data are presented in Figure 2ac (]LIGHT bouts: frequency, intensity and duration, respectively), Figure 3ac ( ]MOD bouts: frequency, intensity and duration, respectively and Figure 4ac (]VIG bouts: frequency, intensity and duration, respectively). Composite activity (Figure 1ac) Boys were more active on weekdays than weekends, but this difference was greater in the summer than the winter (total activity F(1,26) 7.77, p 0.01 (Figure 1a), MOD activity F(1,26) 4.60, p 0.042 (Figure 1b) and ]VIG activity F(1,26) 6.67, p0.016 (Figure 1c)). Girls also tended to be more active on weekdays than weekends, irrespective of activity measure; however, in contrast to the boys, this difference was smaller in the summer compared to the winter for total activity (F(1,29) 6.81, p 0.014, Figure 1a) and MOD activity (F(1,29) 7.96, p 0.009, Figure 1b), but not ]VIG activity (F(1,29) 2.23, p 0.147, Figure 1c). Therefore, seasonal differences in activity level were largest for weekday activity in boys (total, MOD and ]VIG activity) and only present for weekend activity (total, MOD activity) in girls, where activity levels were higher in the summer than the winter. Activity pattern (Figure 2ac, 3ac and 4ac) Season type of day interactions were present for boys for the frequency of ]LIGHT (F(1,26) 9.84, p 0.004, Figure 2a) and ]MOD activity bouts (F(1,26) 4.60, p 0.042, Figure 3a). These largely reflected the composite activity results in that seasonal differences in the frequency of bouts were only present for weekday activity, where boys were more active in the summer than the winter. For girls, the duration of ]LIGHT (Figure 2c) and ]MOD (Figure 3c) activity bouts largely reflected the composite activity data, whereby activity was higher in the summer than the winter on weekend days, but not weekdays (F(1,29) 14.44, pB0.001 and F(1,29) 11.21, p 0.002, respectively). No other interactions were present. Main effects indicated that in boys the duration of ]LIGHT (Figure 2c), ]MOD (Figure 3c) and ]VIG (Figure 4c) bouts were greater in the summer than the winter and on weekdays compared to weekends ( p B0.05). In girls, the frequency (Figure 4a) and intensity of ]VIG bouts (Figure 4b) was greater in the summer than the winter, and the frequency of ]MOD (Figure 3a) and ]VIG (Figure 4a) bouts and the intensity of ]LIGHT (Figure 2b) and ]MOD (Figure 3b) bouts were greater on weekdays compared to weekends ( p B0.05).
A.V. Rowlands et al.
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
368
Figure 1. (a) Total activity, (b) time spent in MOD activity and (c) ]VIG activity by season, type of day and sex. *Summer activity winter activity (pB0.05); $weekday activity weekend activity (pB0.05).
369
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
Intra-individual variability and consistency in physical activity
Figure 2. (a) Frequency, (b) intensity and (c) duration of ]LIGHT bouts by season, type of day and sex. *Summer activity ]winter activity (pB0.05); $weekday activity ]weekend activity (pB0.05).
A.V. Rowlands et al.
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
370
Figure 3. (a) Frequency, (b) intensity and (c) duration of ]MOD bouts by season, type of day and sex. *Summer activity winter activity (pB0.05), $weekday activity weekend activity (pB0.05).
371
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
Intra-individual variability and consistency in physical activity
Figure 4. (a) Frequency, (b) intensity and (c) duration of ]VIG bouts by season, type of day and sex. *Summer activity winter activity (pB0.05), $weekday activity weekend activity (pB0.05).
372
A.V. Rowlands et al.
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
Individual variability in activity (CVs) across season The CVs for activity variables assessed across season are presented in Table II. Regarding composite activity measures, boys’ activity was more variable than girls’ (F(1,62) 17.46, p B0.001) and ]VIG activity was more variable than total or MOD activity (F(1.24,77.06) 10.52, p B0.001). Regarding the activity pattern, the frequency of activity bouts was more variable than the duration or intensity, irrespective of the intensity (]LIGHT: F(1.55,95.91) 9.12, p 0.001; ]MOD: F(1.66 102.65) 35.42, p B0.001; ]VIG: F(1.29,80.11) 47.25, p B0.001). For ]LIGHT activity bouts the variability was greater in boys (F(1,62) 8.86, p 0.004). Sex activity pattern component interactions indicated only the variability of the frequency of bouts was greater in boys than girls for ]MOD (F(1.66 102.65) 4.41, p 0.02) and ]VIG bouts (F(1.29,80.11) 5.84, p 0.012) and, while the frequency of bouts was the most variable component of the activity pattern, the duration was more variable than the intensity for ]MOD bouts in boys and girls and ]VIG bouts in girls (p B0.05).
