Content uploaded by Chris Lonsdale
Author content
All content in this area was uploaded by Chris Lonsdale on May 05, 2021
Content may be subject to copyright.
1
Effectiveness of a scalable school-based intervention to improve
children’s cardiorespiratory fitness: The ‘internet-based Professional
Learning to help teachers promote Activity in Youth’ (iPLAY)
cluster randomized controlled trial
This is the prepublication version of the following manuscript:
Lonsdale C, Sanders T, Parker P, et al. Effect of a Scalable School-Based Intervention on
Cardiorespiratory Fitness in Children: A Cluster Randomized Clinical Trial. JAMA Pediatr.
Published online May 3, 2021. doi:10.1001/jamapediatrics.2021.0417
© 2021. This paper is not the copy of record and may not exactly replicate the authoritative
document published in JAMA Pediatrics
Authors
1. Chris Lonsdale, PhD, Australian Catholic University, 33 Berry Street, North Sydney,
NSW 2060, Australia
2. Taren Sanders, PhD, Australian Catholic University, 33 Berry Street, North Sydney,
NSW 2060, Australia
3. Philip Parker, PhD, Australian Catholic University, 33 Berry Street, North Sydney, NSW
2060, Australia
4. Michael Noetel, PhD, Australian Catholic University, School of Behavioural and Health
Sciences, 1100 Nudgee Rd, Banyo, QLD 4014, Australia
5. Timothy Hartwig, PhD, Australian Catholic University, School of Behavioural and
Health Sciences, 25A Barker Rd, Strathfield NSW 2135, Australia
6. Diego Vasconcellos, PhD, Australian Catholic University, 33 Berry Street, North
Sydney, NSW 2060, Australia
7. Jane Lee, MPH, Australian Catholic University, 33 Berry Street, North Sydney, NSW
2060, Australia
8. Devan Antczak, MSc, Australian Catholic University, , 33 Berry Street, North Sydney,
NSW 2060, Australia
2
9. Morwenna Kirwan, PhD, Macquarie University, Faculty of Medicine and Health
Sciences, Balaclava Rd, Macquarie Park NSW 2109, Australia
10. Philip Morgan, PhD, University of Newcastle, Priority Research Centre for Physical
Activity and Nutrition, University Dr, Callaghan NSW 2308, Australia
11. Jo Salmon, PhD, Deakin University, Institute for Physical Activity and Nutrition, 221
Burwood Highway, Burwood, Victoria 3125, Australia
12. Marj Moodie, PhD, Deakin University, Global Obesity Centre, 221 Burwood Highway,
Burwood, Victoria 3125, Australia
13. Heather McKay, PhD, University of British Columbia, Centre for Hip Health and
Mobility, 7/F, 2635 Laurel Street, Vancouver, BC V5Z 1M9, Canada
14. Andrew Bennie, PhD, Western Sydney University, School of health Sciences, Locked
Bag 1797 Penrith NSW 2751, Australia
15. Ronald C Plotnikoff, PhD, University of Newcastle, Priority Research Centre for
Physical Activity and Nutrition, University Dr, Callaghan NSW 2308, Australia
16. Renata Cinelli, PhD, Australian Catholic University, National School of Education, 25A
Barker Rd, Strathfield NSW 2135, Australia
17. David Greene, PhD, Australian Catholic University, School of Behavioural and Health
Sciences, 25A Barker Rd, Strathfield NSW 2135, Australia
18. Louisa Peralta, PhD, Sydney University, School of Education and Social Work,
Camperdown NSW 2006, Australia
19. Dylan Cliff, PhD, University of Wollongong, School of Education, Northfields Ave,
Wollongong NSW 2522, Australia
20. Gregory Kolt, PhD, Western Sydney University, School of Health Sciences, Locked Bag
1797 Penrith NSW 2751, Australia
21. Jennifer Gore, PhD, University of Newcastle, School of Education, University Dr,
Callaghan NSW 2308, Australia
22. Lan Gao, PhD, Deakin University, School of Health and Social Development, 221
Burwood Highway, Burwood, Victoria 3125, Australia
23. James Boyer, BEd, New South Wales Department of Education, 14-22 Loftus Street
Turrella, NSW 2205, Australia
3
24. Ross Morrison, BAppSci, New South Wales Department of Education, 14-22 Loftus
Street Turrella, NSW 2205, Australia
25. Charles Hillman, PhD, Northeastern University, Bouve College of Health Sciences, 360
Huntington Ave, Boston, MA 02115, United States
26. Tatsu Shigeta, MSc, Northeastern University, Bouve College of Health Sciences, 360
Huntington Ave, Boston, MA 02115, United States
27. Elise Tan, MHealthEc, Deakin University, Global Obesity Centre, 221 Burwood
Highway, Burwood, Victoria 3125, Australia
28. David R Lubans, PhD, University of Newcastle, Priority Research Centre for Physical
Activity and Nutrition, University Dr, Callaghan NSW 2308, Australia
Corresponding Author:
Chris Lonsdale, PhD, Australian Catholic University, 33 Berry Street, North Sydney, NSW
2060, Australia, chris.lonsdale@acu.edu.au, Phone: +61 435 087 411.
Date of revision: January 28, 2021
Word Count: 3,149
4
Key Points
Question: Can an internet-based intervention for teachers improve children’s cardiorespiratory
fitness when delivered across a large number of schools?
Findings: This cluster randomized controlled trial showed that the intervention significantly
improved children’s cardiorespiratory fitness at 12 months and effects strengthened at 24
months.
Meaning: The iPLAY intervention has potential for scale-up in order to benefit children’s health
at a population level.
5
Abstract
Importance: Cardiorespiratory fitness is an important marker of childhood health and a risk
factor for disease later in life. Yet, children's fitness has declined in recent decades. There is
limited evidence that school-based physical activity interventions can be effective at the
population level.
Objective: We delivered an internet-based intervention for teachers across a large number of
schools. Our objective was to test the intervention’s effectiveness on children’s cardiorespiratory
fitness.
Design: From 137 elementary schools providing consent, we assigned 22 to a cluster randomized
controlled trial. In the other schools, we delivered the iPLAY intervention, but did not assess the
primary outcome. Recruitment and baseline testing began in 2016 and ended in 2017. Research
assistants, blinded to treatment allocation, completed follow-up outcome assessments at 12 and
24 months.
Setting: Government-funded elementary schools in New South Wales, Australia.
Participants: Grades 3 and 4 students from a stratified sample of elementary schools designed to
be representative and matched on socioeconomic status and location in New South Wales. Loss
to follow-up was 16% of children (0% schools).
6
Interventions: The iPLAY intervention included standardized, online learning for teachers and
minimal in-person support from a project mentor (9-10 months).
Main Outcome and Measure: Cardiorespiratory fitness (20m shuttle test).
Results: From 1,219 participants, we obtained baseline primary outcome data from 1,188
students (49% girls, mean age = 8.85 years, SD = 0.71 years). At 12 months, 20m shuttles
increased by 3.32 laps [2.44, 4.2] in iPLAY schools and 2.11 laps [1.38, 2.85] in controls
(adjusted difference = 1.2 laps [0.17, 2.24]). By 24 months, the adjusted difference was 2.2 laps
(0.89, 3.55). Cost per student was AUD33 (USD26).
Conclusions and Relevance: This appears to be the first trial to show that a school-based
intervention can benefit children's cardiorespiratory fitness when delivered in a large number of
schools. The intervention’s low cost and sustained effect over 24 months suggests that iPLAY has
potential for scale-up.
Registration
Australia New Zealand Clinical Trials Registry, ACTRN12616000731493,
https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=370583&isReview=true
7
Introduction
Children's cardiorespiratory fitness has declined globally in recent decades 1. Cardiorespiratory
fitness is an important marker of childhood health and a risk factor for disease later in life 2,3.
Meta-analyses of efficacy studies show that school-based physical activity interventions can
improve children’s cardiorespiratory fitness 4–6. However, when potentially scalable versions of
school-based interventions have been tested in randomized controlled trials (RCTs), they have
not improved children’s cardiorespiratory fitness 7,8. There is, therefore, limited evidence that
school-based interventions can benefit children’s cardiorespiratory fitness at the population level
9.
A promising elementary school-based intervention was the Supporting Children’s Outcomes
using Rewards, Exercise, and Skills (SCORES) program 10. A cluster RCT efficacy study
showed that SCORES significantly improved children’s cardiorespiratory fitness. SCORES,
however, relied on university-based researchers to deliver training to teachers, which limits
scalability. Drawing on principles from the Consolidated Framework for Implementation
Research 11 (see eFigure 1), we adapted SCORES so that standardized intervention content could
be delivered via an online platform with minimal in-person support from experienced teachers
employed by the project and no direct contact between schools and the research team. The
adapted intervention is known as “internet-based Professional Learning to help teachers promote
Activity in Youth” (iPLAY).
We tested the hypothesis that, compared with controls, students from schools randomized to the
iPLAY intervention would show greater improvements in cardiorespiratory fitness 12 months
8
after baseline (primary endpoint). Our secondary hypotheses were that: i. positive changes in
cardiorespiratory fitness would be maintained at 24 months; ii. iPLAY would have positive
effects on secondary outcomes that are potential determinants of cardiorespiratory fitness
improvement (e.g., physical activity, fundamental movement skills 12) and potential outcomes of
this primary outcome (e.g., academic achievement 13); and iii. iPLAY costs would compare
favorably with other school-based interventions.
