Skip to main content
The Cochrane Database of Systematic Reviews logoLink to The Cochrane Database of Systematic Reviews
. 2021 Sep 23;2021(9):CD007651. doi: 10.1002/14651858.CD007651.pub3

School‐based physical activity programs for promoting physical activity and fitness in children and adolescents aged 6 to 18

Sarah E Neil-Sztramko 1, Hilary Caldwell 2, Maureen Dobbins 3,4,
Editor: Cochrane Public Health Group
PMCID: PMC8459921  PMID: 34555181

Abstract

Background

Physical activity among children and adolescents is associated with lower adiposity, improved cardio‐metabolic health, and improved fitness. Worldwide, fewer than 30% of children and adolescents meet global physical activity recommendations of at least 60 minutes of moderate to vigorous physical activity per day. Schools may be ideal sites for interventions given that children and adolescents in most parts of the world spend a substantial amount of time in transit to and from school or attending school.

Objectives

The purpose of this review update is to summarise the evidence on effectiveness of school‐based interventions in increasing moderate to vigorous physical activity and improving fitness among children and adolescents 6 to 18 years of age.

Specific objectives are:

• to evaluate the effects of school‐based interventions on increasing physical activity and improving fitness among children and adolescents;

• to evaluate the effects of school‐based interventions on improving body composition; and

• to determine whether certain combinations or components (or both) of school‐based interventions are more effective than others in promoting physical activity and fitness in this target population.

Search methods

We searched CENTRAL, MEDLINE, Embase, CINAHL, PsycINFO, BIOSIS, SPORTDiscus, and Sociological Abstracts to 1 June 2020, without language restrictions. We screened reference lists of included articles and relevant systematic reviews. We contacted primary authors of studies to ask for additional information.

Selection criteria

Eligible interventions were relevant to public health practice (i.e. were not delivered by a clinician), were implemented in the school setting, and aimed to increase physical activity among all school‐attending children and adolescents (aged 6 to 18) for at least 12 weeks. The review was limited to randomised controlled trials. For this update, we have added two new criteria: the primary aim of the study was to increase physical activity or fitness, and the study used an objective measure of physical activity or fitness. Primary outcomes included proportion of participants meeting physical activity guidelines and duration of moderate to vigorous physical activity and sedentary time (new to this update). Secondary outcomes included measured body mass index (BMI), physical fitness, health‐related quality of life (new to this update), and adverse events (new to this update). Television viewing time, blood cholesterol, and blood pressure have been removed from this update. 

Data collection and analysis

Two independent review authors used standardised forms to assess each study for relevance, to extract data, and to assess risk of bias. When discrepancies existed, discussion occurred until consensus was reached. Certainty of evidence was assessed according to GRADE. A random‐effects meta‐analysis based on the inverse variance method was conducted with participants stratified by age (children versus adolescents) when sufficient data were reported. Subgroup analyses explored effects by intervention type.

Main results

Based on the three new inclusion criteria, we excluded 16 of the 44 studies included in the previous version of this review. We screened an additional 9968 titles (search October 2011 to June 2020), of which 978 unique studies were potentially relevant and 61 met all criteria for this update. We included a total of 89 studies representing complete data for 66,752 study participants. Most studies included children only (n = 56), followed by adolescents only (n = 22), and both (n = 10); one study did not report student age. Multi‐component interventions were most common (n = 40), followed by schooltime physical activity (n = 19), enhanced physical education (n = 15), and before and after school programmes (n = 14); one study explored both enhanced physical education and an after school programme. Lack of blinding of participants, personnel, and outcome assessors and loss to follow‐up were the most common sources of bias. 

Results show that school‐based physical activity interventions probably result in little to no increase in time engaged in moderate to vigorous physical activity (mean difference (MD) 0.73 minutes/d, 95% confidence interval (CI) 0.16 to 1.30; 33 studies; moderate‐certainty evidence) and may lead to little to no decrease in sedentary time (MD ‐3.78 minutes/d, 95% CI ‐7.80 to 0.24; 16 studies; low‐certainty evidence). School‐based physical activity interventions may improve physical fitness reported as maximal oxygen uptake (VO₂max) (MD 1.19 mL/kg/min, 95% CI 0.57 to 1.82; 13 studies; low‐certainty evidence). School‐based physical activity interventions may result in a very small decrease in BMI z‐scores (MD ‐0.06, 95% CI ‐0.09 to ‐0.02; 21 studies; low‐certainty evidence) and may not impact BMI expressed as kg/m² (MD ‐0.07, 95% CI ‐0.15 to 0.01; 50 studies; low‐certainty evidence). We are very uncertain whether school‐based physical activity interventions impact health‐related quality of life or adverse events.

Authors' conclusions

Given the variability of results and the overall small effects, school staff and public health professionals must give the matter considerable thought before implementing school‐based physical activity interventions. Given the heterogeneity of effects, the risk of bias, and findings that the magnitude of effect is generally small, results should be interpreted cautiously.

Plain language summary

Do school‐based physical activity interventions increase moderate to vigorous physical activity and improve physical fitness among children and adolescents?

Key messages

School‐based interventions may improve physical fitness but may have little to no impact on body mass index (which is used to assess whether body weight is in a healthy range), although we do not have confidence in the evidence. 

Very few studies have reported on any potential harmful effects. 

Careful consideration is needed about the type of school‐based physical activity programme to be implemented, and future studies should seek to identify the best types of physical activity interventions for school settings.

Why is it important to promote physical activity in children?

It is estimated that as many as 5.3 million deaths worldwide are caused by not getting enough exercise (physical inactivity), and this is a big risk factor leading to most long‐lasting diseases and cancers. This is a topic of concern, particularly because it is known that physical activity patterns in childhood can lead to similar patterns in adulthood. Programmes that encourage children to exercise while at school are thought to be a way to increase activity levels of all children, regardless of other factors such as parent behaviours and social or financial factors of a child’s early lifetime. 

What did we find?

We found 89 studies that looked at the effects of programmes in schools that focused on increasing physical activity, which included 66,752 children and adolescents (between the ages of 6 and 18) from around the world. The length of programme time varied from 12 weeks to 6 years. No two school‐based physical activity programmes used the same combination of intervention parts. How often and how long each part of a programme was run varied a lot across studies.

Across all included studies, only very small changes were noted in the number of students undertaking physical activity or in minutes per day of moderate to vigorous physical activity or sedentary time, although these programmes were found to improve students’ physical fitness. These programmes were found to have little to no impact on measurements used to assess whether body weight is in a healthy range. Not many studies reported on any potential harmful effects, such as injury or psychological harm. 

What are the limitations of the evidence?

We have little confidence in the evidence because studies were done in different ways and interventions were delivered and assessed in different ways. Also, people in the studies may have been aware of which interventions they were getting, and this can sometimes affect the outcomes reported. In addition, not all studies provided data about everything we were interested in.

How up‐to‐date is the evidence? 

The evidence is up‐to‐date to June 2020 (although we did run a new search for studies in February 2021 and found studies that may be included in a future update and are now described in the “Studies awaiting classification” table).

Summary of findings

Summary of findings 1. School‐based physical activity programmes for promoting physical activity and fitness in children and adolescents aged 6 to 18 years.

School‐based physical activity programmes for promoting physical activity and fitness in children and adolescents aged 6 to 18 years
Population: children and adolescents aged 6 to 18 years
Settings: primarily within the school setting
Intervention: educational, health promotion, counselling, and management strategies focused on promotion of physical activity and fitness
Comparison: standard, currently existing physical education programmes in schools
Outcomes Anticipated effects (95% CI) No. of participants
(trials) Certainty of the evidence
(GRADE)
Risk with control Risk with intervention
% of participants physically active
 
[follow‐up: 12 weeks to 12 months]
% physically active ranged from 2% to 50% % physically active ranged from 1.11% lower to 12.22% higher. 6,068
(5)
⊕⊝⊝⊝
very lowa
Moderate to vigorous physical activity (minutes/d)
[follow‐up: 12 weeks to 3 years]
‐3.63 (‐5.03 to ‐2.23) MD 0.73, 95% CI 0.16 to 1.30 20,614
(33)
⊕⊕⊕⊝
moderateb
Sedentary time (minutes/d)
[follow‐up: 12 weeks to 28 months]  27.77 (‐21.34 to 76.88) MD ‐3.78, 95% CI ‐7.80 to 0.24 11,914
(16)
⊕⊕⊝⊝
lowc
Physical fitness (VO₂ max, mL/kg/min)
[follow‐up: 12 weeks to 1 year] ‐1.00 (‐1.59 to ‐0.41) MD 1.19, 95% CI 0.57 to 1.82 3,980
(13)
⊕⊕⊝⊝
lowd
BMI (z‐score)
[follow‐up: 12 weeks to 4 years] ‐0.01 (‐0.08 to 0.06) MD ‐0.06, 95% CI ‐0.09 to ‐0.02 22,948
(21)
⊕⊕⊝⊝
lowe
BMI (kg/m²)
[follow‐up: 12 weeks to 4 years] ‐0.35 (‐1.06 to 0.36) MD ‐0.07, 95% CI ‐0.15 to 0.01 34,337
(50)
Health‐related quality of life
[follow‐up: 15 weeks to 12 months]
Not estimable; insufficient data reported within studies 4,687
(7)
⊕⊝⊝⊝
very lowf
Adverse events
[follow‐up: 12 weeks to 3 years]
Not estimable; only 3 studies reported any adverse events 11,698
(16)
⊕⊝⊝⊝
very lowg
BMI: body mass index; CI: confidence interval; MD: mean difference; min/d: minutes per day; VO₂max: maximal oxygen uptake.
GRADE Working Group grades of evidence.High quality: we are very confident that the true effect lies close to that of the effect estimate.
Moderate quality: we are moderately confident in the effect estimate. The true effect is likely to be close to the estimate of effect, but there is a possibility that it is substantially different.
Low quality: our confidence in the effect estimate is limited. The true effect may be substantially different from the estimate of the effect.
Very low quality: we have very little confidence in the effect estimate. The true effect is likely to be substantially different from the estimate of the effect.

aDowngraded by one level each for inconsistency (large variation in effects across studies), imprecision (wide confidence intervals), and risk of bias (high or unclear in most studies).

bDowngraded by one level for inconsistency (visual inspection of forest plots and I² value from meta‐analysis).

cDowngraded by one level for imprecision of results (wide confidence intervals) and risk of bias (high or unclear in many studies).

dDowngraded by one level for inconsistency (visual inspection of forest plots and I² value from meta‐analysis) and indirectness (estimated vs measured VO₂ peak).

eDowngraded by one level for inconsistency (visual inspection of forest plots and I² value from meta‐analysis), risk of bias (high or unclear in most studies).

fDowngraded by one level for inconsistency (large variation across studies), risk of bias (high or unclear in most studies), publication bias (most studies not reporting on health‐related quality of life or describing full results).

gDowngraded by one level for inconsistency (large variation across studies), publication bias (most studies not reporting on adverse events or methods for monitoring), risk of bias (high or unclear in most studies).

Background

Description of the condition

International public health and health promotion organisations have identified health risks across the lifespan associated with physical inactivity. Recent estimates suggest that 5.3 million deaths per year throughout the world are attributable to physical inactivity (Lee 2012a). Globally, physical inactivity has been identified as the fourth leading risk factor for global mortality (6% of all deaths), following high blood pressure, tobacco use, and high blood glucose, and overweight and obesity are responsible for 5% of deaths globally (Warburton 2017WHO 2008). Physical inactivity is estimated to cause 10% of the burden of disease from breast cancer and colon cancer, as well as 7% from type 2 diabetes, and 6% from coronary heart disease (Lee 2012a). The burden of these and other chronic diseases has rapidly increased in recent decades (WHO 2008). In fact, physical activity was labelled as “today's best buy in public health” almost three decades ago (Morris 1994). Recent estimates suggest that physical inactivity cost healthcare systems $ (INT$) 53.8 billion worldwide in 2013 (Ding Ding 2016). In addition, the literature indicates that an elevated body mass index (BMI) places children and adolescents at greater risk for cardiovascular disease as adults, and that diet and physical activity are important factors in maintaining a healthy BMI range (Hills 2011). Longitudinal data have shown that for each weekday that adolescents of normal weight participated in physical education (PE), the odds of becoming overweight in adulthood decreased by 5% (Menschik 2008).

Previous reports have concluded that the intensity, frequency, and duration of physical activity contribute to overall physical health status and suggest that a 'threshold' must be maintained to produce positive health effects (CDC 1999Shephard 1997Tolfrey 2000). In fact, a positive linear association between duration of physical activity and positive health effects has been established, with longer duration associated with improved physical health (Carson 2017; Janssen 2010; Shephard 1997). Maximal oxygen uptake (VO₂max) is a standard measure associated with fitness levels, with increasing values expected as fitness level improves, and is an important indicator of successful physical activity interventions. Recent analyses have shown a dose‐response relationship between physical activity behaviours (minutes/week of moderate to vigorous physical activity (MVPA)) and fitness levels (measured by VO₂max) and measures of cardio‐metabolic health, including measures of body composition, blood pressure, and blood cholesterol (Nevill 2020; Sriram 2021). 

Current guidelines suggest that children and adolescents should engage in at least 60 minutes of MVPA per day, along with muscle and bone strengthening activities at least 3 days per week (Chaput 2020). Examples of moderate to vigorous activities include brisk walking, jogging, stair climbing, basketball, racquet sports, soccer, dance, lap swimming, skating, strength training, cross‐country skiing, and cycling. In the most recently released global physical activity guidelines, evidence suggests that time spent in sedentary behaviour is associated with poorer health outcomes in both children and adolescents, and it is recommended that children and adolescents limit sedentary time, especially when sedentary time is combined with recreational screen time (Chaput 2020). Research suggests that the best primary strategy for improving the long‐term health of children and adolescents through exercise may involve creating a lifestyle pattern of regular MVPA that will carry over to the adult years (Freedson 1992Telama 2005Twisk 2000). Despite this, the Global Matrix 3.0 Physical Activity Report Card Grades for Children and Youth on physical levels of children and youth from 49 countries revealed a mean letter grade of C for the percentage of children and youth meeting the physical activity recommendation of 60 minutes of MVPA per day, representing 27% to 33% of children and youth (Aubert 2018). Therefore it is of primary importance to identify approaches that will be effective in increasing and sustaining activity levels of children and adolescents in places where they spend long periods of time, such as schools.

Description of the intervention

To ensure sustained progress towards major improvements in chronic disease prevention, the World Health Organization (WHO) has called on public health organisations within and between countries to work collaboratively with key partners, including educators and health professional bodies, educational institutions, consumer groups, researchers, and the private sector, to provide a comprehensive strategy to promote physical activity among children and adolescents (WHO 2008). The school setting is an ideal environment for population‐based physical activity interventions, as no other institution has as much influence on children during their first two decades of life (Naylor 2009; Story 2009). Recognising the unique opportunity that exists to formulate and implement an effective strategy to substantially reduce deaths and disease burden worldwide by improving diet and promoting physical activity, the WHO has adopted the Health Promoting School (HPS) framework to promote healthy living, learning, and working conditions. An HPS encourages moving beyond individual behaviour change and considering and addressing the whole‐school environment (IUHPE 2010). The WHO specifically identified schools as a target setting for promotion of physical activity among children and youth and suggested that schools implement opportunities and programming for physical activity and provide safe and appropriate spaces that facilitate participation in physical activity (WHO 2008).

Schools are considered an ideal setting for promoting physical activity among children due to their broad reach and multiple opportunities to promote physical activity over the course of the school week, including physical education classes, before and after school programmes, recess programming, active school travel, and classroom‐based physical activity (Singh 2017Watson 2017a). To effectively promote MVPA, the Centers for Disease Control and Prevention suggests that schools engage in comprehensive approaches to develop, implement, and evaluate physical activity policies and practices; establish school environments that support physical activity; implement a quality physical activity programme with quality physical education as a cornerstone; implement health education that provides students with the knowledge, attitudes, and skills needed for physical activity; provide students with health, mental health, and social services to promote physical activity and prevent chronic disease; partner with families and community members in development and implementation of physical activity policies; provide a school employee wellness programme that includes physical activity services for all school staff members; and employ qualified persons and provide professional development opportunities for physical education and physical activity (CDC 2011).

Studies conducted to date have generally utilised different combinations of physical activity promotion interventions in schools, including before and after school programmes, multi‐component physical activity interventions, enhanced PE, and schooltime physical activity, such as physically active lessons. Generally, interventions focused on providing students with information about the benefits of physical activity and healthy nutrition, the risks associated with inactivity and unhealthy food choices, and the importance of increasing the amount of time students were engaged in MVPA during the school day, as well as ensuring that they expend greater amounts of energy during physical activity sessions. Interventions targeted the school curriculum (related to PE classes specifically and to the whole curriculum generally), teacher training, educational materials, changes in the format of the school day, and accessibility to exercise equipment. Interventions included training sessions for teachers (to learn more effective ways to promote physical activity and to incorporate it into curricula) and training materials including kits, packages of materials to be used in curricula, and materials to be given to students and parents. Packages included teaching notes on exercise, how the body works, and healthy eating and nutrition. Interventions also targeted PE classes by increasing the level of activity students engaged in during these classes, introducing activities geared to the age and level of development of the child, and putting greater emphasis on games. Students were encouraged to be more active outside of school during leisure time. The curriculum focused on creating a positive self‐image through awareness of physical activity benefits. Curriculum changes were implemented in other courses as well, such as science courses, which incorporated discussions of healthy eating and physical activity. Some interventions included changes in the food provided in school cafeterias, so as to increase the number of healthy food choices. Other strategies included a risk factor assessment of students to identify students with established risk factors and development of a plan to reduce student risk through increased physical activity, healthy nutrition, and goal attainment.

How the intervention might work

Given that school‐aged children spend a significant amount of their wakeful time in transit to or in the school setting, and that almost all children in many countries attend school until they reach adolescence, school‐based physical activity promotion interventions have the potential to reduce population‐wide chronic disease (Macnab 2014). In fact, several published systematic reviews have highlighted the benefits of physical activity among healthy children (Brown 2009; Cesa 2014; Granger 2017; Janssen 2010; Kriemler 2011a; Marker 2016; Poitras 2016). School‐based interventions can target simultaneously children at risk and children not at risk for future chronic disease and can increase both knowledge and behaviour conducive to healthier lifestyles. School‐based strategies targeting all students through curriculum ensure that 100% of students are exposed to the intervention, thereby increasing the reach of these interventions. Increased physical activity is an essential public health and health promotion strategy to improve child health (WHO 2008).

The intent of school‐based physical activity interventions is to increase the overall percentage of children and adolescents engaged in MVPA each day while increasing the duration of MVPA engaged in on a weekly basis. The aim of these interventions is to create a school environment that is conducive to achieving a greater proportion of children and adolescents who meet physical activity guidelines, while increasing time spent engaged in MVPA. Generally, this means that significant changes to the school curriculum are needed to support increased time for physical activity as well as increased levels of activity during this time. School‐based interventions offer an important opportunity to improve knowledge of how to prevent non‐communicable disease, and to provide both knowledge about the importance of physical activity and the opportunity for students to be more active before, during, and after school hours, thereby helping them develop healthier behaviours that may track into adulthood (Hayes 2019).

Why it is important to do this review

This systematic review was first published in 2009, with an update provided in 2013. Although the benefits of physical activity for healthy children have been documented, at the time the original review was conducted no other reviews had systematically examined the effectiveness of various combinations of school‐based interventions in promoting physical activity and fitness among children. Since that time, several new trials have been published (and are included in this update) and methodological advances have been made (e.g. using accelerometers rather than relying on self‐report data). The purpose of this update is to synthesise new data on the effectiveness of school‐based physical activity interventions with data included in the original review. This update includes evaluations of published studies indexed up to and including 1 June 2020.

Given that school‐aged children spend a significant amount of time in the school setting, and that many barriers prevent participation in physical activity outside of the school setting (e.g. resources, availability, cost), it is particularly important to understand the extent to which school‐based physical activity promotion interventions are effective in increasing activity and improving fitness levels. It is important to evaluate how these types of interventions may impact (positively or negatively) students' overall well‐being and health‐related quality of life and to discern any potential adverse events or harms. When school‐based interventions are combined with broader community‐based interventions, it is difficult to ascertain the impact of school‐based strategies. However, in developing comprehensive physical activity or chronic disease prevention strategies, it is crucial to incorporate effective school‐based strategies. Therefore, it is timely, given low worldwide participation in regular physical activity, that a review focused solely on the effectiveness of school‐based physical activity interventions be conducted and regularly updated.

Objectives

The purpose of this review update is to summarise the evidence on effectiveness of school‐based interventions in increasing moderate to vigorous physical activity and improving fitness among children and adolescents 6 to 18 years of age.

Specific objectives are:

  • to evaluate the effects of school‐based interventions on increasing physical activity and improving fitness among children and adolescents;

  • to evaluate the effects of school‐based interventions on improving body composition; and

  • to determine whether certain combinations or components (or both) of school‐based interventions are more effective than others in promoting physical activity and fitness in this target population.

Methods

Criteria for considering studies for this review

Types of studies

In accordance with the last update, we included randomised controlled trials (RCTs) or cluster‐RCTs with a minimum intervention duration of 12 weeks. There are four unique departures from the protocol of this update in comparison to the original review.

  • Due to growing availability and use of technology‐based assessments of physical activity and sedentary time via accelerometers, and known limitations of self‐report physical activity data for children in particular, we limited this review to studies that included a device‐based measure of physical activity or sedentary time (e.g. with accelerometers). This resulted in exclusion of measures related to TV watching time from the review, replaced by overall measures of sedentary time. This change in inclusion criteria resulted in exclusion from this update of several studies that included only self‐report measures that were included in the original review.

  • Studies not primarily aimed at increasing levels of physical activity were excluded from this update. This resulted in exclusion of some studies that were primarily focused on nutrition or reducing/preventing obesity.

  • Outcomes of blood pressure and pulse rate were excluded from this update. Again, this resulted in exclusion of a very small number of studies that included only these outcomes.

  • For included studies, we extracted outcomes related to health‐related quality of life and adverse events when reported. This did not result in any change to inclusion or exclusion criteria.

The review authors are aware that post hoc questions are susceptible to bias (Higgins 2011); however, in light of the growing body of literature on the effectiveness of school‐based physical activity promotion interventions, we believe the changes in inclusion and exclusion criteria were necessary to ensure inclusion in this update of the most rigorous evidence related to the research question.

Types of participants

Studies that included school‐attending children and adolescents between the ages of 6 and 18 years were included in this review. This included all otherwise healthy children and adolescents, whether they were overweight or obese, or were not. We excluded studies in which participants received a physical activity intervention as part of a treatment regimen for a specific critical illness or comorbidity (e.g. diabetes). Study participants were categorised as children (age 6 to < 12) or adolescents (age 12 to 18) based on Centers for Disease Control and Prevention categories (Centers for Disease Control and Prevention, 2021).

Types of interventions

Interventions

We included any school‐based physical activity programme that aimed to increase physical activity and/or fitness among children and adolescents. We defined school‐based physical activity programmes as any that implemented educational, health promotion, counselling, and/or management strategies focused on promotion of physical activity and/or fitness. The range of interventions included changes to school curriculum, training for teachers about incorporating physical activity into school curriculum and routines, and educational materials for teachers, students, and parents. In some instances, the intervention included strategies to engage parents in the intervention, as well as community‐based strategies, mass media, policy development, and environmental changes. We included studies for which interventions were targeted primarily within the school setting. In some instances, interventions were implemented in the community and in the home, in addition to school‐based interventions, although the primary focus needed to be the school setting. Included studies must have fallen within public health practice (meaning the focus was on health promotion from an individual or population‐wide perspective and was not physician or clinic based) and must have been able to be implemented, facilitated, or promoted by staff in local public health units or by public health professionals.

Comparators

The comparison could be no intervention, usual care, or a concomitant intervention. Concomitant interventions had to be the same in both intervention and comparator groups to establish fair comparisons.

Minimum duration of intervention

The minimum duration of follow‐up was 12 weeks. Given the abundance of literature, this criterion was added during the last update to focus on interventions that were most likely to result in meaningful and sustainable changes in the school setting (Dobbins 2013). We extracted data on extended follow‐up periods. We defined extended follow‐up periods as follow‐up of participants that occurred once the original trial, as specified in the trial protocol, had been terminated (Buch 2011; Megan 2012). 

Summary of specific exclusion criteria

We excluded studies in which participants received a physical activity intervention as part of a treatment regimen for a specific critical illness or comorbidity (e.g. diabetes), studies in which the intervention was conducted entirely outside the school setting (e.g. community setting, public place, recreation facility, physician office, camp setting), and studies in which the intervention could be delivered only by a specific health professional (e.g. physician) or fitness expert.

Types of outcome measures

To be included, studies had to report one or more of the following outcomes, presented as post‐intervention measurement and standard deviation or confidence intervals, or as change from baseline with standard deviation or confidence intervals.

Primary outcomes
  • Proportion of students meeting recommendations for moderate to vigorous physical activity (MVPA)

  • Duration of MVPA

  • Sedentary time

Secondary outcomes
  • Physical fitness

  • Body mass index (BMI)

  • Health‐related quality of life

  • Adverse events

Method and timing of outcome measurement

Outcomes were primarily measured at baseline and immediately post intervention. In a small subset of studies, outcomes were measured at 6 months, at 9 months, and at 12 months, and in one study, up to 4 years post intervention.

  • MVPA: assessed by accelerometers during school time or non‐school (or both) time. The proportion of students who met physical activity guidelines was presented as reported or calculated by dividing the number of students engaged in 60 minutes/d of MVPA by the total number of students allocated to either the intervention group or the control group; duration of physical activity was measured as total minutes per day or weeks spent engaged in MVPA.

  • Sedentary time: measured as time spent sedentary in total minutes or hours per day or per week, measured via accelerometers.

  • BMI (kg/m²or BMI z‐score): measured by trained health professionals using calibrated scales; however, differences existed across studies in terms of which clothes were worn by participants during measurement and whether height and weight were measured during school time.

  • Physical fitness: measured in different ways by trained professionals. In some instances, actual maximal oxygen uptake (VO₂max) was measured; in many, a field‐based test such as a shuttle run or a step test was used to estimate VO₂max.

