International Journal of Obesity (2008) 32, 601612

& 2008 Nature Publishing Group All rights reserved 0307-0565/08 $30.00
www.nature.com/ijo

PEDIATRIC HIGHLIGHT
Outcomes of a group-randomized trial to prevent
excess weight gain, reduce screen behaviours and
promote physical activity in 10-year-old children:
Switch-Play
J Salmon1, K Ball1, C Hume1, M Booth2 and D Crawford1
1
Centre for Physical Activity and Nutrition Research, School of Exercise and Nutrition Sciences, Deakin University, Burwood,
Victoria, Australia and 2Centre for Research into Adolescents Health, School of Public Health, University of Sydney, Sydney,
New South Wales, Australia

Objectives: To evaluate the effectiveness of an intervention to prevent excess weight gain, reduce time spent in screen
behaviours, promote participation in and enjoyment of physical activity (PA), and improve fundamental movement skills among
children.
Participants: In 2002, 311 children (78% response; 49% boys), average age 10 years 8 months, were recruited from three
government schools in low socioeconomic areas of Melbourne, Australia.
Design: Group-randomized controlled trial. Children were randomized by class to one of the four conditions: a behavioural
modification group (BM; n 66); a fundamental movement skills group (FMS; n 74); a combined BM/FMS group (BM/FMS;
n 93); and a control (usual curriculum) group (n 62). Data were collected at baseline, post intervention, 6- and 12-month
follow-up periods.
Results: BMI data were available for 295 children at baseline and 268 at 12-month follow-up. After adjusting for food intake and
PA, there was a significant intervention effect from baseline to post intervention on age- and sex-adjusted BMI in the BM/FMS
group compared with controls (1.88 kg m2, Po0.01), which was maintained at 6- and 12-month follow-up periods
(1.53 kg m2, Po0.05). Children in the BM/FMS group were less likely than controls to be overweight/obese between baseline
and post intervention (adjusted odds ratio (AOR) 0.36, Po0.05); also maintained at 12-month follow-up (AOR 0.38,
Po0.05). Compared with controls, FMS group children recorded higher levels and greater enjoyment of PA; and BM children
recorded higher levels of PA and TV viewing across all four time points. Gender moderated the intervention effects for
participation in and enjoyment of PA, and fundamental movement skills.
Conclusion: This programme represents a promising approach to preventing excess weight gain and promoting participation in
and enjoyment of PA. Examination of the mediators of this intervention and further tailoring of the programme to suit both
genders is required.
International Journal of Obesity (2008) 32, 601612; doi:10.1038/sj.ijo.0803805; published online 5 February 2008

Keywords: group-randomized trial; overweight; physical activity; screen behaviours; moderating; generalized estimating
equations

Introduction Australian data indicate that the prevalence of overweight

In the last two decades, there have been worldwide increases
in the prevalence of paediatric overweight and obesity.1
Correspondence: Dr J Salmon, Centre for Physical Activity and Nutrition
Research, School of Exercise and Nutrition Sciences, Deakin University, 221
Burwood Highway, Burwood, Victoria 3125, Australia.
E-mail: jsalmon@deakin.edu.au
Received 4 April 2007; revised 17 December 2007; accepted 21 December
2007; published online 5 February 2008
and obesity almost doubled between 1985 and 1995.2 This
pattern appears to be global, with obesity increasing more
dramatically in industrialized countries compared with
countries that are less economically developed.1 The specific
causes of this rapid increase in the prevalence of overweight
remains the subject of ongoing debate;3 however, it is widely
accepted that insufficient amounts of physical activity and
excess food intake are important contributors to energy
balance and ultimately weight status.
 Switch-Play intervention
J Salmon et al
602
Time spent being sedentary, particularly in television (TV)
viewing, is also recognized as an important contributor to
overweight and obesity among children.4 A number of
studies have found that sedentary behaviours track more
strongly than physical activity over short follow-up periods57
and that TV viewing habits in childhood are associated with
overweight, poor fitness, smoking and raised cholesterol in
adulthood.8 Recent evidence also suggests that sedentary
behaviours increase through adolescence,9 while physical
activity levels have consistently been found to decline
during childhood and throughout the lifespan.10 Health
authorities recommend that children spend no more than
2 h per day in electronic entertainment media or screen
behaviours (TV viewing, computer use and playing electro-
nic games).11,12 In Australia, however, children spend on
average 2.5 h per day watching TV13 and in the United
States, 25% of young people watch TV for more than 4 h per
day.14 These estimates do not include other screen beha-
viours such as computer use or playing electronic games.
Therefore, developing and testing the effectiveness of
strategies to reduce time spent in screen behaviours and to
promote physical activity is important for childrens current
and future behaviour and health.
Interventions to reduce time children spend in sedentary
behaviour, to reduce obesity and to prevent onset of
overweight and obesity have recently been reviewed.15
Some interventions designed to prevent unhealthy weight
gain have reported positive impact.16,17 Robinson17
conducted a school-based randomized controlled trial,
which aimed to decrease the time children spent watching
TV using curriculum-based materials delivered over 6
months to third and fourth grade students. Time spent
watching TV and body mass index (BMI) declined
significantly in the intervention group compared with the
control group over the period of the intervention.
Gortmaker et al.16 also employed a curriculum-delivered
intervention, over a 2-year period, which aimed to (among
other outcomes) prevent the development of obesity,
decrease time spent watching TV and increase physical
activity participation among children in grades 68. After
the 2-year intervention, time spent watching TV and the
prevalence of obesity were reduced among girls in the
intervention compared with control schools, with no effects
for boys. The maintenance of intervention effects for both of
these studies is not reported.
Importantly, while the studies by Robinson17 and
Gortmaker et al.16 reported favourable outcomes for chil-
drens adiposity and TV viewing time, neither intervention
had a significant effect on physical activity. It may be that in
those interventions children did not have the necessary skills
required to replace TV viewing with more active pastimes. A
review of correlates of childrens physical activity found that
perceived physical competence to be active was positively
associated with physical activity levels.18 In addition,
children with greater fundamental movement skill profi-
ciency are more active19 and are less likely to be overweight.
International Journal of Obesity
Improving childrens fundamental movement skills (FMS)
has also been shown to increase physical activity enjoy-
ment,20 which is a known correlate of physical activity
among children.18 However, the effectiveness of improving
FMS and reducing sedentary behaviours on preventing
declines in childrens physical activity and preventing excess
weight gain has not been previously examined in a
controlled trial.
The primary aims of this study were to evaluate the
effectiveness of an intervention designed to prevent excess
weight gain (beyond gains associated with normal growth
and maturation), reduce the likelihood of being overweight
or obese, reduce time spent in recreational screen behaviours
(TV, computer and electronic games) and promote physical
activity participation among 10-year-old children. The
secondary aims were to evaluate intervention effects on
increasing enjoyment of physical activity and improving
FMS. The effects of the intervention were examined between
baseline and post intervention, and over the 2-year period of
the study incorporating 6- and 12-month follow-up data.
Methods and findings of this study are presented in
accordance with the CONSORT guidelines for reporting of
randomized controlled trials.21
Methods
Study design and sample size calculations
The methods of this group-randomized controlled trial
have been reported elsewhere and the study has been
found to have good intervention fidelity.22 A convenience
sample of three government primary schools located on
four campuses in low socioeconomic status (SES) areas
(based on socioeconomic index for areas scores)23 in
metropolitan Melbourne was recruited to the study. Children
attending schools in low SES areas were selected because of
previously shown inverse associations between SES and TV
viewing24 and between SES and adiposity among children.25
All grade 5 (approximately 1011 years old) students
(n 397) in the selected schools were eligible to participate
and were invited to take part in the study (see Figure 1).
School classes (n 17) were randomized to one of the three
intervention groups or a control group by withdrawing a
ticket from a container. No restrictions or stratification
procedures were used.
Sample size calculations were based on BMI (kg m2)
confidence intervals published by Robinson17 and the
observed change of 0.45 kg m2 between groups in that
study. It was estimated that 75 children would be needed in
each arm of the trial (using simple random sampling) to
detect a change in BMI of 0.5 kg m2 over 6 months with a
significance of a 0.05 and power 80%. On the basis of the
within-class intraclass correlations (ICCs) (r 0.0035) from
previous research involving Australian children,26 the design
effect for a class size of approximately 25 children was
1 0.0035  24 1.086. Consequently, the sample size was
 Switch-Play intervention
J Salmon et al
603
3 schools
recruited
(17 classes)