Individual consistency of activity level (ICCs) Weekday to weekend activity by sex and season (Table III). Overall, composite activity measures tracked most highly between weekdays and weekends for girls in the winter (ICC 0.66, p B0.01) and for boys for ]VIG activity in the winter and the summer (ICC 0.58, p B0.01), with little indication of consistency between weekdays and weekends in the summer for the girls (ICC 0.160.36, NS). However, the activity pattern variables were more consistent, with the pattern of ]LIGHT activity bouts showing good tracking between weekdays and weekends in both winter and summer for both boys and girls for all aspects of the activity pattern (ICC 0.51, p B0.05). There was no strong indication for any particular aspect of the activity pattern (frequency, intensity or duration of bouts) to track more strongly than any other between weekdays and weekends. However, for girls, the pattern was more consistent when including bouts of any intensity rather than restricting assessment to bouts of vigorous intensity.
Seasonal tracking of activity by sex and type of day (Table IV) Overall, composite activity measures tracked most highly for girls, particularly when analysing all days (girls ICC 0.7160.899; boys ICC0.6260.682, all pB0.01). As with weekend to weekday measures, there was no strong indication that any particular component of the activity pattern (frequency, intensity or duration of bouts) tracked more strongly than any other. However, for girls, the intensity and duration of bouts was more consistent across season when including bouts of any intensity than when restricting assessment to bouts of vigorous intensity (all days: ]LIGHT ICC 0.7480.841, pB0.01; ]MOD ICC 0.4690.813, p B0.05; ]VIG 0.2480.254, NS), whereas the frequency of bouts was consistent regardless of intensity (all days: ]LIGHT ICC 0.827, p B0.01; ]MOD ICC 0.689, p B0.01; ]VIG ICC 0.819, p B0.01). In addition, tracking tended to be slightly stronger when analysing all days than when looking at weekdays alone, which in turn, showed greater tracking than weekend days alone. Activity pattern tracked more consistently across season (Table IV) than from weekday to weekend within season (Table III).
Intra-individual variability and consistency in physical activity
373
Table II. Individual variability (coefficient of variation, CV) of children’s activity across season. CV of winter and summer physical activity* (%, mean9SD) Boys Composite activity measures$ Total activity Moderate activity ]Vigorous activity
17.0910.0 17.9911.1 23.8916.2
7.997.1 9.597.6 14.1912.5
9.999.6b 4.893.0 7.496.1
6.696.1b 3.492.7 5.393.2
]MOD bouts Frequency Intensity Duration
16.1910.9ab 4.092.5 6.795.9c
10.796.3b 3.692.4 7.896.4c
]VIG bouts Frequency Intensity Duration
26.8918.0ab 4.293.9 8.496.2
17.0914.2b 4.995.3 10.099.3c
Activity pattern measures ]LIGHT bouts$ Frequency Intensity Duration
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
Girls
*Activity at each time point calculated from the mean of all days measured (34 weekdays and 12 weekend days). $Boys]girls (main effect, pB0.005); aboys girls (interaction, pB0.05); bdifferent from intensity and duration components (pB0.001); cdifferent from intensity component (p B0.05).
Discussion This study addressed the hypothesis that some aspects of the activity pattern may be more variable than others when exposed to differing environmental circumstances, primarily different seasons. As expected, and in support of previous research, children were more active in the summer than in the winter (Rifas-Shiman et al. 2001; Rowlands and Hughes 2006) and on weekdays compared to weekends (Rowlands et al. 1999; Gavarry et al. 2003). On an intra-individual level, the most variable component across season was the frequency of activity bouts, particularly bouts of higher intensities. However, children tended to maintain their rank for all components of the activity pattern across time, including frequency of bouts. This indicates that the frequency of activity bouts was the most susceptible to seasonal changes in this sample of children; however a child’s position relative to their peers remained relatively stable. The impact of season on composite measures of weekday and weekend physical activity differed between sexes. The boys showed similar levels of weekend activity across seasons, but weekday activity was higher in the summer than the winter; however, the converse was true for girls whereby weekday activity was relatively stable across season and weekend activity was higher in the summer than the winter. This was reflected in most aspects of the activity pattern, but in particular in the frequency of bouts of activity in boys and in the duration of bouts in girls. Time spent outside is correlated with physical activity (Burdette et al. 2004) and likely explains the seasonal differences in activity. Sex differences in the impact of type of day on activity across season may relate to the type of playground activities typically undertaken by boys and girls at school. Boys undertake more vigorous play during recess than girls (Ridgers et al. 2005) and therefore summer weather may have more impact on the amount of activity that boys undertake during days at school than girls.