  • Health‐related quality of life: quantified with a validated instrument (e.g. Pediatric Quality of Life Inventory). Assessments completed by both students and parent proxies were eligible.

  • Adverse events: any as reported by study authors were noted, as were specific statements of no adverse events occurring.

Search methods for identification of studies

Electronic searches

The search for this update was conducted from the date of our last search (October 2011) to 1 June 2020. Some minor changes were made to the search strategy, such as using validated search filters for RCTs that were not available at the time of our last search and adding terms for sedentary time. Validated search filters were used within MEDLINE Ovid (Lefebvre 2019), Embase Ovid (Glanville 2019a), and Cumulative Index to Nursing and Allied Health Literature (CINAHL) (Glanville 2019b). We searched the following databases.

  • Cochrane Central Register of Controlled Trials (CENTRAL), in the Cochrane Library.

  • MEDLINE Ovid.

  • Embase Ovid.

  • CINAHL EBSCO.

  • PsycINFO Ovid.

  • BIOSIS Web of Science.

  • SPORTDiscus EBSCO.

  • Sociological Abstracts ProQuest.

For detailed search strategies, see Appendix 1. We placed no restrictions on language of publication when searching electronic databases or reviewing reference lists of identified trials.

Searching other resources

We tried to identify other potentially eligible trials or ancillary publications, including trial registries, by handsearching the reference lists of all included trials and relevant systematic reviews and meta‐analyses and health technology assessment reports that were identified in our search. We contacted authors of included trials to request additional information on retrieved trials and to determine if further trials exist, which we may have missed.

We did not use abstracts or conference proceedings for data extraction because this information source does not fulfil the Consolidated Standards of Reporting Trials (CONSORT) requirements, which call for "an evidence‐based, minimum set of recommendations for reporting randomised trials" (CONSORT 2018 Scherer 2007). However, we specified trial details in the Characteristics of studies awaiting classification table, and we contacted study authors to determine whether further publications exist.

Data collection and analysis

Selection of studies

For this update, two review authors (SNS, HC) independently screened abstract, title, or both, for every record retrieved in the literature searches, to determine which trials should be assessed further. We obtained the full text of all potentially relevant records. We resolved disagreements through consensus or by recourse to a third review author (MD). If we could not resolve a disagreement, we categorised the trial as a 'study awaiting classification' and contacted trial authors for clarification. We prepared an adapted PRISMA flow diagram to show the process of trial selection (Figure 1) (Liberati 2009). We listed all articles excluded after full‐text assessment in the Characteristics of excluded studies table and provided reasons for exclusion.

1.

1

Trial flow diagram.

Data extraction and management

For trials that fulfilled our inclusion criteria, two review authors (SNS, HC) independently extracted key participant and intervention characteristics. We reported data on efficacy outcomes and adverse events using standardised data extraction sheets. We resolved disagreements by discussion, or, if required, we consulted a third review author (MD) (for details, see Characteristics of included studiesTable 2Appendix 2Appendix 3). We provided information including trial identifiers for potentially relevant ongoing trials in the Characteristics of ongoing studies table.

1. Overview of study populations.
Trial ID (design) Interventions and comparators Screened/eligible
(N) Randomised
(N) Finishing trial
(N) Randomised finishing trial
(%)
Breheny 2020 I: Daily Mile —/— 1153 1107 96.0
C: usual school routine 1127 1070 94.9
total:  2280 2177 95.5
Ketelhut 2020 I: high‐intensity interval training during PE —/— 22
C: usual PE 24
total: 46
Belton 2019 (cluster‐RCT) I: multi‐component PE, whole‐school and parent‐targeted intervention 564/534 275 123 44.7
C: usual care 259 126 48.6
total: 534 249 46.6
Corepal 2019 
(cluster‐RCT) I: pedometer challenge —/— 142 136 95.8
C: usual school 82 81 98.8
total:  224 217 96.9
Ickovics 2019 (cluster‐RCT) I1: PA school wellness policy —/756 413 330 79.9
I2: PA + nutrition school wellness policy
C1: nutrition school wellness policy 305 265 86.9
C2: delayed control
total: 718 595 82.9
Jago 2019 (cluster‐RCT) I: Action 3:30R after‐school PA club 1139/1125 170 113 66.5
C: — 165 139 84.2
total: 335 252 75.2
Leahy 2019 (cluster‐RCT) I: Burn2Learn, multi‐component high‐intensity interval training —/68 38 32 84.2
C: usual school activities 30 29 96.7
total: 68 61 89.7
Lonsdale 2019a
(cluster‐RCT) I: teacher PE training —/1806 693 630 90.9
C: standard teaching 728 628 86.3
total: 1421 1258 88.5
Müller 2019
(cluster‐RCT) I1: PA only 1009/944 265
I2: PA + health education
I3: PA + health education + nutrition
C1: health education + nutrition 398
C2: no PA
total: 944 663 70.2
Ordóñez Dios 2019 I: 2 x 45‐minute PE sessions per week and daily run —/— 45
C: 2 x 45‐minute PE sessions per week 44
total: 89
Seibert 2019 I: 4 core strategies to increase PA —/— 2495
C: usual PE 2399
total: 4894
Seljebotn 2019
(cluster‐RCT)
 
 
I: physically active lessons, active homework, physically active recess —/473 228 224 98.2
C: continued normal routine, approximately 135 minutes/week of PA 219 218 99.5
total: 447 442 98.9
Zhou 2019 
(cluster‐RCT) I1: modified PE —/—
  204 163 79.9
I2: after school PA programme 200 180 90.0
I3: modified PE and after school PA programme 178 168 94.4
C: regular PE  176 170 96.6
 total:  758 681 89.8
Adab 2018
(cluster‐RCT) I: 30 minutes of additional MVPA on each school day, cooking workshops, a 6‐week healthy eating programme, information sheets for families —/— 1134 660 58.2
C: ongoing Year 2 health‐related activities and education resources, excluding topics related to healthy eating and PA 1328 732 55.1
  total: 2462 1392 56.5
Carlin 2018
(cluster‐RCT) I: brisk walking intervention —/— 101 100 99.0
C: continued with normal PA habits 98 97 99.0
total: 199 197 99.0
Harrington 2018
(cluster‐RCT) I: support for PA, PE, and school sport culture and practices with support of the Youth Sport Trust and a hub school —/1753 867 735 84.8
C: usual practice of PE and sport 885 626 70.7
total: 1752 1361 77.7
Have 2018
(cluster‐RCT) I: active math lessons —/557 294 268 91.2
C: regular classroom instruction 211 182 86.3
total: 505 450 89.1
Pablos 2018
(cluster‐RCT) I: lunchtime extracurricular PA —/210 100 82 82.0
C: continued with daily activities 90 76 84.4
total: 190 158 83.2
Robbins 2018
(cluster‐RCT) I: an after school PA club, counselling, interactive Internet‐based sessions 4192/1543 766 706 92.2
C: no additional after school programming 777 680 87.5
total: 1543 1386 89.8
Siegrist 2018
(cluster‐RCT) I: weekly lifestyle lessons 792/— 331 243 73.4
C: usual activities 257 191 74.3
total: 588 434 73.8
Ten Hoor 2018
(cluster‐RCT) I: strength training and motivational interviewing —/808 353 262 74.2
C: usual curriculum 342 246 71.9
total: 695 508 73.1
Donnelly 2017
(cluster‐RCT) I: Academic Achievement and Physical Activity Across the Curriculum lessons, 160 minutes/week of MVPA —/698 316 244 77.2
C: traditional classroom instruction and typical PE schedule 268 204 76.1
total: 584 448 76.7
Farmer 2017
(cluster‐RCT) I: school‐specific playground action plan —/— 812 344 42.4
C: no change to school play spaces 851 325 38.2
total: 1663 669 40.2
Sutherland 2017
(cluster‐RCT) I: modified Supporting Children's Outcomes using Rewards, Exercise and Skills programme —/1959 571
C: delivered school PA practices according to the curriculum 568
total: 1139
Torbeyns 2017 (RCT) I: cycling desks —/— 28 21 75.0
C: no lifestyle change 28 23 82.1
total: 56 44 78.6
Daly 2016
(cluster‐RCT) I: specialist‐taught PE intervention —/— 457 273 59.7
C: usual PE programme 396 267 67.4
total: 853 540 63.3
de Greeff 2016
(cluster‐RCT) I: physically active mathematics and language lessons —/— 181
C: usual curriculum 195
total: 388 376 96.9
Drummy 2016
(cluster‐RCT) I: teacher‐led activity break —/150 54
C: normal daily routine 53
total: 120 107 89.2
Jarani 2016
(cluster‐RCT) I1: group circuit training‐based PE —/767 261 253 96.9
I2: games‐based PE 251 243 96.8
C: traditional PE school 255 240 94.1
total: 767 736 96.0
Kocken 2016
(cluster‐RCT) I: theory and practical lessons on nutrition and PA —/— 615 367 59.7
C: regular school programme or curriculum on nutrition and PA 497 496 99.8
total: 1112 863 77.6
Lau 2016 (RCT) I: Xbox 260 Kinect gaming sessions after school 152/84 40 40 100.0
C: regular PA and PE class 40 40 100.0
total: 80 80 100.0
Resaland 2016
(cluster‐RCT) I: physically active Norwegian, mathematics, and English lessons on the playground; PA breaks and PA homework 1395/1202 620 593 95.6
C: curriculum‐prescribed PE and PA 582 530 91.1
total: 1202 1123 93.4
Sutherland 2016
(cluster‐RCT) I: 7 PA intervention strategies and 6 implementation strategies —/1468 696 250 35.9
C: only measurement components of the trial: regular PA and PE 537 191 35.6
total: 1233 441 35.8
Tarp 2016
(cluster‐RCT) I: 60 minutes of PA during schooltime, PA homework 869/855 215 194 90.2
C: normal practice 490 438 89.4
total: 705 632 89.6
Cohen 2015
(cluster‐RCT) I: teacher learning, PA policies, school‐community linkages —/— 199 166 83.4
C: usual PE and school sport programmes 261 217 83.1
total: 460 383 83.3
Jago 2015
(cluster‐RCT) I: after school dance classes —/663 284
C: provided data only 287
total: 571 508 89.0
Madsen 2015
(cluster‐RCT) I: nutrition education curriculum, Playworks structured recess before or after school activities, PA and games implemented by teachers —/— 583 446 76.5
C: — 296 230 77.7
total: 879 676 76.9
Muros 2015
(cluster‐RCT) I1: extracurricular PA sessions 242/162 28 28 100.0
I2: PA and nutrition 21 21 100.0
I3: PA and nutrition and extra virgin olive oil during the final month 25 25 100.0
C1: nutrition and lifestyle education sessions 41 41 100.0
C2: usual activities
total: 135 135 100.0
Suchert 2015
(cluster‐RCT) I: multi‐level intervention targeting students, classrooms, schools, and parents —/1489 790 702 88.9
C: no intervention 506 460 90.9
total: 1296 1162 89.7
Andrade 2014
(cluster‐RCT) I: ACTIVITAL individual‐ and environment‐based intervention —/— 700 550 78.6
C: standard curriculum 740 533 72.0
total: 1440 1083 75.2
Jago 2014
(cluster‐RCT) I: Action 3:30 activity club —/— 284 153 53.9
C: schools provided data only 255 157 61.6
total: 539 310 57.5
Kipping 2014
(cluster‐RCT) I: PA education intervention 2242/2221 1064
C: continued standard education provision 1157
total: 2221 1252 56.4
Kobel 2014
(cluster‐RCT) I: teacher training, PA education, active breaks 3159/1968
C: no intervention
total: 1964 1724 87.8
Martinez‐Vizcaino 2014 (cluster‐RCT) I: MOVI‐2 extracurricular PA programme —/— 769 420 54.6
C: standard PE curriculum 823 492 59.8
total: 1592 912 57.3
Nogueira 2014
(cluster‐RCT) I: high‐intensity capoeira sessions 341/339 185 176 95.1
C: usual school activities 154 135 87.7
total: 339 138 40.7
Santos 2014
(cluster‐RCT) I: healthy buddies, healthy living lessons, structured aerobic exercise —/— 340 310 91.2
C: standard curriculum 347 273 78.7
total: 687 583 84.9
Toftager 2014
(cluster‐RCT) I: physical and organisational environmental changes —/— 623 551 88.4
C: — 725 608 83.9
total: 1348 1159 86.0
Fairclough 2013
(cluster‐RCT) I: weekly lesson plans, worksheets, homework tasks, lesson resources 420/318 166 117 70.5
C: normal instruction 152 89 58.6
total: 318 196 61.6
Ford 2013 (RCT) I: accumulated brisk walking programme —/174 77
C: normal school lessons 75
total: 174 152 87.4
Grydeland 2013
(cluster‐RCT) I: structured lessons, PA breaks, PA promotion —/— 784 519 66.2
C: — 1381 945 68.4
total: 2165 1464 67.6
Melnyk 2013
(cluster‐RCT) I: goal‐setting, education, PA homework 1560/807 374 286 76.5
C: Healthy Teens attention control curriculum was intended to promote knowledge of common adolescent health topics and health literacy 433 341 78.8
total: 807 627 77.7
Sacchetti 2013
(cluster‐RCT) I: daily PA in schoolyard and classroom 521/521 247 212 85.8
C: standard programme of PE 250 216 86.4
total: 497 428 86.1
Siegrist 2013
(cluster‐RCT) I: JuvenTUM educational and environmental intervention —/902 486 427 87.9
C: continued with usual school activities 340 297 87.4
total: 826 724 87.7
Aburto 2011
(cluster‐RCT) I1: basic intervention of environmental and policy‐level changes —/— 262 241 92.0
I2: plus intervention adding additional resources and daily morning exercise 264 242 91.7
C: no change to standard practices 338 216 63.9
total: 864 699 80.9
Ardoy 2011
(cluster‐RCT) I1: 4 sessions/week of PE 70/67 26 25 96.2
I2: 4 sessions/week of PE with emphasis on increasing intensity 23 23 100.0
C: 2 sessions/week of PE 18 18 100.0
total: 67 66 98.5
de Heer 2011
(cluster‐RCT) I: Bienstar intervention of health education and 45 to 60 minutes of after school PA 1720/901 292 242 82.9
C1: Grade 4 health workbooks and incentives 251 236 94.0
C2: spillover control group 354 326 92.1
total: 897 804 89.6
Jago 2011
(cluster‐RCT) I: education, social marketing, food environment, PE curriculum, equipment provision —/11158 5571 2060 37.0
C: recruitment and data collection only 5587 2003 35.9
total: 11158 4063 36.4
Jansen 2011
(cluster‐RCT) I: 3 PE sessions/week, additional after school sport and play, classroom education, parent health promotion —/— 1271 1149 90.4
C: continued with usual curriculum 1499 1267 84.5
total: 2770 2416 87.2
Magnusson 2011
(cluster‐RCT) I: students engaged in PA during PE lessons, during recess, and during classes; schools had access to PA equipment to use in school lessons; teaching materials promoting PA were provided —/321 151 138 91.4
C: followed the general PA curriculum 170 116 68.2
total: 321 254 79.1
Okely 2011
(cluster‐RCT) I: PA action plan —/1769 771 566 73.4
C: continuation of usual programmes 747 633 84.7
total: 1518 1199 79.0
Thivel 2011
(cluster‐RCT) I: 120 minutes of additional supervised PE —/— 229 229 100.0
C: habitual 2 hours of PE/week 228 228 100.0
total: 457 457 100.0
Wilson 2011
(cluster‐RCT) I: Active by Choice Today programme, PA homework, in‐school PA, motivational skills training   729 673 92.3
C: General Health Education Programme 693 635 91.6
total: 1422 1308 92.0
Kriemler 2010
(cluster‐RCT) I: 2 additional 45‐minute PE lessons/week, activity breaks, PA homework   305 297 97.4
C: usual mandatory PE lessons 235 205 87.2
total: 540 502 93.0
Neumark‐Sztainer 2010
(cluster‐RCT) I: New Moves curriculum (nutrition and self‐empowerment, motivational interviewing, lunch meetings, parent outreach) 687 182 177 97.3
C: participation in all‐girls PE class 174 159 91.4
total: 356 336 94.4
Angelopoulos 2009
(cluster‐RCT) I: educational intervention covering self‐esteem, body image, nutrition, PA, fitness, and environmental issues, with motivational methods to increase knowledge, skills, self‐efficacy, self‐monitoring, and social influence   321
C: — 325
total: 646
Donnelly 2009
(cluster‐RCT) I: 90 minutes/week of moderate to
vigorous physically active academic
lessons   814 792 97.3
C: regular classroom instruction 713 698 97.9
total: 1527 1490 97.6
Dorgo 2009
(cluster‐RCT) I1: PE manual resistance training programme   141 93 66.0
I2: PE manual resistance training plus
cardiovascular endurance training
C: regular PE programme that followed
the usual school curriculum 232 129 55.6
total: 373 222 59.5
Gentile 2009
(cluster‐RCT) I: 'Switch' programme: promoted
healthy lifestyles targeting family,
school, and community   670
C: no intentional exposure to the Switch programme 653
total: 1323 1029 77.8
Neumark‐Sztainer 2009
(cluster‐RCT) I: after school theatre sessions, booster
sessions, family outreach   56 51 91.1
C: a theatre‐based control condition 52 45 86.5
total: 108 96 88.9
Peralta 2009 (RCT) I: curriculum and peer‐facilitated
lunchtime PA session, parent newsletters   16 16 100.0
C: PA curriculum 17 16 94.1
total: 33 32 97.0
Walther 2009
(cluster‐RCT) I: 1 unit of physical exercise (45 minutes)
with at least 15 minutes of endurance
training/school day, plus
lessons on healthy lifestyle once/
month   112 109 97.3
C: German standards, 2 units (each 45
minutes) of PE/week, 12 units (45 minutes/
unit) of high‐level endurance exercise
training/week plus participation
in competitive sporting events   76 73 96.1
total: 188 182 96.8
Reed 2008
(cluster‐RCT) I: Action Schools!BC whole‐school PA
approach   178 156 87.6
C: regular programme of PE and
school‐based PA   90 81 90.0
total: 268 237 88.4
Salmon 2008
(cluster‐RCT) I: 1, 2, 3 behavioural modification group; fundamental motor skills group; combined behavioural modification and fundamental motor skills group   233 213 91.4
C: usual classroom lessons   62 55 88.7
total: 295 268 90.8
Wang 2008
(cluster‐RCT) I: 'FitKid' after school intervention sessions   603 260 43.1
C: —   584 265 45.4
total: 1187 525 44.2
Webber 2008
(cluster‐RCT) I: health education lessons to enhance
behavioural skills known to influence
PA participation (self‐monitoring, setting
goals for behaviour change)  
C: —  
total: 3502 3378 96.5
Weeks 2008 (RCT) I: directed jumping activity at the beginning
of every PE class   52 43 82.7
C: regular PE warm‐ups and stretching
at the beginning of every PE class   47 38 80.9
total: 99 81 81.8
Barbeau 2007 (RCT) I: after school PA programme   81
C: —   84
total:
Williamson 2007
(cluster‐RCT) I: Healthy Eating and Exercise programme
to increase PA during the school day and at home   313 282 90.1
C: Alcohol/Drug/Tobacco abuse prevention
programme   348 304 87.4
total: 661 586 88.7
Haerens 2006
(cluster‐RCT) I1: a computer‐tailored intervention
to increase MVPA to 60 minutes/d, increase
fruit consumption, increase water
consumption, and reduce fat   2105
I2: group 1 plus parental involvement  
C: no PA and nutrition intervention   735
total: 2840 2434 85.7
Young 2006 (RCT) I: PE curriculum taught 5 days/week
and family   116 111 95.7
C: standard PE class   105 99 94.3
total: 221 210 95.0
Bayne‐Smith 2004 (RCT) I: Physical Activity and Teenage Health
programme, education sessions plus
20 to 25 minutes of PA   310
C: same frequency or duration of PE
classes, but without lecture or discussion   132
total: 442
Simon 2004
(cluster‐RCT) I: an educational component focusing
on PA and sedentary behaviours and
new opportunities for PA during and
after school hours   475 —
 
C: —   479
total: 1046 954 91.2
Trevino 2004
(cluster‐RCT) I: health programming regarding 3
health behaviour messages associated
with diabetes mellitus control and
goal‐setting   969 619 63.9
C: —   1024 602 58.8
total: 1993 1221 61.3
Stone 2003
(cluster‐RCT) I: food service, skills‐based classroom
curricula, family, and PE   879 644 73.3
C: —   825 653 79.2
total: 1704 1297 76.1
Burke 1998
(cluster‐RCT) I1: standard PA and nutrition programme
including classroom lessons,
fitness sessions daily, and nutrition
programme  
I2: I1 plus a PA enrichment programme
for higher‐risk children  
C: no programme  
total: 800 720 90.0
Ewart 1998 (RCT) I: 50‐minute 'Project Heart' aerobic exercise classes   45 44 97.8
C: 50‐minute standard PE classes   54 44 81.5
total: 99 88 88.9
Luepker 1996
(cluster‐RCT) I1: school food service modifications,
PE interventions, and Child and Adolescent
Trial for Cardiovascular Health
curricula   3651 3297 90.3
I2: I1 plus a family‐based programme  
C: usual health curricula, PE, and food
service programmes   1455 722 49.6
total: 5106 4019 78.7
Bush 1989
(cluster‐RCT) I1: 'Know Your Body' curriculum focusing
on nutrition, fitness, prevention of
smoking, a personalised health screening,
and results on a 'health passport'
for parentsg  
I2: 'Know Your Body' curriculum and
health screening, but students do not
receive the results of their screening;
only parents receive the results  
C: health screening only  
total: 892 431 48.3
Walter 1988
(cluster‐RCT) I: special curriculum targeting voluntary
changes in risk behaviour in the
areas of diet, PA, and smoking — / 3388 2075 1104 53.2
C: — 1313 665 50.6
total: 3388 1769 52.2
Grand total All interventions   46 073 28 089  
All comparators 40 566 23 639
All interventions and comparators b 96 740 66 752

—: denotes not reported.

aFollow‐up under randomised conditions until end of trial (= duration of intervention + follow‐up post intervention or identical to duration of intervention); extended follow‐up refers to follow‐up of participants once the original trial was terminated as specified in the power calculation.

bNote that numbers from all interventions and all interventions and comparators are greater than the sum of interventions only and comparators only, as some studies reported only the total number of included participants and did not note numbers within each group.

C: comparator; I: intervention; MVPA: moderate to vigorous physical activity; NA: not applicable; PA: physical activity; PE: physical education; RCT: randomised controlled trial.

We emailed all authors of included trials to enquire whether they would be willing to answer questions regarding their trials. We thereafter sought relevant missing information on the trial from the primary trial author(s), if required.

Dealing with duplicate and companion publications

In the event of duplicate publications, companion documents, or multiple reports of a primary trial, we maximised the information yield by collating all available data, and we used the most complete data set aggregated across all known publications. We listed duplicate publications, companion documents, multiple reports of a primary trial, and trial documents of included trials (such as trial registry information) as secondary references under the study identifier (ID) of the included trial. Furthermore, we listed duplicate publications, companion documents, multiple reports of a trial, and trial documents of excluded trials (such as trial registry information) as secondary references under the study ID of the excluded trial.

Assessment of risk of bias in included studies

For this review, all newly included studies were assessed independently for risk of bias by two review authors (SNS, HC) according to six domains (sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, incomplete outcome data, and selective reporting) using the ‘Risk of Bias 1’ tool in the Cochrane Handbook for Systematic Reviews of Interventions to assign assessment of low, high, or unclear risk of bias (for details, see Appendix 4) (Higgins 2017). Incomplete outcome data were rated separately for (1) device‐based measures of physical activity and sedentary time, and (2) BMI and fitness, due to the large proportion of missing data that is often seen in studies using accelerometers. For studies that used a cluster‐randomised design, four additional domains (recruitment bias, baseline imbalance, loss of clusters, and incorrect analysis) were assessed. All disagreements were resolved through discussion. If adequate information was unavailable, trial authors were contacted to request missing data on ‘Risk of bias’ items.

Two main changes were made to the assessment of risk of bias for this update. First, due to changes to Cochrane recommendations, blinding is now assessed separately for participants and personnel, and outcome assessors. Previously included studies were re‐assessed for this domain. In this update, we do not include assessment for control for confounders or reliability and validity of data collection methods. Second, previously included studies that utilised a cluster‐randomised design were appraised for the four new cluster‐randomised trial domains. Finally, assessment of 'other' risk of bias has been removed.

Risk of bias for an outcome across trials and across domains

These are the main summary assessments that we incorporated into our judgements about quality of evidence in Table 1. We defined outcomes as at low risk of bias when most weight in the meta‐analysis comes from trials at low risk of bias, unclear risk when most weight in the meta‐analysis comes from trials at low or unclear risk of bias, and high risk when a sufficient proportion of information comes from trials at high risk of bias. We defined single studies as having low risk of bias when all but one domain was assessed to be at low risk of bias, or when all but two domains were assessed at low risk of bias, one of which was 'blinding of participants and personnel', as adequate blinding of study participants is nearly impossible to do well in school‐based physical activity interventions. We defined single studies to be at high risk of bias when at least three domains were assessed as having unclear or high risk of bias, or when four domains were assessed as having unclear or high risk of bias, one of which was 'blinding of participants and personnel'. 

Measures of treatment effect

When at least five included trials were available for comparison of a given outcome measured on the same scale (e.g. MVPA in minutes/d), we combined effects into a meta‐analysis. When available, we extracted data on post‐intervention values adjusted for baseline differences, along with confidence intervals or standard deviation. When only changes from baseline were reported, these differences, along with confidence intervals or standard deviations, were extracted. Both were combined in meta‐analyses, with the calculator function in RevMan 5.4 used to calculate adjusted between‐group difference, as outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2021).

When possible, we combined multiple study groups into a single pairwise comparison using formulae for combining continuous data from multiple groups as provided in the Cochrane Handbook for Systematic Reviews of Interventions when separate data were presented for each group, or when between‐group comparisons from multiple group had independent control groups (e.g. when results were presented separately for boys and girls or by grade level) (Higgins 2017). When between‐group differences were reported between multiple intervention groups and the same control group, we did not include these data in the meta‐analysis, so as not to double‐count participants in control groups.

