Background
Reducing sedentary time (ST) and physical inactivity are among the most important global health challenges of the twenty-first Century [
1]. There is a large body of evidence which suggests that decreasing ST is associated with lower health risk in children [
2] and that physical activity (PA) in preschool children improves health outcomes [
3]. Children become more sedentary with age [
4] and furthermore accumulate their sedentary time in increasingly prolonged bouts [
5]. Levels of ST and PA throughout childhood are to a great extent not according to recommendations and both ST and PA track from early childhood into adolescence [
6]. This, highlights the importance of establishing recommended levels of PA and ST during early childhood.
Accumulating evidence suggests that the amount of time children spend sedentary may be associated with increased risk of developing metabolic disease independent of moderate to vigorous PA (MVPA) and obesity [
7‐
9], at least when subjective measures such as screen-time are used. On the contrary, observational studies, with objectively measured ST, investigating the effects of prolonged ST and breaks in ST on health outcomes in the pediatric population have failed to detect an association between breaks in ST or sedentary bout length with metabolic disease risk [
5,
9‐
11]. Owing to the evident association between screen-time and several health outcomes [
12], government authorities have created specific screen-time guidelines, advocating no more than 1 h of screen-time per day, for children aged two- to four-years of age [
13].
Most previous research assessing activity levels in preschool children focus on average activity levels throughout the week. Nevertheless, recent evidence suggests that levels of PA and ST differ noticeably over the course of the day and week [
14]. Describing patterns of PA and ST throughout the day provides information when children are less active and thus susceptible to efforts to increase activity. The present study aims to describe average levels and patterns of PA and ST across the day and week and activity pattern differences between time spent in and outside the preschool, using hour specific accelerometer data, in a population-based sample of four-year old Swedish children. In addition, parent-reported screen-time behaviors and accelerometer measured bouts of ST in different lengths will be assessed, which will provide detailed information on how high- and low-intensity activities are distributed across the day and week.
Discussion
The present study examined average levels and patterns of PA and ST across the day and week and screen-time behaviors in a population-based sample of four-year old Swedish children. Our findings showed that children have different daily patterns of PA and ST on weekdays compared with weekend days. Children were more active and less sedentary during preschool hours on weekdays compared with hours spent outside preschool. Conversely, children were less active and more sedentary from 9 a.m. to 3 p.m. on weekend days compared with 7 a.m. to 9 a.m. and 3 a.m. to 9 p.m. Furthermore, screen-time behavior differed significantly over the week with 86% and 97% of children engaging in more than the recommended maximum of 1 h of screen-time per day on weekdays and weekend days, respectively. The present study provides novel information on how activity and sedentary patterns differ across the day and throughout the week in a population based sample of young Swedish children. These objectively measured time-specific observations may be important for intervention development, targeting periods when children are sedentary and less active. For example, after preschool hours on weekdays and from 9 a.m. to 3 p.m. on weekend days.
Overall, levels of PA and ST observed in the present study are similar to other international studies with objective measured PA and ST. In accordance with findings in the present study, a study on 593 four-year old British children using the Actiheart accelerometer, indicated that children are less active and more sedentary outside preschool hours [
14]. However, preschool hours differ substantially between UK (usually 12–5 p.m.) and Sweden (9 am.-3 p.m.) and only 45% of the children in the study by Hesketh et al. attended preschool full-time compared with 94% in the present study. The study by Hesketh et al. used a different accelerometer and only used 1 day of accelerometer data as a valid measure of children’s habitual PA with no analyses of activity patterns across the week, which makes comparability with the present study somewhat limited. A similar study on 703 Australian three- to five-year old preschool children showed that children are highly sedentary and engage in low levels of MVPA in early afternoon on weekdays and around midday on weekends [
27]. In contrast to findings in the current study, ST was the lowest and participation in MVPA was the highest in children from afternoon until the evening on weekdays and on weekends. However, the study by van Cauwenberghe et al. used a different accelerometer and children attended preschool fewer hours and to less extent compared to the present study. The increase in MVPA and decrease in ST from awakening until startf of preschool at 9 a.m., found in the present study, may indicate that children routinely use active transportation to the preschool. In addition, the sharp increase in ST from 11 a.m., seen in Fig.
2, may to some extent be explained by the fact that most Swedish preschools serve lunch, which is a sedentary activity, at 11 a.m.
A recent review on 1485 preschooler’s activity levels, comprising 91 preschools, measured with accelerometers or direct observation, showed that children three- to five-years of age spent approximately 3% of the preschool day in MVPA and the majority of the day (>50%) sedentary [
28]. This is substantially lower than the 7.6% MVPA, and somewhat higher than the 48% ST, presented in the present study. However, the review included studies from several countries, with somewhat different childcare policies possibly impacting levels of PA, where children on average spend 22 h per week in childcare, which is considerably lower than in the present study.
Objectively measured activity data on Swedish preschool children is sparse. A small Swedish sample of four-year old children showed that children spend 19 out of 445 min (4.3%) in MVPA during preschool hours [
29]. However, the study sample was small (
n = 24) and the authors only analyzed the Vt axis in 15 s epochs from a uni-axial accelerometer, which may oversight preschool children’s sporadic activity patterns [
30].