374
A.V. Rowlands et al.
Table III. Consistency in activity level (assessed by intraclass correlation coefficients (ICC)) across type of day by sex and season.
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
Winter
Summer
Boys
Girls
Boys
Girls
Composite activity measures Total Moderate ]Vigorous
0.40 0.33 0.64$
0.71$ 0.66$ 0.69$
0.48* 0.475 0.58*
0.16 0.36 0.22
Activity pattern measures ]LIGHT bouts Frequency Intensity Duration
0.66$ 0.75$ 0.63$
0.69$ 0.75$ 0.76$
0.62$ 0.51* 0.62$
0.66$ 0.69$ 0.57*
]MOD bouts Frequency Intensity** Duration**
0.41 0.66$ 0.57*
0.55* 0.64$ 0.61$
0.60* 0.76$ 0.48
0.47* 0.43 0.51*
]VIG bouts Frequency** Intensity** Duration**
0.54* 0.14 0.27
0.46* 0.56* 0.30
0.59* 0.55* 0.59*
0.35 0.22 0.58*
*p B0.05; $pB0.01; **natural log of variables used in analysis
However, Ridgers et al. (2007) found no evidence for significant variation in children’s recess activity across season and day. Therefore it is likely that season had more of an impact on activity levels outside of school than activity levels in school. The main focus of this paper was to address intra-individual variability and consistency of aspects of the activity pattern across season and to highlight aspects of activity that were less stable across environment. It was hypothesized that the more responsive the component of the activity pattern to the environment the more susceptible that component may be for manipulation in activity interventions. The environmental change utilized was season with one activity assessment carried out in winter and one in summer in the UK. Each assessment included up to 6 days of measurement including at least 1 weekend day, meeting the recommendations for a representative measure of habitual activity (Trost et al. 2000). The CV for this measure from the two seasons was calculated. This may explain the relatively low CV for girls’ compared to those reported previously for day-to-day intraindividual variability in activity (approximately 20%: Wickel et al. 2007; Mattocks et al. 2007); however the boys’ activity was as variable as previous reports for day-to-day activity (Wickel et al. 2007; Mattocks et al. 2007). The sex difference for intra-individual variability present herein, but not in day-to-day activity (data not shown) indicates that season impacted on the boys’ habitual activity to a greater degree than that of the girls’. Time accumulated in vigorous activity was the most variable across season in both boys and girls and this was due to a greater variability in the frequency of vigorous bouts, as variability in intensity and duration of bouts was similar across all intensities. Previous research has indicated that vigorous activity is the prime factor that differentiates boys’ activity from girls’ (Trost et al. 2002; Rowlands and Eston 2005) and high-active from lowactive children (Rowlands et al. 2008b). It is demonstrated herein that vigorous activity (specifically frequency of vigorous bouts) is the dominant factor responsible for changes in
375
Intra-individual variability and consistency in physical activity
Table IV. Consistency in activity level (assessed by intraclass correlation coefficients (ICC)) across season by type of day and sex. All days
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
Boys
Girls
Weekday Boys
Girls
Weekend Boys
Girls
Composite activity measures Total Moderate ]Vigorous
0.63$ 0.68$ 0.67$
0.86$ 0.72$ 0.90$
0.48* 0.62$ 0.64$
0.77$ 0.55* 0.86$
0.58* 0.62$ 0.47
0.59* 0.60$ 0.53*
Activity pattern measures ]LIGHT bouts Frequency Intensity Duration
0.60$ 0.80$ 0.80$
0.83$ 0.84$ 0.75$
0.62$ 0.77$ 0.69$
0.83$ 0.80$ 0.62$
0.51* 0.59* 0.79$
0.62$ 0.66$ 0.46
]MOD bouts Frequency Intensity** Duration**
0.47* 0.76$ 0.82$
0.69$ 0.81$ 0.47*
0.43 0.77$ 0.79$
0.61$ 0.74$ 0.50*
0.47 0.46 0.54*
0.57* 0.64$ 0.27
]VIG bouts Frequency** Duration** Intensity**
0.67$ 0.77$ 0.59$
0.82$ 0.25 0.25
0.66$ 0.72$ 0.75$
0.78$ 0.47* 0.64$
0.26 0.69$ 0.13
0.70$ 0.02 0.27
*p B0.05; $pB0.01; ** natural log of variables used in analysis.