Unit of analysis issues

We considered the level at which randomisation occurred, such as cluster‐randomised trials and multiple observations for the same outcome. If more than one comparison from the same trial was eligible for inclusion in the same meta‐analysis, we used the end of intervention comparison.

Standard errors for cluster‐RCTs that  were not appropriately adjusted for potential clustering of participants within clusters in analyses were adjusted using the design effect (1 + (M‐1) × ICC), where M is the average cluster size, and ICC is the intraclass correlation coefficient (Higgins 2017). The ICC was estimated at 0.01, as has been previously reported (Murray 2006).

Dealing with missing data

If possible, we obtained missing data from authors of included trials. We carefully evaluated important numerical data such as screened, randomly assigned participants, as well as intention‐to‐treat and as‐treated and per‐protocol populations. We investigated attrition rates (e.g. dropouts, losses to follow‐up, withdrawals), and we critically appraised issues concerning missing data and use of imputation methods (e.g. last observation carried forward). When included trials did not report sufficient data for inclusion in the meta‐analysis (e.g. provided only P values) and we did not receive requested information from trial authors, we did not include these studies in the meta‐analysis; however, we included them in the narrative tables.

Assessment of heterogeneity

We identified heterogeneity (inconsistency) by visually inspecting forest plots and by using a standard Chi² test with a significance level of α = 0.1 (Deeks 2021). In view of the low power of this test, we also considered the I² statistic, which quantifies inconsistency across trials to assess the impact of heterogeneity on the meta‐analysis (Higgins 2002; Higgins 2003). When we found heterogeneity, we attempted to determine possible reasons for this by examining individual trial and subgroup characteristics in the narrative summary.

Assessment of reporting biases

If we included 10 or more trials that investigated a particular outcome, we planned to use funnel plots to assess small‐trial effects. Several explanations may account for funnel plot asymmetry, including true heterogeneity of effect with respect to trial size, poor methodological design (and hence bias of small trials), and publication bias (Sterne 2017).

Data synthesis

We planned to undertake (or display) a meta‐analysis only if we judged participants, interventions, comparisons, and outcomes to be sufficiently similar to ensure an answer that is clinically meaningful. We used random‐effects meta‐analyses to calculate the mean of the effects of included studies (Borenstein 2017aBorenstein 2017bHiggins 2009). We performed statistical analyses according to the statistical guidelines presented in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2021).

When studies could not be included in the meta‐analysis, for example, because they did not report the required data or they measured the outcome in a format that was incompatible with other studies, we summarised results from each study under Results and in table form. We used vote counting based on the direction of effect to determine whether the majority of studies found a positive or negative effect (McKenzie 2021).

Subgroup analysis and investigation of heterogeneity

Given the differences between children and adolescents with respect to school setting, types of interventions, etc., we explored the effects of interventions on these two age groups separately, using the subgroup function in Review Manager 5.4.

Given the substantial heterogeneity in types of interventions included, we conducted secondary analyses to synthesise results with respect to each outcome type by subcategories of intervention types. For this purpose, we broadly categorised each intervention type as primarily focused on (1) before or after school physical activity programmes, for example, after school dance clubs, walk to school programmes; (2) enhanced PE classes that focused on increasing the frequency, duration, intensity, or types of activities above and beyond the usual PE classes; (3) multi‐component interventions that typically included a whole‐school approach and utilised multiple strategies including environmental changes to increase physical activity, and often promoted healthy eating or healthy weight amongst students; and (4) schooltime physical activity interventions, such as active academic lessons or cycling desk interventions that focused on integrating physical activity throughout the school day itself rather than restricting physical activity to recess or PE classes.

Sensitivity analysis

We performed no sensitivity analyses.

Summary of findings and assessment of the certainty of the evidence

We used the GRADE approach to assess overall certainty of evidence for each of the primary and secondary outcome measures. GRADE takes into account issues related to both internal and external validity to state how confident we are in the effect estimates presented. Two review authors (SNS, HC) independently rated the certainty of evidence for each outcome. We resolved differences in assessment by discussion. For each outcome, we rated evidence certainty as very low, low, moderate, or high based on the GRADE domains as described in Chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2021). As only RCTs were included, the starting point for certainty of evidence was high. Then we considered each GRADE domain to determine whether downgrading of certainty was needed based on:

  • risk of bias ‐ based on critical appraisal using the Cochrane 'Risk of bias' tool;

  • inconsistency of results ‐ based on visual inspection of forest plots and I² in meta‐analyses and consistency of effects reported across narrative syntheses;

  • indirectness ‐ based on the validity of outcome measures used and how directly they measured the outcome of interest, for example, studies that measure VO₂ peak as a direct measure of fitness provide greater certainty than studies that report estimated fitness based on functional tests;

  • publication bias ‐ based on a small number of studies or indication of publication bias based on funnel plots; and

  • imprecision ‐ based on width of the confidence intervals, and whether they include the possibility of a small or null effect.

We presented a summary of the evidence in Table 1. This provides key information about the best estimate of the magnitude of effect as absolute differences for each relevant comparison, numbers of participants and trials addressing each important outcome, and a rating of overall confidence in effect estimates for each outcome. We created the 'Summary of findings' table using the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2021), along with Review Manager (RevMan 5.4) table editor (RevMan 2014). We reported the following outcomes, listed according to priority.

  • MVPA (proportion meeting guidelines, duration).

  • Sedentary time.

  • Physical fitness.

  • BMI.

  • Health‐related quality of life.

  • Adverse events.

Results

Description of studies

For a detailed description of trials, see Table 2, Characteristics of included studies, Characteristics of excluded studies, and Characteristics of ongoing studies sections.

Results of the search

A total of 16,219 records related to physical activity interventions with children and adolescents were identified for the last update. Of these, 587 studies were assessed for eligibility, and 44 were deemed to meet the criteria for inclusion in the review. The most common reasons studies were judged as not relevant were data on relevant outcomes not reported, studies not RCTs, and studies not relevant to public health.

For this update, when the new inclusion and exclusion criteria were applied (i.e. only studies in which the primary aim was to increase levels of physical activity and that included an objective measure of physical activity, physical fitness, or body composition), 16 of the original 44 studies were excluded: 11 included only self‐report measures (Araujo‐Soares 2009a; Colin‐Ramirez 2010; Dishman 2004; Haerens 2006; Jones 2008; Kipping 2008; Lubans 2009; McManus 2008; Petchers 1988; Singhal 2010; Verstraete 2006); three did not include any measure of physical activity or physical fitness (Li 2010; Martinez 2008; Stephens 1998); and two did not have a primary objective to increase physical activity (Robinson 1999; Singh 2009). Therefore, this update includes 28 of the 44 studies included in the last update (Angelopoulos 2009; Barbeau 2007; Bayne‐Smith 2004; Burke 1998; Bush 1989; Donnelly 2009; Dorgo 2009; Ewart 1998; Gentile 2009; Haerens 2006; Kriemler 2010; Luepker 1996; Neumark‐Sztainer 2009; Neumark‐Sztainer 2010; Peralta 2009; Reed 2008; Salmon 2008; Simon 2004; Stone 2003; Trevino 2004; Walter 1988; Walther 2009; Wang 2008; Webber 2008; Weeks 2008; Williamson 2007; Wilson 2011; Young 2006).

The search strategy for this update from October 2011 to June 2020 yielded 9968 records. Of these, 978 unique full texts were assessed for eligibility. A total of 61 studies met all relevance criteria and were included in this update (Aburto 2011; Adab 2018; Andrade 2014; Ardoy 2011; Belton 2019; Breheny 2020; Carlin 2018; Cohen 2015; Corepal 2019; Daly 2016; de Greeff 2016; de Heer 2011; Donnelly 2017; Drummy 2016; Fairclough 2013; Farmer 2017; Ford 2013; Grydeland 2013; Harrington 2018; Have 2018; Ickovics 2019; Jago 2011; Jago 2014; Jago 2015; Jago 2019; Jansen 2011; Jarani 2016; Ketelhut 2020; Kipping 2014; Kobel 2014; Kocken 2016; Lau 2016; Leahy 2019; Lonsdale 2019a; Madsen 2015; Magnusson 2011; Martinez‐Vizcaino 2014; Melnyk 2013; Müller 2019; Muros 2015; Nogueira 2014; Okely 2011; Ordóñez Dios 2019; Pablos 2018; Resaland 2016; Robbins 2018; Sacchetti 2013; Santos 2014; Seibert 2019; Seljebotn 2019; Siegrist 2013; Siegrist 2018; Suchert 2015; Sutherland 2016; Sutherland 2017; Tarp 2016; Ten Hoor 2018; Thivel 2011; Toftager 2014; Torbeyns 2017; Zhou 2019). The most common reasons studies were judged as not relevant were (1) studies were not RCTs; (2) no device‐based measure of physical activity or physical fitness was included; and (3) the intervention lasted < 12 weeks. This update includes 89 studies (28 from the last review and 61 from this update). A flow diagram depicting these results is presented in Figure 1.

A total of 84 authors were contacted for missing information, and 126 responses were received (median 1 response, range 0 to 4). In most cases, study authors were able to provide clarification on risk of bias or study characteristics, but few of them provided updated outcome data.

Included studies

In addition to information included in the Characteristics of included studies table, we have presented greater detail for each study in Table 2 (overview of study populations) and in Appendix 2 (baseline characteristics). The following is a succinct overview.

Trial design

All included studies were RCTs, with nine randomising participants at the level of the individual, and 80 using a cluster design, whereby schools or classrooms were the unit of randomisation. Across comparator groups, a majority of comparator groups (n = 67) specified continuation of usual school curriculum, standard PE, or no intervention; others reported offering a delayed intervention (n = 3). Some comparator groups included an intervention unrelated to physical activity, such as nutrition education, theatre play group, or education about alcohol or tobacco use (n = 10), and 13 of the comparator groups were not clearly described. The number of included schools ranged from 1 to 96. Most trials were multi‐centre (n = 81), and only 8 trials were conducted within a single school. A total of 31 trials were conducted in 2 to 10 schools, 27 trials in 11 to 20 schools, and 23 trials in more than 20 schools. A majority of trials were not double‐blinded; only 9 of 89 trials reported blinding participants, personnel, and outcome assessors; 6 trials reported blinding participants and personnel but not outcome assessors; and 24 trials reported blinding outcome assessors but not participants or personnel; the remaining 50 trials did not report blinding at all. Trials were conducted from 1983 to 2018. Duration of intervention ranged from 12 weeks to 6 years. Trials were most commonly 12 weeks to 6 months in duration (n = 31), followed by longer than 6 months to 1 year (n = 29), 1 to 2 years (n = 17), and longer than 2 years (n = 12). A majority of studies evaluated only outcomes immediately following the intervention (n = 72); others collected additional data during post‐intervention follow‐up (n = 17). Post‐intervention follow‐up ranged from 2 weeks to 1 year. No trials described a run‐in period, and no trials reported that they were terminated before the planned end of study.

Participants

Across all studies, 96,740 participants were randomised, with at least 46,073 individuals in intervention groups and 40,566 in comparator groups, as not all studies reported the exact numbers randomised to each group. A total of 66,752 participants completed the trials and were included in the analyses. The number of participants randomised ranged from 33 to 11,158, and the number of participants completing trials ranged from 32 to 4063. The average percentage of participants completing the trials was 69.0%, ranging from 35.7% to 100%. Within intervention groups, the average percentage of participants completing the trial was 60.9%, and within control groups, 58.3%, when reported.

A majority of studies were conducted in children 12 years of age or younger at baseline (n = 56); others included only adolescents between the ages of 12 and 18 (n = 22), and some included both children and adolescents (n = 10). One study did not report the age of participants. Most included studies were conducted in the USA (n = 26), Australia (n = 12), and the UK (n = 9). Other countries included Germany (n = 6), Spain (n = 5), The Netherlands (n = 4), Denmark (n = 3), Norway (n = 3), Northern Ireland (n = 3), Belgium (n = 2), Canada (n = 2), China (n = 2), and France (n = 2), and one study each from Albania, Ecuador, Greece, Iceland, Ireland, Italy, Mexico, New Zealand, South Africa, and Switzerland. A range of ethnic groups was represented across trials; however, ethnicity was not reported in 40 of the 89 included studies. Most studies included both male and female students and reported a roughly even split between genders; one study included male students only, 11 included female students only, and 4 did not report the breakdown of male and female students.

Interventions

All studies had intervention components that were delivered in the school setting. Some projects provided additional interventions in the home, community, local theatre, or after school programmes, or via the computer. All studies included a control group that represented a school or a group of schools from a different community, city, or state that did not receive the school‐based intervention. However, in some studies, control schools received other physical activity promotion interventions provided through other health organisations or venues or by a standard PE curriculum. The duration of interventions varied greatly from a minimum of 12 weeks to 6 years, with 10 studies reporting intervention periods of 3 years or longer (Bush 1989Daly 2016Donnelly 2009Donnelly 2017Ickovics 2019Luepker 1996Simon 2004Stone 2003Walter 1988Wang 2008). Several theoretical frameworks were used to develop the physical activity interventions, with some studies citing more than one framework. In 36 studies, it is unclear if a theoretical framework had been used to design and/or deliver the intervention. The most commonly cited theoretical models were social cognitive theory (n = 20), a socioecological model (n = 11), self‐determination theory (n = 10), and the theory of planned behaviour (n = 6). Studies reported in this review differed in funding levels, numbers of project staff, and resources available to deliver the programmes. Further, although all projects were primarily school based, no projects used the same combination of interventions with the same intensity, making each programme unique; however, some similarities were observed with respect to the ways in which interventions were delivered. Most commonly, interventions were multi‐component, whole‐school interventions that included a combination of educational materials, changes to the school environment, and/or school curriculum; and they targeted students, teachers, and/or parents (n = 40). Other interventions (n = 19) were focused primarily on providing opportunities for MVPA within school time, such as active academic lessons. All but one of these interventions targeted children rather than adolescents. Also common were interventions that enhanced the usual school PE programme (n = 15) by incorporating high‐intensity activity into PE classes or increasing the frequency or duration of PE classes. Finally, other interventions included additional opportunities for physical activity before school activities (such as walking groups), lunchtime physical activity programmes, or after school programmes within the school environment (n = 14). One study used a factorial design, comparing enhanced PE and/or an after school programme to usual school activities.

Comparisons

Across 89 trials, a total of 93 comparison groups were described, as each of four studies reported two comparison groups. Most often, investigators described the comparison group as continuing with normal school activities or regular school physical activity or PE without specifying what that might include (40 studies). Sixteen studies described what the typical physical activity in a school would be, which ranged from one PE class per week to two hours of PE per week. Thirteen comparison groups were simply described as ‘no intervention’ or participating only in data collection, with no indication of whether physical activity or PE was a part of the regular school setting. Ten studies described alternative or sham interventions, such as an alcohol and drug abuse prevention programme, with health screening only. One study used a spillover group as a second comparator group, which comprised students who were eligible but declined to participate in the intervention. The remaining 13 studies provided no description of the comparator group.

Outcomes

A protocol paper or trial registry was available for 59 of the 89 included trials; for the remaining 30 trials, a trial document was not identified. Within the 59 trial documents, a single primary outcome was specified in 38 trials; 17 trials documented multiple primary outcomes, and 4 trial documents did not specify a primary outcome. When a single primary outcome was stated, 17 were measures of MVPA, 8 were measures of BMI, 4 were measures of fitness, and 9 involved other endpoints, including other measures of body composition, academic achievement, feasibility, executive function, diabetes, and screen time.

Of the 55 studies that specified a primary outcome in their trial documents, 38 reported the same primary outcome in the publication, 12 specified a different primary outcome in the publication, and 5 did not specify a primary outcome in the publication at all.

Physical activity

A total of 38 of the 89 included trials reported some measure of activity using accelerometers. A number of different devices and protocols were used. Participants were asked to wear the accelerometer for anywhere from 3 to 9 days, but most often (n = 21), participants were asked to wear the accelerometer for 7 days. Three studies did not report accelerometer wear time. Most studies had participants wear the accelerometer for both weekday and weekend days (n = 29); however others recorded only weekday activity (n = 3), and 6 did not specify whether weekend days were included.

Five studies reported on the proportion of students who were physically active, and all studies used the criterion of reaching more than 60 minutes of MVPA per day. Each of two studies measured activity using the Actiheart and Actigraph accelerometers, and 1 used the GENEactive. The specific cut points used to classify activity as MVPA were not reported in 2 studies using the Actiheart (Adab 2018Kobel 2014); different cut points were used in the other 3 studies using Actigraph and GENEactive accelerometers.

Of the 38 studies that reported on duration of MVPA, most used a model of Actigraph accelerometer (n = 26), and 7 did not report the type of accelerometer model used. Other models include the MTI (n = 2), Actiheart (n = 2), Minimeter (n=1), and GENEactive (n=1). The most commonly used cut points for classifying MVPA were Evenson cut points (17 studies), and 12 studies did not report the cut points used to classify MVPA. Across the remaining 9 studies, a variety of different cut points were reported.

Sedentary time

Sedentary time was measured via accelerometer in 20 studies. The most common cut points used to categorise time spent in sedentary behaviour were Evenson cut points of fewer than 100 counts per minute (n = 9); 6 studies did not specify the cut points used, and 5 studies reported other cut points.

Fitness

Objective physical fitness assessments were reported in 42 studies. Field‐based running tests were used most often, with 13 studies using the Progressive Aerobic Cardiovascular Endurance Run (PACER) test by Leger et al, and an additional 10 studies using a 20‐metre shuttle run protocol but not specifying whether it was the PACER protocol. Studies that used a shuttle run reported outcomes as number of laps completed, estimated VO₂max, age, and sex‐specific z‐scores for number of laps, highest level reached, and/or number of stages completed. Six studies reported using the Anderson 10‐minute interval test with distance run and estimated VO₂max as the outcome variable. One study used a 1‐kilometre run, 1 used a mile run, and 1 used a 9‐minute run protocol. Incremental treadmill tests with gas analysis were used in 3 studies, expressed as VO₂max, and 1 employed a peak power test on a cycle ergometer, expressed as Watts per kilogram of body mass. Last, 2 studies used the Queens College Step Test, 1 used the bench‐stepping test, 1 used the Harvard step test, 1 used the British Athletics Linear Track Test, and 1 used a 6‐minute run test.

Body mass index

Seventy‐one studies reported on BMI using objective measures. The most common expression of BMI, reported in 49 studies, was as kg/m². Three studies used country‐specific z‐scores (England and Germany), and 2 studies used German‐specific BMI percentile values. WHO z‐scores were used in 2 studies, Centers for Disease Control and Prevention z‐scores were used in 3 studies, and percentiles were used in an additional 3 studies. One study used the International Obesity Task Force cutoffs for weight status. Twelve studies used z‐scores but did not specify the source, 6 studies did not describe methods, and 1 study reported percentage body fat.

Health‐related quality of life

Only 7 studies reported some aspect of health‐related quality of life; a summary of the instruments used can be found in Appendix 5. Only one tool ‐ the Child Health Utility 9D ‐ was used in more than 1 study (Breheny 2020Harrington 2018Jago 2019).

Adverse events

Adverse events were not commonly reported in studies. Only 16 of 89 included studies provided any information about adverse events, most commonly to say that no adverse events were noted. Only 3 studies reported data about the number and nature of adverse events that occurred during the study in the intervention or control group.

Excluded studies

For this update, we excluded a total of 736 studies after full‐text review. The most common reason for exclusion was that trials were not randomised trials or trials did not include an objective measure of physical activity or physical fitness. Reasons for exclusion of studies from this update are available in the Characteristics of excluded studies table.

Risk of bias in included studies

For details on the risk of bias of included trials, see Characteristics of included studies and a summary across trials in Figure 2 and Figure 3.

2.

2

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included trials (blank cells indicate that the particular outcome was not measured in some trials).

3.

3

Risk of bias summary: review authors' judgements about each risk of bias item for each included trial (blank cells indicate that the particular outcome was not measured in some trials).

Allocation

All included studies were RCTs, and a majority of included studies provided an adequate description of the methods used to generate the randomisation sequence (e.g. computer randomisation) and the methods used to conceal allocation from participants and personnel. Sixty‐seven of 89 studies adequately described randomisation sequence, 18 of 89 studies were unclear in their reporting, and 4 of 89 studies did not report an adequate method for creating the randomisation sequence and were described as high risk of bias in this domain. Sixty‐three of 89 studies adequately described allocation concealment, 15 of 89 were unclear or did not describe allocation concealment well, and in 11 of 89 studies, allocation concealment was not performed, introducing high risk of bias in this domain.

Blinding

The most notable methodological weakness of these studies is the lack of blinding of participants and personnel, with only 15 out of 89 studies adequately reporting that participants and study personnel were blinded to group allocation. This was primarily done by not informing participants about the overall goal of the study or the presence of other intervention arms. Blinding of participants and personnel was unclear in 17 of 89 studies, and in 57 of 89 studies, blinding of participants or personnel was not done, introducing high risk of bias in this domain. Adequate blinding of outcome assessors was described in 33 of 89 trials and was unclear in 22 of 89 trials. In 34 of 89 trials, outcome assessors were not blinded, introducing high risk of bias in this domain. All outcomes included in this review were objective measures and did not rely on self‐report, therefore reducing but not eliminating the potential for bias.

Incomplete outcome data

We assessed incomplete outcome data separately for physical activity and sedentary time outcomes and for physical fitness and anthropometric data. Among studies that assessed physical activity participation or duration, risk of bias due to incomplete outcome data was deemed low in 20 out of 40 studies, unclear in 4 of 40 studies, and high in 16 of 40 studies. Among studies that assessed sedentary time, 11 of 20 studies were deemed at low risk of bias for incomplete outcome data, 2 of 20 studies were unclear, and 7 of 20 studies had large quantities of missing data that introduced high risk of bias.

For studies that measured physical fitness, 23 of 42 studies were deemed at low risk of bias, 5 of 42 studies were unclear in their reporting of outcome data completion, and 14 of 42 had large quantities of missing data that introduced high risk of bias. Among studies that measured BMI, 40 of 74 studies were deemed at low risk of bias for completion of outcome assessment, 5 of 74 studies were unclear in the quantity of missing data, and 29 of 74 had large quantities of missing data that introduced high risk of bias.

Selective reporting

Forty‐six of 89 studies were deemed at low risk of bias for selective outcome reporting and reported on all of the outcomes specified in trial protocols or published protocol papers. Twenty‐nine studies were unclear in their selective outcome reporting, often because a protocol paper was not published, and 14 studies were deemed at high risk of bias for selective outcome reporting.

Other potential sources of bias

Clusterrandomised trials

A total of 80 included trials were cluster‐randomised trials; therefore, risk of bias was appraised within four additional categories (recruitment bias, baseline imbalances, loss of clusters, and incorrect analysis). With respect to recruitment bias, 35 of 80 trials were deemed at low risk of bias, 13 were unclear in the timing of recruitment and randomisation, and 32 were at high risk of bias, often because schools or classes were aware of their intervention status prior to participant enrolment in the trial. With respect to baseline imbalance, 63 of 80 trials were deemed at low risk of bias, 11 of 80 were unclear, and 6 of 80 were deemed at high risk of bias. For loss of clusters, 54 out of 80 were deemed at low risk of bias, as they retained all clusters in the trials, 4 of 80 were unclear, and 22 were deemed at high risk of bias due to loss of clusters throughout the trial. Finally, with respect to incorrect analysis of cluster‐RCTs, 62 of 80 were deemed at low risk of bias, as they properly accounted for the clustered nature of the data in their statistical analysis, and 18 trials were deemed at high risk of bias for failing to incorporate clustering into their analyses.

Effects of interventions

See: Table 1

See Table 1.

Effects of school‐based physical activity interventions on primary outcomes

Physical activity participation

Overall, we are very uncertain about the effects of school‐based physical activity programmes on the proportion of students meeting physical activity guidelines due to inconsistency of effects between studies, imprecision around the effects, and risk of bias in the included studies contributing to this outcome.

This outcome was reported in only 5 studies with quite different interventions (Analysis 1.1). One study explored the effects of after school dance classes on the proportion of girls meeting physical activity guidelines (Jago 2015). At the end of the 20‐week study, between‐group differences in adherence to guidelines were found to be ‐1.11% (95% confidence interval (CI) ‐1.68 to ‐0.73) in the intervention group compared to the control group. One study found that although both groups had fewer adolescents meeting guidelines at the end of study, the decline was noted to be smaller in the intervention group than in the control group; however confidence intervals were not reported (difference 12.22%; P < 0.01) (Andrade 2014). Another study found an uncertain effect on the odds of meeting the guidelines in the intervention group compared to the control group at end of study (odds ratio (OR) 0.65, 95% CI 0.23 to 1.85) (Harrington 2018). One study found a similar proportion of students meeting the guidelines at 15 or 18 months following a whole‐school physical activity and nutrition intervention (difference at 15 months 0.005%, 95% CI ‐0.101 to 0.140) (Adab 2018). One study found that after one year, differences in the proportion of participants meeting guidelines between intervention and control groups were 10.4%; study authors noted that the difference between groups was not statistically significant but included no measure of variation (Kobel 2014).