397 students assessed for - 38 active non-consent

eligibility (100%)
- 48 passive non-consent

- 300 consented to
assessment of all
components (75.6%)

- 11 consented to limited
assessment (2.7%)

Classes randomized into

treatmentor control conditions

Control BM FMS BM/FMS

BMI data collected at BMI data collected at BMI data collected at BMI data collected at
baseline: n = 62 baseline: n = 66 baseline: n = 74 baseline: n = 93
(49.2% boys) (49.3% boys) (47.4% boys) (48.9% boys)

Control BM FMS BM/FMS

BMI data collected at post BMI data collected at post BMI data collected at post BMI data collected at post
intervention: n = 63 intervention: n = 59 intervention: n = 72 intervention: n = 84

Control BM FMS BM/FMS

BMI data collected at 6- BMI data collected at 6- BMI data collected at 6- BMI data collected at 6-
month follow-up: n = 49 month follow-up: n = 59 month follow-up: n = 60 month follow-up: n = 78

Control BM FMS BM/FMS

BMI data collected at 12- BMI data collected at 12- BMI data collected at 12- BMI data collected at 12-
month follow-up: n = 55 month follow-up: n = 60 month follow-up: n = 69 month follow-up: n = 84

Figure 1 Flow of participants through each stage of the intervention.

inflated by approximately 10%, leading to a required sample
size of n 350.
Approvals
Ethics approval was obtained from the Deakin University
Human Research Ethics Committee and the Victorian
Department of Education and Training. School councils
approved the inclusion of the programme into the school
curriculum to be delivered to randomly assigned interven-
tion classes. All assessment components of the study
required active consent from the child and a parent (an
ethics requirement in Australia).
Intervention
The intervention components were based on principles from
social cognitive theory (for example, self-efficacy and
behavioural capability)27 and behavioural choice theory
(for example, preference, reinforcement),28 using techniques
such as self-monitoring, behavioural contracting to switch
off the TV for increasing durations (from one show per
week increasing to four shows per week maintained for the
duration of the intervention), reinforcement and skill
building. Full details of the intervention have been
previously described.22 There were two intervention compo-
nents: a behavioural modification (BM) condition and an
FMS condition. These intervention components were
delivered in addition to the usual physical education and
sports classes. Each of the intervention conditions consisted
of 19 lessons (4050 min each), which were delivered by
one qualified physical education teacher from March to
November 2002 (1 school year in Australia). Classes were
randomly allocated to a BM only condition, an FMS only
condition, a combined BM/FMS condition (2  19 lessons) or
a control (usual curriculum) group. Children in the BM/FMS

International Journal of Obesity

 Switch-Play intervention
J Salmon et al
604
condition received both the BM and FMS lessons, therefore focused on at least two skills. The six skills were selected on
receiving double the dose of the other intervention groups. the basis that they are commonly used in childrens games,
The BM lessons were delivered in the classroom and sports and physical activities. More detail on the FMS lesson
incorporated: self-monitoring (increasing childrens aware- content is provided in Table 1.
ness of time spent in physical activity and screen
behaviours); the health benefits of physical activity; aware-
ness of the home and community physical activity, and Outcome measures
sedentary behaviour environments; decision-making and Body mass index. Childrens weight and height without
identifying alternatives to screen behaviours that included shoes were measured in private. Two trained staff members
designing their own physical activity games; intelligent TV (not blinded to group assignment) collected childrens
viewing and reducing viewing time; advocacy of reduced height and weight at each of the four time points. On each
screen time through poster displays and role plays; use of occasion, childrens height and weight were measured twice
pedometers; and group games including all children in the and the average reading was used in data analyses. BMI
BM condition at each of the schools (see Table 1). From (kg m2) was calculated and converted as recommended for
Lessons 11 to 14, children completed a weekly contract analysis of longitudinal adiposity data.30 This involves
undertaking to switch off one television programme per subtracting the sexage population median (based on US
week over the 4-week period (that is, they switched off one data)31 from the childs raw BMI score. For convenience,
programme for the week of Lesson 11, two programmes for these BMI units of difference from the sexage population
Lesson 12 and so on). A newsletter was sent home to parents median will hereafter be referred to simply as BMI. Children
of children in the BM or combined BM/FMS condition asking were also categorized as healthy or overweight/obese based
them to sign their childs switch-off contracts each week to on International Obesity Task Force definitions.32
confirm that the nominated programme was turned off, and
after Lesson 14 parents were encouraged to help their child Objectively assessed physical activity. Physical activity was
maintain the switch-off. assessed using Manufacturing Technology Inc. (MTI), Florida,
The FMS lessons were delivered either in the indoor or USA, Actigraph Model, AM7164-2.2C accelerometers.
outdoor physical activity facilities at each school (dependent Children wore the MTI on a belt positioned over the right
on the weather and accessibility). Through games and hip during waking hours, except when bathing or swim-
activities developed for this intervention (based on ming, for 8 days at each of the four measurement points. For
previously published materials),29 these lessons focused each child, the first and last days of accelerometer data were
on mastery of six FMS. The interventionist taught the skills discarded due to incomplete data (administering and
with an emphasis on enjoyment and fun through games and collecting the monitors) and possible reactivity effects on
maximum involvement for all the children. Most lessons the first day. Only children with at least 3 complete days of
accelerometer data, including 1 weekend day (with at least
10 h of data per day), were included in the analyses. Extreme
Table 1 Lesson content for the behavioural modification (BM) and accelerometer counts (for example, 44 000 000 counts per
fundamental motor skills (FMS) conditions
day) were also excluded from the analyses as this indicated a
Lesson BM FMS possible unit malfunction. As well as computing an average
daily movement count, movement count thresholds based
1 Introduction to Switch-Play Run and throw
on an age-specific energy expenditure prediction equation
2 Patterns of sedentary behaviour (SB)a Throw and dodge
3 Self-monitoring SB Run and strike were applied to the data to calculate the average time
4 Physical activity and health Vertical jump and throw spent in moderate- (3.05.9 metabolic equivalent units of
5 Patterns of physical activity behaviour Dodge and kick rest (METs)) and vigorous-intensity (46.0 METs) physical
6 The home environment Run and strike
activity.33 The time spent in each intensity of activity on
7 The community environment Throw and dodge
8 Decision-making Kick and vertical jump each day was summed and divided by the total number of
9 Identifying alternative activities Dodge and throw days on which the accelerometer was worn.
10 Increasing physical activity Throw and kick
11 Intelligent viewing and decreasing SBb Throw and strike
Self-reported screen behaviours. Children completed a
12 Intelligent viewingb and TV advertising Throw and vertical jump
13 Advocacy of decreased SBFrole playsb Run and kick questionnaire at baseline and post intervention, and at
14 Perform advocacy playsb Dodge and strike 6- and 12-month follow-up periods during classroom time
15 Advocacy of decreased SBFposters Vertical jump and strike under the supervision of the investigators (not blinded to
16 Complete advocacy posters Throw and run
group assignment), who were available to assist students
17 Increasing physical activityFpedometers Kick and dodge
18 Switch-Play games Vertical jump, run and kick with any questions or problems. Children reported the usual
19 Present posters to younger grades Dodge, run, strike and kick time (hours/minutes) MondayFriday and SaturdaySunday
a in TV viewing, computer use and playing electronic games.
Sedentary behaviourFTV viewing, computer use and electronic game use.
b
Switch-off challenge from one TV programme in Lesson 11 to four Each of these variables was summed and divided by seven to
programmes in Lesson 14. calculate average minutes per day spent watching TV, using

