Skip to content

Advertisement

  • Research article
  • Open Access
  • Open Peer Review

Associations between children’s independent mobility and physical activity

  • 1Email author,
  • 1,
  • 2,
  • 3 and
  • 4
BMC Public Health201414:91

https://doi.org/10.1186/1471-2458-14-91

  • Received: 19 August 2013
  • Accepted: 23 January 2014
  • Published:
Open Peer Review reports

Abstract

Background

Independent mobility describes the freedom of children to travel and play in public spaces without adult supervision. The potential benefits for children are significant such as social interactions with peers, spatial and traffic safety skills and increased physical activity. Yet, the health benefits of independent mobility, particularly on physical activity accumulation, are largely unexplored. This study aimed to investigate associations of children’s independent mobility with light, moderate-to-vigorous, and total physical activity accumulation.

Methods

In 2011 - 2012, 375 Australian children aged 8-13 years (62% girls) were recruited into a cross-sectional study. Children’s independent mobility (i.e. independent travel to school and non-school destinations, independent outdoor play) and socio-demographics were assessed through child and parent surveys. Physical activity intensity was measured objectively through an Actiheart monitor worn on four consecutive days. Associations between independent mobility and physical activity variables were analysed using generalized linear models, accounting for clustered sampling, Actiheart wear time, socio-demographics, and assessing interactions by sex.

Results

Independent travel (walking, cycling, public transport) to school and non-school destinations were not associated with light, moderate-to-vigorous and total physical activity. However, sub-analyses revealed a positive association between independent walking and cycling (excluding public transport) to school and total physical but only in boys (b = 36.03, p < 0.05). Frequent independent outdoor play (three or more days per week) was positively associated with light and total physical activity (b = 29.76, p < 0.01 and b = 32.43, p = 0.03, respectively). No significant associations were found between independent outdoor play and moderate-to-vigorous physical activity. When assessing differences by sex, the observed significant associations of independent outdoor play with light and total physical activity remained in girls but not in boys. All other associations showed no significant differences by sex.

Conclusions

Independent outdoor play may boost children’s daily physical activity levels, predominantly at light intensity. Hence, facilitating independent outdoor play could be a viable intervention strategy to enhance physical activity in children, particularly in girls. Associations between independent travel and physical activity are inconsistent overall and require further investigation.

Keywords

  • Cross-sectional
  • Active travel
  • Unsupervised outdoor play
  • Movement
  • Young people
  • Actiheart

Background

In Australia, 60% of children aged 9-13 years do not meet current national guidelines for physical activity recommending that children engage in at least 60 minutes of moderate-to-vigorous physical activity every day [1]. Habitual physical activities such as playing outdoors, and walking and cycling for transport provide many opportunities to be physically active throughout the day [2]. Physical activity is vital for children’s bone health, motor skills, physical fitness, healthy weight and protection against chronic diseases later in life [37]. Instilling habitual physical activity in young children is crucial as physical activity behaviours tend to track from childhood to adolescence and adulthood [8].

Children’s independent mobility describes the freedom of those aged under 18 years to play outdoor and travel to places without adult supervision [9, 10]. Independent mobility provides an important source for children’s habitual physical activity. Children’s unsupervised neighbourhood play and exploration brings psychosocial, cognitive and developmental benefits in the form of social interactions with peers, spatial and traffic safety skills for navigating in public spaces, and maturity in regard to decision making [1113]. Independent mobility starts to increase between the ages 8-13 years, often coinciding with the transition from primary to secondary school; this is in response to parents recognising increasing physical and cognitive capabilities as children age [10, 14]. Boys tend to have higher levels of independent mobility compared to girls [1518]. This can be explained by parent’s tendency to be more protective towards girls than boys, and differences in socialising between boys and girls [18]. For example, boys tend to form casual friendships through local outdoor activities such as football which requires independent mobility nearby the home [18]. In contrast, girls tend to form close friendships through talking and socialising at friend’s home, shopping centres or cinemas which often requires motorised transport [18]. For children, independent travel usually involves active modes of travel such as walking and cycling, as well as using public transport [19]. In many developed countries including the United States of America, Australia, Italy, Denmark, Finland and Norway, children’s walking and cycling for transport has dramatically declined [14, 1922]. Long-term trend data from England showed that in 1971, 86% of primary school children were allowed to travel home from school alone, whereas in 1990 this proportion was 35%, and in 2010 it was only 25% [23]. Reasons for the decreases in children’s independent mobility include parental concerns about road safety and stranger danger, greater complexity in families’ daily schedules, and longer travel distances to schools, shops and recreational facilities [14, 15, 24, 25]. The declines in independent mobility are concerning from a child development and public health perspective including contributions to low levels of physical activity in children.

Few studies have investigated the potential of independent mobility to enhance children’s physical activity levels [26]. Three studies focused on associations between parental licences for their child’s independent travel (but not actual travel behaviour) in public spaces and physical activity, finding significant positive associations [15, 16, 27]. Another study [28] examined children’s independent outdoor play in parks, playgrounds and other open spaces; findings showed that children were more physically active when playing solely in the presence of peers compared with adults. However, the small number of existing studies in this research field limits the generalisability of findings. Moreover, existing studies have not differentiated between independent mobility domains (travel, outdoor play) and physical activity intensities (light, moderate-to-vigorous) [15, 16, 27]. Evidence on whether independent mobility enhances physical activity across various intensities is valuable since both light and moderate-to-vigorous physical activity provide health benefits in children [5, 7, 29]. Notably, light physical activity has been ignored in previous studies though it is associated with improvements in blood pressure, insulin and HDL-cholesterol levels in children; health benefits which are particularly important for overweight and obese children in whom metabolic risk factors are increasingly prevalent [30, 31].