We also explored potential mediating pathways that could explain intervention effects (e.g.,
improvements in fundamental movement skills explained the effects of the SCORES
intervention on physical activity and cardiorespiratory fitness 12). Finally, to investigate if
iPLAY was equally effective for all subpopulations, we tested if demographic (e.g., age, sex)
or baseline characteristics (cardiorespiratory fitness, physical activity, or fundamental
movement skill competence) moderated the effects on cardiorespiratory fitness14.
Method
Trial design and participants
From 137 schools providing consent, we assigned 22 to a cluster RCT (see Figure 1). We offered
iPLAY to the other 115 schools (108 accepted), but did not assess the primary outcome in those
schools. We did, however, assess teachers’ adoption and implementation fidelity data in all
schools. In the cluster RCT, we conducted baseline assessments and then allocated schools (1:1)
to the intervention or an attention control condition. We completed follow-up assessments at 12
months (post-intervention) and 24 months. We prospectively registered the trial
Lost to follow-up at 12 months
Schools (n=0)
Students (n=117)
Consent withdrawn (n=2)
Left school (n=77)
Absent during assessments (n=38)
Students assessed at 12 months (n=459)
Students assessed on primary outcome at 12
months (n=440)
Students included in the primary outcome analysis at
12 months (n=555)
Assigned to receive intervention outside of
cluster RCT (n=115 schools offered, n=108
accepted). Primary outcome not assessed.
Lost to follow-up at 24 months
Schools (n=0)
Students (n=67)
Consent withdrawn (n=0)
Left school (n=47)
Absent during assessments (n=20)
Students assessed at 24 months (n=514)
Students assessed on primary outcome at 24
months (n=489)
Students included in the primary outcome analysis at
24 months (n=631)
Assigned to cluster RCT
(n=22 schools)
Enrollment
Figure 1. CONSORT flow diagram indicating participant flow throughout the procedure
Lost to follow-up at 24 months
Schools (n=0)
Students (n=89)
Consent withdrawn (n=0)
Left school (n=66)
Absent during assessments (n=23)
Students assessed at 24 months (n=408)
Students assessed on primary outcome at 24
months (n=375)
Students included in the primary outcome analysis at
24 months (n=555)
Lost to follow-up at 12 months
Schools (n=0)
Students (n=81)
Consent withdrawn (n=0)
Left school (n=62)
Absent during assessments (n=19)
Students assessed at 12 months (n=562)
Students assessed on primary outcome at 12
months (n=553)
Students included in the primary outcome analysis at
12 months (n=631)
Allocated to intervention (n=11 schools)
Received intervention (n=11 schools)
Students consented (n=576)
Median students/school
(n=36; range=11-147)
Students assessed at baseline (n=576)
Students assessed on primary outcome at
baseline (n=557)
Allocated to control (n=11 schools)
Students consented (n=643)
Median students/school
(n=39; range=6-188)
Students assessed at baseline (n=643)
Students assessed on primary outcome at
baseline (n=631)
24 months
12 months
Allocation
Schools ineligible for cluster RCT
(n=108 schools)
Schools for Specific Purposes
(n=100 schools)
Schools from original SCORES
efficacy trial (n=8 schools)
Providing consent
(n=137 schools)
New South Wales government-
funded primary schools
(n=1772 schools)
9
(ACTRN12616000731493) and published our protocol 15. The New South Wales Department of
Education and our university review boards provided ethical approval.
Government-funded elementary schools (Kindergarten–Grade 6) in New South Wales, Australia
were eligible to participate. Schools focused on students with special needs and schools from the
original SCORES efficacy study 10 were not eligible. To recruit schools, we presented at
principals’ meetings, sent emails to schools, and posted on social media.
We invited all teachers in participating schools to join the study with written consent. We invited
students in Grades 3 and 4 (via written parental consent) to participate in outcome assessments.
We collected data at the schools where the students were enrolled.
Interventions
Grounded in the Centers for Disease Control and Prevention’s Comprehensive School Physical
Activity Program framework 16, iPLAY included six components to promote physical activity
participation, enhance student motivation towards physical activity, and develop fundamental
movement skill competence (see eFigure 2). Three components focused on curricular strategies:
1. Quality physical education (delivered according to the SAAFE principles 17), 2. Classroom
energizers (i.e., brief physical activity breaks), and 3. Active homework in which students
completed academic tasks while being physically active. Three further components targeted non-
curricular strategies: 4. Active playgrounds during recess and lunch breaks (e.g., providing
physical activity equipment and implementing playground policies to encourage physical
activity, such as “No hat, play in the shade” rather than “No hat, no play”), 5. Parent engagement
via newsletters (e.g., ideas for active gifts at Christmas) and active school fundraising events, and
10
6. Community physical activity links (e.g., helping schools access funding for after-school
physical activity programs delivered by external agencies).
An iPLAY ‘mentor’ (i.e., an experienced physical education teacher trained by the research team)
facilitated a 2h group workshop for all teachers at each school using standardized multimedia
content hosted on the iPLAY online platform. Mentors also visited schools to provide teachers
with 1h of individualized implementation mentoring. The online platform provided teachers with
4h of additional learning modules to complete independently (see eFigure 3).
Principals in each school identified up to three teachers as iPLAY ‘leaders’ who implemented the
non-curricular intervention components. Leaders completed four additional online modules to
learn how to implement the non-curricular components. Leaders in each school then met with
their school’s mentor to set implementation goals. Subsequently, leaders met with mentors once
per term to review progress, set new goals, and discuss how to support teachers to implement the
curricular components.
The online platform not only delivered content to mentors, leaders, and teachers, but also
provided adoption and implementation feedback. Mentors could oversee leaders’ and teachers’
adoption and implementation and provide support (e.g., answering questions via phone). iPLAY
leaders could also monitor their teachers’ online learning module completion and offer assistance
(e.g., solving implementation barriers).
11
Mentors completed all in-person delivery within 12 months. Teachers and leaders could,
however, continue to access the website after this period in order to complete online training or
download resources.
For schools allocated to the attention control arm we provided professional learning unrelated to
physical activity (e.g., science curriculum).
Outcomes
Trained research assistants assessed students’ cardiorespiratory fitness (primary outcome) by
recording laps completed in the 20m multistage fitness test 18. This test has demonstrated strong
criterion-related validity in this population and is the most widely used field-based measure 2.
See our protocol 15 for secondary outcome details. Briefly, we used wrist-worn accelerometers 19
to measure students’ moderate-to-vigorous physical activity across a week: (i) within school time
(as per school’s schedule), (ii) during lunch and recess (per schedule), (iii) after-school, (iv) on
weekends, and (v) overall. Students self-reported their wellbeing 20 and experiences during
physical education lessons 21–26. We linked student data with their standardized academic test
scores 27 and measured height and weight to calculate body mass index 28. Within each school,
we randomly selected one class at baseline from which to measure students’ fundamental
movement skills 29 and cognitive function 30.
12
Sample size
Based on the SCORES efficacy trial 10, we expected intraclass correlations of 0.01 (school) and
0.09 (class), with d = 0.35 for the between-arm difference. Our power analysis 31 indicated that
1,080 students from 60 classes in 20 schools would provide power of 0.91.
Recruitment, assignment, randomization and blinding
We recruited schools across three cohorts. Using procedures described in our protocol 15, we
assigned 22 of the 137 schools to the cluster RCT using a blocked randomization process
designed to ensure that schools in the trial were: (i) broadly representative of government schools
in New South Wales, and (ii) allocated to trial arms such that most school-level covariates were
balanced.
We planned to assign 20 schools to the cluster RCT; however, our recruitment in the first two
cohorts oversampled students from low socioeconomic (SES) areas–69.9% of students from low
SES schools. Thus, we recruited an additional pair of high SES schools in the final cohort.
Following baseline assessments, an independent statistician, randomized schools using an
algorithm in the R environment 32.
Blinded research assistants collected all student-level outcomes in the cluster RCT. We evaluated
the extent to which this blinding was successful (see eTable 1). Students and teachers were not
blinded to allocation.
13
Statistical methods
Statisticians, blinded to schools’ allocation, tested for between-arm differences in changes in
student outcomes using mixed effects models with random effects for student, teacher, and
school to account for clustering. The primary outcome was a count variable (20m shuttle test
laps) and followed a Poisson distribution; thus, we used a Poisson link function. For secondary
outcomes we ran mixed-effects models with a Gaussian link function. We ran all models in R
using Markov Chain Monte Carlo estimation 33. We analyzed according to intention-to-treat
principles, meaning we used all available data.
We explored pre-registered demographic moderators of intervention effects 15, including self-
reported age, sex, ethnicity, weight status, and family SES (i.e., a principal component of wealth
status and books in the home variables). Pre-registered moderators also included baseline levels
of cardiorespiratory fitness, physical activity, and fundamental movement skill competence.