International Journal of Obesity


the computer and playing electronic games. All of these
items showed testretest ICC equal to or greater than 0.6.
The convergent validity of these items has been shown to be
acceptable in previous research.34
Self-reported enjoyment of physical activity. At baseline and
post intervention, and at 6- and 12-month follow-up
periods, enjoyment of physical activity was assessed with
a modified version of an existing instrument. Using a
five-point Likert scale, children rated how much they enjoy,
or think they might enjoy each of the 36 physical activities
(for example, basketball, skipping rope, riding a bicycle,
walking to and from school and swimming) by circling a
smiley face that ranged from an unhappy face with
downturned mouth (2) to a happy face with an upturned
mouth ( 2). The enjoyment scores of all 36 physical
activities were summed to create a physical activity enjoy-
ment score (72 to 72). The reliability of the combined
physical activity enjoyment score was ICC 0.7.
Fundamental movement skills. Six FMS, including three
object control skills (overhand throw, two-handed strike
and kick) and three locomotor skills (dodge, sprint run and
vertical jump) were assessed at baseline, post intervention
and at 12-month follow-up periods by four trained members
of staff (not blinded to group assignment). An established
protocol was used to assess childrens FMS. Test conditions
where possible, were identical between schools. Each of the
skills that were assessed comprised between five and eight
components and the methods for assessing proficiency in
each of these skills have been described in detail pre-
viously.29 To assess mastery at these skills, children were
placed into groups of up to six students and their
performance video tape recorded. In each of the skills, the
child was given up to four attempts. Five specialist
evaluators, trained to assess childrens mastery of FMS and
blind to group assignment, examined the video tapes of each
of the six skills. Agreement between evaluators was high
(485%).
Each fundamental movement skill examined was
composed of between five and seven components.29 If
children displayed correct performance on all, or all but
one component of a skill, they were classified as having
achieved mastery or near mastery, respectively, for that
skill.35 For example, the kick had seven skill components;
therefore a child displaying mastery of all components of
the kick would receive a score of seven. For the dodge skill,
inconsistent testing conditions at one of the schools (for
example, indoor and outdoor) meant children at that school
were unable to demonstrate their level of mastery of that
skill. Therefore, the dodge was excluded from all analyses.
Data from the remaining five skills were summed (score
range: 034) and z-scores were created to account for the fact
that the number of components varied among the skills.
Switch-Play intervention
J Salmon et al
605
Demographics. Children reported their gender and date of
birth at baseline.
Unintended consequences. To assess unintended adverse body
image-related consequences from the intervention, children
were asked How happy are you with your body weight? and
How happy are you with the shape of your body? at each of
the four assessment points using a five-point Likert scale
depicting a smiley face that ranged from an unhappy face
with downturned mouth (2) to a happy face with an
upturned mouth ( 2). Testretest of these items found
acceptable reliability (ICC40.8). In addition, children were
asked to report (yes/no) whether they had changed their
eating in the last month to try to lose weight or to gain
weight (percent agreement 92%).
Food intake. In determining intervention effects on
preventing excess weight gain among children, it is im-
portant to adjust for the impact of any changes in food
intake. A food-frequency questionnaire consisting of 22
common food/drink items that were identified from
National Nutrition Survey (NNS) data36 for the target age
groups as important contributors to energy and fat intakes
(for example, high-energy drinks, savoury snack foods, sweet
snack foods and takeaway foods), and thus the energy
density of the diet. This questionnaire was administered at
each of the four assessment points. Children were asked how
many times they ate a certain food in the past week, and had
a range of eight response options from four or more times
per day to once a week and not eaten. Weekly consump-
tion scores for individual food/drink items were summed to
give a total consumption score. Previous research has
reported acceptable reliability of this measure.37
Statistical analyses
All analyses were performed using Stata Version 8. Gender
and intervention group differences at baseline were deter-
mined using linear regression and adjusted for clustering by
school class. Generalized estimating equations (GEEs),38 also
known as population-averaged panel-data models, were used
to analyse the initial effects of the intervention between
baseline and post intervention. GEE provides a method of
analysing correlated data, where participants are measured at
several different points over time; and clustered data,
where participants share a common characteristic, such as
belonging to the same school class (as is the case in this
study). In addition, GEE can accommodate missing data,
assuming that they are randomly missing.
For the continuous outcome data (age- and sex-adjusted
BMI; time (min week1) spent in TV viewing, computer use,
electronic games; physical activity counts per day; moderate-
and vigorous-intensity physical activities min day1;
enjoyment of physical activity; and FMS z-scores), the xtgee
function in Stata was employed, adjusting for group*time
interactions and clustering by school class. For the
International Journal of Obesity
 Switch-Play intervention
J Salmon et al
606
categorical data (weight status, eating to lose weight and
eating to gain weight) the xtlogit function adjusting for
group*time interactions and clustering by school class was
used. In all analyses of intervention effects on BMI and risk
of overweight and obesity, statistical tests were initially
unadjusted but then to account for potential variations in
eating and physical activity among children (that is,
confounding), subsequent analyses adjusted for food fre-
quency and moderate-to-vigorous intensity physical activity
(MVPA). Unintended intervention outcomes were deter-
mined by examining intervention effects on childrens self-
reported happiness with their body shape and body weight
(using xtgee), and whether in the last month they had
changed their eating to try to lose weight or to try to gain
weight (using xtlogit).
To determine longer term effects of the intervention, all
analyses were repeated incorporating baseline and post
intervention and 6- and 12-month follow-up data (apart
from FMS data, which was only collected at three time
points). To test if gender was a moderator of the interven-
tion, gender*group interactions were analysed for each of
the outcome variables. Where gender was found to moderate
the effects of the intervention, outcomes were analysed
separately for boys and girls.
Results
Profile of participants
Informed consent was received from 311 children (78%
response rate). Data were unavailable for one child (left the
school) and incomplete for five children, resulting in a final
baseline sample size of N 306. Flow of participants through
all stages of the study is presented in Figure 1. The sample
included 150 boys (mean age 10 years 85 months) and 156
girls (mean age 10 years 84 months). Complete BMI data
were available for 295 children at baseline, 278 children at
post intervention, 246 children at 6-month follow-up and
268 children at 12-month follow-up (12% attrition from
baseline). Reasons for missing data included active or passive
non-consent, student absence from school or student
departed school (27 children left their school over the 2
years of the study).
Table 2 shows descriptive baseline data (raw means) by
childs gender and by intervention group within gender.
Linear regression analyses, adjusted for clustering, found
that compared with girls, boys spent significantly more
time engaged in moderate- and vigorous-intensity physical
activities and playing electronic games, had significantly
higher accelerometer movement counts and FMS z-scores.
Girls reported significantly higher enjoyment of physical
activity than boys. There were significant differences
between intervention groups at baseline in age- and
sex-adjusted BMI among girls; however, these differences
were taken into account with the inclusion of baseline data
in all GEE analyses.
International Journal of Obesity
BMI and weight status
There was a significant intervention effect between baseline
and post intervention on childrens BMI (unadjusted and
adjusted for food frequency and MVPA) among those in the
combined BM/FMS group compared with the control group
(Table 3). That is, after adjusting for food frequency and
MVPA, the BM/FMS group recorded on average 1.88 BMI
units less than the control group. These effects were
maintained with the inclusion of 6- and 12-month follow-
up data. In addition, after adjusting for food frequency and
MVPA, compared with children in the control group, those
in the combined BM/FMS group were over 60% less likely to
be overweight or obese on average between baseline and post
intervention and over the four time points of the study.
There were no significant group*gender interactions for
BMI or for weight status; however, as there were significant
differences in BMI by intervention group among girls at
baseline, separate analyses were performed by gender
adjusting for baseline BMI. There were no significant
intervention effects on boys or girls unadjusted or adjusted
BMI from baseline to post intervention. However, with the
inclusion of 6- and 12-month follow-up data, there were
significant intervention effects on unadjusted BMI among
girls in the FMS and BM/FMS groups compared with the
control group. After adjusting for food frequency and MVPA,
girls in the BM/FMS group recorded on average 0.15 BMI
units less than the control group.
Physical activity
Between baseline and post intervention, there were signifi-
cant average effects over time between the control and BM
groups and between the control and FMS groups in move-
ment counts per day and in vigorous-intensity physical
activity min day1 (Table 3). For example, the BM group
spent almost 3 min day1 more in vigorous-intensity
physical activity compared with the control group; and
children in the FMS group spent 7.8 min day1 more in
vigorous-intensity physical activity compared with children
in the control group. These effects were maintained with
the inclusion of data from all four time points. From baseline
to post intervention, there was also a significant average
difference over time with the FMS group recording
10.4 min day1 more in moderate-intensity physical activity
than the control group, which was maintained with the
inclusion of the 6- and 12-month follow-up data.
Gender was a significant moderator of the intervention for
movement counts per day, moderate-intensity physical
activity and vigorous-intensity physical activity min day1
(Po0.001). Table 4 shows that between baseline and post
intervention, there were significant positive average differ-
ences over time between the BM and control groups and
between the FMS and control groups in movement counts
per day and in vigorous-intensity physical activity min day1
among boys. These differences were maintained with the
inclusion of 6- and 12-month follow-up data. There was also
 Switch-Play intervention
J Salmon et al
607
Table 2 Baseline values for outcome variables by gender and intervention group