This paper aims to extend the current evidence base by
  1. (1)

    investigating associations between children’s self-reported independent mobility (including independent travel and outdoor play) and objectively measured light, moderate-to-vigorous, and total physical activity,

     
  2. (2)

    testing whether associations between children’s independent travel and physical activity would differ when independent travel involves walking and cycling versus walking, cycling and public transport,

     
  3. (3)

    examining differences in all observed associations by sex.

     

As such, information gathered from this study could help design interventions to promote children’s independent mobility with the aim to enhance habitual physical activity at various intensities.

Methods

Study design

This cross-sectional study is based on data drawn from the Australian CATCH/iMATCH projects (Children's Activity, Travel, Connectedness and Health/Independent Mobility, Active Travel and Children’s Health). The interrelated CATCH/iMATCH projects were designed to investigate the role of policy, social, and built environments in influencing children's independent mobility, active travel, and related health outcomes [32]. Ethical clearance for conducting the CATCH/iMATCH projects was obtained from several universities (Central Queensland University, H10/11-170; The University of Melbourne, HREC 1135410; Curtin University, HR 140/2010; Griffith University ENV/31/10/HREC) and State Departments of Education and Training in Australia (Queensland (QLD), 550/27/1042; Victoria (VIC), 2011-001232; Western Australia (WA), D11/0698973).

Participants

A clustered sampling design and convenience sample were used in the CATCH/iMATCH projects. Overall, 375 children aged 8-13 years were recruited from nine public primary schools in Rockhampton (QLD), Brisbane (QLD), Melbourne (VIC) and Perth (WA). The schools were located in neighbourhoods with various levels of urbanisation (inner urban, n = 2; middle suburban, n = 2; outer suburban, n = 2; regional area, n = 3) but similar socio-economic status (SES) characteristics of student populations (representing a middle SES range). The selection of schools was guided by SES information available from Australia’s My School website (http://www.myschool.edu.au) which determined SES via several socio-demographic factors such as parents’ occupational level and level of educational attainment. School principals, teachers, parents and students (Years 3-7) were briefed about the study through informational materials and meetings held at the schools. Written consent by school principals, parents and children was a requirement for study participation.

Data collection

Data collection took place between June 2011 and November 2012 including the Australian winter and summer. Under researchers’ supervision, students completed a paper-based survey at school, either in class or during lunch recess. The paper-based survey administered to children assessed socio-demographics, travel and independent mobility behaviours. A paper-based survey administered to parents included information on socioeconomic status. Furthermore, children wore an Actiheart, a combined heart rate and accelerometer monitor (CamNtech Actiheart, Cambridge, UK), for four consecutive days, except during periods of heavy contact sports and water-based activities (e.g. showering, swimming). The Actiheart monitoring period included weekdays and weekend days, either consecutively from Thursday until Monday morning or from Friday until Tuesday morning. The Actiheart was attached directly onto the skin on the lower left side of the chest using two electrocardiograph (ECG) electrodes (model Red Dot 2238 soft cloth, 3 M) [33, 34]. One electrode was placed at the base of the child’s sternum and the other electrode placed horizontally to the left so the wire of the Actiheart was straight with tension, but not tight. Researchers demonstrated to the students how the Actiheart should be worn. The students then fitted the units and researchers checked the positioning.

Socio-demographic assessment

Sex and age in years was captured through the child surveys. Additional covariates were assessed through parent surveys administered via mail. These included parental education attainment and number of motor vehicles in the household. In this study, neighbourhood urbanisation in state capital city locations (Brisbane, Melbourne, Perth) was defined based on neighbourhood proximity to the central business district (CBD) with the general post office being a marker of the CBD centre. Inner urban, middle suburban, and outer suburban were defined as being within 5 km from CBD, 11-20 km from CBD and beyond 20 km from CBD, respectively. Regional area was defined as a regional town that is not a major state capital city of Australia.

Independent mobility assessment

Independent mobility was assessed through child self-report surveys; complementary parental proxy reports of children’s independent mobility behaviour could not be collected due to space restrictions in parent surveys. In the absence of standardised, validated measures of independent mobility, questions were derived from child surveys used in similar studies [9, 35] to capture (a) travel mode and accompaniment of active travel, and (b) frequency and accompaniment of outdoor play (predictor variables).

Children were asked ‘How do you usually travel to: (1) school; (2) local shops; (3) local friend’s houses; (4) local parks and playgrounds; and (5) organised activities (e.g. at a local sports club, church or recreational centre)?’ Children selected the response option best representing their predominant travel mode (walk, bicycle, take public transport, be driven) and accompaniment status (alone, with other children, with an adult) for each destination. Questions relating to non-school destinations also included the option ‘Don’t go there’. Responses to these five destinations were collapsed to create two variables: (1) usual travel to school; and (2) usual travel to local non-school destinations. Response options for the ‘usual travel to school’ variable were dichotomised into ‘independently mobile’ (i.e. walk/bicycle/take public transport alone or with other children) or ‘not independently mobile’ (i.e. walk/bicycle/take public transport or car with an adult). Using this same classification of independent mobility the ‘usual travel to local non-school destinations’ variable was calculated as the ratio of independently mobile trips from all trips a child undertook to non-school destinations. Based on the median distribution this variable was then dichotomised into ‘higher levels of independent travel’ (i.e. 33-100% of all trips occur independently mobile) or ‘lower levels of independent travel’ (i.e. < 33% of all trips occur independently mobile).

Children were also asked ‘How often do you usually play outdoors in your neighbourhood, for example, on the street, in a nearby park, bush area, or on a playground?’ Children selected from the following response options: ‘5 or more days per week’; ‘3-4 days per week’; ‘1-2 days per week’ and ‘Never’. If children played outdoors they also reported on their usual accompaniment status: alone; without an adult present but with other children; or always with an adult present irrespective of other children. In this sample, all children responded that they play outdoors independently (i.e. alone or with other children). Based on the median distribution the response categories were dichotomised into ‘higher levels of independent outdoor play’ (i.e. played ≥ 3 days per week, alone or with other children) or ‘lower levels of independent outdoor play’ (i.e. never or played 1-2 days per week, alone or with other children).

