Abstract

Background: Evidence on the correlates of objectively measured physical activity and sedentary behaviour in childhood is limited. This study aimed to identify correlates of physical activity and sedentary behaviour among 7-year-old children in England. Methods: Physical activity was measured using Actigraph accelerometry in 480 participants as part of the Gateshead Millennium Study during 2006–07. Twenty-two potential correlates across five domains (demographic and biological; psychological, cognitive and emotional; behavioural; social and cultural; physical environmental) were tested for associations with total volume of habitual physical activity, moderate–vigorous intensity physical activity (MVPA) and sedentary behaviour. Multiple linear regression analysis was used. Results: Seven correlates, including four that are potentially modifiable, were significantly associated with total physical activity, MVPA and sedentary behaviour in final models: gender, child weight status, maternal age, child interest in active play, active commuting to school, parenting practice and season. Four of these variables were significantly associated with all three constructs in final models. The final models explained 18, 18 and 24% of variance in total volume of physical activity, MVPA and sedentary behaviour, respectively. Conclusion: A number of potentially modifiable factors are associated with increased physical activity and/or reduced sedentary behaviour in English children. These could be valuable targets of future interventions.

Introduction

Habitual physical activity and sedentary behaviour in childhood and adolescence are now recognized as important influences on current and future health.1 There is growing concern that objectively measured levels of physical activity among children and youth are often lower than recommendations,2–4 and the development of effective interventions to increase physical activity requires an improved understanding of the correlates of physical activity and sedentary behaviour.5–7

Reviews8–13 have established a valuable conceptual framework to improve our understanding of the influences on physical activity in childhood and adolescence, but have focused largely on physical activity rather than sedentary behaviour to date. Furthermore, reviews of previous studies have actually identified relatively few correlates of physical activity consistently, and so have called for more research.8–13 Although there are studies of the correlates of objectively measured habitual physical activity,14–19 it would appear that there are no studies of the correlates of objectively measured sedentary behaviour in children. In addition, the correlates of physical activity may differ from the correlates of sedentary behaviour,20–22 and these constructs should ideally be considered separately. Correlates of objectively measured physical activity and sedentary behaviour may also differ significantly by age group,23 and so should ideally be studied in relatively large samples with narrow age ranges. The present study therefore aimed to identify significant correlates of objectively measured habitual physical activity and sedentary behaviour in a sample of 7-year-old children.24

Methods

Study participants

A total of 1029 infants had been recruited during 1999–2000 to a birth cohort study, the Gateshead Millennium Study (GMS), which is described in detail elsewhere.24 Briefly, all babies born in pre-specified recruiting weeks between June 1999 and May 2000 to Gateshead-resident mothers were eligible to join the study. This cohort is socio-economically representative of North East England,24 but is ethnically homogenous and is comprised almost entirely of Caucasians—the ethnic majority group in the UK. Data collection for the present study during 2006–07 involved a variety of anthropometric and lifestyle measures. All parents who had not previously indicated that they did not wish to be contacted were invited to take part. The present study was given a favourable ethical opinion from the Gateshead and South Tyneside Local National Health Service Research Ethics Committee, and all participating parents gave informed written consent to participation; all children gave assent to participation.

Objective measurement of physical activity and sedentary behaviour

Physical activity and sedentary behaviour were measured using the Actigraph 7164 and GT1M models (MTI, Fort Walton Beach, Florida). Reviews have concluded repeatedly that the Actigraph measures habitual physical activity and sedentary behaviour in children with high practical utility, reasonably high reliability, high validity relative to criterion measures (energy expenditure and direct observation of movement) and negligible reactivity.25–27

In the present study, the Actigraphs were attached to a waist belt and parents were asked to put the belt on when children woke up and remove it before bed, for a period of 7 consecutive days. Parents were also asked to note periods when the Actigraph was removed in a log sheet. Accelerometry data were reduced manually as described previously.28 In brief, long periods of consecutive ‘zeros’ in the accelerometry record were checked for entries in the log sheet. These periods were rare and usually corresponded to log-sheet records (e.g. showering) which explained the zeros satisfactorily. Accelerometer records which consisted of at least 3 days were included (2 week days plus 1 week end day; days of <6 h excluded), on the a priori grounds that reliability of this amount of Actigraph accelerometry in UK children is adequate.29 Analyses confirmed that for the data from the present study, reliability was high for total volume of physical activity, moderate–vigorous intensity physical activity (MVPA) and sedentary behaviour for the minimum period of accelerometry.28

The Actigraph Accelerometers in the present study were set to summarize activity data in 15 s sampling intervals (epochs), but data were collapsed to 60 s epochs when summarized to allow use of cut-points in accelerometry output as described below. Objectively measured physical activity was measured using two commonly-used constructs: total volume of physical activity [expressed as the mean counts per minute (cpm) over the duration of accelerometry monitoring19] and percent of time spent in MVPA. The epoch chosen does not affect measurements of total volume of activity,25 but tends to misclassify vigorous intensity physical activity as moderate intensity.25 Measurement of the amount of time spent in sedentary behaviour is largely unaffected by epoch.25

In order to express accelerometry output in terms of intensity of activity, it is necessary to apply cut-points to the accelerometry data.25 A recent review found a moderately large body of high quality and consistent evidence from paediatric validation and calibration studies, which suggests that the appropriate cut-point to measure MVPA with the Actigraph lies in the range 3100–3600 cpm.25 The cut-point of Puyau et al.30 (3200 cpm) was used to define MVPA. The Actigraph GTIM model has been shown to have a consistent bias of −9% relative to the older Actigraph model31 and so a +9% correction to GTIM data was made before applying any cut-point to define MVPA and sedentary behaviour. A body of high quality and consistent evidence from paediatric calibration and validation studies suggests that an Actigraph cut-point of around 1100 cpm will measure sedentary behaviour with optimal accuracy,32,33 across a wide paediatric age range.25 The cut-point of 1100 cpm was, therefore, used in the present study to define sedentary behaviour.