Total Controls BM FMS BM/FMS

BMIa (mean, s.d.)

Boys 3.4 (3.6) 4.5 (3.3) 3.3 (3.8) 3.3 (3.2) 2.8 (4.0)
Girls 2.6 (3.5) 2.8 (4.1) 3.1 (3.3) 3.0 (3.7) 1.8 (3.1)w

Overweight/obeseb (%, 95% CI)

Boys 46.9 (38.655.4) 56.3 (37.773.6) 43.8 (26.462.3) 50.0 (32.967.1) 40.0 (25.755.7)
Girls 37.6 (29.845.9) 43.3 (25.562.6) 44.1 (27.262.1) 36.8 (21.854.0) 29.8 (17.344.9)

TV viewing (min week1) (mean, s.d.)

Boys 959.8 (650.4) 823.8 (676.2) 1037.7 (778.3) 976.2 (618.1) 988.2 (548.7)
Girls 866.7 (543.4) 730.0 (520.0) 977.3 (530.6) 892.5 (583.8) 849.8 (530.2)

Computer use (min week1) (mean, s.d.)

Boys 143.3 (218.1) 125.0 (153.7) 200.0 (263.1) 110.4 (164.2) 141.2 (255.0)
Girls 164.2 (255.4) 161.4 (274.4) 147.3 (183.5) 155.4 (266.8) 186.6 (286.7)

Electronic games use (min week1) (mean, s.d.)

Boys 583.5 (645.5)# 520.6 (595.6) 585.8 (620.0) 758.5 (740.0) 480.5 (604.4)
Girls 197.3 (354.1) 256.8 (447.3) 206.4 (362.9) 108.6 (149.4) 227.3 (398.7)

Counts per day (  103) (mean, s.d.)

Boys 543.1 (235.7)# 481.1 (127.3) 543.3 (268.6) 594.7 (354.7) 551.7 (150.1)
Girls 425.3 (138.4) 397.8 (101.7) 433.4 (129.6) 453.6 (153.1) 414.9 (151.4)

Moderate PA (min day1) (mean, s.d.)

Boys 131.2 (35.9)** 124.1 (27.8) 123.1 (34.1) 133.9 (43.9) 140.4 (34.8)
Girls 112.5 (33.5) 107.1 (27.6) 118.3 (28.9) 118.4 (42.0) 107.1 (31.9)

Vigorous PA (min day1) (mean, s.d.)

Boys 25.4 (22.8)# 19.9 (11.0) 24.2 (18.5) 33.6 (39.5) 24.4 (13.6)
Girls 14.2 (8.9) 12.2 (7.1) 13.8 (9.4) 16.1 (8.2) 14.2 (10.0)

PA enjoyment (mean, s.d.)

Boys 0.63 (0.54)* 0.56 (0.59) 0.57 (0.54) 0.84 (0.43) 0.55 (0.54)
Girls 0.78 (0.45) 0.79 (0.33) 0.75 (0.51) 0.88 (0.46) 0.72 (0.46)

FMS z-scores (mean, s.d.)