Physical activity assessment

Light, moderate, vigorous, and total physical activity (outcome variables) were assessed using accelerometer counts derived from the Actiheart unit. The Actiheart accelerometer has demonstrated good validity (R2 = 0.7) when tested with 13-year old children in a laboratory setting [36]. At the conclusion of wearing the Actiheart units, data were downloaded using the CamNtech Actiheart software. Subsequently, the Actiheart data was transferred into Microsoft Excel and data reduction was carried out to extract daily minutes spent in light, moderate and vigorous physical activity out using custom written software (National Instruments Labview). Wear time and non-wear time: Actiheart recordings were classified by the Actiheart software as being ‘OK’, ‘recovered’, ‘interpolated’, ‘lost’ or ‘not worn’. Only accelerometer counts recorded under the classifications ‘OK’, ‘recovered’ or ‘interpolated’ were used for calculating time spent in physical activity; this shows that the Actiheart was detecting a signal and suggests wear time. Actiheart recorded Recordings under the classifications ‘lost’ and ‘not worn’ suggest non-wear time or the units were not detecting a signal; hence these were excluded from analyses. Epoch length: The epoch length was aggregated to 60 seconds to account for differences in epoch length settings across sites. Sixty seconds epochs have shown to be appropriate for measuring moderate-to-vigorous physical activity in children [37] and allowed comparison with similar studies [3841]. Cut-points: For inclusion in analyses, children were required to record accelerometer counts on at least two days for a minimum of 360 minutes per day between the hours of 6.00 am-11.00 pm. This threshold combination was used based on availability of valid Actiheart accelerometer data and sensitivity analyses. The child-specific ActiGraph cut points developed by Evenson et al. [42] were adopted for classifying light, moderate and vigorous physical activity (being 101-2295, 2296-4011 and ≥ 4012 counts per minute, respectively). These values showed acceptable classification accuracy for light, moderate and vigorous physical activity with ROC-AUC (area under the receiver operating characteristic curve) curve values ranging between 0.70-0.84 [42, 43]. Following the protocols of Ridgway et al. [44] a conversion factor of five (i.e. ActiGraph counts = Actiheart counts x 5) was applied to generate comparable cut points for Actiheart data, resulting in final cut points for light, moderate and vigorous physical activity of 21-459, 460-802 and ≥ 803 counts per minute, respectively. Average number of minutes per day spent in light, moderate-to-vigorous (i.e. moderate and vigorous physical activity combined) and total physical activity were extracted.

Statistical analyses

Randomly missing data on independent travel to school and non-school destinations, independent outdoor play and socio-demographics were handled using multiple imputation following Rubin’s [45] conventional method. This involved three steps: (1) creating five data sets with imputed values for the relevant independent mobility and socio-demographic variables with consideration of all variables to be included in analyses; (2) conducting the analyses in all data sets; (3) calculating averages from each data output. Missing values were imputed via 5-fold multiple imputation involving repeated sampling the empirical multivariate distribution using the Amelia package [46] within R statistical programming environment [47]. Reported coefficient means and standard-errors were aggregated across imputed data-sets. Chi-square tests and independent t-tests were conducted to assess differences in socio-demographics, independent mobility and time spent in physical activity (moderate-to-vigorous physical activity was log-transformed) between boys and girls, and between children included and excluded from analyses. Subsequently, a series of regression analyses were carried out using generalized linear models to examine associations between children’s independent mobility and time spent in light, moderate-to-vigorous and total physical activity. Analyses were adjusted for clustering by school location using robust standard errors; the school location represented neighbourhoods with various levels of urbanisation. Included covariates were age, parental education attainment, number of motor vehicles in the household and Actiheart wear time. Given documented sex differences in children’s independent mobility and physical activity [16, 18] an interaction term was entered in the model to test for differences by sex. Furthermore, all analyses were repeated in a sub-sample of children (n = 179) that excluded those who travelled by public transport (n = 12). These sub-analyses tested whether associations between independent travel and physical activity differ when focusing specifically on independent travel by walking and cycling. Associations are presented using unstandardised beta coefficients (b), standard errors, confidence intervals and p-values. The significance level was set at 0.05. Analyses were performed in IBM SPSS Statistics (version 19.0) and StataMP (version 12.0).

Results

Study population

Of approximately 1,534 eligible children invited across all participating schools, 375 (24%) children provided written parental consent and child assent to study participation. Thirteen children dropped out of the study because they either left the school or withdrew from data collection. Actiheart data were non-valid (i.e. < 2 days of accelerometer recordings for at least 360 minutes between 6.00 am-11.00 pm) in 136 children, and imputation of independent mobility and socio-demographic data was not suitable for 35 children because responses such as ‘Other’ and ‘Don’t go there’ could not be classified into available response categories. Hence, these children were excluded from analyses. After data imputation in 65 cases, the final sample included in analyses was 191 (51%) out of 375 recruited children.

There were no significant differences in socio-demographics, levels of independent mobility, or physical activity between children included and excluded in analyses (data not reported), except for neighbourhood urbanisation (p < 0.01). A higher proportion of excluded children were from schools in outer suburban neighbourhoods (64.9% versus 35.1%), and a lower proportion were from schools in inner urban (35.7% versus 64.3%) and regional neighbourhoods (27.6% versus 72.4%). Table 1 shows descriptive characteristics of the study participants included in analyses. Overall, 36% of children commuted independently to school and 43% usually travelled independently to non-school destinations. Sub-analyses of travel modes showed that most children (57%) were driven to school and non-school destinations (data not presented in Table 1). All children engaged in independent outdoor play; 67% of them at higher levels (≥ 3 days per week). There were no significant differences in socio-demographics, independent mobility and physical activity levels between boys and girls.
Table 1

Descriptive characteristics of the study participants included in analyses

 

All

Boys

Girls

p-value

N (%)

191

73 (38.0)

118 (62.0)

 