In our per-protocol analyses we investigated the moderating influence of iPLAY leaders’ and
teachers’ intervention adoption and implementation fidelity on the primary outcome (students’
cardiorespiratory fitness). Adoption focused on the proportion of intervention training
components teachers and leaders completed (e.g., online learning modules). Implementation
fidelity measures focused on the proportion of strategies utilized. See eTables 2-4.
To examine potential mediating processes, we employed mixed-model analyses via a causal
mediation approach.
14
We assessed the cost-effectiveness of the iPLAY intervention using prospectively collected data.
We costed resources related to iPLAY intervention delivery, including staff salaries, equipment
34, and consumables. We valued costs in 2018 Australian dollars (and provide US dollar
equivalents). We calculated the incremental cost-effectiveness ratio as the ratio of the difference
in cost and the difference in the primary outcome per student.
Results
We recruited and collected baseline data for Cohort 1 from July 22-August 25 2016, Cohort 2
from May 15-June 22 2017, and Cohort 3 from November 27-December 7 2017. From the 22
RCT schools, we recruited 1,219 students into the trial (70.0% of enrolled Grade 3 and 4
students). Parents of two students withdrew their consent, without providing reasons for their
decisions. Of the 1,217 participants, 1,188 students completed the primary outcome assessment.
Table 1 shows baseline characteristics.
Our main intention-to-treat analyses found that, compared with controls, students in the iPLAY
intervention schools had greater increases from baseline in cardiorespiratory fitness at 12 and 24
months (see Table 2). The mean adjusted difference at 12 months was 1.2 laps [0.17, 2.24] on the
20m shuttle test. By 24 months, the mean adjusted difference was 2.2 laps [0.89, 3.55]. Blinding
of research assistants to schools’ allocation was largely successful. In four instances, research
assistants became unblinded (see eTable 1).
Table 3 provides full moderator results. eFigure 4 depicts significant effects. At 12 and 24
months, boys showed greater benefit from the intervention compared with girls. At 24 months,
students from lower SES families, younger students, and students with better fundamental
15
movement skills at baseline received more benefit than students from wealthier families, older
students, and students with poorer skills.
eTables 3 and 4 show the proportion of teachers and leaders in the 11 RCT schools allocated to
the iPLAY intervention who adopted and/or implemented each aspect of the intervention at 12
and 24 months. It is beyond the scope of this paper to report the full details from the schools that
completed iPLAY but were not included in the cluster RCT. Briefly, adoption and
implementation fidelity across the cluster RCT intervention schools and 108 non-RCT schools
were similar. For example, in the RCT, 63% of teachers completed all 8 online learning modules
by 24 months. In the non-RCT schools, the completion rate for these modules was 62%.
As shown in eTable 5, per-protocol-analysis in the RCT suggested that students whose teachers
and leaders adopted and implemented iPLAY as per protocol likely received greater
cardiorespiratory fitness benefit at 12 and 24 months than students whose schools did not adopt
the intervention. All these adjusted differences in the per-protocol-analyses were larger than the
main intention-to-treat analyses; however, due to the smaller effective sample sizes, the per-
protocol effects were not reliably different from zero.
As shown in eTable 6-9, secondary outcomes analyses showed significant intervention effects on
physical activity (by accelerometry) during school lunch and recess breaks at 12 and 24 months
(i.e., compared with intervention schools, controls showed greater declines in physical activity
over time 35). However, the intervention did not produce significant effects on physical activity at
other times (during school day, after school, weekend) or overall. There were also no significant
16
effects on self-reported physical activity, wellbeing, academic achievement, or BMI. Within
physical education lessons, iPLAY significantly enhanced students’ perceived support from their
teachers at 24 months, but did not influence other outcomes. Data from the subsample of
students who completed fundamental movement skill and cognitive function assessments showed
no intervention effects on these outcomes.
Students’ moderate-to-vigorous physical activity during lunch/recess partially mediated the
effect of the intervention on students’ cardiorespiratory fitness. At 12 months, the log-odds for
the intervention effect was 0.06 [0.02, 0.09] while the indirect effect was 0.02 [0.01, 0.04]. At 24
months, the log odds for the intervention effect was 0.10 [0.06, 0.13], while the indirect effect
was 0.02 [0.01, 0.03].
The cost of delivering the iPLAY intervention across 11 schools in the cluster RCT was
AUD88,713 (USD69,515), with an average cost of AUD33/student (USD26/student) across the
2,702 from all grades in the schools. At 12 months, the incremental cost-effectiveness ratio was
AUD27 (USD21) for each additional lap achieved. At 24 months, this ratio was AUD15/lap
(USD12/lap). Bootstrapping sensitivity analyses produced similar results (see eFigure 5).
No participants reported any harm.
Discussion
We found that an internet-based professional learning intervention for elementary school
teachers improved students’ cardiorespiratory fitness at 12 months. Effects continued to grow
even after we withdrew implementation support from mentors, with 24-month benefits nearly
17
double the 12-month effect. Meta-analysis of efficacy studies showed that school-based physical
activity interventions can improve students’ cardiorespiratory fitness 4. However, previously
successful school-based interventions were evaluated in small-scale trials with little evidence of
scalability. The few randomized trials of school-based interventions delivered to a large number
of schools have not improved children’s cardiorespiratory fitness 7,8. Thus, our cluster RCT
appears to be the first to show a school-based intervention can improve children’s
cardiorespiratory fitness when delivered across a large number of schools (i.e., 11 schools from
RCT and 108 other schools).
iPLAY is novel because it relies primarily on an internet-based delivery model with minimal in-
person support from experienced teachers (known as iPLAY mentors). Delivering via the internet
allowed us to standardize the learning content, which likely limited ‘program drift’ (i.e.,
deviations from planned intervention protocol) and ‘voltage drop’ (i.e., reduced intervention
benefits) 36. Internet-based delivery also allowed teachers to learn according to a flexible
schedule, which was important given teachers’ busy schedules. Internet-based delivery also
allowed teachers to learn in a distributed fashion over 9-10 months, rather than relying on a small
number of face-to-face workshops (e.g., 1-1.5 days of workshops per year 8).
Cardiorespiratory fitness is an important indicator of health in youth and a predictor of disease
later in life 2,3. Over the past three decades, children’s cardiorespiratory fitness levels have
decreased in many countries. In Australia, for example, children’s cardiorespiratory fitness
decreased by an average of 0.17mL/kg/min per year over this period (estimated VO2PEAK) 1.
Converting our trial participants’ 20m shuttle laps into estimated peak oxygen consumption
18
(VO2PEAK) 37 showed that iPLAY improved cardiorespiratory fitness by 0.45 mL/kg/min (0.06,
0.84) at 24 months. As such, iPLAY could be seen to reverse ~3 years of the decline observed in
Australian population levels of children’s fitness in recent decades. If an intervention like iPLAY
can be implemented effectively across the population, the effect on cardiorespiratory fitness and
corresponding health benefits could be sizable for society 38.
Optimism is also warranted because iPLAY’s cost (AUD33/student = USD26/student) falls well
below the values reported for other school-based physical activity programs (e.g., Child and
Adolescent Trial for Cardiovascular Health intervention 39 = USD104/student [2004] =
AUD198/student [2018]). The Education Endowment Foundation (UK) 40 considers
interventions with a cost per student below 80GBP (USD107) per year to be very low cost.
Despite these promising findings, iPLAY was not universally successful. For example,
intervention effects on device-measured physical activity were limited to recess and lunch time
breaks. This pattern of significant cardiorespiratory fitness effects alongside limited device-
measured physical activity effects is, however, consistent with meta-analyses of efficacy studies
4,6,41. It is possible that iPLAY (and other school-based interventions that improved
cardiorespiratory fitness but not device-measured physical activity) actually did increase
students’ overall physical activity; however, the effects may not have been detectable because
accelerometers did not adequately capture some physical activities. Possible sources of error
include participation in organized sports (where students often remove accelerometers), such as
those promoted in the community physical activity component of the iPLAY intervention.
19
Relative to the controls, iPLAY increased intervention students’ cardiorespiratory fitness over the
study period. However, enthusiasm for the effectiveness of the intervention should be tempered
due to notable differences in CRF between groups at baseline. Intervention students completed
nearly three laps fewer than control students–meaning that iPLAY served to narrow this pre-
existing gap. Further caution is warranted because our moderator analyses showed intervention
effects were not equivalent for all students. A recent meta-analysis of school-based efficacy trials
reported similar findings, with smaller intervention effects for girls and older children, than boys
and younger children 6. As a result, we suggest that targeted online teacher professional learning
needs to be developed and tested to ensure that all students benefit from this type of intervention.
Further research is also needed to determine the extent to which our results generalize to primary
schools outside New South Wales, Australia.
The trial’s strengths include stratified sampling procedures that helped ensure the results are
generalizable to the population. We recruited a high proportion of enrolled students (70%)
compared with other high quality school-based trials (e.g., 60% 42). We retained 84% of the
baseline sample on our primary outcome assessment at 12 months, which is comparable to other
quality school-based trials (e.g., 82-89% at 9-30 months 42,43). Finally, we included a longer
follow-up period (24 months) than many other school-based interventions targeting
cardiorespiratory fitness (e.g., mean = 13 months in a recent meta-analysis 4).