Boys 0.52 (0.83)# 0.28 (0.83) 0.66 (0.68) 0.51 (0.78) 0.59 (0.93)
Girls 0.50 (0.89) 0.95 (0.60) 0.28 (0.89) 0.27 (0.94) 0.60 (0.91)

Abbreviations: BM, behavioural modification; BMI, body mass index; CI, confidence interval; FMS, fundamental movement skills; PA, physical activity. aBMI/sexage
population median.31 bAge- and sex-specific internationally accepted cut points.32 *Po0.05, **Po0.01, #Po0.001 significant differences by gender across
intervention groups (adjusting for clustering). wPo0.05 significant difference by intervention group within gender (adjusting for clustering).

a significant positive average difference over time between
boys in the BM/FMS and control groups in vigorous-intensity
physical activity from baseline to post intervention. These
differences were not maintained with the inclusion of the
6- and 12-month follow-up data. Among girls, there were
significant positive average differences in movement counts
and in moderate-intensity physical activity between the BM
and control groups from baseline to post intervention and
over the four time points (Table 5).
Screen behaviours
There were significant intervention effects between baseline
and post intervention for childrens TV viewing min week1
among children in the BM group compared with those in the
control group (Table 3). However, this was in the undesired
direction, with BM children reporting 229 min week1 more
in TV viewing on average over time compared with the
control group. These effects were maintained with
the inclusion of 6- and 12-month follow-up data. There
were no significant intervention effects on other screen
behaviours.
There was a significant group*gender interaction for time
spent playing electronic games (Po0.001). However, when
the data were reanalysed separately by gender, there were no
significant effects from baseline to post intervention or over
the four time points of the study on electronic games use
among boys or girls (data not shown).
Enjoyment of physical activity
There were significant average differences in physical activity
enjoyment between baseline and post intervention, with
children in the FMS group reporting higher average

International Journal of Obesity

 Switch-Play intervention
J Salmon et al
608
Table 3 Interventiona and maintenance effectsb (coefficient, 95% confidence intervals (CIs)) on screen time (min week1), physical activity (PA; min day1)c,
fundamental movement skills (FMS)d, enjoyment of PA, body mass index (BMI)e and weight status

BM FMS BM/FMS

Baseline to post Baseline to 12-month Baseline to post Baseline to 12-month Baseline to post Baseline to 12-month
intervention a follow-up b intervention a follow-up b intervention a follow-up b

Odds ratios (95% CI)f

Overweight/obese 0.78 (0.39 to 1.57) 0.78 (0.39 to 1.57) 0.76 (0.38 to 1.50) 0.76 (0.39 to 1.50) 0.53 (0.28 to 1.03) 0.53 (0.28 to 1.03)
(unadjusted)
Overweight/obese 0.88 (0.36 to 2.15) 0.65 (0.35 to 2.10) 0.62 (0.26 to 1.48) 0.66 (0.28 to 1.56) 0.36 (0.15 to 0.86)* 0.38 (0.16 to 0.89)*
(adjusted)g

b-coefficients (95% CI)

BMI (unadjusted) 0.40 (1.11 to 0.30) 0.42 (1.07 to 0.23) 0.50 (1.25 to 0.25) 0.45 (1.19 to 0.29) 1.30 (2.29 to 0.31)* 1.30 (2.24 to 0.35)**
BMI (adjusted)f 0.06 (1.23 to 1.12) 0.15 (1.29 to 0.99) 0.86 (1.94 to 0.23) 0.77 (1.80 to 0.26) 1.88 (3.22 to 0.53)** 1.53 (2.82 to 0.24)*
TV viewing (min week1) 229.3 (16.6 to 442.0)* 239.9 (27.6 to 452.2)* 149.4 (20.7 to 319.5) 142.6 (33.6 to 318.9) 137.0 (17.4 to 291.5) 141.9 (15.6 to 299.5)
Computer use 29.5 (46.9 to 105.9) 31.7 (46.2 to 109.6) 9.5 (102.8 to 83.8) 11.8 (107.5 to 83.7) 21.9 (68.7 to 112.4) 21.4 (71.7 to 114.4)
(min week1)
Electronic games 8.5 (180.0 to 163.1) 8.5 (179.7 to 162.7) 28.5 (269.1 to 326.0) 22.8 (277.3 to 323.0)42.4 (236.2 to 151.5) 44.0 (237.6 to 149.6)
(min week1)
Counts per day (  103) 47.0 (24.2 to 69.8)# 47.5 (24.6 to 70.4)# 76.6 (35.2 to 118.0)# 76.1 (33.4 to 118.9)# 40.8 (9.9 to 91.4) 40.1 (9.8 to 90.1)
Moderate PA 5.3 (2.0 to 12.6) 4.3 (3.6 to 12.2) 10.4 (2.8 to 18.1)** 9.5 (1.4 to 17.6)* 7.7 (5.1 to 20.6) 6.7 (6.4 to 19.8)
(min day1)
Vigorous PA (min day1) 2.8 (0.3 to 5.4)* 2.8 (0.2 to 5.4)* 7.8 (3.4 to 12.3)** 7.7 (3.2 to 12.2)** 3.1 (0.58 to 6.7) 3.0 (0.59 to 6.6)
PA enjoyment 0.23 (0.27 to 0.22) 0.02 (0.26 to 0.23) 0.18 (0.00 to 0.35)* 0.19 (0.01 to 0.36)* 0.05 (0.29 to 0.20) 0.04 (0.28 to 0.20)
FMS z-scores 0.47 (0.17 to 1.11) 0.48 (0.15 to 1.11) 0.44 (0.14 to 1.02) 0.45 (0.11 to 1.02) 0.35 (0.36 to 1.06) 0.36 (0.35 to 1.06)

Abbreviations: BM, behavioural modification; MVPA, moderate-to-vigorous intensity physical activity. aGeneralized estimating equation (GEE) coefficient at baseline
and post intervention, adjusted for clustering by school class. bGEE coefficient at baseline and post intervention, 6- and 12-month follow-up periods, adjusted for
clustering by school class. cAccelerometer. dFMS z-scores. eBMI units of difference from US sex- and age-adjusted population median.31 fReferent category: not
overweight/obese. gAdjusted for food-frequency intake (high energy drinks, sweet and savoury snacks, confectionery and fast food) and MVPA. *Po0.05, **Po0.01,
#
Po0.001. Bold values denote statistical significance.

Table 4 Interventiona and maintenance effectsb (coefficient, 95% confidence intervals) on physical activity (PA; counts per day or min day1)c and enjoyment of PA
among boys

BM FMS BM/FMS

Baseline to post Baseline to 12-month Baseline to post Baseline to 12-month Baseline to post Baseline to 12-month
intervention a follow-up b intervention a follow-up b intervention a follow-up b

Counts per day (  103) 61.5 (21.7 to 101.4)** 61.6 (12.4 to 110.9)* 112.8 (59.1 to 166.5)# 114.0 (52.8 to 175.2)# 72.1 (7.7 to 151.9) 66.2 (20.7 to 153.0)
Vigorous PA (min day1) 4.5 (0.91 to 8.0)* 4.4 (0.44 to 8.4)* 13.8 (8.7 to 18.9)# 13.8 (8.4 to 19.1)# 5.7 (0.21 to 11.2)* 4.8 (1.2 to 10.7)
PA enjoyment 0.03 (0.26 to 0.31) 0.03 (0.25 to 0.32) 0.30 (0.10 to 0.51)** 0.32 (0.11 to 0.52)** 0.02 (0.28 to 0.33) 0.03 (0.27 to 0.34)

Abbreviations: BM, behavioural modification; FMS, fundamental movement skills. aGeneralized estimating equation (GEE) coefficient at baseline and post
intervention, adjusted for clustering by school class. bGEE coefficient at baseline and post intervention, 6- and 12-month follow-up periods, adjusted for clustering by
school class. cAccelerometer. *Po0.05, **Po0.01, #Po0.001. Bold values denote statistical significance.