Multiple factors, including measure of PA and ST, data processing and the population studied contribute to the large discrepancies observed across studies. Studies using accelerometers to assess activity in preschool children often use different cut-points to define different intensities of PA and ST. The majority of studies assessing activity with accelerometers in preschool children use uni-axial accelerometers only analyzing the vertical (V
t) axis [
31]. However, a tri-axial accelerometer, such as the GT3X+, may better estimate intensity from free-living daily activities [
32]. Several studies assessing ST in preschool children use a cut-point of <100 cpm for the V
t axis. Conversely, when analyzing tri-axial accelerometer data from the V
m axes, higher cut-points are needed to classify ST [
31]. In short, varying data processing protocols and lack of raw cpm comparison of prevalence estimates in young children’s activity levels make comparability between studies, and compliance with guidelines, challenging [
33].
The present study observed that children spend a large proportion of the day in LPA, both on weekdays and weekend days as well as during and outside preschool hours. However, health benefits from LPA in preschool children are not fully understood. Studies in four-year old children have shown that MVPA, and not LPA, is positively associated with bone density [
34] and lower fat mass [
22]. Indeed, the importance of activity intensity requires further investigation to direct future activity guidelines for young children.
Substantial evidence indicates that girls on average are less active and more sedentary compared with boys [
4]. This was also observed in the present study. However, this study adds to the current literature by emphasizing that these differences also vary across the day and week. Activity levels between the sexes were most apparent outside preschool, especially during weekdays, indicating that girls are relatively more active and less sedentary than boys in preschool, and vice versa less active and more sedentary when interacting with their parents. Substantial evidence indicates that parent to child correlates of PA is stronger within the same sex, that is father’s levels of PA correlate stronger to boy’s activity levels [
35], whereas mother’s levels of PA correlate stronger with girls activity levels [
36]. These activity pattern differences across the day between the sexes may direct future interventions to consider a time-based focus to differently target girls and boys. For example, engaging girls in active play during preschool hours and emphasizing maternal PA on both weekdays and weekend days.
Data on screen-time from the present study showed that the majority of children engaged in more than the recommended maximum of 1 h of screen-time per day [
37]. This is a higher prevalence compared with a recent population based Canadian study showing that 64% of children aged two- to four-years of age exceeded the maximum screen-time recommendations [
38]. However, these differences in prevalence of screen-time may be explained by age differences between the study populations. The high prevalence of screen-time in the present study is alarming, since substantial evidence from a recent systematic review, in children aged zero- to four-years of age, indicates that screen-time, but not objectively measured ST [
2], is consistently negatively associated with health outcomes in a dose-response manner [
12].
Important to consider is that the present study only measured screen-time outside preschool hours. A recent systematic review on the prevalence of screen-time behaviors in children during preschool hours reported that children on average spend 1 h per day sedentary in front of a screen [
39]. Thus, the prevalence of total daily screen-time on weekdays in the present study may be higher than reported. In addition, the present study did not measure other forms of screen time (i.e. mobile phones, tablets etc.) besides TV viewing and computer games. Possibly, the observed screen-time prevalence of 2.5 h per day on weekend days is a more representative prevalence estimate of average screen-time behavior in Swedish four-year old children.
Strengths and limitations
The strengths of the present study a largest study population of Swedish four-year old children entailing objectively measured activity patterns across the day. The study area, including larger cities, medium sized cities and countryside with low population density, covers approximately 37% of the Swedish population and is fairly representative of Sweden with children from all socio-economic strata [
15]. However, results from the present study are only generalizable for the Swedish population. In addition, we used accelerometers to assess PA and ST patterns across the day and both measurements and data processing followed best practices [
40]. We have used cut-points validated in four-year old children developed specifically for the accelerometer (GT3X+) used in the present study [
25]. However, analysis of the Vm axes and higher cut-points to define ST compared with several other studies analyzing the Vt axis limits the comparability with previous research [
31]. Additional limitations with using accelerometers to define ST is their inability to separate sitting from standing [
41]. Hence, we may have misclassified standing time as ST, which according to the latest definition is not defined as time spent sedentary [
42].
We included all children with at least 4 days of valid PA data, including one weekend day, and sensitivity analyses on children with six or more days of PA data did not alter the results. A limitation possibly affecting results presented in the statistical models is the lack of appropriate data at preschool level. Thus, we were unable to adjust your analyses for possible clustering of participants within preschools.
The use of hour-specific data, dividing the day into two time-periods, enables a more detailed description of time-based patterns of children’s activity levels across the day and week. These segments reflect preschool children’s daily differences in activity intensity, both between time spent in and outside the preschool and between weekdays and weekend days. However, time-matched data on children not attending preschool was not available and we were therefore unable to assess what influence preschool attendance had on children’s activity patterns across the day. Nevertheless, this unique hour by hour activity data across the day on both weekdays and weekend days’ pinpoints specific periods throughout the day and week where public health interventions may be more likely to show a beneficial impact on children’s activity levels.