Anthropometrics

Height and weight were measured by trained researchers according to standard protocols using a Leicester portable height measure and a TANITA TBF 300MA body fat analyser (both Chasmors, London UK). Body mass index (BMI) was calculated in the usual way [weight (kg)/height (m)2]. For children, BMI z-scores relative to UK 1990 reference data were calculated to define healthy weight (z-score <1.04); overweight (BMI z-score 1.04 ≤ 1.64); obese (BMI z-score >1.64). For parents, weight status was assigned as healthy weight (BMI ≤ 25.0), overweight (BMI 25.0 ≤ 30.0), obese (BMI >30.0) (table 1).

Table 1

Potential correlates of objectively measured habitual physical activity and sedentary behavioura

Domain, variable (s) Measurement method Use in analysis 
Demographic and biological   
    Gender  Binary variable: female, male 
    Child weight status Researcher measured height, weight. BMI z-score relative to UK 1990 reference data Ordered categorical variable (0–2): healthy weight, overweight, obese 
    Socio-economic status Maternal education, four categories (no qualifications; school qualifications to age 16; post-16 school education; higher education) Ordered categorical variable (1–4): lowest to highest education 
    Maternal age Calculated via self-reported birthday Ordered categorical variable (1–4): youngest to eldest age quartile 
    Maternal weight status Researcher measured height, weight. BMI Ordered categorical variable (0–2): healthy weight, overweight, obese 
    Main carer works outside home Parent self-report Binary variable: no, yes 
Psychological, cognitive and emotional   
    Child interest in physically active play Parent questionnaire Continuous variable (1–5): least to most 
Behavioural domain   
    Child active commuting to school Parent report of usual frequency of walking or cycling vs. car use Ordered categorical variable (1–5): least to most frequent 
    Outdoor play Parent questionnaire Continuous variable (1–5): least to most frequent 
    Participation in after school clubs Parent questionnaire Binary variable: no, yes 
Social and cultural   
    Permissive parenting practice in relation to physical activity Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Authoritarian parenting practice in relation to physical activity Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Authoritative parenting practice in relation to physical activity Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Praise and encouragement in relation to physical activityb Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Rationalec in relation to physical activityb Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Modellingd physical activity behaviourb Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Parental perceived importance of active play Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Parental habitual physical activity Parent recent physical activity questionnaire (RPAQ)40 Ordered categorical variable (1–4): least to most active quartile 
    Parental habitual TV viewing and screen time Parent RPAQ40 Ordered categorical variable (1–4): least to most sedentary quartile 
    Parental regular use of active transport Parent RPAQ40 Binary variable: non-active, active 
Physical environment   
    Perceived availability of safe local places for play Parent questionnaire Continuous variable (1–5): least to most 
    Season Date of physical activity measurement Winter (January, February, November, December) Spring/fall (March, April, September, October) Summer (May, June, July, August) Ordered categorical variable (1–3): winter, spring/fall, summer 
Domain, variable (s) Measurement method Use in analysis 
Demographic and biological   
    Gender  Binary variable: female, male 
    Child weight status Researcher measured height, weight. BMI z-score relative to UK 1990 reference data Ordered categorical variable (0–2): healthy weight, overweight, obese 
    Socio-economic status Maternal education, four categories (no qualifications; school qualifications to age 16; post-16 school education; higher education) Ordered categorical variable (1–4): lowest to highest education 
    Maternal age Calculated via self-reported birthday Ordered categorical variable (1–4): youngest to eldest age quartile 
    Maternal weight status Researcher measured height, weight. BMI Ordered categorical variable (0–2): healthy weight, overweight, obese 
    Main carer works outside home Parent self-report Binary variable: no, yes 
Psychological, cognitive and emotional   
    Child interest in physically active play Parent questionnaire Continuous variable (1–5): least to most 
Behavioural domain   
    Child active commuting to school Parent report of usual frequency of walking or cycling vs. car use Ordered categorical variable (1–5): least to most frequent 
    Outdoor play Parent questionnaire Continuous variable (1–5): least to most frequent 
    Participation in after school clubs Parent questionnaire Binary variable: no, yes 
Social and cultural   
    Permissive parenting practice in relation to physical activity Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Authoritarian parenting practice in relation to physical activity Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Authoritative parenting practice in relation to physical activity Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Praise and encouragement in relation to physical activityb Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Rationalec in relation to physical activityb Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Modellingd physical activity behaviourb Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Parental perceived importance of active play Parent questionnaire, adapted from Vereecken et al.39 Continuous variable (1–5): least to most 
    Parental habitual physical activity Parent recent physical activity questionnaire (RPAQ)40 Ordered categorical variable (1–4): least to most active quartile 
    Parental habitual TV viewing and screen time Parent RPAQ40 Ordered categorical variable (1–4): least to most sedentary quartile 
    Parental regular use of active transport Parent RPAQ40 Binary variable: non-active, active 
Physical environment   
    Perceived availability of safe local places for play Parent questionnaire Continuous variable (1–5): least to most 
    Season Date of physical activity measurement Winter (January, February, November, December) Spring/fall (March, April, September, October) Summer (May, June, July, August) Ordered categorical variable (1–3): winter, spring/fall, summer 

a: Individual questionnaire items are provided in the Supplementary Data

b: Subdomains of the authoritative parenting practice

c: Example of ‘rationale’ questioning: How often do you tell your child that active play and games are good for them?