enjoyment scores over time compared with those in the
control group (Table 3). These effects were maintained with
the inclusion of physical activity enjoyment data across all
four time points. There was a significant group*gender
interaction for enjoyment of physical activity (P 0.001).
Table 4 shows that there were significant positive differences
in physical activity enjoyment between baseline and post
intervention among boys in the FMS group compared with
those in the control group. These effects remained after the
inclusion of the 6- and 12-month follow-up data.
Fundamental movement skills
There were no significant intervention effects on FMS
z-scores between baseline and post intervention or over
the four time points. There was, however, a significant
group*gender interaction for FMS (Po0.001). Table 5 shows
FMS between baseline and post intervention, where com-
pared with girls in the control group, girls in the BM and
the FMS groups recorded significantly higher average FMS
z-scores over time. These positive differences remained with
the inclusion of data from all four time points.
Unintended outcomes
From baseline to post intervention, there were no effects on
unintended outcomes of the intervention such as childrens
happiness with their body shape and body weight, or eating
to gain weight or lose weight in the last month. However,
when a group*gender interaction term was included in the
model, gender appeared to moderate the intervention effects
on satisfaction with body shape and with body weight

International Journal of Obesity


Table 5 Interventiona and maintenance effectsb (coefficient, 95% confidence intervals) on physical activity (PA; counts per day or min day1)c and fundamental
movement skills (FMS) among girls
BM
Baseline to post Baseline to 12-month
interventiona follow-upb
BMI (unadjusted)d 0.01 (0.05 to 0.03) 0.01 (0.07 to 0.04)
BMI (adjusted)e 0.07 (0.03 to 0.17) 0.01 (0.14 to 0.12)
Counts per day (  103) 36.5 (0.29 to 72.7)* 39.3 (9.4 to 69.1)*
Moderate PA (min day1) 12.1 (3.9 to 20.3)** 11.1 (3.8 to 18.4)**
FMS z-scores 0.76 (0.11 to 1.41)* 0.75 (0.11 to 1.40)*
Abbreviations: BM, behavioural modification; BMI, body mass index; MVPA, moderate-to-vigorous intensity physical activity. aGeneralized estimating equations
(GEE) coefficient at baseline and post intervention, adjusted for clustering by school class. bGEE coefficient at baseline and post intervention, 6- and 12-month follow-
up periods, adjusted for clustering by school class. cAccelerometer. dBMI units of difference from US sex- and age-adjusted population median.31 eAdjusted for food-
frequency intake (high energy drinks, sweet and savoury snacks, confectionery and fast food) and MVPA. *Po0.05, **Po0.01, #Po0.001. Bold values denote
statistical significance.
(Po0.01). When stratified by gender, boys in the FMS group
(B 0.56, 95% CI: 0.19 to 0.93, P 0.003) and combined
BM/FMS group (B 0.59, 95% CI: 0.12 to 1.05, P 0.014)
recorded significantly higher satisfaction with their body
shape between baseline and post intervention compared
with boys in the control group. This effect was maintained
over the four time points. Among girls, there were no effects
on the unintended outcomes that were assessed.
Discussion
This intervention aimed to prevent excess weight gain
among 10-year-old children, to prevent declines in physical
activity and to reduce screen behaviours. The intervention
also aimed to increase enjoyment of physical activity and to
improve FMS. Consistent with previous screen behaviour
intervention studies,16,17 this study found favourable out-
comes for childrens BMI and weight status. On average,
between baseline and post intervention, and including
6- and 12-month follow-up data, children in the combined
BM/FMS group recorded significantly lower BMI compared
with children in the control group. These differences were
strengthened with the inclusion of food-frequency and
physical activity data across all the time points. Adjusted
analyses also found that children in the BM/FMS group were
more than 60% less likely to be overweight or obese on
average over time (baseline and post intervention, and over
the four time points) compared with those in the control
group.
Unlike previous screen behaviour interventions,16,17,39
this intervention was not effective in reducing screen
behaviours, with children in the BM group reporting higher
mean time per week watching TV between baseline and post
intervention compared with children in the control group.
This difference may be due to a failure of the intervention to
reduce childrens TV viewing, instead increasing childrens
awareness and engagement with that behaviour (an
undesired outcome). However, it may also be because a
Switch-Play intervention
J Salmon et al
609
FMS BM/FMS
Baseline to post Baseline to 12-month Baseline to post Baseline to 12-month
interventiona follow-upb interventiona follow-upb
0.02 (0.07 to 0.03) 0.07 (1.12 to 0.02)** 0.03 (0.08 to 0.02) 0.07 (0.13 to 0.01)*
0.13 (0.02 to 0.29) 0.08 (0.22 to 0.05) 0.01 (0.50 to 0.13) 0.15 (0.31 to 0.00)*
56.2 (18.6 to 131.0) 58.0 (11.8 to 127.8) 180.3 (17.3 to 53.3) 21.6 (10.3 to 53.5)
11.4 (4.5 to 27.3) 10.2 (4.7 to 25.1) 1.1 (7.5 to 9.8) 0.3 (7.7 to 8.3)
0.78 (0.27 to 1.29)** 0.76 (0.27 to 1.26)** 0.51 (0.16 to 1.18) 0.49 (0.17 to 1.15)
self-reported measure of childrens TV viewing was used. As
children learnt about the pros and cons of TV viewing and
how to monitor their viewing and turn off the TV, their
accuracy in reporting this behaviour may have improved
over time. Baranowski et al.40 call this phenomenon a
response shift bias, suggesting that based on learning
effects, there is a differential favourable shift in the accuracy
of reporting among children in the intervention group
compared with those in the control group. Although the
screen behaviour self-reported measure used in this study
has shown acceptable reliability and concurrent validity
(compared with parental proxy reporting), an objective or
log or diary measure of these behaviours may have yielded
different results.
There were no intervention effects on electronic games or
computer use. This may be because the intervention focused
more strongly on TV viewing. The intervention programme
included self-monitoring of the screen behaviours and
lessons focused on reducing time spent in these behaviours;
however, the behavioural contracts focused only on
switching off the TV. Although children were encouraged
not to exchange TV viewing for computer use (apart from
educational instruction) or playing electronic games, the
intervention programme did not employ a specific BM
technique for reducing electronic games or computer use.
Robinson17 reported separately the intervention effects on
time children spent watching videotapes and playing video
games and found significant reductions in self-reported
video game playing among children in the intervention
group compared with the control group. He did not,