activity on an intra-individual level too. If the frequency of activity bouts is more responsive to the environment on an intra-individual level, then this may be the ideal activity component to be manipulated in an intervention aiming to increase children’s activity levels. It is important to note that, despite the relatively high intra-individual variability shown for the frequency of activity bouts, ICCs indicated that a child’s rank for activity level remained similar both across season and across short-term changes in environment, specifically weekdays while attending school, compared to weekend days. These tracking coefficients exceed tracking coefficients (assessed by Spearman rank order correlations) reported in the literature (Janz et al. 2005; Telama et al. 2005); however, these studies tended to assess tracking over much longer periods of time. The greater consistency across season shown for all days and weekdays in comparison to weekend days is not surprising as weekend days were based on only 12 days of measurement at each time point. Therefore, all days and weekdays can be considered to represent a stable measure of habitual activity at each time point as most children had 4 days of measurement (Trost et al. 2000), whereas this is not the case for weekend days where only 12 days of measurement were available. Interestingly, components of the activity pattern showed greater consistency than composite measures of activity, with all components of bouts of any intensity (]LIGHT) tracking well across type of day and season. The maintenance of a child’s rank in activity level despite overall changes in activity may reflect the focus on class groups within two schools. However, the stability may also reflect an ‘activitystat’ (Rowland 1998; Wilkin et al. 2006) about which a child’s activity naturally fluctuates dependent on the environment. If this is the case, the data indicate that the environment did impact on where these children lay within their natural activity range and that the frequency of activity bouts appeared to be the most responsive component of the activity pattern to the environment. It is therefore possible that interventions focusing on
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
376
A.V. Rowlands et al.
providing an environment that maximizes the opportunities to increase the frequency of activity bouts may enable a child to maintain activity at the upper end of their natural range. This study used season to assess the impact of environment on children’s activity. This was based on the change in weather and hours of daylight between a UK winter and summer and the research demonstrating that season impacts on children’s activity (RifasShiman et al. 2001; Rowlands and Hughes 2006). A second measure of environment was type of day whereby weekdays spent in school were compared to weekend days, considered to be free time. Another approach would be to impose an environmental restriction or facilitate opportunities for activity. This approach would have enabled the assessment of whether compensation for restricted or imposed activity occurs. However, our choices reflect the natural environmental changes that children are exposed to and therefore have ecological validity for the exploration of variability and consistency in activity. Additionally, they did not reflect an imposed condition that may have had novelty value. We felt that this was a necessary first step prior to a study assessing whether children compensate for imposed or restricted activity. Results from the present study could inform a study investigating compensation in order to ensure the activity dose or restriction was outside of the normal range of activity fluctuation and to provide a basis for the manipulation of differing components of the activity pattern. Data from this study are from 64 children in a relatively small region in the southwest of England. In addition, there was no control for socio-economic status. Therefore, results may not be generalizable beyond this sample. However, the activity pattern data recorded does compare favourably with a similar study on children from France (Baquet et al. 2007). Future studies should extend the assessment of the pattern of children’s activity to larger more generalizable samples across more than one location. Of additional interest would be the activity pattern of children living the traditional agrarian lifestyle of the Old Order Amish and Old Order Mennonite communities.
Conclusions In conclusion, the frequency of bouts, particularly those of a higher intensity, was the most variable component of activity in this sample of children on an intra-individual level across season. However, the physical activity pattern was fairly consistent with children tending to maintain their rank for frequency of bouts across both type of day and season. As season appeared to impact on the frequency of bouts on an individual level, it would be of interest to investigate whether this variable can be manipulated in an intervention designed to increase children’s physical activity. However, it is unclear how manipulation of one component of activity may impact on the overall activity pattern. Further research should investigate the dose or restriction of activity that children can absorb before changing activity levels outside of the intervention to compensate, or indeed whether compensation occurs. It is likely that whether compensation occurs and/or the degree of compensation will differ according to the component(s) of activity manipulated; results from this study suggest that the frequency of activity bouts may be more amenable to manipulation than the intensity or duration of bouts. Cross-sectional research has indicated that the frequency of bouts is related to health in children (Stone et al. 2009). Larger scale longitudinal and intervention research is needed to investigate the causal direction in these relationships.