d: Example of ‘modelling’ questioning: I try to be more active so my child is more active

Potential correlates of physical activity and sedentary behaviour

Reviews of the correlates of physical activity in youth8–13 were considered when identifying potential correlates to be considered in the present study. The literature on the correlates of objectively measured sedentary behaviour in childhood is much more limited, and the topic is complicated by variations in the definitions of sedentary behaviour, which are discussed below. All potential correlates were supported by a priori hypotheses of associations with physical activity, sedentary behaviour or both. The social ecologic framework has dominated the literature on correlates of physical activity, with five domains widely accepted:8–13 (i) demographic and biological; (ii) psychological, cognitive and emotional; (iii) behavioural; (iv) social and cultural; (v) physical environmental. In the present study, potential correlates were grouped according to this conceptual framework and measured using a variety of methods as described in table 1.

Data that were collected from the main carer in the form of questionnaires have been described in detail by Parkinson et al.24 Further information on individual questionnaire items is available in the Supplementary Data.

The variables studied for each domain were:

Demographic and biological: gender; child weight status; socio-economic status (SES); maternal age; maternal weight status; main carer works outside the home.

Psychological, cognitive, and emotional: child interest in physically active play.

Behavioural: child active commuting to school; outdoor play; participation in after school clubs.

Social and cultural: permissive/authoritarian/authoritative parenting practice in relation to physical activity (authoritative subdomains: praise and encouragement in relation to physical activity; rationale in relation to physical activity; modelling parenting behaviour); parental-perceived importance of active play; parental habitual physical activity; parental habitual TV viewing and screen time; parental regular use of active transport.

Physical environment: perceived availability of safe local places for play; season.

Statistical analysis and power

All questionnaire data were double entered to minimize errors in data entry. Participants and non-participants were compared using chi-squared tests for categorical variables and two sample t-tests for continuous variables.

No formal power calculation was carried out, as sample size for the present study was fixed by the size of the GMS cohort. Previous studies of the correlates of objectively measured physical activity in youth have generally been able to identify correlates successfully with samples of <300 participants.14–19 For the present study, the sample size was expected to exceed 400, and was deemed likely to be adequate to carry out the analyses detailed below, with >30 study participants per explanatory variable in the multiple regression models.

Minitab release 15.1 was used for all analyses. Physical activity and BMI data were skewed, and the median and the interquartile range (IQR) were calculated for these variables. Mean and standard deviation (SD) are presented for normally distributed data. Simple univariate linear regression analysis was used to examine individual relationships between potential correlates listed in table 1 and the three separate outcome variables (total volume of physical activity; percentage of monitored time in MVPA; percentage of monitored time spent sedentary). As the relationship was considered linear among ordered categorical variables, they were treated as continuous for this analysis. In order to meet the assumptions of linear regression, namely that residuals should be normally distributed, two of the outcome variables required transformation (percentage of time spent in MVPA, square root transformation; total physical activity expressed as mean accelerometer cpm, log10 transformation).

To build the final models, stepwise multiple linear regression was used. All explanatory variables associated significantly with each outcome (P < 0.05) in the univariate analysis were considered. Those variables which remained associated significantly (at P < 0.05) in the final model were considered to be significant correlates of physical activity and sedentary behaviour in the present study.

Results

Characteristics of study participants

A total of 607 children were followed up during 2006–07, and all were invited to participate in physical activity and sedentary behaviour measurements. The remaining children were either untraceable or the family had opted out of the study. One hundred children did not wear (or return) the Actigraphs and/or did not complete accelerometer log sheets, and 27 did not meet the minimum accelerometry requirement, leaving 480 eligible children (244 boys; 236 girls). Mean duration of accelerometry monitoring was 6.4 days (SD 0.9) with mean duration of 11.1 h (SD 1.1) per day. Total physical activity was similar for boys and girls, median (IQR) cpm was 739 (624–888) for boys and 717 (595–855) for girls (P = 0.102).

Questionnaires for the present analysis were completed by the child's birth mother in 450 (of 467) (96%), by the child's birth father in 5, and by grandparents and others in 12 cases.

Table 2 summarizes characteristics from those children included in the present study and provides a comparison with non-participants. Sample attrition produced a final sample characterized by slightly higher SES than the initial sample, but left a sample for the present study, which was representative of North East England in terms of SES. At the inception of the birth cohort from which the present study sample was recruited in 2000, lower SES families were slightly overrepresented (data not shown) and so the slightly differential sample attrition by the time of the present study in 2007 led to a more socio-economically representative sample.24

Table 2

Characteristics of study participants and non-participants, mean (SD), median (IQR) or number (%)