1.1. Analysis.

Comparison 1: PA programme vs no PA programme, Outcome 1: Physical activity participation: all data

Physical activity participation: all data
Study Study population Intervention group Control Group Measurement period Overall effect Comment
Children: before and after school programme
Jago 2015 Year 7 female students After‐school dance classes Provided data only 20 weeks Between group difference in % meeting 60 min/day of MVPA: −1.11% (95% CI −1.68, −0.73)  
      1 year Between group difference in % meeting 60 min/day of MVPA: −1.18% ( (95% CI −1.82, 0.76)  
Children: schooltime PA
Kobel 2014 Pupils at primary school, grades 1 and 2 Teacher training, PA education, and active breaks No intervention 1 year Between‐group difference in % meeting 60 min/day of MVPA: 10.4% (not statistically significant)
Intervention group: 54.7%
Control group: 44.3%
 
           
Children: multi‐component interventions
Adab 2018 Year 1 students (aged 5 to 6 years) 30 min of additional MVPA on each school day, cooking workshops, a 6‐week healthy eating program, and information sheets for families Ongoing year 2 health related activities and education resources, excluding topics related to healthy eating and PA 15 months Between‐group difference in % meeting 60 min/day of MVPA: 0.005% (95% CI −0.101, 0.140)  
      18 months Between‐group difference in % meeting 60 min/day of MVPA: −0.067% (95% CI −0.165, 0.096)  
Adolescents: multi‐component interventions
Andrade 2014 Grades 8 and 9 students ACTIVITAL individual‐ and environmental‐based intervention Standard curriculum 28 months Between‐group difference in % meeting guidelines: 12.22% (P < 0.01)
Intervention group: ‐5.87%
Control: ‐18.09%
 
           
Harrington 2018 Female students in years 7 to 9, 11 to 14 years old Support for PA, PE, and school sport culture and practices with the support of the Youth Sport Trust and a hub school Continued with normal PA habits 7 months (midpoint) Meeting guidelines: OR: 0.78 (0.23, 2.65)  
      14 months Meeting guidelines: OR: 0.65 (0.23, 1.85)  
Physical activity duration

Overall, school‐based physical activity interventions probably have little to no effect on minutes per day of MVPA among children and adolescents (mean difference (MD) 0.73, 95% CI 0.16 to 1.30; 33 studies; Analysis 1.2 moderate‐certainty evidence). These findings should be interpreted with caution due to the inconsistency in outcomes reported based on visual inspection of forest plots and substantial heterogeneity across studies (I² = 75%).

1.2. Analysis.

1.2

Comparison 1: PA programme vs no PA programme, Outcome 2: Physical activity duration (minutes/d): meta‐analysis

Six additional studies provided data that were not included in the meta‐analysis. Most findings were consistent with results from the meta‐analysis (Analysis 1.4). Following one year of PE enhanced with strength training and motivational interviewing among adolescents, a significant between‐group difference was found in the percentage of time spent in MVPA; however, the magnitude of this difference was not reported (Ten Hoor 2018). In a study of Grade 7 students, those who took part in a biweekly after school PA programme, enhanced PE, or both were found to increase the percentage of time spent in MVPA (MD 1.99%, 95% CI 1.68 to 2.30; MD 3.12%, 95% CI 2.76 to 3.48; MD 4.98%, 95% CI 4.62 to 5.34, respectively), which was noted by study authors as statistically significant (Zhou 2019).

1.4. Analysis.

Comparison 1: PA programme vs no PA programme, Outcome 4: Physical activity duration: additional data

Physical activity duration: additional data
Study Study population Intervention group Control Group Measurement period Overall effect Comment
Before and after school programmes
Zhou 2019 Junior high school students, grade 7 I1: Biweekly after school program Regular PE (2 days per week) 32 weeks % of time spent in MVPA: MD 1.99 (95% CI: 1.68, 2.30)
I1: m 4.22 (sd 1.39)
Control: m 2.23 (sd 1.52)
 
 
 
% time reported, not min/day
  I2: Enhanced PE (3 days per week) plus after school program Regular PE (2 days per week) 32 weeks % of time spent in MVPA: MD 4.98 (95% CI: 4.62, 5.34)
I2: m 7.21 (sd 1.84)
Control: m 2.23 (sd 1.52)
Authors note that statistically significant changes were found from baseline to follow‐up in both intervention groups but not the control group however analyses were not described.
Enhanced PE
Ten Hoor 2018 Secondary school students, 11 to 15 years old Strength training and motivational interviewing Usual curriculum 1 year Between group difference in % of time spent in MPVA tested, P = 0.046 No values reported
           
Zhou 2019 Junior high school students, grade 7 I1: Enhanced PE (3 days per week) Regular PE (2 days per week) 32 weeks % of time spent in MVPA: MD 3.12 (95% CI: 2.76, 3.48)
I1: m 5.35 (sd 1.79)
Control: m 2.23 (sd 1.52)
% time reported, not min/day
  I2: Enhanced PE (3 days per week) plus after school program     % of time spent in MVPA: MD 4.98 (95% CI: 4.62, 5.34)
I2: m 7.21 (sd 1.84)
Control: m 2.23 (sd 1.52)
Authors note that statistically significant changes were found from baseline to follow‐up in both intervention groups but not the control group however analyses were not described.
Multi‐component intervention
Corepal 2019 Students age 12‐14 Pedometer challenge Usual school Baseline Intervention: m 33.3 min/day, IQR: 23.6, 49.0)
Control: m 43.6 min/day, IQR: 31.0, 69.3 Feasibility trial, therefore statistical testing not conducted
      22 weeks MD: ‐14.4 min/day
Intervention: m 33.0 min/day, IQR: 20.0, 46.2
Control: m 47.4 min/day, IQR: 32.7, 65.1
No estimates of variance given, only mean and interquartile range
Salmon 2008 Grade 5 students (10 to 11 years old) I1: Behavioral modification group,
I2: fundamental motor skills group,
I3: combined behavioral modification and fundamental motor skills group
Usual classroom lessons 1 school year Adjusted between group difference (vs. control group)
I1: MD 2.8 (95% CI 0.3, 5.4) min/day
I2: MD 7.8 (95% CI 3.4, 12.3) min/day
I3: MD 3.1 (95% CI ‐0.58, 6.7) min/day
Vigorous PA only
      1 year post‐intervention Adjusted between group difference (vs. control group)
I1: MD 2.8 (95% CI 0.2, 5.4) min/day
I2: MD 7.7 (95% CI 3.2, 12.2) min/day
I3: MD 3.0 (95% CI ‐0.59, 6.6) min/day
Vigorous PA only
Schooltime PA
Ford 2013 Primary school students aged 5 to 11 years Accumulated brisk walking program Normal school lessons 15 weeks Change from baseline
MD ‐27.4 (95% CI: ‐91.0, 36.2) counts per min
Intervention group: ‐29.2 (‐72.0, 13.6) counts per min, P = 0.415
Control group: ‐1.8 (‐50.3, 46.7) counts per min, P = 0.772
Weekday counts per min only
           
Magnusson 2011 Children attending grade 2 (born in 1999) Students engaged in PA during PE lessons, recess, and during classes; schools had access to PA equipment to use in school lessons; teaching materials promoting PA were provided Followed the general PA curriculum 1 year (midpoint) Group x time interaction in multivariable model P < 0.0001  
      2 years Group x time interaction in multivariable model P = 0.10  

In one study, changes in vigorous activity were reported from an intervention targeting behavioural modification (MD 2.8 minutes/d, 95% CI 0.3 to 5.4) or fundamental movement skills (MD 7.8 minutes/d, 95% CI 3.4 to 12.3 minutes/d), and combining behavioural modification and fundamental movement skills (MD 3.1 minutes/d, 95% CI ‐0.58 to 6.7) compared to control (Salmon 2008). Grade 2 children who engaged in physical activity during school, lessons, and recess were noted to take part in more MVPA at study midpoint but not at the end of the intervention; values were not reported (Magnusson 2011). Following implementation of a brisk walking programme during the school day, moderate to vigorous accelerometer counts were reported in the intervention group compared to a control group (MD ‐27.4 counts/min, 95% CI ‐91.0 to 36.2) (Ford 2013). Finally, within‐school and between‐school pedometer step challenges among adolescents age 12 to 14 years were found to be feasible, but minutes/d of MVPA appeared stable across groups throughout the intervention periods (MD ‐14.4 minutes/d, no measure of variance reported) (Corepal 2019).

In subgroup analyses, no differences in effects were found between interventions targeting children and adolescents (test for subgroup differences, P = 0.35; Analysis 1.2); however there were subgroup differences by intervention type (test for subgroup difference, P = 0.03; Analysis 1.3).

1.3. Analysis.

1.3

Comparison 1: PA programme vs no PA programme, Outcome 3: Physical activity duration by intervention type (minutes/d): meta‐analysis 

Children

In subgroup analyses, school‐based physical activity interventions targeting children probably do not result in a meaningful change in minutes per day of MVPA (MD 1.01 minutes/d, 95% CI 0.08 to 1.93; 22 studies; Analysis 1.2 moderate‐certainty evidence). These findings should be interpreted with caution due to substantial heterogeneity across studies (I² = 69%). Across studies that were not included in the meta‐analysis, findings were consistent (Analysis 1.4).

Adolescents

School‐based physical activity interventions probably do not result in a meaningful change in minutes per day of MVPA among adolescents (MD 1.84 minutes/d, 95% CI 0.34 to 3.35; 11 studies; Analysis 1.2; moderate‐certainty evidence), with substantial heterogeneity (I² = 81%); when differences were found, they were generally small in magnitude (Analysis 1.4).

Before and after school programmes

A total of 6 included studies implemented before and after school programmes in the school setting. Overall, before and after school programmes probably do not increase time spent in MVPA (MD 0.77 minutes/d, 95% CI ‐1.40 to 2.94; 6 studies; Analysis 1.3; moderate‐certainty evidence). Moderate heterogeneity was found across studies (I² = 50%).

Enhanced PE

Three studies tested the effects of enhanced PE interventions. These interventions probably do not change MVPA (MD ‐0.23 minutes/d, 95% CI ‐1.58 to 1.11; 3 studies; Analysis 1.3; moderate‐certainty evidence); however results should be interpreted with caution due to high heterogeneity (I² = 83%).

Multi‐component interventions

Multi‐component interventions in the school setting probably result in small increases in MVPA among children and adolescents (MD 2.42 minutes/d, 95% CI 0.62 to 4.22; 16 studies; Analysis 1.3; moderate‐certainty evidence); however results should be interpreted with caution due to high heterogeneity (I² = 76%).

Schooltime PA

Schooltime PA interventions probably do result in small increases in MVPA among children and adolescents (MD 5.30 minutes/d, 95% CI 0.89 to 9.72; 8 studies; Analysis 1.3; moderate‐certainty evidence); however results should be interpreted with caution due to high heterogeneity (I² = 73%).

Sedentary time

Identified evidence suggests that school‐based physical activity interventions may have little to no difference in minutes per day of sedentary time (MD ‐3.78 minutes/d, 95% CI ‐7.80 to 0.24; 16 studies; Analysis 1.5; low‐certainty evidence). These findings should be interpreted with caution due to imprecision of the effect estimate and risk of bias of included studies.

1.5. Analysis.

1.5

Comparison 1: PA programme vs no PA programme, Outcome 5: Sedentary time (minutes/d): meta‐analysis

Four additional studies provided data that were not included in the meta‐analysis. A majority of findings were consistent with meta‐analysis results, finding little or no effect (Analysis 1.7). Three studies reported sedentary time as an outcome within enhanced PE interventions in adolescents. In a study of teacher PE training, the between‐group mean difference in time spent sedentary at 7 months was 0.92% (95% CI ‐0.28 to 2.13) and was 0.02% (95% CI ‐0.99 to 0.95) at 14 months (Lonsdale 2019a). After one year of strength training and motivational interviewing, study authors note no statistically significant decrease in the percentage of time spent sedentary; however only P values were reported (Ten Hoor 2018). In a study of Grade 7 students, the change in percentage of time spent sedentary among those who took part in a biweekly after school PA programme, enhanced PE, or both was found to be 1.34% (95% CI ‐0.73 to 3.41), 1.11% (95% CI ‐1.09 to 3.31), and 0.11% (95% CI ‐2.31 to 2.54), respectively (Zhou 2019). Finally, within‐school and between‐school pedometer step challenges in adolescents age 12 to 14 years were found to be feasible, but sedentary time appeared stable across groups throughout intervention periods (MD 1.2 minutes/d; no measure of variance reported) (Corepal 2019).

1.7. Analysis.

Comparison 1: PA programme vs no PA programme, Outcome 7: Sedentary time: additional data

Sedentary time: additional data
Study Study population Intervention group Control Group Measurement period Overall effect Comment
Before and after school programmes
Zhou 2019 Junior high school students, grade 7 I1: Biweekly after school program Regular PE (2 days per week) 32 weeks % of time spent sedentary: MD 1.34 (95% CI: ‐0.73, 3.41)
I1: m 75.74 (sd 8.81)
Control: m 74.40 (sd 10.80)
% time reported, not min/day
  I2: Enhanced PE (3 days per week) plus after school program Regular PE (2 days per week) 32 weeks % of time spent sedentary: MD 0.11 (95% CI: ‐2.31, 2.54)
I2: m 74.51 (sd 11.95)
Control: m 74.40 (sd 10.80)
Authors note that statistically significant changes were found from baseline to follow‐up in the after school program only, however analyses were not described or presented
Enhanced PE
Lonsdale 2019a Grade 8 students Teacher PE training Standard teaching 7‐8 months % time spent sedentary: MD 0.92 (95% CI −0.28, 2.13)  % time reported, not min/day
      14‐15 months % time spent sedentary:  MD 0.02 (95% CI −0.99, 0.95)   
Ten Hoor 2018 Secondary school students, 11 to 15 years old Strength training and motivational interviewing Usual curriculum 1 year Between group difference in % of time spent sedentary, P = 0.715 % time reported, not min/day
           
Zhou 2019 Junior high school students, grade 7 I1: Enhanced PE (3 days per week) Regular PE (2 days per week) 32 weeks % of time spent sedentary: MD 1.11 (95%CI: ‐1.09, 3.31)
I1: m 75.51 (sd 9.64)
Control: m 74.40 (sd 10.80)
% time reported, not min/day
  I2: Enhanced PE (3 days per week) plus after school program Regular PE (2 days per week) 32 weeks % of time spent sedentary: MD 0.11 (95% CI: ‐2.31, 2.54)
I2: m 74.51 (sd 11.95)
Control: m 74.40 (sd 10.80)
Authors note that statistically significant changes were found from baseline to follow‐up in the after school program only group, however analyses were not described.
Multi‐component intervention
Corepal 2019 Students 12‐14 years old Pedometer challenge Usual school Baseline Intervention: m 449.6 min/day, IQR: 416.5, 524.3
Control: m 466.3 min/day, IQR: 410.0, 534.9
Feasibility trial therefore no statistical analyses performed
      22 weeks MD 1.2 min/day
Intervention: m 454.7 min/day, IQR: 405.7, 517.8
Control: m 453.5 min/day IQR: 399.8, 529.6
Variance cannot be calculated as only mean and interquartile range reported

In subgroup analyses, no differences in effects were found between children and adolescents (test for subgroup differences, P = 0.58; Analysis 1.5) or by intervention type (test for subgroup difference, P = 0.58; Analysis 1.6).

1.6. Analysis.

1.6

Comparison 1: PA programme vs no PA programme, Outcome 6: Sedentary time (minutes/d) by intervention type: meta‐analysis

Children

Overall, school‐based physical activity programmes may not reduce sedentary time among children (MD ‐3.35 minutes/d, 95% CI ‐9.30 to 2.60; 11 studies; Analysis 1.5; low‐certainty evidence), with low heterogeneity of 37% (Analysis 1.5).

Adolescents

School‐based physical activity programmes may result in little to no difference in sedentary time of adolescents (MD ‐5.67 minutes/d, 95% CI ‐11.48 to 0.14; 5 studies; Analysis 1.5; low‐certainty evidence), with no heterogeneity observed (I² = 0%) (Analysis 1.5).

Before and after school programmes

Before and after school programmes in the school setting may not decrease sedentary time (MD 2.01 minutes/d, 95% CI ‐15.28 to 19.31; 2 studies; Analysis 1.6; low‐certainty evidence). Moderate heterogeneity was found across studies (I² = 46%).

Enhanced PE

Only one study explored the impact of enhanced PE on sedentary time (MD ‐11.18 minutes/d, 95% CI ‐21.96 to ‐0.40; 1 study; Analysis 1.6; low‐certainty evidence).

Multi‐component interventions

Multi‐component interventions in the school setting may result in small decreases in sedentary time (MD ‐4.60 minutes/d, 95% CI ‐9.08 to ‐0.12; 11 studies; Analysis 1.6; low‐certainty evidence). No heterogeneity was noted (I² = 0%).

Schooltime PA

Schooltime PA interventions may not decrease sedentary time (MD ‐3.26 minutes/d, 95% CI ‐19.05 to 12.52; 2 studies; Analysis 1.6; low‐certainty evidence); however results should be interpreted with caution due to the small number of studies and high heterogeneity (I² = 70%).

Effects of school‐based physical activity interventions on secondary outcomes

Fitness

Evidence suggests that school‐based physical activity programmes may improve physical fitness assessed by measured or estimated VO₂max (MD 1.19 mL/kg/min, 95% CI 0.57 to 1.82; 13 studies; Analysis 1.8; low‐certainty evidence). These findings should be interpreted with caution due to inconsistency in effect estimates (based on the high level of heterogeneity in the meta‐analysis (I² = 90%) and visual inspection of forest plots) and indirectness in measuring fitness using estimated VO₂max in most studies.

1.8. Analysis.

1.8

Comparison 1: PA programme vs no PA programme, Outcome 8: Physical fitness (mL/kg/min): meta‐analysis

Twenty‐nine additional studies provided data that were not included in the meta‐analysis. Most results were consistent with the direction of the pooled effect (Analysis 1.10).

1.10. Analysis.

Comparison 1: PA programme vs no PA programme, Outcome 10: Physical fitness: additional data

Physical fitness: additional data
Study Study population Intervention group Control Group Measurement period Overall effect Comment
Before or after school programme
Carlin 2018 Female students, aged 11 to 13 years old Brisk walking intervention Continued with normal PA habits 12 weeks No significant changes were observed between group Data not shown
      6 months No significant changes were observed between group  
de Heer 2011 Children in grades 3 to 5 with no condition that would endanger their own or others’ safety Bienstar intervention of health education and 45 min to 60 min of after school PA Grade 4 health workbooks and incentives 12 weeks Shuttle run performance: MD 3.87 (SE 1.51) laps, P = 0.012  
           
Wang 2008 Grade 3 students 'FitKid' after‐school intervention sessions 3 years Heart rate response to step test, group*time interaction P < 0.01  
           
Zhou 2019 Junior high school students, grade 7 I1: Biweekly after school program Regular PE (2 days per week) 32 weeks 20‐m shuttle run performance (change from baseline): MD 8.86, 95% CI: 5.68, 12.04)
I1: m 12.38 (95% CI 10.2, 14.56)
Control: m 3.52 (95% CI 1.18, 5.85)
Authors note that statistically significant changes were found between both intervention and control groups but analyses were not described
  I2: Enhanced PE (3 days per week) plus after school program Regular PE (2 days per week) 32 weeks 20‐m shuttle run performance (change from baseline): MD 22.26 laps (95% CI: 19.15, 25.37)
I2: m 25.78 (95% CI: 23,7, 27.86)
Control: m 3.52 (95% CI: 1.18, 5.85)
 
Enhanced PE
Ketelhut 2020 Grade 3 students High intensity interval training Regular PE 12 weeks Between‐group difference in aerobic fitness z‐score: 7.7 (95% CI 2.3, 13.2)  
           
Kriemler 2010 Grades 1 and 5 students 2 additional 45‐minute PE lessons/week, activity breaks, and PA homework Usual, mandatory PE lessons 9 months Adjusted shuttle run performance: MD −0.12 (95% CI −0.21, −0.03) stages, P = 0.009  
           
Ordóñez Dios 2019 Children age 11‐12 years Daily run added to regular PE Regular PE 12 weeks 1km time
Between‐group difference in change from baseline: ‐0.55 minutes, 95% CI: ‐0.75, ‐0.35
 
           
Thivel 2011 Children in grades 1 or 2 120 min of additional supervised PE Habitual 2 H of PE/week 6 months Shuttle run performance
Between‐group difference in change from baseline: m 0.36, 95% CI: 0.23, 0.49 stages
 
           
Zhou 2019 Junior high school students, grade 7 I1: Enhanced PE (3 days per week) Regular PE (2 days per week) 32 weeks 20‐m shuttle run performance (change from baseline): MD 14.33 laps (95%CI: 11.16, 17.50)
I1: m 17.85 (95% CI 15.68, 20.02)
Control: m 3.52 (95% CI 1.18, 5.85)
Authors note that statistically significant changes were found between both intervention and control groups but analyses were not described
  I2: Enhanced PE (3 days per week) plus after school program     20‐m shuttle run performance (change from baseline): MD 22.26 laps (95% CI: 19.15, 25.37)
I2: m 25.78 (95% CI: 23,7, 27.86)
Control: m 3.52 (95% CI: 1.18, 5.85)
 
Multi‐component intervention
Aburto 2011 Students in grades 4 and 5 I1: basic intervention of environmental and policy‐level change
I2: plus intervention adding additional resources and daily morning exercise
No change to the standard practices 18 months From baseline to follow‐up, there were no significant changes in either group in distance run during the 9 minute run test (P > 0.05)  
           
Andrade 2014 Grades 8 and 9 students ACTIVITAL individual‐ and environmental‐based intervention Standard curriculum 28 months Shuttle run performance: MD −0.19 (95% CI −0.54, 0.16) min  
           
Burke 1998 I1: standard PA and nutrition program including classroom lessons, fitness sessions daily, and nutrition program,
I2: I1 plus a PA enrichment program for higher‐risk children
No program 9 months Shuttle run performance
Girls: Number of laps increased in both intervention groups vs. control group (P = 0.0001)
Boys: Number of laps increased in both intervention groups vs. control group (P = 0.0008)
 
           
Cohen 2015 Grades 3 and 4 students Teacher learning, PA policies, school‐community linkages Usual PE and school sport programs 12 months Shuttle run performance: MD 5.4 (95% CI 2.3, 8.6) laps, P = 0.003 20 m shuttle run test
           
Jago 2011 Students enrolled in grade 6 Education, social marketing, food environment, PE curriculum, and equipment provision Recruitment and data collection only 2.5 years Between‐group difference in change in shuttle run performance: MD 0.2 laps, 95% CI: ‐0.5, 0.9)  
           
Jansen 2011 Grades 3 to 5 3 PE sessions/week, additional after school sport and play, classroom education, and parent health promotion Continued with usual curriculum 1 school year Shuttle run performance: MD 0.57 (95% CI 0.13, 1.01) laps  
Grades 6 to 8       Shuttle run performance: MD 0.04 (95% CI ‐0.45, 0.53) laps  
Madsen 2015 Grades 3, 4, and 5 students Nutrition education curriculum, Playworks structured recess before or after school activities, PA and games implemented by teachers 2 years Mile run time: MD 0.2 (95% CI ‐0.8, 0.4) minutes  
           
Reed 2008 Grades 4 and 5 children Action Schools!BC whole‐school PA approach Regular program of PE and school‐based PA 11 months Shuttle run performance, adjusted for baseline values: MD 6 laps, 95% CI: 1.6, 10.4  
           
Seibert 2019 Grade 5, 9 to 10 years old Physically active lessons (45 min) 2‐3d/week on days without PE, physically active homework and physically active recess Normal routine, 135 min/week of PA 10 months Progressive Aerobic Cardiovascular Endurance Run score (change from baseline): MD ‐1.1, 95% CI: ‐2.2, 0.01)
Fitness z‐scores (change from baseline: MD ‐0.10, 95% CI: ‐0.15, ‐0.04
Change favours the control group
           
Toftager 2014 Physical and organizational environmental changes 2 years Shuttle run distance: MD 6 (95% CI ‐20, 31) metres, P = 0.43  
           
Trevino 2004 All grade 4 children Health programming regarding 3 health behavior messages associated with diabetes mellitus control and goal setting 7 months Fitness score:
MD 1.87 (95% CI ‐1.44, 5.17) points, P = 0.04
Harvard Step Test used
           
Schooltime PA
Breheny 2020 Years 3 (aged 7–8 years) and 5 (aged 9–10 years) Daily mile, 15‐minutes of PA incorporated into the school day Usual school day 12 months Linear track test:
MD ‐37.4 (95% CI ‐74.7, ‐0.19) metres
Difference favours control group
           
de Greeff 2016 Grades 2 and 3 students Physically active mathematics and language lessons Usual curriculum 2 years Shuttle run performance:
MD adjusted for baseline values 0.05 stages, SE: 0.14
 
           
Donnelly 2017 Grades 2 and 3 students Academic Achievement and Physical Activity Across the Curriculum lessons, 160 min/week of MVPA Traditional classroom instruction and typical PE schedule 3 years Progressve Aerobic Cardiorespiratory Endurance Run test performance: MD 1.3 laps, 95% CI: ‐0.5, 3.1  
           
Have 2018 Grade 1 students Active math lessons Regular classroom instruction 10 months Between group difference in intermittent shuttle run test performance: MD 10.0 (SE 13.9) metres, P > 0.05  
           
Leahy 2019 Grade 11 students Burn2Learn, multi component high intensity interval training Usual school activities 14 weeks Shuttle run performance: MD 8.9 (95% CI 1.7, 16.2) laps, P = 0.01  
           
Magnusson 2011 Children attending grade 2 Students engaged in PA during PE lessons, recess, and during classes; schools had access to PA equipment to use in school lessons; teaching materials promoting PA were provided Followed the general PA curriculum 2 years Load achieved on a cycling test: MD 0.37 (95% CI ‐0.27, 1.01) w/kg, P = 0.18  
           
Resaland 2016 Grade 5 and 6 Physically active Norwegian, mathematics, and English lessons on the playground; PA breaks and PA homework Curriculum‐prescribed PE and PA 7 months Intermittent shuttle run test performance: MD 6.9 (95% CI −8.9, 22.6) metres, P = 0.387  
           
Seljebotn 2019 Grade 5 students Physically active lessons, active homework, and physically active recess Continued normal routine, approximately 135 min/week of PA 10 months No significant differences found (no data reported)  
           
Tarp 2016 Grades 6 and 7 students 60 min of PA during school time, PA homework Normal practice 20 weeks Shuttle run test performance: MD 9.4 (95% CI ‐3.7, 22.4) metres, P = 0.16  
           
Torbeyns 2017 Grades 3 and 4 students Cycling desks No lifestyle change 22 weeks Shuttle run test performance (change from baseline): MD 0.5 stages, 95% CI: ‐0.5, 1.5  
           

Four studies explored the impact of before and after school physical activity interventions. One study found an improvement in shuttle run performance with MD of 3.87 laps (standard error (SE) 1.51; P = 0.012) laps following 12 weeks of health education and after school physical activity (de Heer 2011). In a second study, the authors noted a statistically significant improvement in heart rate response to a step test after the ‘FitKid’ after school physical activity programme, although absolute values were not reported (Wang 2008). In a study of a brisk walking intervention delivered to adolescents, authors reported that the intervention had little to no effect on fitness (Carlin 2018); however no values for physical fitness were reported. In another study, a bi‐weekly after school programme on its own or combined with enhanced PE yielded statistically significant improvement in shuttle run performance in both intervention groups (MD 8.86 laps, 95% CI 5.68, 12.04 and 22.26 laps; 95% CI 19.15 to 25.37, respectively) (Zhou 2019).