Age in years, mean (SD)

10.64 (0.89)

10.76 (0.91)

10.57 (0.87)

0.15

Level of urbanisation, n (%)

    

 Inner urban

54 (28.3)

22 (30.1)

32 (27.1)

 

 Middle suburban

35 (18.3)

14 (19.2)

21 (17.8)

 

 Outer suburban

47 (24.6)

14 (19.2)

33 (28.0)

 

 Regional area

55 (28.8)

23 (31.5)

32 (27.1)

 

Parental education attainment 1 , n (%)

   

0.48

 High school

63 (32.9)

25 (34.2)

38 (32.2)

 

 Trade/apprenticeship/certificate/diploma

72 (37.9)

24 (32.9)

48 (40.7)

 

 University degree

56 (29.2)

24 (32.9)

32 (27.1)

 

Motor vehicles in household, n (%)

   

0.79

 0

3 (1.6)

1 (1.4)

2 (1.7)

 

 1

45 (23.6)

20 (27.4)

25 (21.2)

 

 2

109 (57.0)

40 (54.8)

69 (58.5)

 

 ≥ 3

34 (17.8)

12 (16.4)

22 (18.6)

 

Usual travel to school, n (%)

   

0.88

 IM

69 (36.1)

27 (37.0)

42 (35.6)

 

 Non-IM

122 (63.9)

46 (63.0)

76 (64.4)

 

Usual travel to non-school destinations, n (%)

   

0.47

 Higher IM

82 (42.9)

35 (47.9)

47 (39.8)

 

 Lower IM

109 (57.1)

38 (52.1)

71 (60.2)

 

Usual outdoor play, n (%)

   

0.77

 Higher IM

126 (66.5)

46 (63.0)

80 (67.8)

 

 Lower IM

65 (33.5)

27 (37.0)

38 (32.2)

 

Average daily minutes in light PA

    

Mean (SD)

381.57 (95.72)

375.87 (86.7)

384.67 (89.90)

0.55

Average daily minutes in MVPA

    

Mean (SD)

57.22 (45.90)

64.60 (48.59)

52.65 (43.80)

0.08

Median (IQR)2

45.30 (44.80)

52.00 (50.30)

42.40 (40.30)

0.07

Average daily minutes in total PA

    

Mean (SD)

438.90 (95.3)

440.47 (110.0)

437.32 (118.50)

0.85

Abbreviations: IM independent mobility, IQR interquartile range, MVPA moderate-to-vigorous physical activity, PA physical activity, SD standard deviation.

1Highest education attainment was reported by one parent completing a survey.

2The distribution of MVPA was positively skewed; hence means, SD, medians and IQR are presented. MVPA was log-transformed prior to conducting independent t-test.

Associations between independent mobility and physical activity

Table 2 presents associations between independent mobility and daily time spent in light, moderate-to-vigorous, and total physical activity. Compared to adult-accompanied travel, children’s independent travel (walking, cycling, public transport) to school and non-school destinations were not significantly associated with light, moderate-to-vigorous and total physical activity. However, children who frequently engaged in independent outdoor play (≥ 3 days per week) accumulated more daily minutes of light physical activity (b = 29.76, p < 0.01) and total physical activity (b = 32.43, p = 0.03) than those who played outdoors less often. The frequency of independent outdoor play was not significantly associated with moderate-to-vigorous physical activity.
Table 2

Associations between independent mobility and mean daily minutes spent in light, moderate-to-vigorous and total physical activity 1

 

Light PA

MVPA

Total PA

 

b

SE2

95% CI

p-value

b

SE

95% CI

p-value

b

SE

95% CI

p-value

Usual travel to school

            

IM (n = 68)

-0.55

9.05

-19.12 – 18.02

0.95

-1.48

5.11

-12.86 – 9.90

0.80

-2.53

11.35

-25.92 – 20.85

0.83

Non-IM (n = 123)

 

 

 

Usual travel to non-school destinations

            

Higher IM (n = 82)

-13.24

8.95

-35.85 – 9.37

0.25

4.81

7.76

-11.17 – 20.78

0.56

-4.10

15.44

-39.96 – 31.76

0.82

Lower IM (n = 109)

 

 

 

Usual outdoor play

            

Higher IM (n = 127)

29.73

9.57

9.39 – 50.06

<0.01

3.96

7.75

-12.01 – 19.93

0.63

32.43

13.53

3.23 – 61.62

0.03

Lower IM (n = 64)

 

 

 

Abbreviations: b unstandardised beta coefficient, 95% CI 95% confidence interval, p p-value, SE robust standard error, IM independent mobility, MVPA moderate-to-vigorous physical activity, PA physical activity.

1Adjusted for clustering by school location (representing levels of neighbourhood urbanisation), age, sex, parental education attainment, number of motor vehicles in the household, Actiheart wear time.

2Standard errors, confidence intervals and p-values were calculated from imputed beta coefficients using Rubin’s [38] conventional method.

Note, independent travel to school and non-school destinations included walking, cycling and public transport.

We tested whether the associations between independent travel and physical activity would differ when excluding travel involving public transport. Findings differed only slightly; independent outdoor play remained significantly associated with light physical activity (b = 26.62, p < 0.05) but not with total physical activity (b = 28.26, p = 0.13). All other associations remained non-significant.

Differences by sex

Analyses by sex showed no significant associations of independent travel (walking, cycling, public transport) to school and non-school destinations with light, moderate-to-vigorous and total physical activity. The observed positive associations between frequent independent outdoor play (≥ 3 days per week) and light and total physical activity were significant in girls (b = 41.41, p < 0.01 and b = 46.02, p = 0.01, respectively) but not in boys.

We tested whether the associations between independent travel and physical activity would differ when excluding travel by public transport. Results differed slightly. Independent walking and cycling to school was positively associated with total physical activity but only in boys (b = 36.03, p < 0.05), not in girls. Positive associations of frequent independent outdoor play (≥ 3 days per week) with light and total physical activity were significant in girls only (b = 48.46, p < 0.001 and b = 41.95, p = 0.03, respectively). All other associations were non-significant.