Limitations
Our trial did not have sufficient sample size at the cluster level to examine the influence of
between-school moderators (e.g., school size and location) on intervention effects. We were
20
similarly underpowered to detect effects on some secondary outcomes (e.g., fundamental
movement skills).
Conclusions
This appears to be the first trial to show that a school-based intervention can benefit children’s
cardiorespiratory fitness when delivered across a large number of schools. Our internet-based
intervention for teachers improved children’s cardiorespiratory fitness at 12 months. The
intervention also built capacity in schools such that these benefits continued to grow for at least
another year after we withdrew in-person support. This evidence, alongside the intervention’s
low cost, suggests that iPLAY has potential for scale-up in order to benefit children’s health at a
population level.
Acknowledgements
We thank staff from the New South Wales branch of the Australian Council for Physical
Education and Recreation for helping recruit mentors. We also thank the New South Wales
Office of Sport for providing data regarding local sport and recreation clubs in each school’s
local area. We thank staff from the New South Wales Department of Education for their input in
the design phase of the intervention. Funding for this study came from the National Health and
Medical Research Council (Australia) Partnership Project Grant (Ref: APP1114281) and the
New South Wales Department of Education. Jo Salmon is funded by an NHMRC Leadership
Level 2 Fellowship (Ref: APP1176885). David R Lubans is funded by an NHMRC Senior
Research Fellowship (Ref: APP1154507). The National Health and Medical Research Council
played no role in the design and conduct of the study; collection, management, analysis, or
interpretation of the data; preparation, review, or approval of the manuscript; or decision to
21
submit the manuscript for publication. Staff from the other funding agency, the New South
Wales Department of Education, played an important role in designing the intervention (i.e., as
part of the design and conduct of the study). No staff from the New South Wales Department of
Education played a role in collection, management, analysis, or interpretation of the data. Two
staff from the Department (James Boyer and Ross Morrison) reviewed and approved the
manuscript; they also agreed to submit the manuscript for publication. These two staff have been
named as co-authors. The corresponding author had full access to all the data in the study and
takes responsibility for the integrity of the data and the accuracy of the data analysis. Philip
Parker, Marj Moodie, Lan Gao, Elise Tan, and Tatsu Shigeta designed and/or conducted the
analysis. The corresponding author had final responsibility for the decision to submit this
manuscript for publication.
22
References
1. Tomkinson GR, Lang JJ, Tremblay MS. Temporal trends in the cardiorespiratory fitness of
children and adolescents representing 19 high-income and upper middle-income countries
between 1981 and 2014. Br J Sports Med. 2019;53(8):478-486.
2. Raghuveer G, Hartz J, Lubans DR, et al. Cardiorespiratory Fitness in Youth: An Important
Marker of Health: A Scientific Statement From the American Heart Association.
Circulation. Published online July 20, 2020:CIR0000000000000866.
3. García-Hermoso A, Ramírez-Vélez R, García-Alonso Y, Alonso-Martínez AM, Izquierdo
M. Association of Cardiorespiratory Fitness Levels During Youth With Health Risk Later in
Life: A Systematic Review and Meta-analysis. JAMA Pediatr. Published online August 31,
2020. doi:10.1001/jamapediatrics.2020.2400
4. Pozuelo-Carrascosa DP, García-Hermoso A, Álvarez-Bueno C, Sánchez-López M,
Martinez-Vizcaino V. Effectiveness of school-based physical activity programmes on
cardiorespiratory fitness in children: a meta-analysis of randomised controlled trials. Br J
Sports Med. 2018;52(19):1234-1240.
5. García-Hermoso A, Alonso-Martínez AM, Ramírez-Vélez R, Pérez-Sousa MÁ, Ramírez-
Campillo R, Izquierdo M. Association of Physical Education With Improvement of Health-
Related Physical Fitness Outcomes and Fundamental Motor Skills Among Youths: A
Systematic Review and Meta-analysis. JAMA Pediatr. Published online April 6,
2020:e200223.
6. Hartwig TB, Sanders T, Vaconcellos D, et al. School-based interventions modestly increase
physical activity and cardiorespiratory fitness but are least effective for youth who need
them most: an individual participant pooled analysis of 20 controlled trials. Br J Sports
Med. Published online January 13, 2021. doi:10.1136/bjsports-2020-102740
7. Stavnsbo M, Aadland E, Anderssen SA, et al. Effects of the Active Smarter Kids (ASK)
physical activity intervention on cardiometabolic risk factors in children: A cluster-
randomized controlled trial. Prev Med. 2020;130:105868.
8. Luepker RV, Perry CL, McKinlay SM, et al. Outcomes of a field trial to improve children’s
dietary patterns and physical activity. The Child and Adolescent Trial for Cardiovascular
Health. CATCH collaborative group. JAMA. 1996;275(10):768-776.
9. Reis RS, Salvo D, Ogilvie D, et al. Scaling up physical activity interventions worldwide:
stepping up to larger and smarter approaches to get people moving. Lancet.
2016;388(10051):1337-1348.
10. Cohen KE, Morgan PJ, Plotnikoff RC, Callister R, Lubans DR. Physical activity and skills
intervention: SCORES cluster randomized controlled trial. Med Sci Sports Exerc.
2015;47(4):765-774.
23
11. Damschroder LJ, Aron DC, Keith RE, Kirsh SR, Alexander JA, Lowery JC. Fostering
implementation of health services research findings into practice: a consolidated framework
for advancing implementation science. Implement Sci. 2009;4:50.
12. Cohen K, Morgan P, Plotnikoff R, Barnett L, Lubans D. Improvements in fundamental
movement skill competency mediate the effect of the SCORES intervention on physical
activity and cardiorespiratory fitness in children. J Sci Med Sport. 2015;19:e90.
13. Álvarez-Bueno C, Hillman CH, Cavero-Redondo I, Sánchez-López M, Pozuelo-Carrascosa
DP, Martínez-Vizcaíno V. Aerobic fitness and academic achievement: A systematic review
and meta-analysis. J Sports Sci. 2020;38(5):582-589.
14. Yildirim M, van Stralen MM, Chinapaw MJM, et al. For whom and under what
circumstances do school-based energy balance behavior interventions work? Systematic
review on moderators. Int J Pediatr Obes. 2011;6(2-2):e46-57.
15. Lonsdale C, Sanders T, Cohen KE, et al. Scaling-up an efficacious school-based physical
activity intervention: Study protocol for the “Internet-based Professional Learning to help
teachers support Activity in Youth” (iPLAY) cluster randomized controlled trial and scale-
up implementation evaluation. BMC Public Health. 2016;16(1). doi:10.1186/s12889-016-
3243-2
16. Centers for Disease Control and Prevention (CDC). School health guidelines to promote
healthy eating and physical activity. MMWR Recomm Rep. 2011;60(RR-5):1-76.
17. Lubans DR, Lonsdale C, Cohen K, et al. Framework for the design and delivery of
organized physical activity sessions for children and adolescents: Rationale and description
of the “SAAFE” teaching principles. Int J Behav Nutr Phys Act. 2017;14(1).
doi:10.1186/s12966-017-0479-x
18. The Cooper Institute. Fitnessgram and Activitygram Test Administration Manual-Updated
4th Edition. Human Kinetics; 2010.
19. Hildebrand M, VAN Hees VT, Hansen BH, Ekelund U. Age group comparability of raw
accelerometer output from wrist- and hip-worn monitors. Med Sci Sports Exerc.
2014;46(9):1816-1824.
20. Roberts C, the International HBSC Study Group, Freeman J, et al. The Health Behaviour in
School-aged Children (HBSC) study: methodological developments and current tensions.
International Journal of Public Health. 2009;54(S2):140-150. doi:10.1007/s00038-009-
5405-9
21. Vlachopoulos SP, Katartzi ES, Kontou MG. Fitting Multidimensional Amotivation into the
Self-Determination Theory Nomological Network: Application in School Physical
Education. Meas Phys Educ Exerc Sci. 2013;17(1):40-61.
22. Ahn I, Patrick H, Chiu MM, Levesque-Bristol C. Measuring Teacher Practices That Support
Student Motivation: Examining the Factor Structure of the Teacher as Social Context
24
Questionnaire Using Multilevel Factor Analyses. J Psychoeduc Assess. 2019;37(6):743-
756.
23. Lam S-F, Jimerson S, Wong BPH, et al. Understanding and measuring student engagement
in school: the results of an international study from 12 countries. Sch Psychol Q.
2014;29(2):213-232.
24. McAuley E, Duncan T, Tammen VV. Psychometric properties of the Intrinsic Motivation
Inventory in a competitive sport setting: a confirmatory factor analysis. Res Q Exerc Sport.
1989;60(1):48-58.
25. Standage M, Duda JL, Ntoumanis N. A test of self-determination theory in school physical
education. Br J Educ Psychol. 2005;75(3):411-433.
26. Solmon MA, Lee AM. Development of an instrument to assess cognitive processes in
physical education classes. Res Q Exerc Sport. 1997;68(2):152-160.
27. Australian Curriculum Assessment and Reporting Authority. NAPLAN Achievement in
Reading, Persuasive Writing, Language Conventions and Numeracy: National Report for
2014. http://www.nap.edu.au/_resources/
National_Assessment_Program_Literacy_and_Numeracy_national_report_for_ 2014.pdf.