Variable Included in analysis Excluded from analysis P for difference 
N 480 127  
Gender (%)   0.254 
    Male 244 (51) 54 (45)  
    Female 236 (49) 66 (55)  
Body mass index 16.3 (15.2–17.8) 16.4 (15.2–18.3) 0.092 
Body mass index z- score 0.32 (−0.34 to 1.06) 0.40 (−0.30 to 1.28) 0.299 
Weight status (%)   0.192 
    Overweight 57 (12%) 10 (8)  
    Obese 63 (13%) 24 (20)  
Socio-economic status, from maternal educational attainment (%)   <0.001 
    1 (no qualifications) 57 (12) 41 (32)  
    2 (education to age 16) 254 (53) 53 (42)  
    3 (post-16 education) 62 (13) 24 (19)  
    4 (higher education) 107 (22) 9 (7)  
Variable Included in analysis Excluded from analysis P for difference 
N 480 127  
Gender (%)   0.254 
    Male 244 (51) 54 (45)  
    Female 236 (49) 66 (55)  
Body mass index 16.3 (15.2–17.8) 16.4 (15.2–18.3) 0.092 
Body mass index z- score 0.32 (−0.34 to 1.06) 0.40 (−0.30 to 1.28) 0.299 
Weight status (%)   0.192 
    Overweight 57 (12%) 10 (8)  
    Obese 63 (13%) 24 (20)  
Socio-economic status, from maternal educational attainment (%)   <0.001 
    1 (no qualifications) 57 (12) 41 (32)  
    2 (education to age 16) 254 (53) 53 (42)  
    3 (post-16 education) 62 (13) 24 (19)  
    4 (higher education) 107 (22) 9 (7)  

Levels of objectively measured physical activity were relatively low, and levels of objectively measured sedentary behaviour relatively high. Median (IQR) time spent in MVPA was 3.9% (2.6–5.8), equivalent to 20–30 min MVPA per day. Median (IQR) monitoring time spent sedentary was 77.8% (73.7–81.7) equivalent to >9 h day−1 sedentary (no trunk movement as defined by accelerometry). The remaining time (1–2 h of accelerometry wear time during the waking day) was categorized as light intensity physical activity, defined as physical activities with an energy cost below approximately three times resting energy expenditure.25

Correlates of objectively measured total volume of habitual physical activity (accelerometry cpm)

The results of the univariate analysis are shown in table 3. Of seven variables significantly associated with total volume of physical activity in the univariate analysis, four remained significant in the final model (table 4). In the final model, total volume of physical activity was significantly lower in: the overweight and obese; those with parent-perceived low interest in physically active play; those who used non-physically active methods of transport to school; and in winter and spring/fall compared to summer. A total of 18.1% of variance in total volume of physical activity was explained in the final model.

Table 3

Univariate analysis of correlates associated with total physical activity, MVPA and sedentary behavioura

Variable  Total volume of physical activity (cpm) Percentage of monitored time in MVPA % Monitored time in sedentary behaviour 
 n P P P 
Demographic and biological domain     
    Gender 480 0.100 0.003 <0.001 
    Child weight status 476 0.003 <0.001 0.020 
    Socio-economic status 475 0.169 0.632 0.125 
    Maternal age 466 0.049 0.138 0.045 
    Maternal weight status 428 0.564 0.414 0.675 
    Main carer works outside home 460 0.063 0.368 0.039 
Psychological, cognitive and emotional domain     
    Child interest in physically active play 467 <0.001 <0.001 <0.001 
Behavioural domain     
    Active commuting to school 466 <0.001 <0.001 <0.001 
    Outdoor play 467 <0.001 0.001 <0.001 
    Participation in after school clubs 463 0.036 0.062 0.125 
    Permissive parenting practice 467 0.646 0.844 0.938 
    Authoritarian parenting practice 467 0.286 0.426 0.172 
    Authoritative parenting practice 467 0.569 0.633 0.614 
        Praise and encouragement  0.327 0.294 0.371 
        Rationale  0.286 0.208 0.642 
        Modelling behaviour  0.092 0.178 0.045 
    Parental perceived importance of active play 467 0.129 0.230 0.141 
    Parental habitual physical activity 425 0.273 0.490 0.836 
    Parental habitual TV viewing and screen time 441 0.310 0.374 0.587 
    Parental use of active transport method 445 0.073 0.128 0.011 
Physical environment domain     
    Availability of safe places for play 467 0.074 0.052 0.022 
    Season 480 <0.001 <0.001 <0.001 
Variable  Total volume of physical activity (cpm) Percentage of monitored time in MVPA % Monitored time in sedentary behaviour 
 n P P P 
Demographic and biological domain     
    Gender 480 0.100 0.003 <0.001 
    Child weight status 476 0.003 <0.001 0.020 
    Socio-economic status 475 0.169 0.632 0.125 
    Maternal age 466 0.049 0.138 0.045 
    Maternal weight status 428 0.564 0.414 0.675 
    Main carer works outside home 460 0.063 0.368 0.039 
Psychological, cognitive and emotional domain     
    Child interest in physically active play 467 <0.001 <0.001 <0.001 
Behavioural domain     
    Active commuting to school 466 <0.001 <0.001 <0.001 
    Outdoor play 467 <0.001 0.001 <0.001 
    Participation in after school clubs 463 0.036 0.062 0.125 
    Permissive parenting practice 467 0.646 0.844 0.938 
    Authoritarian parenting practice 467 0.286 0.426 0.172 
    Authoritative parenting practice 467 0.569 0.633 0.614 
        Praise and encouragement  0.327 0.294 0.371 
        Rationale  0.286 0.208 0.642 
        Modelling behaviour  0.092 0.178 0.045 
    Parental perceived importance of active play 467 0.129 0.230 0.141 
    Parental habitual physical activity 425 0.273 0.490 0.836 
    Parental habitual TV viewing and screen time 441 0.310 0.374 0.587 
    Parental use of active transport method 445 0.073 0.128 0.011 
Physical environment domain     
    Availability of safe places for play 467 0.074 0.052 0.022 
    Season 480 <0.001 <0.001 <0.001 

a: Numbers for analysis ranged from 428 to 480, due to missing data on individual independent variables