Five studies investigated the effects of enhanced PE on the fitness of school‐age children. Following 12 weeks of adding a daily run to regular physical activity, between‐group difference in 1‐km run time was MD ‐0.55 minutes (95% CI ‐0.75 to ‐0.35) (Ordóñez Dios 2019). An additional 120 minutes/week of supervised PE resulted in a between‐group difference in shuttle run stages of MD 0.36 stages (95% CI 0.23 to 0.49) (Thivel 2011). One study found that two additional 45‐minute PE lessons per week resulted in a between‐group difference in stages on the shuttle run test favouring the control group (MD ‐0.12 stages, 95% CI ‐0.21 to ‐0.03) (Kriemler 2010). After 12 weeks of high‐intensity interval training, between‐group difference in z‐scores in Grade 3 students was MD 7.7 (95% CI 2.3 to 13.2) (Ketelhut 2020). In another study, enhanced PE on its own or combined with a biweekly after school programme yielded statistically significant improvement in shuttle run performance in both intervention groups (MD 14.33 laps, 95% CI 11.16 to 17.50; and MD 22.26 laps, 95% CI 19.15 to 25.37, respectively) (Zhou 2019).

Eleven studies explored the effects of multi‐component interventions. After 12 months of teacher learning, physical activity policies, and school community linkages, MD of 5.4 laps (95% CI 2.3 to 8.6) was found in children (Cohen 2015). A one‐year multi‐component intervention resulted in between‐group differences among students in Grades 3 to 5 of 0.57 laps (95% CI 0.13 to 1.01) and in Grades 6 to 8 of 0.04 laps (95% CI ‐0.45 to 0.53) (Jansen 2011). Following a 9‐month intervention, study authors reported that the number of laps completed by children on the shuttle run significantly increased in the intervention group versus the control group, but values were not reported (Burke 1998). After 11 months of a whole‐school physical activity programme, between‐group difference in shuttle run laps in intervention versus control groups was +6 laps (95% CI 1.6 to 10.4) (Reed 2008). In one study, authors reported that the difference in change in shuttle run performance at 2.5 years was 0.2 laps (95% CI ‐0.5 to 0.9) following a multi‐component intervention for children consisting of education, social marketing, changes to the food environment, and PE curriculum (Jago 2011). One study reported that physically active lessons, physically active homework, and physically active recess resulted in within‐group differences in laps in the control group (m 5.8, SE 0.4) and in the intervention group (m 4.7, SE 0.4), with a between‐group MD of ‐1.1 (95% CI ‐2.2 to 0.01) favouring the control group (Seibert 2019). Following an intervention with environmental and policy‐level changes, study authors reported no statistically significant differences in distance covered during a nine‐minute run test; however absolute values were not reported (Aburto 2011). Following two years of nutrition education, Playworks structured recess, and before and after school activities, the between‐group difference in mile run time was MD 0.2 minutes (95% CI ‐0.8 to 0.4) (Madsen 2015). After 28 months of individual‐ and environmental‐based interventions, between‐group difference in time during a shuttle run for adolescents was MD ‐0.19 minutes (95% CI ‐0.54 to 0.16) (Andrade 2014). Following seven months of health programming and health messaging targeting diabetes control and goal‐setting, between‐group difference in Harvard Step Test scores was MD 1.87 points (95% CI ‐1.44 to 5.17; P = 0.04) (Trevino 2004). Last, the mean difference in shuttle run test distance between adolescents in schools that underwent physical and organisational environmental changes and a control group was MD 6 metres (95% CI ‐20 to 31 after two years of follow‐up) (Toftager 2014).

Finally, ten studies reported on the impact of schooltime PA interventions. Following 12 months of the Daily Mile programme, results favoured the control group (MD ‐37.4 metres, 95% CI ‐74.7 to ‐0.19) (Breheny 2020). Study authors noted that shuttle run performance increased by 0.8 laps in both intervention and control groups after two years of physically active math and language lessons (MD 0.05 stages, SE 0.14; not statistically significant) (de Greeff 2016). After three years of physically active lessons, mean difference in shuttle run test was 1.3 laps (95% CI ‐0.5 to 3.1) (Donnelly 2017). One study found that active math lessons delivered over one 10‐month school year resulted in mean difference in shuttle run test distance of 10 metres (SE 13.9; P > 0.05) (Have 2018). An intervention including 60 minutes of physical activity during school time and physical activity homework found a between‐group difference in shuttle run distance of MD 9.4 metres (95% CI ‐3.7 to 22.4) at 20 weeks (Tarp 2016). An intervention consisting of 7 months of physical activity lessons, homework, and breaks found MD 6.9 metres (95% CI ‐8.9 to 22.6) (Resaland 2016). After 22 weeks of using cycling desks in the classroom, the mean difference in 20‐metre shuttle run was 0.5 stages (95% CI ‐0.5 to 1.5) (Torbeyns 2017). After two years of active physical activity, PE lessons, classroom physical activity, and additional physical activity equipment and teaching materials, the mean difference in maximal cycling test output was 0.37 watts/kg (95% CI ‐0.27 to 1.01) (Magnusson 2011). Physically active lessons, active homework, and recess did not produce a statistically significant effect, although no data were reported (Seljebotn 2019). Finally, 14 weeks of high‐intensity interval training resulted in a mean difference of 8.9 laps (95% CI 1.7 to 16.2) at 14 weeks (Leahy 2019).

In subgroup analyses, no differences in effects were found between children and adolescents (test for subgroup differences, P = 0.08; Analysis 1.8). Differences in effects by intervention type were noted (test for subgroup difference, P < 0.001; Analysis 1.9).

1.9. Analysis.

1.9

Comparison 1: PA programme vs no PA programme, Outcome 9: Physical fitness (mL/kg/min) by intervention type: meta‐analysis

Children

Overall, evidence suggests that school‐based physical activity programmes probably improve physical fitness among children (MD 1.47 mL/kg/min, 95% CI 0.84 to 2.09; 9 studies; Analysis 1.8 moderate‐certainty evidence). However, this should be interpreted with caution, as only 9 of 31 included studies reported sufficient data to be included in the meta‐analysis. Moderate heterogeneity was also noted across trials (I² = 64%).

Adolescents

Generally, school‐based physical activity programmes probably result in little to no difference in physical fitness (Analysis 1.10). Pooled analysis from studies that reported VO₂max revealed no difference (MD 0.58 mL/kg/min, 95% CI ‐0.18 to 1.35; 4 studies; moderate‐certainty evidence; moderate heterogeneity (I² = 87%).

Before and after school programmes

Before and after school programmes in the school setting probably improve physical fitness (MD 1.38 mL/kg/min, 95% CI 0.34 to 2.41; 5 studies; Analysis 1.9 moderate‐certainty evidence). High heterogeneity was found across studies (I² = 88%).

Enhanced PE

Studies that enhanced PE as part of the intervention probably resulted in improvements in physical fitness (MD 1.99 mL/kg/min, 95% CI 0.76 to 3.21; 4 studies; Analysis 1.9 moderate‐certainty evidence); however high heterogeneity was found across studies (I² = 82%).

Multi‐component interventions

Multi‐component interventions in the school setting probably do not change physical fitness (MD ‐0.33 mL/kg/min, 95% CI ‐0.73 to 0.08; 3 studies; Analysis 1.9 moderate‐certainty evidence); however results should be interpreted with caution due to the small number of studies. No heterogeneity was noted (I² = 0%).

Schooltime PA

Only one study that used schooltime PA to increase fitness was included in the meta‐analysis (MD 2.70, 95% CI 1.04 to 4.36; 1 study; Analysis 1.9); thus, results should be interpreted with caution.

Body mass index

Overall, evidence suggests that school‐based physical activity programmes may result in a very small decrease in BMI z‐score among children and adolescents (MD ‐0.06, 95% CI ‐0.09 to ‐0.02; 21 studies; Analysis 1.11 low‐certainty evidence) and may not decrease BMI (MD ‐0.07 kg/m², 95% CI ‐0.15 to 0.01; 50 studies; Analysis 1.13 low‐certainty evidence). These results should be considered with caution, as substantial heterogeneity and risk of bias were found across studies.

1.11. Analysis.

1.11

Comparison 1: PA programme vs no PA programme, Outcome 11: BMI: meta‐analysis [z‐scores]

1.13. Analysis.

1.13

Comparison 1: PA programme vs no PA programme, Outcome 13: BMI: meta‐analysis [kg/m2]

Nine additional studies provided data that could not be included in the meta‐analysis. Findings were mixed. After a 9‐month multi‐component intervention, study authors reported that BMI decreased among boys in the intervention group versus the control group, but not among girls; values were not reported (Burke 1998). After 3 years of a physical activity or physical activity and nutrition wellness policy, study authors found no differences in children's BMI; however values were not reported (Ickovics 2019). Between‐group differences were reported in children's BMI/age‐sex population median values following an intervention targeting behavioural modification (MD ‐0.40, 95% CI ‐1.11 to 0.30), fundamental movement skills (MD ‐0.50, 95% CI ‐1.25 to 0.25), or both (MD ‐1.30, 95% CI ‐2.29 to ‐0.31) (Salmon 2008). After 3 years of physically active lessons, the difference in BMI percentile between intervention and control groups was MD ‐2.3 (95% CI ‐4.8 to 0.2) (Donnelly 2017). After 1 year of active breaks during class time, the between‐group difference in BMI percentile was 0.5 (95% CI ‐0.5 to 1.5) (Kobel 2014). A 2‐year intervention to increase schooltime PA also yielded a between‐group difference in BMI that was not statistically significant, and effect estimates were not reported (Williamson 2007). A 12‐week brisk walking intervention for adolescents produced no difference, but effect estimates were not reported (Carlin 2018). A 16‐week intervention consisting of enhanced PE or enhanced PE with a focus on increasing intensity had no impact on BMI in adolescents, with no values reported (Ardoy 2011). Last, a 1‐year multi‐component intervention for adolescents resulted in a mean difference in BMI percentile of MD 1.09 (95% CI ‐0.64 to 2.82) (Suchert 2015).

In subgroup analyses, no differences in effects were found between children and adolescents for BMI z‐scores (test for subgroup differences, P = 0.23; Analysis 1.11) nor for BMI (test for subgroup differences, P = 0.19; Analysis 1.13). In subgroup analyses by intervention type, no differences in effects were found between intervention types for BMI z‐scores (test for subgroup differences, P = 0.61; Analysis 1.12) nor for BMI (test for subgroup differences, P = 0.80; Analysis 1.14).

1.12. Analysis.

1.12

Comparison 1: PA programme vs no PA programme, Outcome 12: BMI by intervention type: meta‐analysis [z‐scores]

1.14. Analysis.

1.14

Comparison 1: PA programme vs no PA programme, Outcome 14: BMI by intervention type: meta‐analysis [kg/m2]

Children

School‐based physical activity interventions for children may decrease BMI z‐scores; MD ‐0.06 (95% CI ‐0.11 to ‐0.01; 16 studies; substantial heterogeneity of 88%; Analysis 1.11; low‐certainty evidence). These interventions may also result in a small decrease in BMI (MD ‐0.11 kg/m², 95% CI ‐0.19 to ‐0.02; 38 studies; substantial heterogeneity of 84%; Analysis 1.13; low‐certainty evidence).

Adolescents

School‐based physical activity interventions for adolescents may not decrease BMI z‐scores (MD ‐0.03, 95% CI ‐0.05 to ‐0.00; 5 studies; I² = 0%; Analysis 1.11 low‐certainty evidence) nor BMI (MD 0.05 kg/m², 95% CI ‐0.16 to 0.25; 12 studies; I² =88%; Analysis 1.13 low‐certainty evidence).

Before and after school programmes

Before and after school programmes in the school setting may not decrease BMI z‐scores (MD ‐0.02, 95% CI ‐0.05 to 0.01; 2 studies; Analysis 1.12 low‐certainty evidence) nor BMI (MD ‐0.12 kg/m², 95% CI ‐0.25 to 0.01; 9 studies; Analysis 1.14 low‐certainty evidence). Very little heterogeneity was found across studies (I² = 0%, 7%, respectively).

Enhanced PE

Studies that enhanced PE as part of the intervention may not decrease BMI z‐scores (MD ‐0.08, 95% CI ‐0.29 to 0.13; 1 study; Analysis 1.12 low‐certainty evidence) nor BMI (MD ‐0.04 kg/m², 95% CI ‐0.32 to 0.24; 10 studies; Analysis 1.14 low‐certainty evidence). Results should be interpreted with caution, as only one study reported changes in BMI z‐scores, and high heterogeneity was found across studies for BMI (I² = 92%).

Multi‐component interventions

Multi‐component interventions in the school setting may result in small decreases in BMI z‐scores (MD ‐0.06, 95% CI ‐0.11 to ‐0.01; 17 studies; Analysis 1.12 low‐certainty evidence) but not in BMI (MD ‐0.10 kg/m², 95% CI ‐0.24 to 0.03; 20 studies; Analysis 1.14 low‐certainty evidence). In both analyses, high heterogeneity was found across studies (I² = 87%, 93%, respectively).

Schooltime PA

Only one study reported on the effect of schooltime PA on BMI z‐score (MD ‐0.03, 95% CI ‐0.08 to 0.02; 1 study; Analysis 1.12 low‐certainty evidence). Schooltime PA may not decrease BMI (MD ‐0.05 kg/m², 95% CI ‐0.14 to 0.04; 11 studies; Analysis 1.14 low‐certainty evidence). Low heterogeneity was found across studies.

Health‐related quality of life

Seven included studies reported on health‐related quality of life (Analysis 1.16). Given the limited data reported across heterogeneous interventions and populations, as well as the risk of bias in included studies and possible reporting bias in studies that did not report results for this outcome, we are very uncertain about the effects of school‐based physical activity interventions on health‐related quality of life. A full description of the scales used to assess health‐related quality of life can be found in Appendix 5.

1.16. Analysis.

Comparison 1: PA programme vs no PA programme, Outcome 16: Health‐related quality of life: all data

Health‐related quality of life: all data
Study Study population Intervention group Control Group Measurement period Overall effect Comment
Children
Adab 2018 Year 1 students (aged 5 to 6 years) 30 min of additional MVPA on each school day, cooking workshops, a 6‐week healthy eating program, and information sheets for families Ongoing year 2 health related activities and education resources, excluding topics related to healthy eating and PA 15 months
 
30 months
MD −0.630 (95% CI −4.385, 3.124) points
 
MD 1.248 (95% CI −2.301, 4.796) points
Measured using Pediatric quality of life inventory
Breheny 2020 Year 3 (aged 7‐8 years) and 5 (9‐10 years) students Daily Mile, 15 minutes of running/walking within the school grounds during the school day, not to replace PE Usual school  12 months MD 0.010 (95% CI ‐0.002, 0.04) Measured using Child Health Utility 9D
Jago 2015 Year 7 female students After‐school dance classes Provided data only Baseline
T1
T2
Baseline MD 0.01 points, P = 0.309
T1 MD 0.0 points, P = 0.667
T2 MD 0.0 points, P = 0.382
Meausured using European Quality of Life‐5 Dimensions Youth survey
Jago 2019 Year 3 and 4 students, 7 to 9 years old Action 3:30R after school PA club End of study No difference in utility scores or z‐scores between groups Measured using KIDSCREEN‐10
Resaland 2016 Grade 5 and 6 Physically active Norwegian, mathematics, and English lessons on the playground; PA breaks and PA homework Curriculum‐prescribed PE and PA End of study No significant differences found (no data shown) Measured using KIDSCREEN‐10
Adolescents
Harrington 2018 Female students in years 7 to 9, 11 to 14 years old Support for PA, PE, and school sport culture and practices with the support of the Youth Sport Trust and a hub school Continued with normal PA habits End of study No significant differences found (no data shown) Measured using Child Health Utility 9D
Leahy 2019 Grade 11 students Burn2Learn, multi component high intensity interval training Usual school activities 14 weeks MD −2.1 (95% CI −4.0, −0.3) points, P = 0.02 Measured using Strengths and Difficultlies Questionnaire; lower score indicates fewer perceived difficulties

One study reported a decrease in perceived psychological difficulties among adolescents after 14 weeks of high‐intensity interval training compared to those in a control group (MD ‐2.1 points, 95% CI ‐4.0 to ‐0.3) as measured by the Strengths and Difficulties Questionnaire (Leahy 2019). A school‐based physical activity and healthy eating programme noted a mean difference of 1.248 (95% CI ‐2.301 to 4.796) in paediatric quality of life as measured by the Pediatric Quality of Life inventory (Adab 2018). An after school dance programme noted a mean difference in health‐related quality of life of 0.0 points (P = 0.667) when using the European Quality of Life 5 Dimensions Youth Survey (Jago 2015). The Daily Mile intervention resulted in a between‐group difference of 0.010 points (95% CI ‐0.002 to 0.04) after 12 months on the Child Health Utility 9D, where higher scores indicate poorer health (Breheny 2020). The remaining three studies reported no statistically significant differences between groups and provided no data (Harrington 2018; Jago 2019Resaland 2016).

Adverse events

Of the 89 trials included, only 16 reported anything related to adverse events (Analysis 1.17). Based on limited data on adverse events reported, including inconsistency between studies, high risk of bias, and the possibility of reporting bias in studies that did not report results for this outcome, the evidence is of very low certainty; we cannot confidently conclude whether there are or are not potential safety concerns related to school‐based physical activity interventions. Of the studies that noted adverse events, 13 simply stated that no adverse events occurred as part of the intervention. Often minimal detail was given as to how adverse events were tracked or recorded. Adverse events were reported in three studies. In one study, a minor adverse event occurred when an intervention participant made contact with another participant while doing a handstand (Nogueira 2014). Another study reported adverse event rates across both study groups of 2.4% at baseline and 1.7% at end of study related to a blood draw for data collection (Jago 2011). The most commonly reported adverse event was dizziness and was not deemed to be related to the intervention itself. Finally, one study reported 24 adverse events such as musculoskeletal injuries; 20 were deemed to be mild, three moderate, and one serious, for an overall adverse event rate of 0.0006 events per programme hour (Wang 2008).

1.17. Analysis.

Comparison 1: PA programme vs no PA programme, Outcome 17: Adverse events: all data

Adverse events: all data
Study Participants with at least one adverse event
(N) Participants discontinuing trial due to an adverse event
(N)
Andrade 2014 0 0
Breheny 2020 0 0
Cohen 2015 0 0
Ford 2013 0 0
Harrington 2018 0 0
Ickovics 2019 0 0
Jago 2011 Baseline: 205 events
End of Study: 141 events
0
Jago 2015 0 0
Ketelhut 2020 0 0
Leahy 2019 0 0
Martinez‐Vizcaino 2014 0 0
Müller 2019 0 0
Nogueira 2014 1 0
Okely 2011 0 0
Salmon 2008 0 0
Wang 2008 Year 1: 24 events 0

Assessment of reporting bias

To assess the potential for reporting bias, we created funnel plots for MVPA, sedentary time, fitness, and BMI reported in kg/m² and z‐scores (Figure 4Figure 5Figure 6Figure 7Figure 8). Because we used a random‐effects meta‐analysis, 95% confidence intervals are not calculated via RevMan 5.4. Based on our interpretation of the funnel plots, it appears there may be some degree of reporting bias in studies that report on minutes per week of MVPA. This reporting bias may lead to overestimation of the magnitude of the effect; however given the overall null findings of the meta‐analysis, this does not change our conclusions.

4.

4

Funnel plot of comparison: 1.2 Physical activity duration: meta‐analysis.

5.

5

Funnel plot of comparison: 1.5 Sedentary time: meta‐analysis.

6.

6

Funnel plot of comparison: 1.8 Physical fitness: meta‐analysis.

7.

7

Funnel plot of comparison: 1.11 BMI: meta‐analysis [z‐scores].

8.

8

Funnel plot of comparison: 1.13 BMI: meta‐analysis [kg/m2].

Trials ongoing and awaiting classification

Within our search, we identified 16 trials that are awaiting classification (Characteristics of studies awaiting classification), as well as 12 studies that are ongoing (Characteristics of ongoing studies). Within the studies awaiting classification, 12 are marked as ‘complete’ in the clinical trials registry, but no publications can be found; one has been published only as a protocol paper (Friedrich 2015); two have published conference abstracts but with insufficient information to determine eligibility (O'Malley 2011Telford 2019); and one has published baseline results only (Salmon 2011a). Within the 9 ongoing studies, trial start dates ranged from 2014 to 2018, with planned end dates from 2020 to 2022. Three studies did not indicate a planned end date.

Discussion

Summary of main results

The objective of this updated review was to assess, analyse, and draw conclusions about the effectiveness of school‐based interventions in promoting physical activity and fitness among school‐attending children and adolescents aged 6 to 18 years. Our primary outcomes were physical activity and sedentary time, with secondary outcomes of fitness, body composition, health‐related quality of life, and adverse effects. Finally, through subgroup analyses, we sought to identify which types of interventions may be most effective for improving physical activity, fitness, and body mass index (BMI) in this population.

The results of this update do not differ greatly from those reported in the original review in 2009 and in the update in 2013. Overall, school‐based physical activity interventions may improve physical fitness (low‐certainty evidence) but probably have minimal impact on time engaged in moderate to vigorous physical activity (MVPA) (moderate‐certainty evidence) and may result in little to no decrease in sedentary time (low‐certainty evidence). Although school‐based physical activity interventions may result in a small decrease in BMI z‐scores (low‐certainty evidence), they may not impact BMI measured as kg/m² (low‐certainty evidence). In this version of the review, only objective measures of physical activity, sedentary time, fitness, and BMI were included. This is important progress, as the advantages of objectively measured physical activity and sedentary time outweigh the advantages of self‐report measures of these outcomes. In addition, the commercialisation of these devices means the costs of these devices are no longer as prohibitive as they once were. The original review and the 2013 update found that school‐based physical activity interventions had a small positive impact on duration of MVPA and television viewing; however, these systematic reviews primarily measured activity using self‐report measures completed by children, parents, or teachers, which may have introduced substantial bias into the results.

We are uncertain as to the effects of school‐based physical activity interventions on the proportion of children or adolescents who met the physical activity guidelines recommendation of 60 minutes of daily MVPA. Some studies report that multi‐component interventions increase the proportion of adolescents meeting guidelines; however, only 2 studies reported on this outcome, and more work is needed to increase the certainty of these findings. When MVPA was reported in minutes/d, little to no difference was seen in the duration of MVPA among children and adolescents. When separated by the type of intervention implemented, some evidence suggests that schooltime physical activity programmes and multi‐component interventions may result in larger increases in MVPA. This is a new finding from our 2013 review, which did not examine effects separately by the type of intervention implemented due to the smaller number of studies. Except for multi‐component interventions, which may result in a small decrease in sedentary time, school‐based physical activity programmes do not appear to be effective in reducing sedentary time among both children and adolescents.

In contrast, school‐based physical activity interventions may have a small to moderate effect on physical fitness among both children and adolescents. In particular, enhanced physical education (PE) and before and after school programmes may result in the largest gains in fitness. Interventions that focused specifically on increased exercise intensity (such as high‐intensity interval training ‐ Leahy 2019 ‐ and an enhanced PE intervention with a specific focus on increased exercise intensity ‐ Ardoy 2011) led to the largest effect sizes. 

Although BMI was the most reported outcome, school‐based physical activity interventions may result in a small decrease when measured as z‐scores, and little to no difference when measured as kg/m². Although many of the multi‐component interventions did include additional components such as nutrition education or changes to the school food environment, a more specific focus on diet and nutrition both inside and outside the school environment may be needed to change body weight trajectories (Ho 2012Ho 2013). Notably, in this update, interventions that were targeted primarily at improving body composition without an explicit focus on physical activity or physical fitness were excluded. 

The 2013 review found limited evidence that positive effects are maintained in the longer term, although only a small number of studies measured outcomes beyond the end of the intervention. In this version, several studies reported long‐term follow‐up, but evidence to suggest that changes were maintained long term remains limited. One limitation of this update is that we explored impact only at the immediate post‐intervention time period due to wide heterogeneity in follow‐up times across interventions. The wide variety of study designs, lengths of intervention, and lengths of follow‐up make it challenging to further comment on sustainable effects of the interventions. As more data become available, future updates may have access to sufficient data to pool effects from studies with longer follow‐up time periods.

Very few studies reported on differences in response to interventions between boys and girls, and these results were mixed in with results of studies that did report these differences. As such, we did not seek to carry out a subgroup analysis on differential effects in boys versus girls. Recent estimates suggest there are meaningful differences in physical activity levels between boys and girls, and that although prevalence of insufficient physical activity decreased from 2001 to 2016 in boys, no such change occurred among girls (Guthold 2020). Future studies in this field should examine results separately for boys and girls to determine if interventions have similar effects in individuals of both genders.

Finally, few studies reported information about adverse events, or how these were identified and captured. Adverse events reported were generally muscle soreness or injury related to physical activity and bruising related to a data collection blood draw, for example. One systematic review on physical activity and health outcomes related to physical activity interventions overall reported that no included studies reported any harm or injury associated with physical activity participation (Poitras 2016). These types of potential harms should be explored and reported clearly in future school‐based trials.

One aspect not often considered is the potential for adverse effects on quality of life or harms related to the stigma of participating in physical activity with their peers at school. Also missing from most studies was consideration of factors related to health equity. If physical activity or physical education programming does not meet the needs of individual students or certain subgroups of students, participation may be limited, which may be reflected in variation in findings across studies. In a review of studies about meaningful experiences in physical education and sport, identified themes were social interaction, fun, challenge, competition, motor competence, and personally relevant experiences (Beni 2017). Negative experiences reported in physical education classes in childhood were related to embarrassment and lack of enjoyment of fitness testing and sport, and positive memories of school physical education, such as enjoyment of class activities, time spent with friends or outside, or being allowed to move more after sitting in class all day, were associated with positive attitudes and intentions to be physically active in adulthood in another study (Ladwig 2018). These studies suggest that students need to identify personal meaning in their school physical activity opportunities; a “one‐size‐fits‐all” approach may not be appropriate to encourage physical activity participation.