28. Flegal KM, Cole TJ. Construction of LMS parameters for the Centers for Disease Control
and Prevention 2000 growth charts. Natl Health Stat Report. 2013;(63):1-3.
29. Ulrich DT. Test of Gross Motor Development Examiner’s Manual. 2nd ed. Pro Ed; 2000.
30. Paxton JL, Barch DM, Racine CA, Braver TS. Cognitive control, goal maintenance, and
prefrontal function in healthy aging. Cereb Cortex. 2008;18(5):1010-1028.
31. Judd CM, Westfall J, Kenny DA. Experiments with More Than One Random Factor:
Designs, Analytic Models, and Statistical Power. Annu Rev Psychol. 2017;68:601-625.
32. Team RC. R development core team. RA Lang Environ Stat Comput. 2013;55:275-286.
33. Hamra G, MacLehose R, Richardson D. Markov chain Monte Carlo: an introduction for
epidemiologists. Int J Epidemiol. 2013;42(2):627-634.
34. Hartwig TB, Del Pozo-Cruz B, White RL, et al. A monitoring system to provide feedback
on student physical activity during physical education lessons. Scand J Med Sci Sports.
2019;29(9):1305-1312.
35. Farooq MA, Parkinson KN, Adamson AJ, et al. Timing of the decline in physical activity in
childhood and adolescence: Gateshead Millennium Cohort Study. Br J Sports Med.
2018;52(15):1002-1006.
36. Chambers DA, Glasgow RE, Stange KC. The dynamic sustainability framework: addressing
the paradox of sustainment amid ongoing change. Implement Sci. 2013;8:117.
25
37. Léger LA, Mercier D, Gadoury C, Lambert J. The multistage 20 metre shuttle run test for
aerobic fitness. J Sports Sci. 1988;6(2):93-101.
38. Rose G. Sick individuals and sick populations. Int J Epidemiol. 2001;30(3):427-432;
discussion 433-4.
39. Brown HS 3rd, Pérez A, Li Y-P, Hoelscher DM, Kelder SH, Rivera R. The cost-
effectiveness of a school-based overweight program. Int J Behav Nutr Phys Act. 2007;4:47.
40. Sutton Trust-Education Endowment Foundation. Teaching and Learning Toolkit. Technical
Appendix and Process Manual (Working Document v.01).; 2019.
https://educationendowmentfoundation.org.uk/public/files/Toolkit/Toolkit_Manual_2018.p
df
41. Love R, Adams J, Sluijs EMF. Are school‐based physical activity interventions effective
and equitable? A meta‐analysis of cluster randomized controlled trials with accelerometer‐
assessed activity. Obes Rev. 2019;20(6):859-870.
42. Adab P, Pallan MJ, Lancashire ER, et al. Effectiveness of a childhood obesity prevention
programme delivered through schools, targeting 6 and 7 year olds: cluster randomised
controlled trial (WAVES study). BMJ. 2018;360:k211.
43. Kriemler S, Zahner L, Schindler C, et al. Effect of school based physical activity
programme (KISS) on fitness and adiposity in primary schoolchildren: cluster randomised
controlled trial. BMJ. 2010;340:c785.
26
Table 1. Baseline Characteristics
Variable
N
(n intervention)
Control
Mean
Control SD
Intervention
Mean
Intervention
SD
Total
Mean
Total SD
Student level variables
Cardiorespiratory fitness (20 m shuttle laps)
1188 (557)
25.83
14.95
22.84
13.43
24.43
14.33
Girls
1209 (569)
49.06%
NA
49.74%
NA
49.38%
NA
Age (years)
1190 (554)
8.89
0.70
8.81
0.73
8.85
0.71
BMI (kg/m2)
1186 (552)
18.14
3.09
18.23
3.58
18.18
3.32
BMI-z
1174 (547)
0.44
1.02
0.42
1.11
0.43
1.06
Obese/Overweight status
1174 (547)
27.11%
NA
28.88%
NA
27.94%
NA
Immigrant status
1208 (568)
15.62%
NA
26.58%
NA
20.78%
NA
Indigenous [Aboriginal/Torres Strait
Islander]
1206 (569)
9.06%
NA
8.61%
NA
8.85%
NA
Wealth status
1196 (557)
2.45
0.84
2.45
0.84
2.45
0.84
Books in home
1196 (557)
3.05
1.14
2.99
1.23
3.02
1.18
School level variables
Index of Community Socio-Educational
Advantage score
22 (11)
986.3
58.3
1006
56
996.3
56.7
School size
(students in Kindergarten-Year 6)
22 (11)
294.18
266.9
253.63
190.49
273.91
227.23
Previous participation in the Live Life Well
at School Program
22 (11)
90.91%
NA
100%
NA
95.45%
NA
NAPLAN Numeracy school mean score
22 (11)
445.95
54.48
436.91
49.23
441.42
50.5
NAPLAN Literacy school mean score
22 (11)
452.73
54.62
447.86
44.94
450.3
49.5
27
Urban location
22 (11)
63.64%
NA
72.73%
NA
68.18%
NA
Length of school day (mins)
22 (11)
361.31
2.24
358.2
4.61
359.75
3.88
Time allocated to recess/lunch breaks per
day (mins)
22 (11)
73.70
5.58
68.72
8.19
71.21
7.21
Schools with a dedicated physical education
teacher (%)
20 (10)
30.00%
NA
0.00%
NA
15.00%
NA
Note: BMI = body mass index; NAPLAN = National Assessment Program – Literacy and Numeracy. NAPLAN Literacy = combination of
reading, writing, spelling and grammar scores. PE = Physical Education. Live Life Well at School was a state-wide physical activity and
nutrition program delivered from 2008 to 2015. Wealth status measured on scale from 1 ("very wealthy") to 5 ("not at all wealthy"). Books in
the home measured on scale from 1 ("none or very few (0-10)") to 5 ("enough to fill 3 or more bookcases (more than 200 Books)").
28
Table 2. Primary Outcome Analysis (Cardiorespiratory Fitness as measured by Laps on 20m Shuttle Test)
Model
Follow-up
N
(n intervention)
Change from
Baseline:
Control
Change from
Baseline:
Intervention
Adjusted difference
in 20m shuttle test
laps (intervention v
control)
Main Analysis: Controlling for schools'
remoteness and Index of Community
Socio-Educational Advantage score
12 Months
993 (440)
2.11 [1.38, 2.85]
3.32 [2.44, 4.2]
1.2 [0.17, 2.24]
24 Months
864 (375)
4.42 [3.41, 5.43]
6.64 [5.3, 7.98]
2.22 [0.89, 3.55]
Sensitivity Analysis 1: Model without
controlling for schools' remoteness or
Index of Community Socio-Educational
Advantage score
12 Months
993 (440)
1.74 [1.17, 2.31]
2.75 [2.1, 3.4]
1.01 [0.14, 1.88]
24 Months
864 (375)
3.63 [2.92, 4.34]
5.5 [4.61, 6.4]
1.87 [0.73, 3.02]
Sensitivity Analysis 2a: Model
controlling for schools' urban status and
Index of Community Socio-Educational
Advantage score and imputing missing
data
12 Months
1021 (459)
2.01 [1.27, 2.76]
3.36 [2.46, 4.26]
1.35 [0.27, 2.43]
24 Months
970 (408)
4.56 [3.53, 5.59]
7.22 [5.8, 8.65]
2.66 [1.23, 4.08]
Note: Main analysis = mixed effect model with a Poisson link function. ICC student = 0.86, ICC teacher = 0.06, ICC School = 0.04.
Sensitivity Analyses 2a,b included imputation for students who were absent during a primary outcome assessment, but provided other
assessment data at that time point. The median absolute deviation in shuttle runs for the control group was 8 laps (at baseline). This
means that the intervention effect size was equivalent to 15% and 28% of the median absolute deviation at 12 and 24 months,
respectively.