Table 4

Final models: associations between total physical activity, MVPA and sedentary behaviour

Variable Total physical activity (cpm)
 
Percentage of monitored time in MVPA
 
Percentage of monitored time in sedentary behaviour
 
 β-coeff. 95% CI P-value β-coeff. 95% CI P-value β-coeff. 95% CI P-value 
Gender – – – 0.150 (0.054 to 0.246) 0.002 −1.961 (−2.948 to −0.974) <0.001 
Child weight status −0.021 (−0.035 to −0.006) 0.005 −0.150 (−0.217 to −0.082) <0.001 0.764 (0.071 to 1.456) 0.031 
Maternal age – – – – – – 0.550 (0.107 to 0.992) 0.015 
Child perceived interest in physically active play 0.041 (0.025 to 0.058) <0.001 0.228 (0.149 to 0.306) <0.001 −2.371 (−3.175 to −1.567) <0.001 
Child active commuting to school 0.012 (0.005 to 0.018) 0.001 0.052 (0.020 to 0.085) 0.001 −0.600 (−0.929 to −0.272) <0.001 
Modelling behaviour – – – – – – 1.025 (0.133 to 1.917) 0.025 
Season 0.046 (0.033 to 0.058) <0.001 0.147 (0.087 to 0.208) <0.001 −2.296 (−2.907 to −1.684) <0.001 
(Constant) 2.579 (2.508 to 2.650)  0.642 (0.304 to 0.980)  89.214 (84.733 to 93.695)  
Adjusted R2   18.1%   17.8%   24.0% 
Variable Total physical activity (cpm)
 
Percentage of monitored time in MVPA
 
Percentage of monitored time in sedentary behaviour
 
 β-coeff. 95% CI P-value β-coeff. 95% CI P-value β-coeff. 95% CI P-value 
Gender – – – 0.150 (0.054 to 0.246) 0.002 −1.961 (−2.948 to −0.974) <0.001 
Child weight status −0.021 (−0.035 to −0.006) 0.005 −0.150 (−0.217 to −0.082) <0.001 0.764 (0.071 to 1.456) 0.031 
Maternal age – – – – – – 0.550 (0.107 to 0.992) 0.015 
Child perceived interest in physically active play 0.041 (0.025 to 0.058) <0.001 0.228 (0.149 to 0.306) <0.001 −2.371 (−3.175 to −1.567) <0.001 
Child active commuting to school 0.012 (0.005 to 0.018) 0.001 0.052 (0.020 to 0.085) 0.001 −0.600 (−0.929 to −0.272) <0.001 
Modelling behaviour – – – – – – 1.025 (0.133 to 1.917) 0.025 
Season 0.046 (0.033 to 0.058) <0.001 0.147 (0.087 to 0.208) <0.001 −2.296 (−2.907 to −1.684) <0.001 
(Constant) 2.579 (2.508 to 2.650)  0.642 (0.304 to 0.980)  89.214 (84.733 to 93.695)  
Adjusted R2   18.1%   17.8%   24.0% 

Correlates of objectively measured MVPA

The results of the univariate analysis are shown in table 3. Of seven variables significantly associated with MVPA in univariate analysis, five remained significantly associated with MVPA in the final model (table 4). MVPA was significantly lower among: girls; overweight and obese children; in spring/fall and winter compared to summer; in those who used non-active means of transport to school, and children perceived by their parents to have low interest in physically active play. A total of 17.8% of the variance in MVPA was explained in the final model.

Correlates of objectively measured sedentary behaviour

Results of the univariate analysis are shown in table 3. Eleven variables were significantly associated with sedentary behaviour in univariate analyses; seven remained significant in the final model (table 4). Levels of objectively measured sedentary behaviour were significantly higher among: girls; overweight and obese children; those with older mothers; in spring/fall and winter compared to summer; those with low interest in physically active play; those using modelling parenting practices towards physical activity; those who did not use active forms of commuting to school. In the final model, 24.0% of variance in objectively measured sedentary behaviour was explained.

Discussion

We present data in the under-researched area of correlates of objectively measured physical activity in children, including novel data on the correlates of objectively measured sedentary behaviour. The present study found four variables that were significantly associated in final models with all three constructs of habitual total physical activity, habitual MVPA and habitual sedentary behaviour; namely active commuting to school; child interest in physical activity; child weight status; season. This suggests that at least some common factors may influence all three constructs in English children.

While the two constructs of physical activity might have been expected to have similar correlates, empirical evidence on this issue is scarce since studies have generally not considered objectively measured total physical activity and MVPA as separate constructs, and older studies which used more traditional subjective methods were unable to consider the data in this manner due to limitations in subjective methodology. In the current study, while gender remained significant in the final model for MVPA, it did not remain in the final model for total volume of physical activity. This distinction may reflect gender-specific differences in activity patterns and may imply that analysis using more than one construct may lead to a more informed picture of behaviour. It is widely accepted that sedentary behaviour should be considered separately from physical activity,20–22 as the correlates of physical activity and sedentary behaviour may differ. In the present study, correlates of objectively measured habitual physical activity and sedentary behaviour were broadly similar. However, maternal age and parental modelling in the present study were significant correlates only of sedentary behaviour and not physical activity. This may simply reflect a decrease in activity of older mothers, or be a reflection of family size, with the potential for younger children requiring more parental involvement.