however, assess their computer use. These behaviours have
different correlates34 and therefore future interventions need
to tailor programmes to target reductions in specific
sedentary behaviours.
The significant effects on physical activity were a unique
aspect of this intervention, with previous screen behaviour
interventions not reporting a significant impact on
childrens physical activity.16,17,39 The average effect across
time between the control and the FMS groups in the current
study was approximately 10 min day1 in moderate-intensity
International Journal of Obesity
 Switch-Play intervention
J Salmon et al
610
and 8 min day1 in vigorous-intensity physical activity,
which equates to 13% of the daily average time spent in
MVPA at baseline. There were more modest intervention
effects on physical activity among children in the BM group.
A previous review of childrens physical activity interven-
tions concluded that curriculum-based interventions are not
as effective in promoting childrens physical activity as those
that include some focus on physical education and that
also focus on family.41 The FMS intervention focused on
movement skills using a programme that emphasized games
and fun. Although the BM intervention was primarily
classroom-based, it did also incorporate physical
activity elements (children designed their own physical
activity games, which were played at the end of each BM
lesson) and a strong family focus.
Although intervention effects on BMI and weight status
for the whole sample were found only among children in the
combined BM/FMS group, girls in the FMS and BM/FMS
groups also had more favourable BMI outcomes compared
with the control group. It is logical that the combined
intervention had the most favourable outcomes on BMI and
weight status given that these children received double the
intervention dose compared with the others. Nevertheless,
the strongest effects on physical activity participation and
enjoyment and FMS were among children in the BM and
FMS groups. It may be that some of the messages on
behaviour change may have been lost in the 38 lessons
delivered over 10 months to children in the BM/FMS group.
Further, although BMI and weight status analyses adjusted
for food intake for all of the groups over time, that measure
may not have had the required accuracy or responsiveness to
changes in diet to explain potential differences in food
intake between groups over time. In addition, there may
have been behavioural changes that occurred during
non-assessment periods of the intervention that could have
affected the childrens weight but that were not captured at
the time of the assessment.
Gender was a significant moderator of the effects of the
intervention for physical activity, enjoyment of physical
activity and FMS. There were significant intervention effects
on physical activity for boys and girls in the BM group and
for boys in the FMS and BM/FMS groups. In addition, boys in
the FMS group reported higher average levels of physical
activity enjoyment over time compared with boys in the
control group. The FMS intervention focused on skills that
may have been more appealing to boys than girls, which
may be why boys in the FMS and BM/FMS groups showed
favourable physical activity outcomes. In contrast, the BM
intervention focused on play activities, which may have
been more appealing to girls. A recent review identified eight
school-based interventions that were developed to promote
physical activity among girls.42 Five of these studies reported
significant effects on girls moderate- or vigorous-intensity
physical activity, or physical activity during physical educa-
tion classes or on the number of pedometer steps per day.
A consistent feature of these effective interventions was the
International Journal of Obesity
promotion of noncompetitive activities, a focus on enjoy-
ment and fun, goal setting and self-monitoring, and the
social aspects of the activities. It may be important to
incorporate these strategies in future physical activity
interventions that include girls.
Enjoyment has been found to mediate the effects of a
school-based physical activity intervention among adole-
scent girls.43 Further analyses are required to test whether
enjoyment mediated the positive effects on physical activity
among boys. Among girls in the BM and FMS groups, there
were significant positive effects on FMS over time compared
with the control group. Baseline data in this study and
population data35 suggest that girls have poorer FMS than
boys; therefore, girls in the FMS group may have been more
receptive to the intervention. Although girls in the BM group
did not receive the FMS intervention, higher FMS z-scores in
that group compared with the girls in the control group may
be explained by their higher overall movement counts and
moderate-intensity physical activity.
There were a number of limitations of this study. With
the intervention groups being randomized by class, there
was potential for contamination between intervention and
control groups. However, this was assessed in process
evaluation, and there was a much lower level of awareness
of the Switch-Play intervention among parents with
children in the control group (44%) compared with parents
whose children participated in the intervention (80%). The
use of self-reported measures to assess screen behaviours was
a limitation. Childrens pubertal staging was not assessed;
therefore, children may have been at different stages of
maturation between groups, which may have affected
weight outcomes. A limitation of the BM intervention was
that much of the targeted behaviour change in screen
behaviours was to occur in the home. Previously published
process evaluation results suggests that many parents were
not aware of all of the intervention requirements (that is, to
reduce childrens TV viewing time and to increase their
physical activity), with most reporting the intervention was
about promoting childrens physical activity.22 This may be
why we had no effect on screen behaviours in this study. An
intervention with a stronger focus on supporting children
and their families to reduce screen time may have resulted in
more favourable outcomes in those behaviours.
The study was underpowered according to the original
sample size estimates. Although several intervention effects
were identified, these effects may have been stronger, or
additional intervention effects may have been significant,
had we been successful in recruiting more children. Our
initial sample size calculations were based on previously
published observed change scores in BMI between groups,
and inflated by 10% to accommodate the within-class
correlations of BMI. A more appropriate method of sample
size calculation for a group-randomized trial would assume
that the number of clusters is known but that the number of
participants in each cluster is not: N K*m*(1ICC)/
(1(m*ICC)), where K is the number of clusters, m is the