29
Table 3. Moderator Effects on Primary Outcome (Cardiorespiratory Fitness as measured by Laps on 20m Shuttle Test)
Moderator
Follow-up
Estimate
-90% CI
+90% CI
Baseline cardiorespiratory fitness
12 Month
0.0471
-0.0187
0.1151
24 Month
-0.01
-0.0659
0.0459
Socioeconomic status
12 Month
0.0069
-0.0274
0.0426
24 Month
-0.0626
-0.0996
-0.0251
Age
12 Month
-0.0479
-0.0962
0.0028
24 Month
-0.0619
-0.1117
-0.012
Immigrant status
12 Month
0.0457
-0.0799
0.1695
24 Month
0.1092
-0.017
0.2403
Indigenous status
12 Month
-0.0246
-0.119
0.0711
24 Month
0.0751
-0.0202
0.1715
Gender
12 Month
-0.0766
-0.1436
-0.0089
24 Month
-0.1493
-0.221
-0.0794
Obese/overweight status
12 Month
-0.0121
-0.1017
0.0749
24 Month
-0.0902
-0.1797
0.0026
MVPA
12 Month
0.0059
-0.0337
0.0436
24 Month
0.0086
-0.0316
0.0473
Fundamental movement skills
12 Month
0.011
-0.01
0.03
24 Month
0.04
0.02
0.06
Moderator Values
Follow-up
Estimate
-90% CI
+90% CI
Boys
12 months
2.12
0.95
3.42
Girls
12 months
0.02
-1.11
1.14
30
Boys
24 months
4.29
2.78
5.89
Girls
24 months
-0.28
-1.66
1.10
Age 8
24 months
3.78
2.00
5.54
Age 10
24 months
0.15
-1.82
2.10
Low socioeconomic status
24 months
4.44
2.89
6.18
High socioeconomic status
24 months
-0.10
-1.60
1.45
Low fundamental movement skills
24 months
-2.79
-5.22
-0.471
High fundamental movement skills
24 months
4.21
1.76
6.71
Note: As specified in our pre-registered protocol, we explored moderation effects with significance set at p < 0.1. As a result,
90% CI are provided. Where significant moderator effects were found, we have displayed estimates of effects at specific levels
of the moderator. MVPA = Moderate-to-Vigorous Physical Activity. Low socioeconomic status = mean -1 SD. High
socioeconomic status mean +1 SD. Low fundamental movement skills = 1st quartile. High fundamental movement skills = 3rd
quartile.
1
Effectiveness of a scalable school-based intervention to
improve children’s cardiorespiratory fitness: The ‘internet-
based Professional Learning to help teachers promote Activity
in Youth’ (iPLAY) cluster randomized controlled trial
Online Only Supplementary Material
1
Online Only Supplementary Material Contents
eFigure 1. Consolidated Framework for Implementation Research… p.2
eFigure 2. iPLAY Intervention Components…. p.3
eFigure 3. Online platform screenshots… p.4
eFigure 4. Summary of Significant Moderator Effects… p.8
eFigure 5. Sensitivity analysis for incremental cost-effectiveness ratios… p.9
eTable 1. iPLAY Data Collector Blinding Evaluation… p.10
eTable 2. iPLAY intervention components and implementation measures… p.12
eTable 3. Intervention Adoption Rates… p.14
eTable 4. Intervention Implementation Rates… p.15
eTable 5. Per Protocol Analyses of Intervention Effects on Primary Outcome… p.16
eTable 6. Main Sample Secondary Outcomes Baseline Values… p.17
eTable 7. Main Sample Secondary Outcome Analyses… p.18
eTable 8. Sub-sample Variables Baseline Values… p.20
eTable 9. Sub-sample Variables Outcome Analyses… p.21
2
eFigure 1. Consolidated Framework for Implementation Research
3
eFigure 2. iPLAY Intervention Components
4
eFigure 3. Online platform screenshots
eFigure 3a - Overview of teachers’ online learning
5
eFigure 3b - Example teacher module
6
eFigure 3c - Overview of leaders’ online learning
7
eFigure 3d - Example leader module
8
eFigure 4. Summary of Significant Moderator Effects
Main Effect
Gender
Boys
Girls
Main Effect
SES
Age
Gender
FMS
Low SES
High SES
8 Years
10 Years
Boys
Girls
Low FMS
High FMS
12 months 24 months
0.5−5.2−0.0 2.5 5.0 0.5−5.2−0.0 2.5 5.0
20m Shuttle Run Laps
9
eFigure 5. Sensitivity analysis for incremental cost-effectiveness ratios
eFigure S5a: Sensitivity analysis for the incremental cost-effectiveness ratios (ICERs) at 12 months
derived from bootstrapping. The ICER for the base case is represented by the orange plot. Values
are presented in AUD (2018).
eFigure S5b: Sensitivity analysis for the incremental cost-effectiveness ratios (ICERs) at 24 months
derived from bootstrapping. The ICER for the base case is represented by the orange plot. Values
are presented in AUD (2018).
10
eTable 1. iPLAY Data Collector Blinding Evaluation
School
Code
Actual Trial
Arm
Research assistants’
responses to: “Was this
school intervention or
control?” a
If you know the school was intervention or control, how did you find out?
Blinding
Status
1
Intervention
I know it was intervention
Research Assistant A: “iPLAY poster in front office window of school viewed upon entry.”
Unblinded
2
Control
I don’t know
N/A
Blinding
maintained
3
Control
I don’t know
N/A
Blinding
maintained
4
Intervention
I don’t know
N/A
Blinding
maintained
5
Intervention
I don’t know
N/A
Blinding
maintained
6
Control
I don’t know
N/A
Blinding
maintained
7
Intervention
I don’t know
N/A
Blinding
maintained
8
Intervention
I don’t know
N/A
Blinding
maintained
9
Control
I don’t know
N/A
Blinding
maintained
10
Control
I don’t know
N/A
Blinding
maintained
11
Control
I don’t know
N/A
Blinding
maintained
12
Intervention
I don’t know
N/A
Blinding
maintained
13
Intervention
I know it was intervention.
Research Assistant D: “iPLAY materials posted at school and a teacher disclosed that they had
received the intervention.”
Unblinded
14
Control
I don’t know
N/A
Blinding
maintained
15
Control
I don’t know
N/A
Blinding
maintained
16
Control
I don’t know
N/A
Blinding
maintained
17
Intervention
I know it was intervention.
Research Assistant C: “This school had iPLAY resources displayed in the window.”
Unblinded
11
18
Intervention
I don’t know
N/A
Blinding
maintained
19
Intervention
I know it was intervention.
Research Assistant D: “A teacher at the school mentioned that they had received iPLAY training.”
Unblinded
20
Control
I don’t know
N/A
Blinding
maintained
21
Control
I don’t know
N/A
Blinding
maintained
22
Intervention
I don’t know
N/A
Blinding
maintained
Note: a Response options included: 1. I don’t know, 2. I know it was control, 3. I know it was intervention.
12
eTable 2. iPLAY intervention components and implementation measures
Curricular
Components
Intended Implementation
Implementation Measurement
Quality PE and
school sport
● Teachers will deliver 150 minutes
of planned PE or school sport
each week.
● Lessons will be delivered
according to the SAAFE principles
(Supportive, Active, Autonomous,
Fair and Enjoyable).
● Students will spend >40% of
PE/sport lesson time being
physically active (i.e., in MVPA).
● Classroom teachers self-reported
delivery of PE and School Sport at
the start of each online learning
module (i.e., up to 8 times).
● Mentors observed and rated each
teacher’s delivery using the
SAAFE checklist once during the
intervention.
● Monitored using the class activity
tracking system provided to each
school.
Classroom
movement breaks
● Teachers will deliver 2 x 3-minute
classroom energizer activities per
day (30 minutes per week)
● Teachers self-reported at the start
of each online learning module.
Physically active
homework
● Teachers will provide one
physically active homework activity
per week (except in schools that
have a ‘no homework’ policy)
● Teachers self-reported at the start
of each of each online learning
module.
Non-Curricular
Components
Intended Implementation
Implementation Measurement
Active
playgrounds
● Children will spend >40% of
recess and lunch breaks in MVPA.
● Leaders rated via the website their
implementation of active
playground strategies. Ratings
occurred up to three times during
the intervention (during meetings
with mentors).
● Student physical activity during
breaks measured via
accelerometry at each assessment
time-point (baseline, 12 months,
24 months), but was not measured
during the intervention.
13
Community
physical activity
links
● Schools will utilize the ‘Sporting
Schools’ funding to offer after-
school physical activity program at
least once per week across two
school terms.
● During the intervention at least
one teacher in each school will
complete accreditation/training
procedures with a recognized
sporting body that will allow them
to deliver the Sporting Schools’
program in their school.
● Principals reported on all non-
curricular sport and recreation in
each school.
● Teachers reported the sport
accreditation/training they
completed.
Parent and
caregiver
engagement
● Schools will deliver 1 x newsletter
item per fortnight, which will
include a link to the parent portion
of the iPLAY website.
● Schools will deliver 2 x iPLAY
update presentations to parents
per year during existing parent-
teacher events.
● Schools will organize one
physically active school
fundraising event each year.
● Leaders recorded via the
website the frequency of
newsletter distribution and
parent meetings.
● Leaders self-reported school
fundraiser events.
14
eTable 3. Intervention Adoption Rates
Proportion Adopted
Core Learning Components
12 months
24 months
Leader Adoption
Leader online learning - 5 modules
100% schools
100% schools
Leader action planning meetings - 4 modules
36% schools
91% schools
Total leader modules - 9 modules
36% schools
91% schools
Leader adoption as per protocol
(i.e., at least one leader at the school completed all
professional learning modules and attended at least one
action plan meeting)
89% schools
100% schools
Teacher Adoption
Teacher workshop module
100% teachers
100% teachers
Teacher online learning - 8 modules
20% teachers
63% teachers
Teacher school-based reflection - 3 modules
11% teachers
63% teachers
Total teacher modules - 12 modules
10% teachers
61% teachers
Teacher adoption as per protocol (i.e., the teacher completed
at least 50% of the 12 professional learning modules) a
48% teachers
71% teachers
Additional Teacher Learning Components
Downloaded resources
82% teachers
82% teachers
Downloaded mobile app
19% teachers
19% teachers
Used class activity monitoring system b
58% teachers
59% teachers
Used posters, water bottles or lanyards
100% schools
100% schools
Note: a We used 50% as the per protocol threshold for adoption. Adult online learners typically do not complete their
courses. Thus, we deemed a 50% completion rate to be considerable. b As noted in our protocol, we designed the
class activity tracker to upload data so that the research team could access and track each teachers’ usage of the
system. Department of Education firewalls in most schools prevented these uploads and we, therefore, could not
access the data. Instead, we report teachers’ self-reported intentions to use the class activity. Delays with our
software development company meant that the mobile app was not available for download until partway through
Cohort 3’s participation in the trial (August 2018). This delay likely accounts for the low rate of the app’s adoption.