Informal subgroup analysis, in the present study, showed the three final regression models to be fairly robust. In addition to those variables, which remained significant in the final models, the subgroup analysis also indicated that outdoor play may be associated with physical activity and negatively associated with sedentary behaviour in some subgroups. This relationship was strongest among boys.

Identifying correlates of objectively measured physical activity and sedentary behaviour should inform future interventions aimed at promotion of physical activity, promotion of reduction in sedentary behaviour and prevention of obesity.5,8 In the present study, some significant correlates of physical activity and sedentary behaviour were non-modifiable (gender, season, maternal age), though some are at least potentially modifiable (active vs. non-active commuting to school; child-perceived interest in physically active play; weight status).

Previous comparable studies are scarce. In the UK, for example, the current authors are aware of only one previous study of the correlates of objectively measured physical activity (not sedentary behaviour), in 11- to 13-year-old children,34 an older group than those recruited to the present study. The explanatory variables included in the present study explained relatively little of the variation in objectively measured physical activity, but this was consistent with other evidence. Previous studies which have used objective measures of physical activity have typically explained <20% of variation in physical activity in their final models.19–23 Studies of correlates of objectively measured physical activity may tend to find fewer correlates than studies of subjectively measured physical activity, in part because of correlated error in studies which depend on subjective methods to derive both explanatory and outcome variables.19,35 However, the higher accuracy of objective measurements is essential for studies of the amount and intensity of physical activity.25 As our data on correlates of objectively measured sedentary behaviour are novel, comparison is made with a study from the USA, which made subjective measures of sedentary behaviour.20 They identified correlates of subjectively measured sedentary time that were different to those of physical activity,20 and tended to be non-modifiable, socio-demographic factors such as low maternal education or lower family income,20 neither of which were in the final model in our analyses. Correlates of both sedentary behaviour and physical activity are likely to differ between such varied geographic regions, as well as potential differences between studies which use different approaches to the measurement of physical activity.

The present study is consistent with many recent studies of objectively measured physical activity in finding that gender was significantly associated with both physical activity (MVPA) and sedentary behaviour. The finding that SES was not significantly associated with habitual physical activity supports recent evidence from all of the objective measurement studies in the UK: all four studies to date, which all used accelerometry, have found no marked differences in total volume of physical activity between socio-economic groups in UK children and adolescents.25 Socio-economic differences in physical activity in UK children and adolescents in the ‘expected’ direction (lower physical activity in lower SES groups) has been observed in studies using subjective methods only25 and it is likely that such variation is not real, but arises from biases inherent in subjective methodology. Seasonality in objectively measured physical activity is understudied, and probably varies markedly by geographic location, but in the UK similar seasonal variations have been shown: increased physical activity during summer months has been reported in both younger and older British children than those studied here.3,36 The present study therefore supports the hypothesis that there are a few core influences (e.g. gender, seasonality) on objectively measured physical activity in the UK which are consistent in their associations across a wide age range from preschool to adolescence.3,25,36 Active commuting to school was shown to be strongly associated with increased overall physical activity and MVPA (and decreased sedentary behaviour) in the present study. This lends further support to existing evidence regarding children who walk or cycle to school37 and emphasizes the potential value of active commuting as a potential target of future interventions.

The principal strengths of the present study were: the recruitment of a sample which was socio-economically representative of North East England;24 the use of validated objective measurements; the inclusion of three constructs (two of physical activity, plus sedentary behaviour); the inclusion of a large number of potential correlates across all domains in the social-ecological model; the relatively large and homogenous (narrow age range; almost exclusively from the majority ethnic group) sample.

The present study was designed prospectively with a number of aims; one of these was to examine the correlates of habitual physical activity and sedentary behaviour in childhood. There are limitations to the present study, in that it was cross-sectional in design, and thus can make no firm claims regarding the causal nature of relationships observed. Generalisability of study findings to other settings or other age groups is unclear and requires further study, but as noted above some correlates do appear to be consistently associated with objectively measured volume of physical activity across a wide paediatric age range in the UK, and others do not (e.g. SES). One limitation of the present study was the inability to examine whether correlates of physical activity differed by subgroup. Exploratory analyses of differences in associations between subgroups (defined by weight status and by gender) were undertaken, but these were probably underpowered. Future larger studies might consider whether associations between exposure and outcome variables vary between subgroups within the population. Practical constraints meant that it was not possible to measure all of the potential correlates of physical activity and sedentary behaviour: correlates which are identified in any study may depend on the precise construct of sedentary behaviour which is being measured. In the present study, sedentary behaviour was defined as time spent with no trunk movement.26,34 As previously mentioned, family size may explain the impact of maternal age on sedentary behaviour and would be a valuable inclusion in future analyses. For future studies of determinants of physical activity and sedentary behaviour there is also value in analysing the data more comprehensively, according to a social ecological approach, as discussed by Spence and Lee,38 but this type of approach was beyond the scope of this manuscript. Future follow-up of the cohort will allow us to track the children and see how the correlates of physical activity and sedentary behaviour change, for example, as interest in active play declines, or weight status changes.

Conclusions

The present study identified a number of correlates of physical activity and sedentary behaviour in English children. Four variables were associated with both habitual physical activity and sedentary behaviour (child's weight status, child's-perceived interest in active play, child's active commuting to school and season). Future interventions aimed at increasing physical activity, and decreasing sedentary behaviour, in mid-childhood should be informed by the evidence of the present study.

Supplementary data

Supplementary data are available at EURPUB online.