Intra-individual variability and consistency in physical activity
377
Acknowledgements The authors would like to thank Mr David Childs, who developed the software used to analyse the ActiGraph data. Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
References Bailey RC, Olson J, Pepper SL, Porszasz J, Barstow TJ, Cooper DM. 1995. The level and tempo of children’s physical activities: An observational study. Med Sci Sports Exerc 27:10331041. Berman N, Bailey R, Barstow TJ, Cooper DM. 1998. Spectral and bout detection analysis of physical activity patterns in healthy, prepubertal boys and girls. Am J Hum Biol 10:289297. Baquet G, Stratton G, Van Praagh E, Berthoin S. 2007. Improving physical activity assessment in prepubertal children with high-frequency accelerometry monitoring: A methodological issue. Prev Med 44:143147. Bassett DR, Tremblay MS, Esliger DW, Copeland JL, Barnes JD, Huntingdon GE. 2007. Physical activity and body mass index of children in an Old Order Amish Community. Med Sci Sports Exerc 39:410415. Blaak EE, Westerterp KR, Bar-Or O, Wouters LJM, Saris WHM. 1992. Total energy expenditure and spontaneous activity in relation to training in obese boys. Am J Clin Nutr 55:777782. Bouchard C, Rankinen T. 2006. Are people physically active because of their genes? Pres Council Phys Fit Sports Research Digest 7:18. Burdette HL, Whitaker RC, Daniels SR. 2004. Parental report of outdoor playtime as a measure of physical activity in preschool-aged children. Arch Pediatr Adolesc Med 158:353357. Chu EYW, Hu Y, Tsang AMC, McManus AM. 2006. The influence of the distinguished pattern of locomotion to fitness and fatness in prepubertal children. Children and Exercise XXII, 23rd Pediatric Work Physiology Meeting Conference Book, P-B-39. Dale D, Corbin CB, Dale KS. 2000. Restricting opportunities to be active during school time: Do children compensate by increasing physical activity levels after school? Res Q Exerc Sport 71:240248. Dehghan M, Ahktar-Danesh N, Merchant AT. 2005. Childhood obesity, prevalence and prevention. Nutr J 4:24. Dencker M, Andersen LB. 2008. Health-related aspects of objectively measured daily physical activity in children. Clin Physiol Funct Imaging 28:13344. Eisenmann JC, Wickel EE. 2009. The biological basis of physical activity in children: Revisited. Pediatr Exerc Sci. In press. Gavarry O, Giacomoni M, Bernard T, Seymat M, Falgairette G. 2003. Habitual activity in children during school and free days. Med Sci Sports Exer 35:525531. Janz KF, Burns TL, Levy SM. 2005. Tracking of activity and sedentary behaviours in childhood. The Iowa bone development study. Am J Prev Med 29:171178. Kesaniemi YA, Danforth Jr E, Jensen MD, Kopelman PG, Lefervre P, Reeder BA. 2001. Doseresponse issues concerning physical activity and health: An evidence-based symposium. Med Sci Sports Exerc 33:S531S538. Mattocks C, Leary S, Ness A, Deere K, Saunders J, Kirkby J, Blair SN, Tilling K, Riddoch C. 2007. Intraindividual variation of objectively measured physical activity in children. Med Sci Sports Exerc 39:622629. Rifas-Shiman SL, Gillman MW, Field AE, Frazier AL, Berkey CS, Tomeo CA, Colditz GA. 2001. Comparing physical activity questionnaires for youth: Seasonal vs annual format. Am J Prev Med 20:282285. Ridgers N, Stratton G, Fairclough SJ. 2005. Assessing physical activity during recess using accelerometry. Prev Med 41:102107. Ridgers N, Stratton G, Fairclough SJ, Twisk JW. 2007. Long term effects of a playground markings and physical structures on children’s recess activity levels. Prev Med 44:393397. Rowland TW. 1998. The biological basis of physical activity. Med Sci Sports Exerc 30:392399. Rowlands AV. 2007. Accelerometer assessment of physical activity in children: An update. Pediatr Exerc Sci 19:252266. Rowlands AV. 2009. Methodological approaches for investigating the biological basis for physical activity in children. Pediatric Exerc Sci. In press. Rowlands AV, Eston RG, Ingledew DK. 1999. The relationship between activity levels, body fat and aerobic fitness in 810 year old children. J Appl Physiol 86:14281435.