Discussion

This study investigated associations between children’s independent mobility and light, moderate-to-vigorous, and total physical activity. In this sample, children’s independent travel (walking, cycling, using public transport) to school and non-school destinations was not related to higher levels of physical activity compared to adult-accompanied travel. These findings were unexpected and contrary to those found in similar studies [15, 16, 27]. There are several possible explanations for these findings. Firstly, previous studies did not examine physical activity intensities (light, moderate-to-vigorous) as an outcome but overall physical activity or physical activity types (walking, outdoor play). Secondly, the leisure-time activities children pursue at non-school destinations will likely also influence their physical activity levels [48], besides the commuting itself. However, detailed information on what leisure-time activities children specifically engaged in at non-school destinations was not available for analysis in this study but is worth examining in future research. Thirdly, associations between independent travel and physical activity can differ when focusing specifically on children’s independent walking and cycling for transport. For example, this study revealed a significant positive association between independent walking and cycling to school and total physical activity; however, only in boys. This suggests that significant increases in physical activity generated by independent travel require active modes of travel. Moreover, boys rather than girls may increase physical activity levels from independent travel, possibly because parents tend to grant boys more independence for travels in local areas whereas parents tend to be more restrictive towards girls [18].

Children who frequently engaged in independent outdoor play (≥ 3 days per week) accumulated significantly more minutes of light and total physical activity than those who engaged in it less frequently. In contrast, no significant associations were observed between children’s frequency of independent outdoor play and moderate-to-vigorous physical activity. The difference in associations across physical activity intensities may be explained by built-environmental and social-environmental factors. For example, wide open spaces such as green areas, parks, playgrounds and quiet streets have eroded over time, and public play areas have become less accessible on foot for children due to increased traffic volume and car-oriented urban planning [49]. Additionally, parental concerns about road safety coupled with fears about stranger danger have led to many parents restricting children’s freedom to play in open spaces beyond the close confines of the home garden and yards [49]. These various environment impediments for accessing outdoor areas outside of the home may reduce children’s engagement in outdoor play at higher intensities [50].

Detailed analyses by sex showed that the observed positive associations between independent outdoor play and light and total physical activity remained significant in girls but not in boys. This suggests that independent outdoor play is particularly important for physical activity accumulation in girls. A reason may be that parents’ tend to be more protective towards girls than boys; hence, restrictions of independent mobility usually affect girls more than boys [18, 51]. As such, for girls, independent outdoor play in the home garden and yards might provide an important source for physical activity engagement, and an opportunity for talking and socialising with friends [51]. This may be less relevant for boys who, compared to girls, accumulate physical activity more often through organised sports, particularly around the ages 9-11 [51, 52].

Strengths, limitations and future research

Strengths of this study include objective physical activity measurement across various physical activity intensities and examination of diverse children’s independent mobility behaviours (outdoor play, travel to school and non-school destinations), rather than independent mobility licences that have been previously examined [15, 16, 27]. The distinction between ‘licences’ and ‘actual behaviours’ is important and requires more attention in future studies as children may be allowed to travel independently to some places but end up being driven because it is faster, more convenient and more comfortable than walking or cycling. Other methodological strengths include the use of multiple imputation techniques to account for randomly missing data, sensitivity analyses to help determine thresholds for valid Actiheart data, adjustment for a range of potential confounders, and testing for differences by independent travel modes and sex. This study also had several limitations. Social and environmental determinants such as perceptions of neighborhood safety, distances to travel destinations, weather, street connectivity, availability of cycling and walking trails were not considered in this study though these will likely influence children’s levels of independent mobility and physical activity [52, 53]. Our analyses accounted for neighbourhood urbanisation though. Other limitations of this study include the cross-sectional study design, the reliance on non-standardised children’s self-report measures for assessing independent mobility, the low response rate and low adherence to wearing an Actiheart monitor. The latter resulted in a short physical activity monitoring period which may not have fully recorded children’s usual physical activity patterns. Moreover, the use of a small convenience sample limits the generalisability of findings. Although a 60 seconds epoch length has shown to be appropriate for measuring moderate-to-vigorous physical activity in children [37], a lower epoch length may have better captured children’s intermittent and spontaneous physical activity patterns [54]. We could not use a shorter epoch length as epoch lengths differed across the study sites. Overall, we suggest that associations between children’s independent mobility and physical activity receive further attention in future research given the potential of independent mobility to enhance habitual physical activity. This may be particularly important for children who do not obtain physical activity through organised sports. The low response rate found in this study demonstrates the common challenge of recruiting children into behavioural risk factor studies; however, little guidance exists on recruitment methods [55]. More evidence should be gathered on successful recruitment strategies applied in pediatric populations [56]. Furthermore, the application of standardised multiple imputation techniques rather than simple complete cases analysis is advisable in studies with large amounts of missing data [57]. Imputation of missing accelerometer data is emerging but not yet standardised [58], hence it was not carried out in this study.

Conclusions

In this study, children’s independent travel in the form of walking, cycling and public transport was not correlated with significant gains in daily physical activity. However, independent travel including solely walking and cycling showed a positive association with total daily physical activity but only in boys. Frequent independent outdoor play was significantly associated with accumulation of light intensity physical activity. Hence, promoting independent outdoor play can be a viable intervention strategy to enhance physical activity in children, especially in girls. Further investigations of independent travel by transport mode, the distances away from home children are allowed to travel and play outdoors independently, and the physical and social environments that support these behaviours are needed to inform future interventions.