15
eTable 4. Intervention Implementation Rates
Proportion Implemented
Leader Implementation - Non-curricular Components
12 months
24 months
Active playgrounds - leader reports of implementing
recommended strategies
45% schools
45% schools
School mean >40% of total break time spent in MVPA
(accelerometers)
0% schools
0% schools
Sporting Schools funding used (principal report)
73% schools
73% schools
At least one teacher complete accreditation to with a
recognized sporting body (teacher report)
64% schools
64% schools
Parent newsletter distribution (leader report)
45% schools
45% schools
Parent info sessions (leader report)
27% schools
27% schools
One physically active school fundraiser (leader report)
9% schools
9% schools
Leader implementation as per protocol (i.e., the school
implemented at least 50% implemented of non-
curricular strategies)
52% schools
51% schools
Teacher Implementation - Curricular Components
12 months
24 months
150 mins of PE/sport/week (teacher reported - median
across modules)
42% teachers
47% teachers
Mean SAAFE rating > 3.0 rating (mentor rated)
93% teachers
91% teachers
10 classroom energizers per week (mean of teacher
reports)
37% teachers
38% teachers
1 weekly active homework activity (excluding schools
with no homework policy) (mean of teacher reports)
49% teachers
46% teachers
Teacher implementation as per protocol (i.e., teacher
implemented 50% of strategies)
30% teachers
22% teachers
Teacher Evaluations of Workshop
Theoretical portion of the workshop (1-5 rating)
4.6 (0.7)
NA
Practical portion of the workshop (1-5 rating)
4.6 (0.8)
NA
How engaging was your mentor? (1-5 rating)
4.5 (0.8)
NA
Note: As noted in our protocol, we designed a class activity tracker to upload data so that the research team could
access and track each teachers’ usage of the system. One of our implementation criteria was the proportion of
physical education lesson time spent in moderate-to-vigorous physical activity should be greater than 40%.
Department of Education firewalls in most schools prevented uploads from our tracker system and we, therefore,
could not access the data.
16
eTable 5. Per Protocol Analyses of Intervention Effects on Primary Outcome
Variable
Follow-up
Condition
Met protocol
(n students)
Did not meet protocol
(n students)
Adjusted difference in 20m shuttle test laps
between treated vs non-treated
Teachers
Teacher Adoption
12 Months
Intervention
165
294
2.16 [-2.14,6.46]
Control
0
562
24 Months
Intervention
236
172
3.35 [-0.41,7.11]
Control
0
514
Teacher
Implementation
12 Month
Intervention
136
323
2.66 [-2.46,7.78]
Control
0
562
24 Months
Intervention
91
317
9.24 [-0.9,19.38]
Control
0
514
Leaders
Leader Adoption
12 Months
Intervention
409
50
0.88 [-0.86,2.62]
Control
0
562
24 Months
Intervention
408
0
NA. All leaders met protocol.
Control
0
514
Leader
Implementation
12 Month
Intervention
239
220
1.54 [-1.84,4.91]
Control
0
562
24 Months
Intervention
209
199
3.7 [-2.47,9.86]
Control
0
514
Note: Teacher adoption = student's teacher completed at least 50% of modules. Teacher implementation = student's teacher implemented at
least 50% of strategies. Leader adoption = student's school had a leader who completed all 5 learning modules + 50% of action plans.
Leader implementation = student's school implemented at least 50% of strategies. We conducted per-protocol analysis using an instrumental
variable approach with shuttle runs coded as continuous and with the application of cluster-robust standard errors.
17
eTable 6. Main Sample Secondary Outcomes Baseline Values
Variable
N
(n intervention)
Control
Mean
Control SD
Intervention
Mean
Intervention
SD
Total Mean
Total SD
School MVPA (mins/day)
926 (424)
42.53
16.09
41.97
15.29
42.27
15.72
Lunch/Recess Breaks MVPA (mins/day)
926 (424)
18.66
8.52
14.28
6.99
16.65
8.15
After School MVPA (mins/day)
926 (424)
33.17
15.22
35.95
16.82
34.45
16.02
Weekend MVPA (mins/day)
925 (423)
77.74
36.37
76.53
33.42
77.19
35.04
Total MVPA (mins/day)
925 (423)
85.90
28.03
87.30
28.79
86.54
28.37
Self-reported PA usual week (days/week)
1195 (558)
5.03
1.97
4.88
2.03
4.96
2.00
Self-reported PA this week (days/week)
1197 (558)
5.17
1.97
5.19
2.01
5.18
1.99
Self-reported PA team sport participation in the
past year
1197 (558)
70.74%
NA
63.62%
NA
67.42%
NA
Self-reported PA individual sport participation in
the past year
1196 (558)
65.36%
NA
61.47%
NA
63.55%
NA
Self-reported PA active travel (days/week)
1197 (558)
2.27
2.15
2.79
2.18
2.52
2.18
NAPLAN Numeracy
890 (378)
-0.45
0.92
-0.54
0.89
-0.49
0.91
NAPLAN Literacy
889 (378)
-0.39
0.95
-0.46
0.92
-0.42
0.94
Wellbeing
1172 (546)
40.94
5.84
41.10
5.52
41.01
5.69
PE Psychological needs support from teacher
1197 (558)
4.06
0.73
4.14
0.70
4.10
0.72
PE Concentration
1196 (558)
4.48
0.63
4.47
0.60
4.48
0.61
PE Effort
1196 (558)
4.60
0.60
4.63
0.57
4.61
0.59
PE Learning strategy use
1196 (558)
3.95
0.84
3.92
0.87
3.94
0.85
PE Enjoyment
1197 (558)
4.54
0.72
4.61
0.70
4.58
0.71
18
eTable 7. Main Sample Secondary Outcome Analyses
Variable
Follow-up
N
(n intervention)
Change from Baseline:
Control
Change from Baseline:
Intervention
Adjusted Difference
(intervention v control)
Secondary Outcomes: Main Sample
School MVPA (mins/day)
12 Month
753 (332)
-5.20 [-7.18, -3.17]
-4.22 [-6.46, -1.98]
0.98 [-2.04, 4.01]
24 Month
577 (260)
-6.07 [-8.27, -3.86]
-4.87 [-7.31, -2.42]
1.21 [-2.05, 4.41]
Lunch/Recess Breaks MVPA (mins/day)
12 Month
753 (332)
-4.78 [-5.63, -3.93]
-1.10 [-2.08, -0.15]
3.68 [2.40, 4.98]
24 Month
577 (260)
-4.82 [-5.77, -3.86]
-1.36 [-2.41, -0.30]
3.46 [2.04, 4.88]
After School MVPA (mins/day)
12 Month
753 (332)
-2.75 [-4.54, -0.96]
-3.70 [-5.70, -1.70]
-0.95 [-3.67, 1.71]
24 Month
577 (260)
-7.38 [-9.40, -5.35]
-9.77 [-11.94, -7.52]
-2.39 [-5.37, 0.59]
Weekend MVPA (mins/day)
12 Month
753 (332)
-8.91 [-13.09, -4.70]
-4.05 [-8.75, 0.66]
4.85 [-1.42, 11.20]
24 Month
576 (259)
-16.80 [-21.43, -12.26]
-17.91 [-23.10, -12.90]
-1.11 [-8.10, 5.79]
Total MVPA (mins/day)
12 Month
753 (332)
-5.58 [-8.79, -2.32]
-5.00 [-8.68, -1.35]
0.58 [-4.23, 5.45]
24 Month
576 (259)
-12.77 [-16.29, -9.21]
-13.87 [-17.76, -9.94]
-1.09 [-6.54, 4.18]
Self-reported PA usual week (days/week)
12 Month
1000 (443)
-0.06 [-0.23, 0.12]