Funding

The Gateshead Millennium Study is supported by a grant from the UK National Prevention Research Initiative (incorporating funding from the British Heart Foundation, Cancer Research UK, Diabetes UK, English Department of Health, Economic and Social Research Council, Food Standards Agency, Medical Research Council, Research and Development Office for the Northern Ireland Health and Social Services, Chief Scientist Office of the Scottish Government Health Directorate, Welsh Assembly Government, and World Cancer Research Fund). The cohort was first set up with funding from the Henry Smith Charity and Sport Aiding Medical Research for Kids (SPARKS). The funding body had no role in the decision to publish or the content of the article.

Conflicts of interest: None declared.

Key points

  • Some factors influence both physical activity and sedentary behaviour among children—four were conserved across all three constructs.

  • Several modifiable or potentially modifiable correlates were identified which will help to identify children most at risk of high levels of sedentary behaviour.

  • Public health professionals could use the correlates identified by the present study to design interventions to increase physical activity and reduce sedentary behaviours in children.

Acknowledgements

The authors acknowledge the support of an external reference group in conducting the present study. The authors appreciate the support of Gateshead Health National Health Service Foundation Trust, Gateshead Education Authority and local schools. The authors warmly thank the research team for their effort. Thanks are especially due to the Gateshead Millennium Study families and children for their participation in the study.

References

1
Strong
WB
Malina
RM
Blimkie
CJR
, et al.  . 
Evidence based physical activity for school-age youth
J Pediatr
 , 
2005
, vol. 
146
 (pg. 
732
-
7
)
2
Reilly
JJ
Jackson
DM
Montgomery
C
, et al.  . 
Total energy expenditure and physical activity in young Scottish children: mixed longitudinal study
Lancet
 , 
2004
, vol. 
363
 (pg. 
211
-
2
)
3
Riddoch
CJ
Mattocks
C
Deere
K
, et al.  . 
Objective measurement of levels and patterns of physical activity
Arch Dis Child
 , 
2007
, vol. 
92
 (pg. 
963
-
9
)
4
Troiano
RP
Berrigan
D
Dodd
KW
, et al.  . 
Physical activity in the United States measured by accelerometer
Med Sci Sports Exerc
 , 
2008
, vol. 
40
 (pg. 
181
-
8
)
5
Brodersen
NH
Steptoe
A
Williamson
S
Wardle
J
Sociodemographic, developmental, environmental, and psychological correlates of physical activity and sedentary behavior at age 11 to 12
Ann Behav Med
 , 
2005
, vol. 
29
 (pg. 
2
-
11
)
6
Finn
K
Johannsen
N
Specker
B
Factors associated with physical activity in preschool children
J Pediatr
 , 
2002
, vol. 
140
 (pg. 
81
-
5
)
7
Heitzler
CD
Martin
SL
Duke
J
Huhman
M
Correlates of physical activity in a national sample of children aged 9-13 years
Prev Med
 , 
2006
, vol. 
42
 (pg. 
254
-
60
)
8
Sallis
JF
Prochaska
JJ
Taylor
WC
A review of correlates of physical activity of children and adolescents
Med Sci Sports Exerc
 , 
2000
, vol. 
32
 (pg. 
963
-
75
)
9
Ferreira
I
van der Horst
K
Wendel-Vos
W
, et al.  . 
Environmental correlates of physical activity in youth - a review and update
Obes Rev
 , 
2006
, vol. 
8
 (pg. 
129
-
54
)
10
Van der Horst
K
Paw
M
Twisk
JWR
Van Mechelen
W
A brief review on correlates of physical activity and sedentariness in youth
Med Sci Sports Exerc
 , 
2007
, vol. 
39
 (pg. 
1241
-
50
)
11
Hinkley
T
Crawford
D
Salmon
J
, et al.  . 
Preschool children and physical activity - a review of correlates
Am J Prev Med
 , 
2008
, vol. 
34
 (pg. 
435
-
41
)
12
Kohl
HW
Hobbs
KE
Development of physical activity behaviors among children and adolescents
Pediatrics
 , 
1998
, vol. 
101
 (pg. 
549
-
54
)
13
Lindquist
CH
Reynolds
KD
Goran
MI
Sociocultural determinants of physical activity among children
Prev Med
 , 
1999
, vol. 
29
 (pg. 
305
-
12
)
14
Moore
LL
Lombardi
DA
White
MJ
, et al.  . 
Influence of parents physical-activity levels on activity levels of young-children
J Pediatr
 , 
1991
, vol. 
118
 (pg. 
215
-
9
)
15
Sallis
JF
Alcaraz
JE
McKenzie
TL
, et al.  . 
Parental behavior in relation to physical-activity and fitness in 9-year-old children
Am J Dis Child
 , 
1992
, vol. 
146
 (pg. 
1383
-
8
)
16
Trost
SG
Pate
RR
Ward
DS
, et al.  . 
Correlates of objectively measured physical activity in preadolescent youth
Am J Prev Med
 , 
1999
, vol. 
17
 (pg. 
120
-
6
)
17
Sallis
JF
Simonsmorton
BG
Stone
EJ
, et al.  . 
Determinants of physical-activity and interventions in youth
Med Sci Sports Exerc
 , 
1992
, vol. 
24
 (pg. 
S248
-
S257
)
18
Trost
SG
Saunders
R
Ward
DS
Determinants of physical activity in middle school children
Am J Health Behav
 , 
2002
, vol. 
26
 (pg. 
95
-
102
)
19
McMinn
AM
van Sluijs
EMF
Wedderkopp
N
, et al.  . 
Sociocultural correlates of physical activity in children and adolescents: findings from the Danish arm of the European Youth Heart Study
Pediatr Exerc Sci
 , 
2008
, vol. 
20
 (pg. 
319
-
32
)
20
Gordon-Larsen
P
McMurray
RG
Popkin
BM
Determinants of adolescent physical activity and inactivity patterns
Pediatrics
 , 
2000
, vol. 
105
 pg. 
e83
 