number per cluster required if no clustering effect and ICC is
the intracluster correlation coefficient.44 Use of this formula
is likely to result in a larger sample size estimate. However, it
is also important to note that although this study was not
designed to support analyses at the level of the unit of
assignment, in using GEE to analyse the data, we are able to
model the effects of the intervention on subgroups (popula-
tion-averaged approach).45
The strengths of this study are that it used objectively
measured height, weight and physical activity, and
childrens FMS were evaluated by specialist staff blinded
to intervention group. The intervention was delivered
to children living in low SES areas. This is an important
subgroup in the population in which to intervene, particu-
larly given that the current sample had a substantially higher
proportion of children in the unhealthy weight range at
baseline (47% boys and 38% girls) compared with the recent
population estimates in New South Wales, Australia, which
found 32% of boys and 25% of girls in Year 6 were
overweight or obese.46 A further strength is the 6- and
12-month follow-up periods to determine maintenance of
intervention effects. The intervention was theoretically
based targeting potential mediators, and gender as a
moderator of the intervention was also examined.
The intervention approach used here holds promise as a
strategy to prevent obesity in children. The combined
BM/FMS intervention had the greatest effect on childrens
weight status and the FMS intervention had the greatest
effect on childrens physical activity. However, further work
is required to determine the feasibility of delivering a
programme such as this in a real world setting, and its
applicability in other countries. It may be necessary to scale
down this programme in order for its adoption into the
already crowded school curricula.
Acknowledgements
This study was funded by the Victorian Health Promotion
Foundation. Jo Salmon is supported by a National Heart
Foundation of Australia and sanofi-aventis Career Develop-
ment Award. Kylie Ball is supported by a National Health and
Medical Research Council/National Heart Foundation of
Australia Career Development Award. David Crawford is
supported by a Victorian Health Promotion Foundation
Senior Research Fellowship.
References
1 Wang Y, Lobstein T. Worldwide trends in childhood overweight
and obesity. Int J Pediatr Obes 2006; 1: 1125.
2 Booth ML, Chey T, Wake M, Norton K, Hesketh K, Dollman J et al.
Change in the prevalence of overweight and obesity among
young Australians, 19691997. Am J Clin Nutr 2003; 77: 2936.
3 Crawford D, Ball K. Behavioural determinants of the obesity
epidemic. Asia Pac J Clin Nutr 2002; 11 (Suppl 8): S718S721.
Switch-Play intervention
J Salmon et al
611
4 Ludwig DS, Gortmaker SL. Programming obesity in childhood.
Lancet 2004; 364: 226227.
5 Janz KF, Burns TL, Levy SM. Tracking of activity and sedentary
behaviours in childhood: the Iowa Bone Development Study.
Am J Prev Med 2005; 29: 171178.
6 Janz KF, Dawson JD, Mahoney LT. Tracking physical fitness and
physical activity from childhood to adolescence: the Muscatine
Study. Med Sci Sports Exerc 2000; 32: 12501257.
7 Pate RR, Trost SG, Dowda M, Ott AE, Ward DS, Saunders R et al.
Tracking of physical activity, physical inactivity, and health-
related physical fitness in rural youth. Pediatr Exerc Sci 1999; 11:
364376.
8 Hancox RJ, Milne BJ, Poulton R. Association between child and
adolescent television viewing and adult health: a longitudinal
birth cohort study. Lancet 2004; 364: 257262.
9 Hardy LL, Bass SL, Booth ML. Changes in sedentary behavior
among adolescent girls: a 2.5-year prospective cohort study.
J Adolesc Health 2007; 40: 158165.
10 Sallis JF. Age-related decline in physical activity: a synthesis
of human and animal studies. Med Sci Sports Exerc 1998; 32:
15981600.
11 American Academy of Pediatrics. Children, adolescents, and
television. Pediatrics 2001; 107: 423426.
12 Commonwealth of Australia and Department of Health and
Ageing. Australias Physical Activity Recommendations for Children
and Young People 2004. Department of Health and Ageing:
Canberra.
13 AC Nielsen Media International. Australian TV Trends 2001 2001.
AC Nielsen Company: Sydney.
14 Andersen R, Crespo C, Bartlett S, Cheskin L, Pratt M. Relationship
of physical activity and television watching with body weight
and level of fatness among children. Results from the Third
National Health and Nutrition Examination Survey. JAMA 1998;
279: 938942.
15 de Mattia L, Lemont L, Meurer L. Do interventions to limit
sedentary behaviours changes behaviour and reduce childhood
obesity? A critical review of the literature. Obes Rev 2007; 8:
6981.
16 Gortmaker SL, Peterson K, Wiecha J, Sobol AM, Dixit S, Fox MK
et al. Reducing obesity via a school-based interdisciplinary
intervention among youth. Arch Pediatr Adolesc Med 1999; 153:
409418.
17 Robinson TN. Reducing childrens television viewing to
prevent obesity: a randomized controlled trial. JAMA 1999; 282:
15611566.
18 Sallis JF, Prochaska JJ, Taylor WC. A review of correlates of
physical activity of children and adolescents. Med Sci Sport Exerc
2000; 32: 963975.
19 Hume C, Okely A, Bagley S, Telford A, Booth M, Crawford D et al.
Does weight status influence associations between childrens
fundamental movement skills and physical activity? Res Q Exerc
Sport (in press, accepted 26 March 2007).
20 Okely AD, Booth M. Relationship of enjoyment of physical
activity and preferred activities to fundamental movement skill
proficiency in young children (abstract). Int J Behav Med 2000; 7:
S151.
21 Begg C, Cho M, Eastwood S, Horton R, Moher D, Olkin I et al.
Improving the quality of reporting of randomized controlled
trials. The CONSORT statement. JAMA 1996; 276: 637639.
22 Salmon J, Ball K, Crawford D, Booth M, Telford A, Hume C et al.
Reducing sedentary behaviour and increasing physical activity
among 10-year-old children: overview and process evaluation of
the Switch-Play intervention. Health Promot Int 2005; 20: 717.
23 Australian Bureau of Statistics. 1996 Census of Population and
Housing: Socio-Economic Indexes for Areas Catalogue No. 2039 1998.
Australian Bureau of Statistics: Canberra.
24 Gorely T, Marshall SJ, Biddle SJ. Couch kids: correlates of
television viewing among youth. Int J Behav Med 2004; 11:
152163.
International Journal of Obesity
 Switch-Play intervention
J Salmon et al
612
25 Booth M, Macaskill P, Lazarus R, Baur L. Sociodemographic
distribution of measures of body fatness among children and
adolescents in New South Wales, Australia. Int J Obes 1999; 23:
456462.
26 Wake M, Salmon L, Waters E, Wright M, Hesketh K.
Parent-reported health status of overweight and obese Australian
primary school children: a cross-sectional population survey. Int J
Obes 2002; 26: 717724.
27 Bandura A. Social Foundations of Thought and Action: A Social
Cognitive Theory 1986. Prentice Hall: Englewood Cliffs, NJ.
28 Rachlin H. Judgement, Decision, and Choice: A Cognitive/Behavioral
Synthesis 1989. WH Freeman: New York.
29 Department of Education Victoria. Fundamental Motor Skills: A
Manual for Classroom Teachers 1996. Department of Education
Victoria: Melbourne.
30 Cole TJ, Faith MS, Pietrobelli A, Heo M. What is the best
measure of adiposity change in growing children: BMI,
BMI %, BMI z-score or BMI centile? Eur J Clin Nutr 2005; 59:
419425.
31 Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, Flegal KM,
Guo SS, Wei R et al. CDC growth charts: United States. US
Department of Health and Human Services, Centers for Disease
Control and Prevention, National Center for Health Statistics Adv
Data 2000; 314: 128.
32 Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a
standard definition for child overweight and obesity worldwide:
International survey. BMJ 2000; 320: 12401243.
33 Trost SG, Pate RR, Sallis JF, Freedson PS, Taylor WC,
Dowda M et al. Age and gender differences in objectively
measured physical activity in youth. Med Sci Sport Exerc 2002;
34: 350355.
34 Salmon J, Timperio A, Telford A, Carver A, Crawford D. Family
environment and childrens television viewing and low level
physical activity. Obes Res 2005; 13: 19391951.
International Journal of Obesity
35 Booth M, Macaskill P, McLellan L, Phongsavan P, Okely T,
Patterson J et al. NSW Schools Fitness and Physical Activity Survey
1997. NSW Department of School Education: Sydney.
36 Australian Bureau of Statistics. Nutrient Intakes and Physical
Measurements. National Nutrition Survey 1995 1998. Australian
Bureau of Statistics: Canberra.
37 Campbell KJ, Crawford D, Ball K. Family food environment and
dietary behaviors likely to promote fatness in 56 year-old
children. Int J Obes 2006; 30: 12721280.
38 Liang KY, Zeger SL. Longitudinal data analysis using generalized
linear models. Biometrika 1986; 73: 1322.
39 Robinson TN, Killen JD, Kraemer HC, Wilson DM, Matheson DM,
Haskell WL et al. Dance and reducing television viewing to
prevent weight gain in African-American girls: the Stanford
GEMS pilot study. Ethn Dis 2003; 13: S65S77.
40 Baranowski T, Klesges LM, Cullen KW, Himes JH. Measurement of
outcomes, mediators, and moderators in behavioral obesity
prevention research. Prev Med 2004; 38: S1S13.
41 Timperio A, Salmon J, Ball K. Evidence-based strategies to
promote physical activity among children, adolescents and
young adults: review and update. J Sci Med Sport 2004; 7: 2029.
42 Dishman RK, Motl RW, Saunders R, Felton G, Ward DS, Dowda M
et al. Enjoyment mediates effects of a school based physical
activity intervention. Med Sci Sport Exerc 2005; 37: 478487.
43 Campbell MJ. Cluster randomized trials in general (family)
practice research. Stat Methods Med Res 2000; 9: 8194.
44 Murray DM. Design and Analysis of Group-Randomized Trials 1998.
Oxford University Press: New York.
45 Booth M, Okely AD, Denney-Wilson E, Hardy L, Yang B, Dobbins
T. NSW Schools Physical Activity and Nutrition Survey (SPANS) 2004:
Summary Report 2006. NSW Department of Health: Sydney.
46 Baranowski T, Anderson C, Carmack C. Mediating variable
framework in physical activity interventions. How are we doing?
How might we do better? Am J Prev Med 1998; 15: 266297.