Abbreviations

HDL-cholesterol: 

High-density lipoprotein cholesterol

CATCH project: 

Children's activity, travel, connectedness and health project

iMATCH project: 

Independent mobility, active travel and children’s health project

QLD: 

Queensland

VIC: 

Victoria

WA: 

Western Australia, SES, Socio-economic status

ECG: 

Electrocardiograph

ROC: 

Receiver operating curve

PA: 

Physical activity

MVPA: 

Moderate-to-vigorous physical activity

IM: 

Independent mobility

SD: 

Standard deviation

IQR: 

Interquartile range

b: 

unstandardised beta coefficient

SE: 

robust standard error

95% CI: 

95% confidence interval

p: 

p-value.

Declarations

Acknowledgements

This research was supported through the Australian Research Council (ARC) funded projects Children's Active Travel, Connectedness and Health (CATCH, Discovery Project DP1094495) and Independent Mobility and Active Travel in Children (iMATCH, LP 100100344). The former project includes financial support from the Merri Community Health Services Victoria, the Moreland City Council, Queensland Health and Queensland Transport which is gratefully acknowledged. The CATCH/iMATCH projects were led by Central Queensland University (Dr Mitch Duncan), Curtin University (Professor Carey Curtis), Griffith University (Dr Matthew Burke), University of Melbourne (Professor Carolyn Whitzman), University of New South Wales (Dr Paul Tranter) and Merri Community Health Services (Dr Christine Armit). Stephanie Schoeppe was funded through an Australian Postgraduate Award (Industry) financed through the iMATCH project. We would like to thank Andrea Cook, Sigrid Pitkin, Maryanne Tadic, Courtney Babb, Oscar Thomson, Kate Ringvall, Kala Wati, Farinaz Moghtaderiesfahani, Saba Al-Saleem, Carlie Hill and Cameron Jennings from the CATCH/iMATCH team for their contributions to participant recruitment and/or data collection. In addition, we are grateful to all participating schools, children and their parents for their collaboration.

Authors’ Affiliations

(1)
Central Queensland University, School of Human, Health and Social Sciences, Centre for Physical Activity Studies, Building 18, Bruce Highway, Rockhampton, QLD 4702, Australia
(2)
McCaughey VicHealth Centre for Community Wellbeing, The University of Melbourne, 207 Bouverie Street, Carlton, VIC 3010, Australia
(3)
Human Potential Centre, Auckland University of Technology, Mail #A-24, Private Bag 92006, Auckland, 1020, New Zealand
(4)
Health and Social Sciences, Central Queensland University, School of Human, University Drive, Branyan, Bundaberg, QLD, 4670, Australia