0.00 [-0.19, 0.19]
0.06 [-0.20, 0.32]
24 Month
890 (384)
0.00 [-0.17, 0.18]
0.04 [-0.17, 0.25]
0.04 [-0.23, 0.31]
Self-reported PA this week (days/week)
12 Month
1001 (444)
0.05 [-0.13, 0.22]
-0.18 [-0.37, 0.02]
-0.23 [-0.49, 0.04]
24 Month
890 (384)
-0.05 [-0.23, 0.13]
-0.11 [-0.31, 0.10]
-0.06 [-0.34, 0.21]
Self-reported PA team sport participation
in the past year
12 Month
1001 (444)
-0.00 [-0.04, 0.04]
0.04 [-0.01, 0.08]
0.04 [-0.02, 0.10]
24 Month
889 (384)
0.03 [-0.01, 0.08]
0.06 [0.01, 0.11]
0.03 [-0.04, 0.09]
Self-reported PA individual sport
participation in the past year
12 Month
1001 (444)
-0.04 [-0.09, 0.01]
-0.05 [-0.10, 0.00]
-0.01 [-0.08, 0.06]
24 Month
890 (384)
-0.05 [-0.10, 0.00]
-0.07 [-0.12, -0.01]
-0.02 [-0.09, 0.06]
Self-reported PA active travel (days/week)
12 Month
1001 (444)
-0.11 [-0.30, 0.08]
-0.02 [-0.23, 0.19]
0.09 [-0.19, 0.38]
24 Month
890 (384)
0.21 [0.02, 0.41]
0.05 [-0.17, 0.27]
-0.16 [-0.46, 0.1
19
Variable
Follow-up
N
(n intervention)
Change from Baseline:
Control
Change from Baseline:
Intervention
Adjusted Difference
(intervention v control)
BMI (raw)
12 Month
992 (435)
0.79 [0.69, 0.90]
0.74 [0.63, 0.86
-0.06 [-0.21, 0.10]
24 Month
878 (372)
1.62 [1.52, 1.73
1.56 [1.43, 1.68
-0.07 [-0.23, 0.10]
BMI-z
12 Month
992 (435)
0.02 [-0.01, 0.05]
-0.00 [-0.04, 0.03]
-0.02 [-0.07, 0.02]
24 Month
878 (372)
0.04 [0.01, 0.07]
-0.00 [-0.04, 0.03]
-0.04 [-0.09, 0.01]
NAPLAN Numeracy
12 Month
771 (313)
1.07 [1.02, 1.13]
1.01 [0.95, 1.07]
-0.06 [-0.14, 0.02]
12 Month
773 (312)
0.88 [0.84, 0.93]
0.91 [0.85, 0.97]
0.03 [-0.05, 0.10]
Well-being
12 Month
971 (432)
0.12 [-0.35, 0.59]
0.22 [-0.31, 0.75]
0.10 [-0.61, 0.82]
24 Month
869 (372)
-0.85 [-1.34, -0.37]
-0.43 [-0.99, 0.13]
0.42 [-0.32, 1.17]
PE Psychological needs support from
teacher
12 Month
1001 (444)
-0.09 [-0.16, -0.02]
-0.09 [-0.16, -0.01]
0.00 [-0.10, 0.11]
PE Psychological needs support from
teacher
24 Month
890 (384)
-0.36 [-0.43, -0.29]
-0.19 [-0.27, -0.11]
0.17 [0.06, 0.27]
PE Concentration
12 Month
1001 (444)
-0.09 [-0.15, -0.03]
-0.06 [-0.12, 0.01]
0.03 [-0.06, 0.12]
24 Month
890 (384)
-0.24 [-0.30, -0.18]
-0.24 [-0.31, -0.17]
-0.00 [-0.10, 0.09]
PE Effort
12 Month
1001 (444)
-0.07 [-0.13, -0.01]
-0.06 [-0.13, 0.00]
0.00 [-0.09, 0.09]
24 Month
890 (384)
-0.23 [-0.30, -0.17]
-0.27 [-0.34, -0.20]
-0.04 [-0.13, 0.06]
PE Learning strategy use
12 Month
1001 (444)
-0.17 [-0.25, -0.09]
-0.15 [-0.24, -0.06]
0.02 [-0.11, 0.15]
24 Month
890 (384)
-0.42 [-0.51, -0.33]
-0.46 [-0.55, -0.36]
-0.04 [-0.17, 0.09]
PE Enjoyment
12 Month
1001 (444)
-0.13 [-0.21, -0.06]
-0.07 [-0.16, 0.01]
0.06 [-0.05, 0.17]
24 Month
890 (384)
-0.33 [-0.41, -0.25]
-0.32 [-0.41, -0.23]
0.01 [-0.10, 0.13]
Note: PA = Physical Activity; MVPA = Moderate-to-Vigorous Physical Activity. NAPLAN = National Assessment Program – Literacy and Numeracy; PE =
Physical Education. Well-being can range from 10 (very low well-being) to 50 (very high well-being). PE Psychological needs support from the teacher, PE
concentration, effort, PE learning strategy use, and PE enjoyment can range from 1 (low) to 5 (high).
20
eTable 8. Sub-sample Variables Baseline Values
Variable
N
(n intervention)
Control
Mean
Control SD
Intervention
Mean
Intervention
SD
Total Mean
Total SD
Fundamental Movement Skills
359 (188)
1.90
0.60
1.82
0.67
1.86
0.64
Response Accuracy AX
174 (91)
44.69
26.83
56.36
24.34
50.79
26.15
Response Accuracy AY
167 (87)
39.06
24.09
46.50
22.81
42.94
23.66
Response Accuracy BX
167 (88)
47.63
27.47
60.27
27.25
54.29
28.00
Cognitive Reaction Time AX
174 (91)
411.74
147.18
389.87
127.71
400.31
137.38
Cognitive Reaction Time AY
167 (87)
548.13
176.48
492.05
124.87
518.92
153.92
Cognitive Reaction Time BX
167 (88)
453.63
188.66
413.75
105.91
432.62
151.67
Overall Response Accuracy AX
173 (91)
48.51
19.50
56.55
17.99
52.74
19.10
Overall Response Accuracy AY
166 (87)
60.81
18.62
68.72
14.95
64.96
17.21
Overall Response Accuracy BX
167 (88)
55.85
17.06
58.24
15.60
57.11
16.30
Note: AX/AY/BX refers to the cognitive test condition (cue and probe).
21
eTable 9. Sub-sample Variables Outcome Analyses
Variable
Follow-up
N
(n intervention)
Change from Baseline:
Control
Change from Baseline:
Intervention
Adjusted Difference
(intervention v control)
Fundamental Movement Skills
12 Month
270 (137)
0.03 [-0.08, 0.13]
0.06 [-0.04, 0.17]
0.04 [-0.12, 0.19]
24 Month
184 (79)
0.01 [-0.12, 0.14]
-0.08 [-0.21, 0.05]
-0.09 [-0.27, 0.09]
Response Accuracy AX
12 Month
146 (74)
5.15 [0.11, 10.35]
3.85 [-0.89, 8.74]
-1.30 [-8.44, 5.85]
24 Month
116 (59)
4.91 [-0.63, 10.61]
2.42 [-3.06, 7.63]
-2.49 [-10.39, 5.24]
Response Accuracy AY
12 Month
142 (72)
3.20 [-2.78, 8.96]
6.07 [0.66, 11.40]
2.86 [-5.17, 10.70]
24 Month
111 (56)
6.63 [0.34, 13.22]
9.47 [3.64, 15.36]
2.84 [-6.35, 11.77]
Response Accuracy BX
12 Month
137 (71)
5.62 [-0.48, 12.02]
2.88 [-3.26, 9.11]
-2.74 [-11.14, 5.62]
24 Month
114 (57)
5.46 [-1.53, 12.51]
1.66 [-5.17, 8.16]
-3.81 [-13.12, 5.75]
Cognitive Reaction Time AX
12 Month
146 (74)
-55.23 [-83.05, -28.26]
-38.45 [-65.07, -12.43]
16.78 [-21.44, 55.02]
24 Month
116 (59)
-44.62 [-75.24, -14.52]
-34.97 [-63.62, -6.83]
9.65 [-32.03, 51.26]
Cognitive Reaction Time AY
12 Month
142 (72)
-59.79 [-97.48, -20.15]
-23.11 [-59.39, 13.73]
36.67 [-15.57, 89.43]
Cognitive Reaction Time AY
24 Month
111 (56)
-122.56 [-164.03, -
81.15]
-73.76 [-115.47, -31.96]
48.79 [-11.54, 109.73]
Cognitive Reaction Time BX
12 Month
137 (71)
-55.72 [-94.04, -19.21]
-33.88 [-69.55, 0.20]
21.84 [-27.68, 76.10]
24 Month
114 (57)
-72.06 [-112.43, -33.36]
-67.95 [-103.74, -29.87]
4.11 [-49.53, 58.91]
Overall Response Accuracy AX
12 Month
146 (74)
3.47 [-0.59, 7.33]
3.16 [-0.81, 7.01]
-0.31 [-5.92, 5.24]
24 Month
116 (59)
5.55 [1.10, 9.92]
5.44 [1.29, 9.54]
-0.11 [-6.12, 5.72]
Overall Response Accuracy AY
12 Month
141 (72)
4.13 [0.20, 8.05]
3.18 [-0.46, 6.96]
-0.95 [-6.24, 4.43]
24 Month
111 (56)
0.30 [-3.70, 4.33]
-0.28 [-4.28, 3.67]
-0.59 [-6.23, 5.32]
Overall Response Accuracy BX
12 Month
136 (71)
1.95 [-2.69, 6.67]
0.09 [-4.55, 4.67]
-1.86 [-8.44, 4.87]
24 Month
114 (57)
-0.73 [-5.64, 4.28]
1.45 [-3.37, 6.32]
2.18 [-4.66, 8.93]
Note: AX/AY/BX refers to the cognitive test condition (cue and probe).