Most studies did not comment on aspects related to implementation of interventions, such as uptake or adherence to the interventions and fidelity of delivery. It is unknown if interventions were successfully delivered within the schools, which can often be challenging. Without an understanding of fidelity of delivery, any additional minutes of MVPA that resulted from taking part in the school‐based intervention could be compensated for by a decrease in MVPA outside of school time; thus, no overall change in full‐day MVPA was observed. Also, failure to properly implement the programme and poor adherence to the intervention at the student level may occur. Finally, although patient‐oriented research and community engagement initiatives are becoming increasingly prevalent in adult research literature, youth engagement in both design and implementation of these interventions may prove to be a useful strategy to promote uptake and adherence. This could in turn result in not only more meaningful improvements in these short‐term behaviours but also a long‐term commitment to physical activity to improve health and reduce chronic disease risk into adulthood.

Overall completeness and applicability of evidence

A comprehensive search of randomised controlled trials (RCTs) and cluster‐RCTs was conducted, and it is unlikely that many studies were missed by our search strategy. Studies included in this review are applicable to public health and education in high‐income countries, as most studies were conducted in the USA, the UK, and Australia; different strategies may be needed and different effects may be found in low‐ to middle‐income countries. Most studies included in this review were conducted in school‐age children (ages 6 to 12); a smaller number were conducted in adolescents. Although ten trials were conducted exclusively in female students, only one study was conducted among boys only, and few trials explored the effects of sex and gender in the analysis.

The types of interventions included varied widely, and no two studies implemented the same interventions. This makes it very challenging to draw conclusions as to the most effective components of interventions that can elicit changes in the outcomes of interest. The benefit of the wide variety of interventions is that researchers and policy makers can search a variety of protocols to determine what might be effective for future interventions or programming.

Although RCTs are considered the gold standard design for exploring efficacy, it is possible that inclusion of non‐randomised intervention studies may have provided more information on outcomes that were important to this review, namely, health‐related quality of life and adverse events. Given the large number of included trials, it is not feasible to include non‐randomised evidence in this update. However, we wish to acknowledge the potential for this information to be included in studies that were not captured in this review.

Quality of the evidence

As outlined, several factors limited the certainty of results in this review. The most common reason why certainty was downgraded was inconsistency, assessed through visual inspection of forest plots and I² values from the meta‐analysis. This criterion for downgrading was applied to all outcome measures except for sedentary time. Inconsistency in findings is not surprising, given the large differences in target populations, components of interventions studied (including dose and type of physical activity), ways in which outcomes were assessed, and time periods of follow‐up. Nonetheless, this inconsistency limits our confidence in the effects of school‐based physical activity interventions overall. 

The quality of the evidence was also limited by high or uncertain risk of bias in the included studies. Although all included trials were individually randomised or cluster‐randomised trials, a number of methodological limitations were present. In particular, blinding of participants and personnel and of outcome assessors often was not reported, introducing the potential for performance and detection bias in assessment of results. Due to the nature of school‐based physical activity interventions, it is near impossible to blind participants and personnel to the assigned intervention. However, blinding of outcome assessors and data analysts is important and could be improved upon in future studies. Attrition bias was also prevalent across studies, particularly for measures of physical activity and sedentary time. Some loss to follow‐up is unavoidable in school‐based interventions, and study authors often reported the proportion of students lost to follow‐up due to moving and changing schools. However, often a greater proportion of missing data was related to physical activity or sedentary time measured by accelerometers. Adherence to wear time protocols for these devices may be poor within some populations, and data loss or technical issues were possible. When there is differential incomplete outcome data by intervention group within trials, these findings are particularly susceptible to bias.

Although this review included only objectively measured physical activity and sedentary time, accelerometers and pedometers are not without limitations. When accelerometers are used to measure physical activity and sedentary time, the accelerometer model, epoch length, non‐wear time, definition of a valid day, wear time criteria, and cut points can differ, causing wide variation between studies and limiting the precision of the effect estimate (as was seen in studies of proportions of participants physically active and sedentary time). For example, a longer epoch length will underestimate MVPA. When studies have minimum wear‐time criteria, the data may be more indicative of habitual physical activity, but sample size will be reduced. The cut points used will also have an impact on physical activity and sedentary time duration (Cain 2013). 

Very few of the included studies measured maximal oxygen uptake (VO₂max) directly using gas exchange; this was most often predicted by field tests. This resulted in downgrading of the certainty of evidence for physical fitness due to indirectness. For the most part, studies did use reliable, valid, field‐based measures of aerobic fitness; however the usefulness of these tests is largely determined by participants’ motivation to try their hardest, thus reducing the change that a true measure of fitness was achieved.

Our confidence in the evidence related to both health‐related quality of life and adverse events was limited by potential publication bias, with overall certainty of findings downgraded accordingly. Although many studies reported that no adverse events were reported, most studies did not adequately describe the approach taken to monitor for any adverse events. Although 7 studies reported on health‐related quality of life, others listed these outcomes in protocol papers or in trial registries but did not publish the findings; thus the potential of publication bias remains. 

Finally, very few studies reported on the extent to which interventions were implemented as specified. Without adequate process evaluation data, it is not known to what degree students participated in the intervention, and this could have an influence on the impact of the intervention on our outcomes of interest. Information about implementation is important, to understand trial fidelity and for scaling up of interventions in the future. About half of the trials were informed by various theoretical models, including social cognitive theory, socioecological model, self‐determination theory, and the theory of planned behaviour. These theories are intended to promote physical activity at the individual level and may not be as relevant to public health interventions, such as the school‐based studies included in this review (King 2002).

Potential biases in the review process

It is possible that biases were introduced in the review process; however, several steps were taken to minimise this. A comprehensive search strategy, with updates from our previous search strategy to include new terms (such as sedentary time), was used to identify over 9000 citations from 2011 to the present. We did not place limitations by language or publication status. Although efforts were undertaken to minimise this bias (multiple review authors were involved in interpreting results and provided comments on drafts of this update), it is possible that we have interpreted the results to be more positive than they actually are. Readers of this review are cautioned therefore to carefully examine results across studies.

Agreements and disagreements with other studies or reviews

Overall, our findings are similar to those of other systematic reviews and meta‐analyses that have addressed similar questions. A recent systematic review and meta‐analysis examined effects of school‐based physical activity interventions on physical activity and sedentary time, including only cluster‐randomised controlled trials (Love 2019). When studies measured changes in MVPA during the actual intervention period (i.e. during PE class only in an intervention of enhanced PE), there was moderate evidence of effect, whereas when changes in MVPA were examined over the whole school day, effects were inconclusive, and when changes in MVPA were examined across the entire day (both in and out of school time), no effect was seen (Love 2019). It is interesting to note that these study authors also explored the effectiveness of interventions by gender and socioeconomic status and found no difference in effect in terms of either of these variables.

Other reviews have focused more closely on specific intervention types, with similar findings to those presented in this review. A 2019 review of 22 studies that implemented active breaks within the classroom found a small but not significant increase in minutes of MVPA compared to a control group (+3.29 minutes/d, 95% CI ‐0.15 to 8.75) (Masini 2020). However, in this review, only physical activity during class time was included, as opposed to full‐day physical activity measures that were included in this review. Bedard and colleagues found a small reduction in sedentary time when schools took part in an active classroom intervention; however all studies were found to have moderate to high risk of bias (Bedard 2019). A review of active transport interventions found low‐quality evidence to suggest that active transportation can increase transportation‐related MVPA, but with no associated change in physical fitness among children aged 4 to 11 years (Jones 2019).

Authors' conclusions

Implications for practice.

Following are suggestions for public health practitioners, decision‐makers, and policy makers. School‐based physical activity interventions as they have been designed and delivered to date probably have little to no impact on overall time spent in MVPA and may have little to no impact on time spent sedentary. Some evidence suggests that multi‐component interventions that address the whole‐school environment and incorporate physical activity throughout the school day (e.g. physically active lessons, physical activity breaks) may have the strongest impact on time spent in MVPA. Although not the focus of this review, an additional focus on physical activity outside the school environment may help to increase overall physical activity levels. Public health organisations can support schools in providing implementation, assessment, and evaluation.

Although school‐based physical activity interventions may improve physical fitness, specific focus on targeting higher‐intensity activity is warranted.

Finally, school‐based physical activity programmes may have only a very small impact on BMI z‐scores and little to no impact on BMI in kg/m². If the primary goal is to promote healthy body weight, it is likely that another type of intervention may be needed to attain meaningful improvements.

Implications for research.

Across outcomes, the certainty of evidence was downgraded due to inconsistency of findings across interventions. This may be attributed to (1) variability in strategies used and in the frequency, intensity, and duration of interventions; (2) use of various theoretical models to guide the intervention; (3) use of a variety of instruments and tools to assess physical activity or physical fitness (or both); and (4) follow‐up periods of different durations. Full reporting on components of the interventions delivered (e.g. by using the Template for Intervention Description and Replication (TIDieR)) may be helpful in further understanding heterogeneity across studies to identify critical components of success (Hoffmann 2014).

Lack of change in leisure‐time physical activity or physical fitness, in turn, has been attributed most often to issues of (1) inadequate dose (Tolfrey 2000); (2) poor compliance (Baranowski 1990); (3) inattention to the multiplicity of risk factors for physical inactivity and subsequent overly simplistic, uni‐dimensional interventions; (4) methodological errors in measuring fitness (e.g. assessing heart rate only after, as opposed to during, activity); and (5) failure to control for potentially confounding variables (Tolfrey 2000), particularly in cluster‐randomised trials. Future studies should ensure that each of these aspects has been carefully considered in both design and delivery of interventions, which may help to enhance understanding and explain heterogeneity across trials.

What's new

Date Event Description
22 September 2021 New search has been performed Search has been updated to June 2020 to include new studies. Inclusion criteria have been updated to include only studies with objective measures of physical activity and sedentary time. Previously included outcomes of television viewing time, blood pressure, and blood cholesterol have been removed. New outcomes have been added: sedentary time, body mass index z‐scores, health‐related quality of life, and adverse events. Appraisal of risk of bias has been updated according to the updated Cochrane Handbook for Systematic Reviews of Interventions, including additional considerations for cluster‐randomised controlled trials. Meta‐analyses have been conducted, with subgroup analyses by age (children vs adolescents) and by intervention type. The GRADE approach to interpretation of findings has been incorporated.
Overall conclusions with respect to the impact of interventions on fitness and BMI have not changed. After limiting the included studies to only those that use objective measures of physical activity we now conclude that school‐based physical activity interventions probably have minimal impact on time engaged in moderate to vigorous physical activity.
22 September 2021 New citation required and conclusions have changed Overall conclusions with respect to the impact of interventions on fitness and BMI have not changed. However, after limiting the included studies to only those that use objective measures of physical activity we now conclude that school‐based physical activity interventions probably have minimal impact on time engaged in moderate to vigorous physical activity.

History

Review first published: Issue 1, 2009

Date Event Description
21 October 2011 New search has been performed Searches have been run for the update period (July 2007 to October 2011); 30 new project accounts have been identified and are included in the updated review
21 October 2011 New citation required but conclusions have not changed This update has not impacted the conclusions and recommendations of the original review. One change to note is that the physical health status outcome blood cholesterol level (mg/dL) is no longer statistically significant
29 April 2010 Amended Change in scope: 3 new relevance criteria have been added and applied to all included studies: (1) randomised controlled trials; (2) interventions implemented a minimum of 12 weeks; and (3) interventions aimed at the general population. This has resulted in exclusion from the update of 12 studies from the original review: 9 because they were not RCTs (Alexandrov 1988; Berenson 1993; Graf 2005; Klepp 1994; Lionis 1991; Manios 1999; Marcus 1987; Plotnikoff 1999; Sallis 1997); 2 because the intervention was shorter than 12 weeks (Eliakim 1996; Fardy 1996); and 1 because the study sample included overweight or obese children only (Carrel 2005a)
19 January 2010 Amended Review author ‐ RL LaRocca ‐ has been added
21 May 2008 Amended Review has been converted to new review format

Notes

Portions of the background and methods sections, the appendices, additional tables, and figures are based on a standard template established by the Cochrane Metabolic and Endocrine Disorders Group.

Acknowledgements

We gratefully acknowledge the support and contribution of the Effective Public Health Practice Project in the initial systematic review of this literature in 1999, from which this updated systematic review stems, as well as the contributions of Elena Goldblatt, City of Hamilton Public Health Librarian, who developed the original search strategies. We also gratefully acknowledge Paula Robeson, Daiva Tirilis, Heather Husson, Kara DeCorby, and Rebecca Larocca, who contributed as authors to previous versions of this review, which were published in the Cochrane Library.

We also thank members of the Cochrane Public Health editorial team, in particular, Miranda Cumpston, Alix Hall, and Irma Klerings, as well as external referees Nicole Nathan and Emily Darlington, for their comments on the updated review. 

Appendices

Appendix 1. Search strategies

Cochrane Central Register of Controlled Trials (the Cochrane Library)
Update search, via Cochrane Library: October 1 2011 to June 1, 2020 (search conducted 17 February 2021)
Details from previous searches can be found in the last update Dobbins 2013.
#1    Exercise OR physical education OR physical training OR physical activity OR physical inactivity OR physical fitness OR fitness OR sedentary OR lifestyle OR sport* OR walk* OR danc* 
#2    child* OR adolescen*    
#3    school*    
#4    #1 and #2 and #3 with Publication Year from 2011 to 2020, with Cochrane Library publication date Between Oct 2011 and Jun 2020, in Trials    
MEDLINE
Update search, via Ovid: October Week 1 2011 to June Week 1 2020 (search conducted 17 February 2021)
Details from previous searches can be found in the last update Dobbins 2013.
1. randomised controlled trial.pt.
2. controlled clinical trial.pt.
3. randomized.ab.
4. placebo.ab.
5. clinical trials as topic.sh.
6. randomly.ab.
7. trial.ti.
8. 1 or 2 or 3 or 4 or 5 or 6 or 7
9. exp animals/ not humans.sh.
10. 8 not 9
11. exercise/ or circuit‐based exercise/ or gymnastics/ or high‐intensity interval training/ or physical conditioning, human/ or plyometric exercise/ or resistance training/ or running/ or swimming/ or walking/
12. physical inactivity.mp.
13. exp Motor Activity/
14. exp "Physical Education and Training"/
15. Phys* ed*.mp.
16. exp Physical Fitness/
17. sedentary.ab. or sedentary.ti.
18. screen time.mp.
19. exp Sedentary Lifestyle/
20. exp Life Style/
21. (("lifestyle" or life‐style) adj5 activ$).tw.
22. (("lifestyle" or life‐style) adj5 physical$).tw.
23. leisure activities/ or recreation/
24. exp Walking/
25. exp Sports/
26. exp Dancing/
27. walk$.tw.
28. sport$.tw.
29. cycl$.tw.
30. dancing.mp.
31. exercise therapy/ or plyometric exercise/ or resistance training/
32. (exercise$ adj aerobic$).tw.
33. (physical$ adj5 (fit$ or train$ or activ$ or endur$)).tw.
34. (exercis$ adj5 (train$ or physical$ or activ$)).tw.
35. exp Child/
36. Child*.mp.
37. exp Adolescent/
38. Youth.mp.
39. Adolescen*.mp.
40. Teen*.mp.
41. 35 or 36 or 37 or 38 or 39 or 40
42. Schools/
43. School$.tw.
44. school‐based.mp.
45. elementary school.mp.
46. middle school.mp.
47. high‐school.mp.
48. grade school.mp.
49. 42 or 43 or 44 or 45 or 46 or 47 or 48
50. 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34
51. 10 and 41 and 49 and 50
52. limit 51 to ed=20111001‐20200601
Embase
Update search, via Ovid Week 41 2011 to Week 23 2020 (search conducted 17 February 2021)
Details from previous searches can be found in the last update Dobbins 2013.
1. exercise/ or aerobic exercise/ or anaerobic exercise/ or aquatic exercise/ or circuit training/ or dynamic exercise/ or endurance training/ or exercise intensity/ or high intensity interval training/ or leg exercise/ or muscle exercise/ or pilates/ or plyometrics/ or resistance training/
2. physical activity/ or cycling/ or jogging/ or jumping/ or lifting effort/ or running/ or swimming/ or walking/ or weight bearing/ or weight lifting/
3. physical inactivity.mp. or physical inactivity/
4. exp fitness/
5. exp physical education/
6. Phys* ed*.mp.
7. exp sport/
8. Exertion.mp.
9. recreation/ or dancing/ or recreational game/
10. exp lifestyle/
11. exp dancing/
12. walking/
13. walking/
14. sedentary lifestyle/
15. sedentary behavio?r.af.
16. screen time.mp.
17. kinesiotherapy/ or dynamic exercise/ or leg exercise/ or movement therapy/ or muscle training/ or pilates/ or plyometrics/
18. exp Exercise therapy/
19. (physical$ adj5 activ$).tw.
20. (physical$ adj5 fit$).tw.
21. (physical$ adj5 lifestyle$).tw.
22. (physical$ adj5 train$).tw.
23. walk.tw.
24. (aerobics or physical activity or physical inactivity).af.
25. (fitness adj (class$ or regime$ or program$)).af.
26. (aerobics or physical training or physical education).af.
27. (fitness adj (class$ or regime$ or program$)).af.
28. (aerobics or physical training or physical education).af.
29. dance therapy.af.
30. or/1‐29
31. child/
32. child*.mp.
33. adolescent/
34. adolescen*.mp.
35. juvenile/
36. teen*.mp.
37. exp middle school student/ or exp high school/ or exp high school student/ or exp primary school/ or exp school/
38. high school/ or middle school/ or primary school/
39. school health service/ or exp school health education/
40. school*.tw.
41. 31 or 32 or 33 or 34 or 35 or 36
42. 37 or 38 or 39 or 40
43. randomised controlled trial/
44. randomisation/
45. random allocation.mp.
46. double blind procedure/
47. single blind procedure/
48. clinical trial/
49. (clinic$ adj trial$1).tw.
50. ((singl$ or doubl$ or treb$ or tripl$) adj (blind$3 or mask$3)).tw.
51. placebo/
52. Placebo$.tw.
53. Randomly allocated.tw.
54. (allocated adj2 random).tw.
55. 43 or 44 or 45 or 46 or 47 or 48 or 49 or 50 or 51 or 52 or 53 or 54
56. case report.tw.
57. letter.pt.
58. 56 or 57
59. 55 not 58
60. 30 and 41 and 42 and 59
61. limit 60 to em=201141‐202023
BIOSIS
Update search, via Web of Science: October 2011 to June 2020 (search conducted 17 February,  2021)
Details from previous searches can be found in the last update Dobbins 2013.
1. TS=("physical activity" OR exercise)
2. TS=("physical fitness" OR fitness)
3. TS=(sedentary or screen‐time)
4. TS=(school* OR "physical education" OR student*)
5. TS=(child* OR adolescent*)
6. #1 OR #2 OR #3
7. TS=(random* and trial)
8. TS=(control* and trial)
9. #7 OR #8
10. #4 and #5 AND #6 AND #9
CINAHL
Update search, via EBSCO Host Research Databases: October 2011 to June 2020 (search conducted 17 February, 2021)
Details from previous searches can be found in the last update Dobbins 2013.
S1 (MH "Exercise+") or (MH "Aerobic Exercises+")
S2 ""physical inactivity""
S3 (MH "Life Style, sedentary")
S4 ("physical education") or (MH "Physical Education and Training") or ("phys* ed*)
S5 ("physical activity" or (MH "Physical activity")
S6 ("physical fitness") or (MH "Physical Fitness+") or ("fitness")
S7 ("walk") or (MH "walking") or (MH "Sports")
S8 ""sport*""
S9 S1 OR S2 OR S3 OR S4 OR S5 OR S6 OR S7 OR S8
S10 ((MH "Schools+") or (MH "Schools, Elementary") or (MH "Schools, Middle") or ("school*"))
S11 (MH "Child")
S12 "adolescen*"
S13 "teen*"
S14 "youth*"
S15 "child*"
S16 S10 OR S11 OR S12 OR S13 OR S14 OR S15
S17 MH randomised controlled trials
S18 MH double‐blind studies
S19 MH single‐blind studies
S20 MH random assignment
S21 MH pretest‐posttest design
S22 MH cluster sample
S23 TI (randomised OR randomised)
S24 AB (random*)
S25 TI (trial)
S26 MH (sample size) AND AB (assigned OR allocated OR control)
S27 MH (placebos)
S28 PT (randomised controlled trial)
S29 AB (control W5 group)
S30 MH (crossover design) OR MH (comparative studies)
S31 AB (cluster W3 RCT)
S32 MH animals+
S33 MH (animal studies)
S34 TI (animal model*)
S35 S32 OR S33 OR S34
S36 MH (human)
S37 S35 NOT S36
S38 S17 OR S18 OR S19 OR S20 OR S21 OR S22 OR S23 OR S24 OR S25 OR S26 OR S27 OR S28 OR S29 OR S30 OR S31
S39 S38 NOT S37
S40 S9 AND S16 AND S39 Limiters ‐ Publication from: 20111001 ‐ 20200631
SPORTDiscus
Updated search, via EBSCO Host Research Databases: October 2011 to June 2020 (search conducted 17 February, 2021)
Details from previous searches can be found in the last update Dobbins 2013.
S1. control group OR randomi* control* trial OR effect* OR random sample* or control subject* 
S2. physical activity OR physical inactivity OR exercise OR fitness OR sport* or danc* OR walk* or physical education OR obesity OR body weight 
S3. child* OR adolescen*
S4. School*
S5. S1 AND S2 AND S3 AND S4 Limit 20111001 ‐ 20200601
PsycINFO
Updated search, via Ovid: October Week 3 2011 to June Week 1 2020 (search conducted 17 February, 2021)
Details from previous searches can be found in the last update Dobbins 2013.
1. Randomized controlled trial.mp.
2. random allocation.mp.
3. double blind method.mp.
4. single blind method.mp.
5. Clinical trial.mp. or exp Clinical Trials/
6. (clinic$ adj trial$1).tw.
7. ((singl$ or doubl$ or treb$ or tripl$) adj (blind$3 or mask$3)).tw.
8. placebo/
9. Placebo$.tw.
10. (allocated adj2 random).tw.
11. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10
12. case report.tw.
13. historical article.mp.
14. letter.mp.
15. 12 or 13 or 14
16. 11 not 15
17. exercise/ or aerobic exercise/ or weightlifting/
18. physical activity/ or exercise/
19. physical inactivity.mp.
20. exercise.mp.
21. physical activity.mp.
22. motor activity.mp.
23. exp Physical Education/
24. phys* ed*.mp.
25. exp physical fitness/
26. physical endurance/
27. exp aerobic exercise/
28. exp SEDENTARY BEHAVIOR/
29. screen time/
30. sedentary.ab. or sedentary.ti.
31. screen time.mp.
32. exp lifestyle/
33. lifestyle changes/
34. (("lifestyle" or life‐style) adj5 activ$).tw.
35. (("lifestyle" or life‐style) adj5 physical$).tw.
36. leisure time/ or recreation/
37. walking.mp. or exp WALKING/
38. sports/ or baseball/ or basketball/ or football/ or judo/ or martial arts/ or soccer/ or swimming/ or tennis/ or weightlifting/
39. exp Dance Therapy/ or exp Dance/ or dancing.mp.
40. walk$.tw.
41. sport$.tw.
42. cycl$.tw.
43. exercise therapy.mp.
44. (exercise$ adj aerobic$).tw.
45. (physical$ adj5 (fit$ or train$ or activ$ or endur$)).tw.
46. (exercis$ adj5 (train$ or physical$ or activ$)).tw.
47. 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46
48. child*.mp.
49. Adolescen*.mp.
50. Teen*.mp.
51. Youth.mp.
52. 48 or 49 or 50 or 51
53. school*.tw.
54. exp Elementary Schools/
55. exp Middle Schools/
56. exp High Schools/
57. secondary school*.mp.
58. primary school*.mp.
59. grade school*.mp.
60. 53 or 54 or 55 or 56 or 57 or 58 or 59
61. 16 and 47 and 52 and 60
62. limit 61 to up=20111021‐20200601
Sociological abstracts
Update search, via ProQuest: October 21 2011 to June 01 2020 (search conducted 17 February, 2020)
Details from previous searches can be found in the last update Dobbins 2013.
Control group OR randomi* control* trial OR effect* OR random sample* OR control subject*
AND
Physical activity OR physical inactivity OR exercise OR physical fitness OR fitness OR sport* OR danc* OR walk* OR physical education OR obesity OR body weight
AND
Child* OR Adolescen*
AND
School*
pd(20110101‐20200601)

 

Appendix 2. Baseline characteristics

Trial ID Interventions and comparators Trial period Ethnic groups (%) Sex (% female) Age, years(mean (SD)/range) BMI, kg/m²(mean (SD))
Breheny 2020 I: Daily Mile 2016 to 2018 52 White British
16 South Asian
8 Black African Caribbean
24 Other/unknown
48 8.9 (1.0)
C: usual school routine
 