References

  1. 2007 Australian National Children's Nutrition and Physical Activity Survey. [http://www.commcarelink.health.gov.au/internet/main/publishing.nsf/Content/phd-nutrition-childrens-survey]
  2. Loprinzi PD, Cardinal BJ, Loprinzi KL, Lee H: Benefits and environmental determinants of physical activity in children and adolescents. Obes Facts. 2012, 5: 597-610. 10.1159/000342684.View ArticlePubMedGoogle Scholar
  3. Hallal PC, Victora CG, Azevedo MR, Wells JCK: Adolescent physical activity and health: a systematic review. Sports Med. 2006, 36: 1019-1030. 10.2165/00007256-200636120-00003.View ArticlePubMedGoogle Scholar
  4. Janz KF, Letuchy EM, Eichenberger-Gilmore JM, Burns TL, Torner JC, Willing MC, Levy SM: Early physical activity provides sustained bone health benefits later in childhood. Med Sci Sports Exerc. 2010, 42: 1072-1078.View ArticlePubMedPubMed CentralGoogle Scholar
  5. Dencker M, Andersen LB: Health-related aspects of objectively measured daily physical activity in children. Clin Physiol Funct Imaging. 2008, 28: 133-144. 10.1111/j.1475-097X.2008.00788.x.View ArticlePubMedGoogle Scholar
  6. Twisk JWR: Physical activity guidelines for children and adolescents a critical review. Sports Med. 2001, 31: 617-627. 10.2165/00007256-200131080-00006.View ArticlePubMedGoogle Scholar
  7. Biddle SJ, Gorely T, Stensel DJ: Health-enhancing physical activity and sedentary behaviour in children and adolescents. J Sports Sci. 2004, 22: 679-701. 10.1080/02640410410001712412.View ArticlePubMedGoogle Scholar
  8. Telama R: Tracking of physical activity from childhood to adulthood: a review. Obes Facts. 2009, 2: 187-195. 10.1159/000222244.View ArticlePubMedGoogle Scholar
  9. Hillman M, Adams J, Whitelegg J: One false move…a study of children's independent mobility. 1990, London: Policy Studies InstituteGoogle Scholar
  10. Whitzman C, Mizrachi D: Vertical living kids: Creating supportive high rise environments for children in Melbourne, Australia. 2009, Melbourne: The University of MelbourneGoogle Scholar
  11. Tranter P, Whitelegg J: Children's travel behaviours in Canberra: car-dependent lifestyles in a low density city. J Transp Geogr. 1994, 2: 265-273. 10.1016/0966-6923(94)90050-7.View ArticleGoogle Scholar
  12. Pooley C, Whyatt D, Walker M, Davies G, Coulton P, Bamford W: Understanding the school journey: integrating data on travel and environment. Environ Plan A. 2010, 42: 948-965. 10.1068/a41405.View ArticleGoogle Scholar
  13. Oliver M, Witten K, Kearns RA, Mavoa S, Badland HM, Carroll P, Drumheller C, Tavae N, Asiasiga L, Jelley S, Kaiwai H, Opit S, Lin EJ, Sweetsur P, Barnes HM, Mason N, Ergler C: Kids in the city study: research design and methodology. BMC Public Health. 2011, 11: 587-598. 10.1186/1471-2458-11-587.View ArticlePubMedPubMed CentralGoogle Scholar
  14. Carver A, Veitch J, Salmon J, Hume C, Timperio A, Crawford D: Chidren's independent mobility - is it influenced by parents' perceptions of safety?. 2010, Melbourne: Deakin UnisversityGoogle Scholar
  15. Mackett R, Brown B, Gong Y, Kitazawa K, Paskins J: Children's independent movement in the local environment. Built Environ. 2007, 33: 454-468. 10.2148/benv.33.4.454.View ArticleGoogle Scholar
  16. Page AS, Cooper AR, Griew P, Davis L, Hillsdon M: Independent mobility in relation to weekday and weekend physical activity in children aged 10-11 years: The PEACH Project. Int J Behav Nutr Phys Act. 2009, 6: 1-9. 10.1186/1479-5868-6-1.View ArticleGoogle Scholar
  17. Carver A, Timperio A, Hesketh K, Crawford D: Are children and adolescents less active if parents restrict their physical activity and active transport due to perceived risk?. Soc Sci Med. 2010, 70: 1799-1805. 10.1016/j.socscimed.2010.02.010.View ArticlePubMedGoogle Scholar
  18. Brown B, Mackett R, Gong Y, Kitazawa K, Paskins J: Gender differences in children's pathways to independent mobility. Child Geogr. 2008, 6: 385-401.View ArticleGoogle Scholar
  19. Fyhri A, Hjorthol R, Mackett RL, Fotel TN, Kyttä M: Children's active travel and independent mobility in four countries: development, social contributing trends and measures. Transp Policy. 2011, 18: 703-710. 10.1016/j.tranpol.2011.01.005.View ArticleGoogle Scholar
  20. Karsten L: It all used to be better? Different generations on continuity and change in urban children's daily use of space. Child Geogr. 2005, 3: 275-290.View ArticleGoogle Scholar
  21. Wridt PJ: A historical analysis of young people’s use of public space, parks and playgrounds in New York City. Child Youth Environ. 2004, 14: 86-106.Google Scholar
  22. Prezza M, Pilloni S, Morabito C, Sersante C, Alparone FR, Giuliani MV: The influence of psychosocial and environmental factors on children's independent mobility and relationship to peer frequentation. J Community Appl Soc Psychol. 2001, 11: 435-450. 10.1002/casp.643.View ArticleGoogle Scholar
  23. Shaw B, Watson B, Frauendienst B, Redecker A, Jones T, Hillman M: Children’s independent mobility: a comparative study in England and Germany (1971-2010). 2013, London: Policy Studies InstituteGoogle Scholar
  24. Bringolf-Isler B, Grize L, Mader U, Ruch N, Sennhauser FH, Braun-Fahrlander C: Personal and environmental factors associated with active commuting to school in Switzerland. Prev Med. 2008, 46: 67-73. 10.1016/j.ypmed.2007.06.015.View ArticlePubMedGoogle Scholar
  25. McDonald NC, Deakin E, Aalborg AE: Influence of the social environment on children's school travel. Prev Med. 2010, 50 (Suppl 1): 65-68.View ArticleGoogle Scholar
  26. Schoeppe S, Duncan MJ, Badland H, Oliver M, Curtis C: Associations of children's independent mobility and active travel with physical activity, sedentary behaviour and weight status: a systematic review. J Sci Med Sport. 2012, 16: 312-319.View ArticlePubMedGoogle Scholar
  27. Wen LM, Kite J, Merom D, Rissel C: Time spent playing outdoors after school and its relationship with independent mobility: a cross-sectional survey of children aged 10-12 years in Sydney, Australia. Int J Behav Nutr Phys Act. 2009, 6: 1-8. 10.1186/1479-5868-6-1.View ArticleGoogle Scholar
  28. Floyd MF, Bocarro JN, Smith WR, Baran PK, Moore RC, Cosco NG, Edwards MB, Suau LJ, Fang K: Park-based physical activity among children and adolescents. Am J Prev Med. 2011, 41: 258-265. 10.1016/j.amepre.2011.04.013.View ArticlePubMedGoogle Scholar
  29. Ekelund U, Anderssen SA, Froberg K, Sardinha LB, Andersen LB, Brage S: Independent associations of physical activity and cardiorespiratory fitness with metabolic risk factors in children: The European youth heart study. Diabetologia. 2007, 50: 1832-1840. 10.1007/s00125-007-0762-5.View ArticlePubMedGoogle Scholar