21
Schmitz
KH
Lytle
LA
Phillips
GA
, et al.  . 
Psychosocial correlates of physical activity and sedentary leisure habits in young adolescents: the teens eating for energy and nutrition at school study
Prev Med
 , 
2002
, vol. 
34
 (pg. 
266
-
78
)
22
Salmon
J
Owen
N
Crawford
D
, et al.  . 
Physical activity and sedentary behavior: a population-based study of barriers, enjoyment, and preference
Health Psychol
 , 
2003
, vol. 
22
 (pg. 
178
-
88
)
23
Jago
R
Baranowski
T
Thompson
D
, et al.  . 
Sedentary behavior, not TV viewing, predicts physical activity among 3- to 7-year-old children
Pediatr Exerc Sci
 , 
2005
, vol. 
17
 (pg. 
364
-
76
)
24
Parkinson
KN
Pearce
MS
Dale
A
, et al.  . 
Cohort Profile: The Gateshead Millennium Study
Int J Epidemiol
 , 
2010
 
doi: 10.1093/ije/dyq015 [Epub ahead of print: 23 March 2010]
25
Reilly
JJ
Penpraze
V
Hislop
J
, et al.  . 
Objective measurement of physical activity and sedentary behaviour: review with new data
Arch Dis Child
 , 
2008
, vol. 
93
 (pg. 
614
-
9
)
26
Trost
SG
State of the art reviews: measurement of physical activity in children and adolescents
Am J Lifestyle Med
 , 
2007
, vol. 
1
 (pg. 
299
-
314
)
27
de Vries
SI
Bakker
I
Hopman-Rock
M
, et al.  . 
Clinimetric review of motion sensors in children and adolescents
J Clin Epidemiol
 , 
2006
, vol. 
59
 (pg. 
670
-
80
)
28
Basterfield
L
Adamson
AJ
Pearce
MS
, et al.  . 
Stability of habitual physical activity and sedentary behavior monitoring by accelerometry in 6-8 year olds
J Phys Act Health
 , 
2010
 
in press
29
Penpraze
V
Reilly
JJ
MacLean
CM
, et al.  . 
Monitoring of physical activity in young children: How much is enough?
Pediatr Exerc Sci
 , 
2006
, vol. 
18
 (pg. 
483
-
91
)
30
Puyau
MR
Adolph
AL
Vohra
FA
Butte
NF
Validation and calibration of physical activity monitors in children
Obes Res
 , 
2002
, vol. 
10
 (pg. 
150
-
7
)
31
Corder
K
Brage
S
Ramachandran
A
Snehalatha
C
, et al.  . 
Comparison of two Actigraph models for assessing free-living physical activity in Indian adolescents
J Sports Sci
 , 
2007
, vol. 
25
 (pg. 
1607
-
11
)
32
Reilly
JJ
Coyle
J
Kelly
L
, et al.  . 
An objective method for measurement of sedentary behavior in 3- to 4-year olds
Obes Res
 , 
2003
, vol. 
11
 (pg. 
1155
-
8
)
33
Sirard
JR
Trost
SG
Pfeiffer
KA
, et al.  . 
Calibration and evaluation of an objective measure of physical activity in preschool children
J Phys Act Health
 , 
2005
, vol. 
3
 (pg. 
345
-
57
)
34
Mattocks
C
Deere
K
Leary
S
, et al.  . 
Early life determinants of physical activity in 11 to 12 year olds: cohort study
Br Med J
 , 
2008
, vol. 
336
 (pg. 
26
-
9
)
35
Sallis
JF
Taylor
WC
Dowda
M
, et al.  . 
Correlates of vigorous physical activity for children in grades 1 through 12: Comparing parent-reported and objectively measured physical activity
Pediatr Exerc Sci
 , 
2002
, vol. 
14
 (pg. 
30
-
44
)
36
Fisher
A
Reilly
JJ
Montgomery
C
, et al.  . 
Seasonality in physical activity and sedentary behavior in young children
Pediatr Exerc Sci
 , 
2005
, vol. 
17
 (pg. 
31
-
40
)
37
Cooper
AR
Anderson
LB
Wedderkopp
N
Page
AS
Physical activity levels of children who walk, cycle, or are driven to school
Am J Prev Med
 , 
2005
, vol. 
29
 (pg. 
179
-
84
)
38
Spence
JC
Lee
RE
Toward a comprehensive model of physical activity
Psychol Sport Exerc
 , 
2003
, vol. 
4
 (pg. 
7
-
24
)
39
Vereecken
CA
Keukelier
E
Maes
L
Influence of mother's educational level on food parenting practices and food habits of young children
Appetite
 , 
2004
, vol. 
43
 (pg. 
93
-
103
)
40
Besson
H
Brage
S
Jakes
RW
, et al.  . 
Estimating physical activity energy expenditure, sedentary time, and physical activity intensity by self-report in adults
Am J Clin Nutr
 , 
2010
, vol. 
91
 (pg. 
106
-
14
)

Author notes

*Gateshead Millennium Study Core Team: Ashley Adamson, Anne Dale, Robert Drewett, Ann Le Couteur, Paul McArdle, Kathryn Parkinson, Mark Pearce, John Reilly, Charlotte Wright.

Comments

0 Comments