Ketelhut 2020 I: high‐intensity interval training during PE 50 10.8 (0.6) 19.6 (4.6)
C: usual PE 41.7 10.7 (0.7) 19.7 (4.0)
Belton 2019 I: multi‐component PE, whole‐school and parent‐targeted intervention 2013 to 2015 51 12.8 (0.4) Male: 20.0 (3.1)
Female: 20.8 (3.5)
C: usual care 50 12.8 (0.4) Male: 19.3 (3.2)
Female: 20.3 (2.8)
Corepal 2019 I: pedometer challenge 2015 to 2016 53
C: usual school
Ickovics 2019 I1: PA school wellness policy 2011 to 2015 47 Hispanic
35 Non‐Hispanic black
18 Non‐Hispanic white 62 10.9 (0.6)
I2: PA + nutrition school wellness policy 50
C1: nutrition school wellness policy 50
C2: delayed control 50
Jago 2019 I: Action 3:30R after school PA club 2017 to 2018 49 8.4/8‐10 16.9 (2.5)
C: — 17.2 (2.4)
Leahy 2019 I: Burn2Learn, multi‐component high‐intensity interval training 2017 71 Australian
19 European
6 Asian
4 other 46 16.2 (0.4) 22.2 (3.0)
C: usual school activities
Lonsdale 2019a I: teacher PE training 2015 to 2016 58 English and European
9 Aboriginal or Torres Strait Islander
33 Other
48 13.0 (0.6)
C: standard teaching 57 English and European
10 Aboriginal or Torres Strait Islander
32 Other
41 12.9 (0.5)
Müller 2019 I1: PA only 2015 to 2016 "Colored children (mixed race ancestry), usually Afrikaans speaking, and black African children, mainly Xhosa speaking" 50 9.2 (0.9) 17.1
I2: PA + health education
I3: PA + health education + nutrition
C1: health education + nutrition 50
C2: no PA
Ordóñez Dios 2019 I: 2 x 45‐minute PE sessions per week and daily run 19.1 (3.1)
C: 2 x 45‐minute PE sessions per week 20.1 (3.8)
Seibert 2019 I: 4 core strategies to increase PA 49 11.1 (—) 22.0 (—)
C: usual PE 46 11.2 (—) 21.7 (—)
Seljebotn 2019
 
I: physically active lessons, active homework, physically active recess 2014 to 2015 48 9‐10 17.6 (3.0)
C: continued normal routine, approximately 135 minutes/week of PA 51 17.3 (2.9)
Zhou 2019
  I1: enhanced PE 2015 to 2016 —
  47 12.7 (0.6) —
 
I2: after school programme
I3: enhanced PE and after school programme
C: regular PE 
Adab 2018 I: 30 minutes of additional MVPA on each school day, cooking workshops, a 6‐week healthy eating programme, information sheets for families 2011 to 2015 45 White British
31 South Asian
8 Black African Caribbean
16 other 51 6.3 (0.3)
C: ongoing Year 2 health‐related activities and education resources, excluding topics related to healthy eating and PA 47
Carlin 2018 I: brisk walking intervention 2014 100 12.4 (0.6) 20.3 (3.9)
C: continued with normal PA habits 100 19.3 (4.3)
Harrington 2018 I: support for PA, PE, and school sport culture and practices with support of the Youth Sport Trust and a hub school 2015 to 2016 77 White European
12 South Asian
11 other 100 12.8 (0.8)
C: usual practice of PE and sport 100
Have 2018 I: active math lessons 2012 to 2013 52 7.2 (0.3) 16.1 (1.7)
C: regular classroom instruction 48 15.7 (2.2)
Pablos 2018 I: lunchtime extracurricular PA 50 10.7 (0.7)
C: continued with daily activities 54
Robbins 2018 I: an after school PA club, counselling, interactive Internet‐based sessions 2011 to 2016 56 Black
44 Non‐Black
16 Hispanic or Latino
84 not Hispanic or Latino
100 12.1 (0.8) 22.9 (6.0)
C: no additional after school programming 64 Black
36 Non‐Black
13 Hispanic or Latino
87 not Hispanic or Latino
100 23.6 (6.1)
Siegrist 2013 I: weekly lifestyle lessons "Mainly Caucasian" 40 11.1 (0.6) 19.1 (3.5)
C: usual activities 47
Ten Hoor 2018 I: strength training and motivational interviewing 2015 to 2016 52 13.0 (0.5) 19.7 (3.5)
C: usual curriculum 47
Donnelly 2017 I: Academic Achievement and Physical Activity Across the Curriculum lessons, 160 minutes/week of MVPA 85 not Hispanic or Latino
11 Hispanic or Latino
2 unknown
2 refused or missing 49 8.1 (0.6) 17.4 (3.1)
C: traditional classroom instruction and typical PE schedule 54 8.1 (0.6)
Farmer 2017 I: school‐specific playground action plan 2011 to 2013 19 Māori
12 Pacific
7 Asian
46 New Zealand
16 unknown 47 8.0 (1.2) 17.4 (2.8)
C: no change to school play spaces 14 Māori
11 Pacific
9 Asian
52 New Zealand
13 unknown 53 7.9 (1.1) 17.4 (2.7)
Sutherland 2017 I: modified Supporting Children's Outcomes using Rewards, Exercise, and Skills programme 2014 to 2015 51 10.2
C: delivered school PA practices according to the curriculum 10.1
Torbeyns 2017 I: cycling desks 2014 to 2015 50 14.3 (0.6) 19.7 (3.5)
C: no lifestyle change 46 20.1 (3.7)
Daly 2016 I: specialist‐taught PE intervention 2005 to 2009 92 Caucasian
6 Asian
1 Indigenous Australian or Polynesian descent
1 unknown 48 8.1 (0.4) Girls: 17.2 (0.2)
Boys: 16.9 (0.2)
49 Girls: 17.2 (0.2)
Boys: 17.0 (0.2)
C: usual PE programme
de Greeff 2016 I: physically active mathematics and language lessons 55 8.1 (0.7) 17.0 (2.8)
C: usual curriculum 59 16.9 (2.6)
Drummy 2016 I: teacher‐led activity break 9.5/9‐10 19.4 (3.5)
C: normal daily routine 18.3 (2.4)
Jarani 2016 I1: group circuit training‐based PE 50 8.4 (1.6) 17.4 (3.2)
I2: games‐based PE 49 8.3 (1.6) 17.9 (3.3)
C: traditional PE school 44 8.3 (1.6) 17.8 (3.4)
Kocken 2016 I: theory and practical lessons on nutrition and PA 2009 to 2011 87 Western
13 non‐Western 52 9.2 (0.6)
C: regular school programme or curriculum on nutrition and PA 85 Western
15 non‐Western 51 9.1 (0.6)
Lau 2016 I: Xbox 260 Kinect gaming sessions after school Asian 28 9.2 (0.5) 19.4 (3.6)
C: regular PA and PE class 35 19.8 (3.6)
Resaland 2016 I: physically active Norwegian, mathematics, and English lessons on the playground; PA breaks and PA homework 2014 to 2015 47 10.2 (0.3) 18.0 (3.0)
C: curriculum‐prescribed PE and PA 49 18.1 (3.0)
Sutherland 2016 I: 7 PA intervention strategies and 6 implementation strategies 2012 to 2014 5 Aboriginal and/or Torres Strait Islander 52 12 19.9 (3.6)
C: only measurement components of the trial, regular PA and PE 8 Aboriginal and/or Torres Strait Islander 51 20.2 (3.8)
Tarp 2016 I: 60 minutes of PA during school time, PA homework 2013 to 2014 Boys:
98 Danish
2 European
Girls:
96 Danish
2 European
1 other
49 12.9 (0.6) 19.8 (2.9)
C: normal practice Boys:
94 Danish
2 European
4 other
Girls:
91 Danish
2 European
7 other
52 19.3 (3.0)
Cohen 2015 I: teacher learning, PA policies, school‐community linkages 2012 to 2013 14 Aboriginal or Torres Strait Islander
86 Australian
1 Asian
5 European
7 other 54 8.5 (0.6)
C: usual PE and school sport programmes 54
Jago 2015 I: after school dance classes 2013 to 2014 100 11‐12 19.5 (3.4)
C: provided data only 100 19.5 (3.7)
Madsen 2015 I: nutrition education curriculum, Playworks structured recess before or after school activities, PA and games implemented by teachers 2011 to 2013 6 White
9 Black
55 Latino
15 mixed
16 other 49
C: — 6 White
13 Black
45 Latino
14 mixed
22 other 56
Muros 2015 I1: extracurricular PA sessions 2012 54 10.7 (0.5) 19.7 (3.7)
I2: PA and nutrition 44
I3: PA and nutrition and extra virgin olive oil during the final month 42
C1: nutrition and lifestyle education sessions 54
C2: usual activities 51
Suchert 2015 I: multi‐level intervention targeting students, classrooms, schools, and parents 2014 to 2015 47 13.7 (0.7)
C: no intervention 49
Andrade 2014 I: ACTIVITAL individual‐ and environment‐based intervention 2009 to 2012 66 12.9 (0.8) 19.8 (3.4)
C: standard curriculum 59 19.7 (2.9)
Jago 2014 I: Action 3:30 activity club 2012 to 2013 59 10 (0.6) 18.8 (3.4)
C: schools provided data only 18.4 (3.3)
Kipping 2014 I: PA education intervention 49 9.5 (0.3)
C: continued standard education provision 52
Kobel 2014 I: teacher training, PA education, active breaks 2010 to 2014 31 migration background 53 7.1 (0.6) 15.98 (2.14)
C: no intervention 49
Martinez‐Vizcaino 2014 I: MOVI‐2 extracurricular PA programme 2010 to 2011 55 9.5 (0.7) 19.0 (3.68)
C: standard PE curriculum 49
Nogueira 2014 I: high‐intensity capoeira sessions Female:
97 Caucasian
3 Asian or Black
Male:
96 Caucasian
4 Asian
41 10.6 (0.6) 18.5 (3.1)
C: usual school activities 50
Santos 2014 I: healthy buddies, healthy living lessons, structured aerobic exercise 2009 to 2010 25 First‐Nations
75 non‐First‐Nations 48 9.3
C: standard curriculum 31 First‐Nations
69 non‐First‐Nations 8.8
Toftager 2014 I: physical and organisational environmental changes 2010 to 2012 91 Native Danish parents 51 12.5 (0.6) 18.9 (3.0)
C: — 92 Native Danish parents 48 12.5 (0.6) 18.8 (3.0)
Fairclough 2013 I: weekly lesson plans, worksheets, homework tasks, lesson resources 2010 to 2011 95 White British 10.6 (0.3) 17.9 (3.0)
C: normal instruction 10.7 (0.3) 18.1 (3.7)
Ford 2013 I: accumulated brisk walking programme 48 5‐11 17.1 (3.3)
C: normal school lessons 16.4 (2.9)
Grydeland 2013 I: structured lessons, PA breaks, PA promotion 2007 to 2009 54 11.2 (0.3) 17.8 (2.5)
C: — 60 17.9 (2.6)
Melnyk 2013 I: goal‐setting, education, PA homework 2010 to 2012 65 Hispanic or Latino 55 14.7 (0.7) 29.7 (7.1)
C: Healthy Teens attention control curriculum was intended to promote knowledge of common adolescent health topics and health literacy 49
Sacchetti 2013 I: daily PA in schoolyard and classroom 2006 to 2009 48 9‐11 18.0 (2.9)
C: standard programme of PE 48 17.8 (2.9)
Siegrist 2013 I: JuvenTUM educational and environmental intervention 2006 to 2007 48 8.4 (0.7) 17.4 (2.9)
C: continued with usual school activities 17.3 (3.0)
Aburto 2011 I1: basic intervention of environmental and policy‐level change 2006 to 2007 Hispanic 50 10.2 (0.7) 19.8 (3.8)
I2: plus intervention adding additional resources and daily morning exercise 55
C: no change to standard practices 50
Ardoy 2011 I1: 4 sessions/week of PE 2007 35 13 (0.7) 22.3 (5.1)
I2: 4 sessions/week of PE with emphasis on increasing intensity 30
C: 2 sessions/week of PE 44
de Heer 2011 I: Bienstar intervention of health education and 45 to 60 minutes of after school PA 2008 Predominantly Hispanic 46 9.2 (1.0) 20.3 (4.4)
C1: Grade 4 health workbooks and incentives 45 20.0 (4.4)
C2: spillover control group 51 19.5 (4.2)
Jago 2011 I: education, social marketing, food environment, PE curriculum, equipment provision 2006 to 2009 59 Hispanic
20 Black
21 White 53 11.3 (0.6)
C: recruitment and data collection only 52
Jansen 2011 I: 3 PE sessions/week, additional after school sport and play, classroom education, parent health promotion 2006 to 2007 Grades 3 to 5:
14 Dutch
9 Surinam
6 Antillean
27 Moroccan
22 Turkish
4 Capeverdean
18 other or missing
Grades 6 to 8:
14 Dutch
11 Surinam
5 Antillean
22 Moroccan
24 Turkish
5 Capeverdea
19 other or missing
Grades 3‐5: 51
Grades 6‐8: 53
9.2 (1.0) Grades 3‐5: 17.1 (2.8)
Grades 6‐8: 19.1 (3.8)
C: continued with the usual curriculum Grades 3 to 5:
7 Dutch
11 Surinam
4 Antillean
36 Moroccan
20 Turkish
4 Capeverdean
17 other or missing
Grades 6 to 8:
8 Dutch
13 Surinam
3 Antillean
35 Moroccan
24 Turkish
5 Capeverdean
15 other or missing
Grades 3‐5: 51
Grades 6‐8: 49
Grades 3‐5: 17.1 (2.8)
Grades 6‐8: 19.8 (4.1)
Magnusson 2011 I: students engaged in PA during PE lessons, during recess, and during classes; schools had access to PA equipment to use in school lessons; teaching materials promoting PA were provided 2006 to 2008 52 7.4 (0.2) 15.8 (1.1)
C: followed the general PA curriculum 59 16.3 (1.0)
Okely 2011 I: PA action plan 2009 to 2010 100 13.6
C: continuation of usual programmes 100
Thivel 2011 I: 120 minutes of additional supervised PE 2003 51 6‐10 Normal weight: 15.6 (1.1)
C: habitual 2 hours of PE/week Obese: 20.6 (2.6)
49 Normal weight: 15.5 (1.1)
Obese: 20.2 (1.8)
Wilson 2011 I: Active by Choice Today programme: PA homework, in‐school PA, motivational skills training 76 African American 56 11.3 (0.6) 22.8 (6.2)
C: General Health Education Programme 69 African American 52 11.4 (0.6) 22.9 (5.9)
Kriemler 2010 I: 2 additional 45‐minute PE lessons/week, activity breaks, PA homework 2005 to 2006 Migrant families
Grade 1: 34
Grade 5: 25 Grade 1: 49
Grade 5: 55 Grade 1: 6.9 (0.3)
Grade 5: 11 (0.5) 17.1 (2.5)
C: usual, mandatory PE lessons Migrant families
Grade 1: 26
Grade 5: 24 Grade 1: 55
Grade 5: 46 Grade 1: 6.9 (0.3)
Grade 5: 11.3 (0.6) 17.0 (2.6)
Neumark‐Sztainer 2010 I: New Moves curriculum (nutrition and self‐empowerment, motivational interviewing, lunch meetings, parent outreach) 2007 to 2009 32 African American
27 White
17 Asian
13 Hispanic
8 mixed/other
3 American Indian 100 15.7 25.9 (7.1)
C: participation in all‐girls PE class 24 African American
22 White
30 Asian
16 Hispanic
7 mixed/other
2 American Indian 100 15.8 25.5 (6.5)
Angelopoulos 2009 I: educational intervention covering self‐esteem, body image, nutrition, PA, fitness, and environmental issues with motivational methods to increase knowledge, skills, self‐efficacy, self‐monitoring, and social influence 2005 to 2006 90 Greek
10 Immigrant 57 10.3 (0.4) 20.3 (3.6)
C: — 88 Greek
12 Immigrants 54 10.3 (0.4) 20.1 (3.4)
Donnelly 2009 I: 90 minutes/week of moderate to vigorous physically active academic lessons 77 Caucasian
6 African American
10 Hispanic
2 Native American
1 Asian
4 multi‐ethnic 52 Grade 2: 7.7 (0.3)
Grade 3: 8.7 (0.4) 17.9 (3.1)
C: regular classroom instruction Grade 2: 7.8 (0.4)
Grade 3: 8.7 (0.4) 18.0 (3.7)
Dorgo 2009 I1: PE manual resistance training programme 45 15.9 (1.2) 24.4 (6.1)
I2: PE manual resistance training plus cardiovascular endurance training 15.2 (1.2) 24.8 (5.9)
C: regular PE programme that followed the usual school curriculum 15.8 (1.1) 24.9 (5.4)
Gentile 2009 I: 'Switch' programme: promoted healthy lifestyles targeting family, school, and community 2005 to 2006 90 White 53 9.6 18.4 (3.3)
C: no intentional exposure to the 'Switch' programme 50 9.6 18.5 (3.5)
Neumark‐Sztainer 2009 I: after school theatre sessions, booster sessions, family outreach 2006 to 2007 54 African American
13 Asian
7 White
3 Hispanic
23 other or mixed 10.3 (1.1) 20.9(5.1)
C: theatre‐based control condition (i.e. children participated in a play focused on environmental health issues using a prepared script)
Peralta 2009 I: curriculum and peer‐facilitated lunchtime PA session, parent newsletters 2007 0 12.5 (0.4) 22.8 (4.1)
C: PA curriculum 0 20.4 (4.1)
Walther 2009 I: 1 unit of physical exercise (45 minutes) with at least 15 minutes of endurance training/school day, plus lessons on healthy lifestyle once/month 47 11.1 18.0 (2.6)
C: German standards, 2 units (each 45 minutes) of PE/week, 12 units (45 minutes/unit) of high‐level endurance exercise training/week plus participation in competitive sporting events 42 11.1 18.2 (2.8)
Reed 2008 I: Action Schools!BC whole‐school PA approach 2003 to 2004 49 10.2 (0.6) 18.8 (3.5)
C: regular programme of PE and school‐based PA 50 10.2 (0.6) 19.1 (3.7)
Salmon 2008 I1: behavioural modification group 2002 to 2003 51 10 (0.4)
I2: fundamental motor skills group 53
I3: combined behavioural modification and fundamental motor skills group 51
C: usual classroom lessons 51
Wang 2008 I: 'FitKid' after school intervention sessions 2002 to 2006 61 Black
31 White
2 Asian
2 Hispanic
5 other 52 8.5 (0.6) 19.4 (4.7)
C: — 19.3 (4.4)
Webber 2008 I: health education lessons to enhance behavioural skills known to influence PA participation (self‐monitoring, setting goals for behaviour change) 2003 to 2006 46 White
24 African American
12 Hispanic
18 other 100 12 20.7
C: — 100 20.9
Weeks 2008 I: directed jumping activity at the beginning of every PE class 58 Girls: 13.7 (0.4)
Boys: 13.8 (0.4) Girls: 19.5 (3.5)
Boys: 20.3 (3.6)
C: regular PE warm‐ups and stretching at the beginning of every PE class
49 Girls: 13.7 (0.5)
Boys: 13.8 (0.4) Girls: 19.5 (2.1)
Boys: 21.5 (5.1)
Barbeau 2007 I: after school PA programme 100 Black 100 9.5/8‐12 20.9 (5.0)
C: — 100 Black 100 9.5/8‐12 20.9 (5.6)
Williamson 2007 I: Healthy Eating and Exercise programme to increase PA during the school day and at home 95 White
2 Black
3 other 50 9.2 (4.1)
C: Alcohol/Drug/Tobacco abuse prevention programme
Haerens 2006 I1: a computer‐tailored intervention to increase MVPA to 60 minutes/d, increase fruit consumption, increase water consumption, and reduce fat 2003 to 2005 37 13.1 (0.8) Girls: 20.2
Boys: 19.2
I2: group 1 plus parental involvement Girls: 20.2
Boys: 19.3
C: no PA and nutrition intervention Girls: 19.1
Boys: 18.5
Young 2006 I: PE curriculum taught 5 days/week and family 83 African American 100 13.8 (0.5) 25.0 (6.9)
C: standard PE class 100 25.2 (6.7)
Bayne‐Smith 2004 I: Physical Activity and Teenage Health programme, education sessions plus 20 to 25 minutes of PA 1994 to 1996 13 White
46 African American
29 Hispanic
12 Asian American 100 16.2 (1.3) 22.8 (4.1)
C: same frequency or duration of PE classes, but without lecture or discussion 5 White
45 African American
28 Hispanic
12 Asian American 100 15.9 (1.2) 23.6 (5.0)
Simon 2004 I: an educational component focusing on PA and sedentary behaviours and new opportunities for PA during and after school hours 2002 to 2006 23 recomposed family 54 11.6 (0.6) 18.7 (3.7)
C: — 26 recomposed family 48 11.7 (0.7) 18.9 (3.9)
Trevino 2004 I: health programming regarding 3 health behaviour messages associated with diabetes mellitus control and goal‐setting 2001 to 2002 6 Asian
7 African American
83 Mexican American
5 other 50 9.8 (0.5) 20.6 (5.1)
C: — 6 Asian
13 African American
77 Mexican American
4 other 49 9.8 (0.5) 20.3 (4.8)
Stone 2003 I: food service, skills‐based classroom curricula, family, and PE 1993 to 2000 "American Indian" 48 7.6 (0.6)
C: —
Burke 1998 I1: standard PA and nutrition programme including classroom lessons, fitness sessions daily, and nutrition programme 1993 to 1994 49 11 Low‐risk girls: 17.5
High‐risk girls: 19.2
I2: I1 plus a PA enrichment programme for higher‐risk children
C: no programme Low‐risk boys: 16.8
High‐risk boys: 21.1
Ewart 1998 I: 50‐minute 'Project Heart' aerobic exercise classes 1991 to 1994 30 White
70 African American 100 24.8 (5.8)
C: 50‐minute standard PE classes 100 24.1 (5.0)
Luepker 1996 I1: school food service modifications, PE interventions, and Child and Adolescent Trial for Cardiovascular Health curricula 1991 to 1994 69 White
13 African American
14 Hispanic
4 other 48 8.8 17.6 (0.1)
I2: I1 plus a family‐based programme
C: usual health curricula, PE, and food service programmes 17.6 (0.1)
Bush 1989 I1: 'Know Your Body' curriculum focusing on nutrition, fitness, prevention of smoking, a personalised health screening, and results on a 'health passport' for parents 1983 to 1988 "Predominantly black" 54 10.5 14.3
I2: 'Know Your Body' curriculum and health screening, but students do not receive the results of their screening; only parents receive the results
C: health screening only 14
Walter 1988 I: special curriculum targeting voluntary changes in risk behaviour in the areas of diet, PA, and smoking 1980 to 1985 Bronx:
32 White
43 Black
25 other
Westchester:
90 White
3 Black
7 other
Bronx: 49
Westchester: 46 Bronx: 9.0
Westchester: 8.9 Bronx: 11.8 (0.3)
Westchester: 11.3 (0.3)
C: — Bronx:
29 White
49 Black
22 other
Westchester:
79 White
15 Black
6 other
Bronx: 46
Westchester: 53 Bronx: 9
Westchester: 8.9 Bronx: 11.8 (0.3)
Westchester: 11.9 (0.3)