  30. Ekelund U, Luan J, Sherar LB, Esliger DW, Griew P, Cooper A: Moderate to vigorous physical activity and sedentary time and cardiometabolic risk factors in children and adolescents. JAMA. 2012, 307: 704-712. 10.1001/jama.2012.156.View ArticlePubMedPubMed CentralGoogle Scholar
  31. Cliff DP, Okely AD, Burrows TL, Jones RA, Morgan PJ, Collins CE, Baur LA: Objectively measured sedentary behavior, physical activity, and plasma lipids in overweight and obese children. Obesity. 2013, 21: 382-385. 10.1002/oby.20005.View ArticlePubMedGoogle Scholar
  32. The Children's Active Travel, Connectedness, and Health (CATCH): Independent Mobility, Active Travel and Children’s Health (iMATCH) projects. [http://research.humanities.curtin.edu.au/projects/catch/]
  33. Brage S, Brage N, Ekelund U, Luan J, Franks PW, Froberg K, Wareham NJ: Effect of combined movement and heart rate monitor placement on physical activity estimates during treadmill locomotion and free-living. Eur J Appl Physiol. 2006, 96: 517-524. 10.1007/s00421-005-0112-6.View ArticlePubMedGoogle Scholar
  34. The Actiheart Guide to Getting Started. [http://www.camntech.com/images/products/actiheart/The_Actiheart_Guide_to_Getting_Started.pdf]
  35. Villanueva K, Giles-Corti B, Bulsara M, Timperio A, McCormack G, Beesley B, Trapp G, Middleton N: Where do children travel to and what local opportunities are available? The relationship between neighborhood destinations and children’s independent mobility. Environ Behav. 2012, 45: 679-705.View ArticleGoogle Scholar
  36. Corder K, Brage S, Wareham NJ, Ekelund U: Comparison of PAEE from combined and separate heart rate and movement models in children. Med Sci Sports Exerc. 2005, 37: 1761-1767.View ArticlePubMedGoogle Scholar
  37. Edwardson CL, Gorely T: Epoch length and its effect on physical activity intensity. Med Sci Sports Exerc. 2010, 42: 928-934.View ArticlePubMedGoogle Scholar
  38. Van Sluijs EMF, Fearne VA, Mattocks C, Riddoch C, Griffin SJ, Ness A: The contribution of active travel to children's physical activity levels: cross-sectional results from the ALSPAC study. Prev Med. 2009, 48: 519-524. 10.1016/j.ypmed.2009.03.002.View ArticlePubMedGoogle Scholar
  39. Carver A, Timperio AF, Hesketh KD, Ridgers ND, Salmon JL, Crawford DA: How is active transport associated with children's and adolescents' physical activity over time?. Int J Behav Nutr Phys Act. 2011, 8: 126-131. 10.1186/1479-5868-8-126.View ArticlePubMedPubMed CentralGoogle Scholar
  40. Smith L, Sahlqvist S, Ogilvie D, Jones A, Griffin SJ, van Sluijs E: Is active travel to non-school destinations associated with physical activity in primary school children?. Prev Med. 2012, 54: 224-228. 10.1016/j.ypmed.2012.01.006.View ArticlePubMedGoogle Scholar
  41. Cooper AR, Andersen LB, Wedderkopp N, Page AS, Froberg K: Physical activity levels of children who walk, cycle, or are driven to school. Am J Prev Med. 2005, 29: 179-184. 10.1016/j.amepre.2005.05.009.View ArticlePubMedGoogle Scholar
  42. Evenson KR, Catellier DJ, Gill K, Ondrak KS, McMurray RG: Calibration of two objective measures of physical activity for children. J Sports Sci. 2008, 26: 1557-1565. 10.1080/02640410802334196.View ArticlePubMedGoogle Scholar
  43. Trost SG, Loprinzi PD, Moore R, Pfeiffer KA: Comparison of accelerometer cut points for predicting activity intensity in youth. Med Sci Sports Exerc. 2011, 43: 1360-1368.View ArticlePubMedGoogle Scholar
  44. Ridgway CL, Brage S, Sharp SJ, Corder K, Westgate KL, van Sluijs EM, Goodyer IM, Hallal PC, Anderssen SA, Sardinha LB, Andersen LB, Ekelund U: Does birth weight influence physical activity in youth? A combined analysis of four studies using objectively measured physical activity. PLoS ONE. 2011, 6: 1-8.Google Scholar
  45. Rubin DB: Multiple Imputation for Nonresponse in Surveys. 1987, New York: Wiley & SonsView ArticleGoogle Scholar
  46. Honaker J, King G, Blackwell M: Amelia II: a program for missing data. J Stat Softw. 2011, 45: 1-47.View ArticleGoogle Scholar
  47. A language and environment for statistical computing. [http://www.r-project.org]
  48. Dunton GF, Kawabata K, Intille S, Wolch J, Pentz MA: Assessing the social and physical contexts of children's leisure-time physical activity: an Ecological Momentary Assessment study. Am J Health Promot. 2012, 26: 135-142. 10.4278/ajhp.100211-QUAN-43.View ArticlePubMedGoogle Scholar
  49. Hillman M: Children's rights and adults' wrongs. Child Geogr. 2006, 4: 61-67.View ArticleGoogle Scholar
  50. Gleave J, Cole-Hamilton I: A world without play: A literature review on the effects of a lack of play on children’s lives. [http://www.playengland.org.uk/media/371031/a-world-without-play-literature-review-2012.pdf]
  51. Valentine G, McKendrick J, Children's outdoor play: Exploring parental concerns about children's safety and the changing nature of childhood. Geoforum. 1997, 28: 219-235. 10.1016/S0016-7185(97)00010-9.View ArticleGoogle Scholar
  52. Davison KK, Lawson CT: Do attributes in the physical environment influence children's physical activity? A review of the literature. Int J Behav Nutr Phys Act. 2006, 3: 1-17. 10.1186/1479-5868-3-1.View ArticleGoogle Scholar
  53. Ding D, Sallis JF, Kerr J, Lee S, Rosenberg DE: Neighborhood environment and physical activity among youth: a review. Am J Prev Med. 2011, 41: 442-455. 10.1016/j.amepre.2011.06.036.View ArticlePubMedGoogle Scholar
  54. Loprinzi PD, Cardinal BJ: Measuring children's physical activity and sedentary behaviors. J Exerc Sci Fit. 2011, 9: 15-23. 10.1016/S1728-869X(11)60002-6.View ArticleGoogle Scholar
  55. Trapp G, Giles-Corti B, Martin K, Timperio A, Villanueva K: Conducting field research in a primary school setting: methodological considerations for maximising response rates, data quality and quantitiy. Health Educ J. 2012, 71: 590-596. 10.1177/0017896911411766.View ArticleGoogle Scholar
  56. Schoeppe S, Oliver M, Badland HM, Burke M, Duncan MJ: Recruitment and retention of children in behavioral health risk factor studies: REACH strategies. Int J Behav Med. 2013, in pressGoogle Scholar
  57. Graham JW: Missing data analysis: making it work in the real world. Annu Rev Psychol. 2009, 60: 549-576. 10.1146/annurev.psych.58.110405.085530.View ArticlePubMedGoogle Scholar
  58. Lee PH: Data imputation for accelerometer-measured physical activity: The combined approach. Am J Clin Nutr. 2013, 97: 965-971. 10.3945/ajcn.112.052738.View ArticlePubMedGoogle Scholar
  59. Pre-publication history

    1. The pre-publication history for this paper can be accessed here:http://0-www.biomedcentral.com.brum.beds.ac.uk/1471-2458/14/91/prepub

Copyright

© Schoeppe et al.; licensee BioMed Central Ltd. 2014

This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Advertisement