Ann Hum Biol Downloaded from informahealthcare.com by 46.16.226.10 on 05/20/14 For personal use only.
378
A.V. Rowlands et al.
Rowlands AV, Eston RG. 2005. Comparison of accelerometer and pedometer measures of physical activity in boys and girls, aged 810 yrs. Res Q Exerc Sport 76:251257. Rowlands AV, Hughes DR. 2006. Variability of physical activity patterns by school time, holiday time and season in 810 y old boys. Res Q Exerc Sport 77:391395. Rowlands AV, Pilgrim E, Eston RG. 2008a. Patterns of habitual activity across weekdays and weekend days in 911 year-old children. Prev Med 46:317324. Rowlands AV, Pilgrim EL, Stone MR, Eston RG Frequency, intensity and duration of activity bouts in children. In: Jurimae T, Armstrong N, Jurimae J, editors. Children and exercise XXIV. London: Routledge, pp. 142145. Shephard RJ, Jequier JC, Lavallee R, Labarre R, Rajic M. 1980. Habitual physical activity: Effects of sex, milieu, season and required activity. J Sports Med 20:5560. Stamatakis E, Primatesta P, Chinn S, Rona R, Falascheti E. 2005. Overweight and obesity trends from 1974 to 2003 in English children: What is the role of socioeconomic factors? Arch Dis Child 90:9991004. Stevens J. 1996. Applied multivariate statistics for the social sciences. Hillsdale, NJ: Lawrence Erlbaum Associates. Stone MR, Rowlands AV, Middlebrooke AR, Jawis MN, Eston RG. 2009. The pattern of physical activity in relation to health outcomes in children. Int J Pediatr Obes. DOI: 10.1080/17477160902846179. Strong W B, Malina RM, Blimkie C J, Daniels SR, Dishman RK, Gutin B, Hergenroeder AC, Must A, Nixon PA, Pivarnik JM, Rowland T, Trost S, Trudeau F. 2005. Evidence based physical activity for school-age youth. J Pediatr 146:732737. Tabachnick BG, Fidell LS. 1996. Using multivariate statistics. New York: Harper Collins College. Telama R, Yang X, Viikari J, Va¨lima¨ki I, Wanne O, Raitakari O. 2005. Physical activity from childhood to adulthood. A 21 year tracking study. Am J Prev Med 28:267273. Thorburn AW, Proietto J. 2000. Biological determinants of spontaneous activity. Obes Reviews 1:8794. Tremblay MS, Barnes JD, Copeland JL, Esliger DW. 2005. Conquering childhood inactivity: Is the answer in the past? Med Sci Sports Exerc 37:11871194. Trost SG, Pate RR, Freedson PS, Sallis JF, Taylor WC. 2000. Using objective physical activity measures with youth: How many days of monitoring are needed? Med Sci Sports Exerc 32:426431. Trost SG, Pate RR, Sallis JF, Freedson PS, Taylor WC, Dowda M, Sirad J. 2002. Age and gender differences in objectively measured physical activity in youth. Med Sci Sports Exerc 34:350355. Trost SG, Ward DS, Moorehead SM, Watson PD, Riner W, Burke JR. 1998. Validity of the computer science and applications (CSA) activity monitor in children. Med Sci Sports Exerc 30:629633. van Sluijs EMF, McMinn AM, Griffin SJ. 2008. Effectiveness of interventions to promote physical activity in children and adolescents: Systematic review of controlled trials. Br J Sports Med 42:653657. Wickel EE, Eisenmann JC. 2007. Contribution of youth sports to total daily physical activity level among 6- to 12yr-old boys. Med Sci Sports Exerc 39:14931500. Wickel EE, Eisenmann JC, Pangrazi RP, Vincent SD, Raustorp A, Tomson LM, Cuddihy TF. 2007. Do children take the same number of steps every day? Am J Hum Biol 19:537543. Wilkin TJ, Mallam KM, Metcalf BS, Jeffrey AN, Voss LD. 2006. Variation in physical activity lies with the child, not his environment: Evidence for an ‘activitystat’ in young children (EarlyBird 16). Int J Obes 30:10501055.
This paper was first published online on iFirst on 9 May 2009.