Appendix 3. Definition of endpoint measurement

Trial ID Physical Activity
Belton 2019 Actigraph accelerometer worn on the hip during all waking hours for 9 days; analysed using Evenson cut points for MVPA; reported in minutes/d
Corepal 2019 ActiGraph GT3X/+ accelerometers were worn for a minimum of 8 hours/d for at least 3 days. Activity counts were recorded using 1 second epochs, and were reintegrated in 60 second epochs before Evenson cut points were applied 
Jago 2019 ActiGraph wGT3X‐BT accelerometers worn for 7 consecutive days. Students who provided ≥ 3 valid days (500 minutes) of data were included in the analysis. MVPA was estimated using Evenson cut points. Total physical activity was derived from counts per minute
Lonsdale 2019a ActiGraph accelerometers attached at the right hip worn for 5 weekdays and 2 weekend days. Accelerometers assessed students’ moderate (38.26‐66.85 counts) and vigorous (> 66.86 counts) intensity during leisure time
Seljebotn 2019 Actigraph accelerometers were worn on the right hip for 7 days during all waking time. Data were collected in 10 second epochs, and MVPA was calculated using Evenson cut points
Zhou 2019
  Actigraph GT3X+ on the right hip during waking hours for 7 consecutive days. Accelerometry data reduction followed procedures developed for Chinese children 
Adab 2018 Actiheart accelerometer worn consecutively for 5 days, including a weekend. MVPA recorded as minutes/24 h of at least moderate intensity
Harrington 2018 A GENEActiv accelerometer was worn 24 hours/d for 7 days on their non‐dominant wrist at all time points. Devices were initialised with a sampling frequency of 100 Hz and were set to start recording at midnight on the first day of data collection and to stop recording at midnight 7 days later. Hildebrand cut points were used to estimate MVPA
Have 2018 Total daily PA were assessed using accelerometer (ActiGraph, GT3X and GT3X+, ActiGraph LLC, Pensacola, FL, USA). PA data were collected for 8 days, with a valid measurement of total PA defined as a minimum of 4 days with at least 10 hours of recorded activity each day. Total PA was expressed as mean counts per minute and as mean daily minutes in moderate to vigorous physical activity defined using Evenson cut points
Robbins 2018 ActiGraph GT3X+ accelerometers worn on an elastic belt at the right hip for 7 consecutive days, including 5 weekdays and 2 weekend days. An imputation approach based on all available data in hour blocks on all 7 days was implemented and wear time was standardised to 14 hours per day
Ten Hoor 2018 Measured using accelerometer (Actigraph GT3x, Actigraph, Pensacola, FL, USA) worn on the lower back for 5 consecutive days during all waking hours. Actilife software (v6.13.3) was used to generate activity counts per minute. Only students who had worn the accelerometer at least 8 hours per day during waking hours (i.e. time awake and time to bed) for a minimum of 3 days were included in the analyses. MVPA cut points were determined as proposed by Mattocks and colleagues
Farmer 2017 All children wore an accelerometer (ActiGraph GT3X, Actigraph Corp, Pensacola, FL, USA) 24 hours a day for 7 days, positioned over the right hip. Accelerometers were initialised using ActiLife in uniaxial mode using 15 second epochs. Data were cleaned and scored using an automated script developed in MATLAB (MathWorks, Natick, MA, USA) that removes the appropriate sleep period for each day for each child individually, to avoid sleep being misclassified as sedentary time. A day was considered valid if there were at least 8 valid awake hours. Non‐wear time (awake hours only) was defined as at least 20 minutes of consecutive zeros. Participants were excluded from analysis if less than 3 valid days of wear was obtained. Activity intensities were calculated using the Evenson cut points developed for children aged 5 to 8 years
Sutherland 2017 Accelerometers were used with non‐wear time defined as 30 minutes of consecutive zeros. Counts were collected in 15‐second epochs. The Evenson cut points were used to categorise the intensity of PA (moderate or vigorous)
Daly 2016 PA was measured by accelerometers (Actigraph GT1M, Pensacola, FL, USA) worn simultaneously, positioned on a belt around the waist. MVPA was defined as counts > 2296 per minute
Drummy 2016 Physical activity was measured using an Actigraph accelerometer (GT1M, Actigraph LLC, Pensacole, FL, USA) set to 5 second epochs. Children were asked to wear accelerometers over a 7‐day period (5 weekdays and 2 weekend days), only to be removed when sleeping, bathing, swimming, and showering
Kocken 2016 MVPA was measured using a 1‐dimensional accelerometer; the ActiGraph. Counts per minute were collected every 15 seconds. The Actigraph was worn on the child’s right hip during at least 3 days and was removed when water was involved and during sleeping time
Lau 2016 MVPA was measured using the ActiGraph GT3X+ accelerometer for 7 continuous days. Non‐wear time was determined as zero accelerometer counts for any continuous period of 20 minutes. Wear‐time validation criterion was set at least 480 minutes/d for 4 days during the 7 assessment days. Cut points for MVPA (> 2296 counts per minute) developed by Evenson were applied to calculate MVPA time
Resaland 2016 Physical activity was measured by ActiGraph accelerometers (ActiGraph GT3X+, LLC, Pensacola, Florida, USA). Children were instructed to wear the accelerometer on the right hip at all times over 7 consecutive days, except during water‐based activities or while sleeping. Wear time ≥ 480 minutes/day for ≥ 4 days was applied as a criterion for a valid measurement. Periods ≥ 20 minutes of zero counts were defined as non‐wear time. Evenson cut points for MVPA were used (2296 counts per minute)
Sutherland 2016 Accelerometer non‐wear time was defined as 30 minutes of consecutive zeroes. Counts were collected in 15‐second epochs and counts per minute calculated by dividing total accelerometer counts by minutes of wear time. The Evenson cut points were used to categorise the intensity of physical activity (moderate or vigorous)
Tarp 2016 Physical activity levels were assessed by accelerometer (GT3X and GT3X+ devices by ActiGraph LLC, Pensacola, FL, USA). Devices were worn on the right hip every day during a 7‐day period. The epoch was set to 2 seconds, but files were downloaded in 10 second epochs. A sequence of more than 30 minutes of consecutive zeroes was considered non‐wear time and was not included in analyses. To be included, students had to obtain a minimum of 4 days with at least 10 hours of valid registration at both time points. Evenson cut points were used to calculate time spent in MVPA
Cohen 2015 ActiGraph GT3Xþ accelerometers (ActiGraph, LLC, Fort Walton Beach, FL) were used. Children were required to wear accelerometers during waking hours for 7 consecutive days, except while bathing and swimming. Data were collected and stored in 10 second epochs with a frequency of 30 Hz. Valid wear time for total physical activity was defined as a minimum of 3 weekdays and a weekend day with at least 8 hours (480 minutes/d) of total wear time recorded. Non‐wear time was defined as strings of consecutive zeroes equating to 20 minutes. Evenson cut points were used to calculate time spent in sedentary (< 25 counts), light (26 to 573 counts), moderate (574 to 1002 counts), and vigorous (> 1003 counts) activity
Jago 2015 MVPA was assessed using an Actigraph GT3X+ accelerometer for 7 days. A valid day of accelerometer data was defined as a minimum of 500 minutes of data between 05:00 and 11.59 PM. Periods > 60 minutes in which zero values were recorded were interpreted as ‘non‐wear’ time. For valid days, mean minutes engaged in MVPA (≥ 2296 counts per minute) were derived
Andrade 2014 Physical activity was assessed using accelerometers (type GT‐256 and GT1M Actigraph, Florida USA) in a sub‐sample of adolescents selected using a random number. A syntax in Stata was used for data reduction and to compute registered time, and time spent in moderate to vigorous physical activity (≥ 760 counts/min). Accelerometers were worn for 5 weekdays and measurements were excluded with less than 540 minutes of registered time per day. The proportion of adolescents who met the recommended 60 minutes of MVPA per day was calculated
Jago 2014 Physical activity was assessed using an ActiGraph accelerometer (Model GT3X+; ActiGraph LLC, FL, USA) set to collect data at 30 Hz for a maximum of 5 days including a weekend day. Periods ≥ 60 minutes of zero values were defined as accelerometer “non‐wear”. Participants were included if they provided at least 2 weekdays of valid accelerometer data (at least 500 minutes of data between 6 am and 11 pm). Mean minutes of MVPA on a weekday was derived using a cut point ≥ 2296 counts per minute
Kipping 2014 The ActiGraph GT3X+ accelerometers were used for 5 days of data collection (3 weekdays and 2 weekend days) during the day (except when bathing, swimming, or participating in contact sports such as karate). Time spent in MVPA was any time spent in activities that were at least 2296 counts per minute
Kobel 2014 Objective measurements of physical activity were performed using the Actiheart® activity sensor continuously over a period of at least 4 to 6 successive days (2 weekend days and 2 to 4 weekdays)
Fairclough 2013 Physical activity was objectively assessed for 7 consecutive days using ActiGraph GT1M accelerometers and 5 second epochs. Sustained 20 minute periods of zero counts were considered non‐wear time. Valid wear time was at least 540 minutes on weekdays and 480 minutes on weekend days for a minimum of 3 days. Cut points ≥ 2160 counts per minutes and ≥ 4806 counts per minute classified moderate and vigorous intensity physical activity
Ford 2013 Physical activity levels were quantified using MTI accelerometers (Manufacturing Technologies Inc., Shalimar, FL) using a 1 minute epoch setting
Grydeland 2013 Children wore accelerometers (GT1M/CSA model 7164; ActiGraph, Fort Walton Beach, FL, USA) for 5 consecutive days and were instructed to wear the monitor continuously all awake hours except when doing water activities. Output was sampled every 10 seconds for 2 weekdays and 2 weekend days with valid wear time set at a minimum of 3 days and at least for 8 hours each day
Magnusson 2011 Accelerometers (Actigraph™ GT1M monitors) were worn during waking hours for 7 consecutive days ‐ 5 weekdays and 2 weekend days ‐ at a sampling epoch of 60 seconds. MVPA was defined as activity above 2000 CPM
Okely 2011 Participants wore an Actigraph accelerometer (7164 and GT1M models; Fort Walton Beach, FL) for 7 consecutive days attached to an adjustable elastic belt over the right hip. Data were collected in 30 second epochs. Thirty‐second activity counts were uploaded to determine the amount of time spent in light (LPA; 1.5‐2.9 METs) moderate (MPA; 4‐6.9 METs), and vigorous (VPA; ≥ 7 METs) physical activity
Wilson 2011 Assessments of MVPA were obtained with omni‐directional Actical accelerometers (Mini‐Mitter, Bend, OR) over 7 consecutive days. Data were recorded in 1 minute epochs and were converted into time spent MVPA (3 to 9 METS) based on Actical‐specific activity count thresholds where MVPA = 1500 to 6500 and VPA ≥ 6500
Kriemler 2010 PA was monitored with an accelerometer, which was worn continuously around the hip for 5 weekdays ‐ at baseline and at the end of the intervention
Donnelly 2009 Accelerometers were worn over 4 consecutive days, which included 2 weekdays and 2 weekend days
Peralta 2009 Weekday MVPA (minutes/d). PA was measured over 7 consecutive days using MTI 7164 Actigraph accelerometers worn on belts at the right hip. Average minutes of moderate (MPA), vigorous (VPA), and MVPA were calculated using a composite method
Salmon 2008 PA was assessed using Manufacturing Technology Inc. AM7164‐2.2C accelerometers. Children wore the MTI on a belt positioned over the right hip during waking hours, except when bathing or swimming, for 8 days at each of the 4 measurement points
Webber 2008 MET‐weighted minutes of MVPA using accelerometers worn for 7 consecutive days except while bathing, swimming, or sleeping
Haerens 2006 Children wore the accelerometer for 6 days above the right hip bone, underneath the clothes. Accelerometers were set to measure activity counts in an epoch time of 1 minute. Cut points > 3200 moderate to vigorous minutes were used
Trial ID Sedentary time
Corepal 2019 ActiGraph GT3X/+ accelerometers were worn for a minimum of 8 hours/d for at least 3 days. Activity counts were recorded using 1 second epochs and reintegrated to 60 second epochs before Evenson cut points were applied 
Jago 2019 ActiGraph wGT3X‐BT accelerometers worn for 7 consecutive days. Students who provided ≥ 3 valid days (500 minutes) of data were included in the analysis. Sedentary time was derived based on a cut point < 100 CPM
Lonsdale 2019a ActiGraph accelerometers attached at the right hip worn for 5 weekdays and 2 weekend days. Accelerometers assessed students’ sedentary behaviour (< 1.67 counts per 1 second) during leisure time
Seljebotn 2019 Actigraph accelerometers were worn on the right hip for 7 days during all waking time. Data were collected in 10 second epochs, and sedentary time was calculated using Evenson cut points
Zhou 2019
  Actigraph GT3X+ accelerometer was worn on the right hip during waking hours for 7 consecutive days. Data reduction followed procedures developed for Chinese children 
Adab 2018 Actiheart accelerometer worn consecutively for 5 days, including a weekend. Sedentary time reported in hours/d
Harrington 2018 A GENEActiv accelerometer was worn 24 hours/d for 7 days on the non‐dominant wrist at all time points. Devices were initialised with a sampling frequency of 100 Hz and were set to start recording at midnight on the first day of data collection and to stop recording at midnight 7 days later. Hildebrand cut points were used to estimate sedentary time
Ten Hoor 2018 Measured using accelerometer (Actigraph GT3x, Actigraph, Pensacola, FL, USA) worn on the lower back for 5 consecutive days during all waking hours. Actilife software (v6.13.3) was used to generate activity counts per minute. Only students who had worn the accelerometer at least 8 hours per day during waking hours (i.e. time awake and time to bed) for a minimum of 3 days were included in the analyses. Sedentary time cut points were determined as proposed by Mattocks and colleagues, and were reported as % of time spent sedentary
Daly 2016 PA was measured by accelerometers (Actigraph GT1M, Pensacola, FL, USA) worn simultaneously, positioned on a belt around the waist. Sedentary activity was defined as < 100 counts per minute
Kocken 2016 Sedentary time was measured using a 1‐dimensional accelerometer; the ActiGraph. Counts per minute were collected every 15 seconds. The Actigraph was worn on the child’s right hip during at least 3 days and the ActiGraph was removed when water was involved and during sleeping time
Resaland 2016 Sedentary time was measured by ActiGraph accelerometers (ActiGraph GT3X+, LLC, Pensacola, Florida, USA). Children were instructed to wear the accelerometer on the right hip at all times over 7 consecutive days, except during water‐based activities or while sleeping. Wear time ≥ 480 minutes/d for ≥ 4 days was applied as a criterion for a valid measurement. Periods ≥ 20 minutes of zero counts were defined as non‐wear time. Evenson cut points for sedentary time (0 to 100 counts per minute) were used
Jago 2015 Sedentary time was assessed using an Actigraph GT3X+ accelerometer worn for 7 days. A valid day of accelerometer data was defined as a minimum of 500 minutes of data between 05:00 and 11.59 pm. Periods > 60 minutes in which zero values were recorded were interpreted as ‘non‐wear’ time. Sedentary cut points used were not reported
Andrade 2014 Sedentary time was assessed using accelerometers (type GT‐256 and GT1M Actigraph, Florida USA) in a sub‐sample of adolescents selected using a random number. A syntax in Stata was used for data reduction and to compute registered time and time spent in sedentary activity (≤ 100 counts/min). Accelerometers were worn for 5 weekdays and measurements were excluded with less than 540 minutes of registered time per day
Kipping 2014 The ActiGraph GT3X+ accelerometers were used for 5 days of data collection (3 weekdays and 2 weekend days) during the day (except when bathing, swimming, or participating in contact sports such as karate). Time spent sedentary was time spent in activities between 0 and 100 counts per minute
Toftager 2014 Accelerometers were worn during all waking hours for 7 consecutive days except when doing water activities. Strings of 60 minutes or longer of consecutive zeroes, allowing for 2 epoch periods of non‐zero interruptions, were interpreted to represent non‐wear time. Valid data were at least 3 days with at least 10 hours (600 minutes) of activity per day. Evenson activity cut points were used to calculate sedentary time (> 100 CPM) expressed as minutes per day
Fairclough 2013 Sedentary time was objectively assessed for 7 consecutive days using ActiGraph GT1M accelerometers and 5 second epochs. Sustained 20 minute periods of zero counts were considered non‐wear time. Valid wear time was at least 540 minutes on weekays and 480 minutes on weekend days for a minimum of 3 days. Cut points of 100 counts per minute were classified as sedentary time
Grydeland 2013 Children wore accelerometers (GT1M/CSA model 7164; ActiGraph, Fort Walton Beach, FL, USA) for 5 consecutive days and were instructed to wear the monitor continuously all awake hours except when doing water activities. Output was sampled every 10 seconds for 2 weekdays and 2 weekend days with valid wear time set at a minimum of 3 days and for at least 8 hours each day
Okely 2011 Participants wore an Actigraph accelerometer (7164 and GT1M models; Fort Walton Beach, FL) for 7 consecutive days attached to an adjustable elastic belt over the right hip. Data were collected in 30 second epochs. Thirty second activity counts were uploaded to determine the amount of time spent in sedentary activity
Webber 2008 Minutes of sedentary time using accelerometers worn for 7 consecutive days except while bathing, swimming, or sleeping
Haerens 2006 Children wore the accelerometer for 6 days above the right hip bone, underneath the clothes. Accelerometers were set to measure activity counts in an epoch time of 1 minute. Cut points < 800 were used for sedentary time
Trial ID Physical fitness
Breheny 2020 British Athletics Linear Track Test, children encouraged to run as far as they could in 2 minutes on a pre‐measured 50 metre linear track
Ketelhut 2020 6‐minute run test
Leahy 2019 20 metre PACER shuttle run test. The last successful stage was recorded and was converted into the number of 20 metre laps
Müller 2019 20 metre shuttle run test, adhering to a standard test protocol. Most schoolchildren wore school or street shoes, and some ran barefoot. The number of fully completed laps was recorded and was converted to VO2max values, according to a standard protocol
Ordóñez Dios 2019 Cardiorespiratory capacity was evaluated using a 1 km test
Seibert 2019 PACER 20 metre shuttle run test, terminated when the participant fails to complete the 20 metre run in the allotted time twice. PACER score expressed in number of laps completed and converted to a z‐score for age and sex
Seljebotn 2019 Aerobic fitness was assessed by the Andersen test, a 10‐minute interval running test. Results are expressed as distance run in 10 minutes on a 20 metre course
Zhou 2019
  20 metre shuttle‐run test was used to assess cardiorespiratory fitness
Carlin 2018 Queens College Step Test was used; participants wore a heart rate monitor during the step test, with heart rate recorded at baseline, and at 10 seconds, 15 seconds, and 20 seconds following completion of the step test and was used to estimate VO2max as mL/kg/min
Have 2018 Aerobic fitness was assessed using the Andersen test. Children were instructed to run as far as possible in 10 minutes back and forth between 2 lines 20 metres apart. The test score was total distance in metres run by each child
Pablos 2018 20 metre shuttle‐run test was used to determine the maximal oxygen consumption (VO2max). Indirect incremental multi‐stage field test over a distance of 20 metres to exhaustion using the pace set by a CD emitting beep signals at preset intervals. The initial speed was set at 8.5 km/hr for the first minute and was increased by 0.5 km/hr each subsequent minute
Robbins 2018 The Progressive Aerobic CV Endurance Run (PACER) test, a 15 or 20 metre shuttle run was used. Participants ran from 1 line to another on a flat surface, according to audio cues, which increase in pace until participants can no longer complete laps in the time allotted. The number of laps completed is converted to estimated VO2 for analysis
Donnelly 2017 The Progressive Aerobic Cardiovascular Endurance Run (PACER) by Leger was used. Participants ran a 20 metre shuttle course with 1 minute stages, paced by an audible beep. The number of laps completed constituted the PACER score
Torbeyns 2017 The 20 metre shuttle run test involved running continuously between 2 points that are 20 metres apart from side to side. These runs are synchronised with a prerecorded CD, which beeps at set intervals. As the test proceeds, intervals between successive beeps decrease, forcing the athlete to increase speed over the course of the test, until it is impossible to keep in sync with the recording (or, on extremely rare occasions, until the athlete completes the test). The recording is structured into 21 ‘levels’, each of which lasts around 62 seconds. The interval of beeps is calculated as requiring speed at the start of 8.5 km/hr, increasing by 0.5 km/hr with each level thereafter. The highest level reached was used as the outcome measure
de Greeff 2016 20 metre shuttle run, cardiorespiratory endurance, in number of completed stages from the EUROFIT fitness battery
Jarani 2016 The Andersen intermittent shuttle run test was used to estimate maximal oxygen uptake. Children had to run as fast as they could to cover the longest possible distance during the 10 minute test run, and this distance was the test result. To estimate child’s VO2max, the equation: VO2max = 18.38 + (0.03301 × distance) – (5.92 × sex) [(boys = 0; girls = 1)] was used
Lau 2016 Assessed using the Progressive Aerobic Cardiovascular Endurance Run 20 metre shuttle run performance test by Leger. The number of laps completed for all participants was recorded and maximal shuttle run speed was calculated accordingly. Aerobic fitness was estimated using the most current cross‐validated regression model for predicting VO2max
Resaland 2016 Aerobic fitness was measured with an Andersen intermittent practical running field test administered according to standard procedures: Children ran from one end line to another (20 metres apart) in an intermittent to‐and‐fro movement, with 15 second work periods and 15 second breaks (standing still), for a total duration of 10 minutes. We recorded the distance covered as the outcome for the analysis. To enable comparing of aerobic fitness level across studies, VO2peak was calculated using the equation suggested by Aadland
Tarp 2016 Cardiorespiratory fitness was assessed by the Andersen test, a 10‐minute intermittent running test with total distance in metres used as the test result
Cohen 2015 Cardiorespiratory fitness was assessed using the Leger 20 metre multi‐stage fitness test. Participants were required to run back and forth between 2 lines over a 20 metre distance within a set time limit. Running speed started at 8.5 km/hr and increased by 0.5 km/hr each minute. The test was completed when a participant failed to reach the line for 2 consecutive shuttles. Scores were recorded as the level and shuttle reached, which was converted to the number of 20 metre laps completed
Madsen 2015 Participants completed the 1 mile run as a measure of cardiorespiratory fitness, with results presented as minutes to completion
Muros 2015 Maximal oxygen uptake (VO2max) was estimated using a 20 metre incremental‐maximum shuttle run field test, employing the equation proposed by Ruiz. The shuttle run test involves running to and from between 2 lines placed 20 metres apart. Participants start at an initial velocity of 8.5 kph and increase their speed by 0.5 kph for every 20 metres covered as indicated by an audio recording played on a validated CD‐ROM. The test concludes when the subject is unable to reach the line on 2 consecutive occasions at the speed demanded by the audio recording
Suchert 2015 The 20 metre shuttle run test (by Leger et al) was used. In this field test, participants run back and forth at a distance of 20 metres in a given time interval indicated by pre‐recorded audio signals. The required running pace starts with 8.0 km/hr and continuously increases by 0.5 km/hr each minute. The test stops either when students abandon by themselves or when they fail to reach the line by the sound for the second time. The total number of completed laps was used for statistical analyses. In addition, maximal oxygen consumption (VO2max) was estimated using the quadratic model by Mahar
Andrade 2014 The EUROFIT test battery was used to assess cardiorespiratory endurance via the 20 metre shuttle run test
Nogueira 2014 The 20 metre shuttle run test (aka, the beep test) was used and VO2max was estimated according to the velocity associated with the level reached by the participant, by using the algorithm VO2max = 31.025 + (3.238 × velocity) − (3.248 × age) + (0.1536 × age × velocity). Participants ran on an indoor surface between 2 points marked on the ground 20 metres apart; once the participant was unable to meet the required speed on successive laps, the level achieved was recorded, and associated velocity was entered into the algorithm along with age
Toftager 2014 Aerobic fitness was measured using the Andersen test (20 metre shuttle run) expressed as metres completed
Fairclough 2013 VO2peak was assessed using an individually calibrated continuous incremental treadmill (H/P/ Cosmos, Traunstein, Germany) test to volitional exhaustion, under ambient conditions, using an online gas analysis system (Jaeger Oxycon Pro; Viasys Healthcare, Warwick, UK)
Aburto 2011 Participants ran around a calibrated track for 9 minutes, and the distance travelled was recorded in metres
Ardoy 2011 A 20 yard or metre shuttle run was used to assess cardiorespiratory fitness
de Heer 2011 The Progressive Aerobic Cardiovascular Endurance Run (PACER) test requires participants to run up and down a 20 metre court. At each side of the court, a beep sounds to signal the student to turn around and run back. The test increases in speed every minute and is completed when a student fails to reach the other side in time for the signal for the second time. Total number of laps completed was reported
Jago 2011 Fitness was assessed by the 20 metre shuttle test (20 metre) using standard procedures by Leger
Jansen 2011 Fitness was measured using the 20 metre shuttle run following the EUROFIT protocol
Magnusson 2011 Cardiorespiratory fitness (W/kg) was measured with a Monark ergometer bike using the study protocol from the European Youth Heart study. This maximal ergometer test is run such that every 3 minutes, the weight on the wheel is increased by 20 to 25 W, depending on the participant’s weight. Each participant keeps a steady pace on the bike until exhaustion, or until he or she can no longer keep a steady pace
Thivel 2011 The 20 metre shuttle run test developed by Léger was used. Children were instructed to run as long as possible between 2 lines 20 metres apart at an increasing speed imposed by emitted tones at intervals. The speed began at 8 km/hr and increased by 0.5 km/hr every minute. As soon as a child was not able to complete a whole stage, the test was stopped; the child’s score corresponded to the last fully completed stage
Kriemler 2010 The Leger 20 metre shuttle run test was used, with results reported as number of laps completed
Walther 2009 All participants underwent a graded treadmill test with spirometry until exhaustion, according to a modified Bruce protocol for children starting at 1.7 mph and 0 degrees
Reed 2008 Leger's 20 metre incremental shuttle run was used, which was designed for children and provides age and sex reference normative data. Children ran 20 metre laps at 8.5 km/hr‐1. Running speed then increased by 0.5 km hr‐1 each minute. Children continued running until they could no longer maintain preset and standardised pace. Total laps were recorded
Wang 2008 Biological measurements were made in a mobile laboratory that was brought to the school sites. Fitness level was assessed by HR at completion of the bench‐stepping test. Low HR at the end of 3 minutes of stepping indicates better CVF
Barbeau 2007 Oxygen consumption (VO₂) was measured using a Sensormedics Vmax 229 cardiopulmonary system (Yorba Linda, CA). The treadmill protocol began with a 4 minute warm‐up at 0% grade and 2.0 mph. The speed was then increased by 0.5 mph every 2 minutes until reaching 3.0 mph, at which time the grade increased to 2% for 2 minutes, then increased an additional 3% every 2 minutes until reaching 20% grade or exhaustion
Bayne‐Smith 2004 Measured fitness level as recovery from Queens College step test. Subjects stepped up and down a step for 3 minutes at 22 steps per minute. HRs were counted for 15 seconds beginning 5 seconds after stepping ended
Trevino 2004 Outcome was measured as physical fitness score using a modified Harvard step test. Baseline HR was recorded. Child then stepped on and off a stool with both feet for 5 minutes. The student was paced at 30 cycles per minute. A physical fitness score was calculated from the total time of exercise (in seconds) multiplied by 100 and was divided by the sum of 3 HR values measured at 0, 1, and 2 minutes after exercise
Burke 1998 Physical fitness was measured by laps completed in the Leger Shuttle Run, in which children ran 20 metre laps in time to a tape recording of beeps at a predetermined pace, continuing until they were unable to keep pace with the recording

Appendix 4. 'Risk of bias' assessment

'Risk of bias' domains
Random sequence generation (selection bias due to inadequate generation of a randomised sequence)
For each included trial, we described the method used to generate the allocation sequence in sufficient detail to allow an assessment of whether it should produce comparable groups.
  • Low risk of bias: trial authors achieved sequence generation using computer‐generated random numbers or a random numbers table. Drawing lots, tossing a coin, shuffling cards or envelopes, and throwing dice are adequate if an independent person performed this who was not otherwise involved in the trial. We will consider use of the minimisation technique as equivalent to being random

  • Unclear risk of bias: insufficient information about the sequence generation process

  • High risk of bias: the sequence generation method was non‐random or quasi‐random (e.g. sequence generated by odd or even date of birth; sequence generated by some rule based on date (or day) of admission; sequence generated by some rule based on hospital or clinic record number; allocation by judgment of the clinician; allocation by preference of the participant; allocation based on the results of a laboratory test or a series of tests; or allocation by availability of the intervention)


Allocation concealment (selection bias due to inadequate concealment of allocation prior to assignment)
We described for each included trial the method used to conceal allocation to interventions prior to assignment, and we assessed whether intervention allocation could have been foreseen in advance of or during recruitment or changed after assignment.
  • Low risk of bias: central allocation (including telephone, interactive voice‐recorder, Internet‐based and pharmacy‐controlled randomisation); sequentially numbered drug containers of identical appearance; sequentially numbered, opaque, sealed envelopes

  • Unclear risk of bias: insufficient information about allocation concealment

  • High risk of bias: open random allocation schedule (e.g. a list of random numbers) used; assignment envelopes used without appropriate safeguards; alternation or rotation; date of birth; case record number; any other explicitly unconcealed procedure


We also evaluated trial baseline data to incorporate assessment of baseline imbalance into the 'Risk of bias' judgement for selection bias (Corbett 2014). Chance imbalances may also affect judgements on risk of attrition bias. In the case of unadjusted analyses, we distinguished between trials that we rated as being at low risk of bias on the basis of both randomisation methods and baseline similarity, and trials that we judged as being at low risk of bias on the basis of baseline similarity alone (Corbett 2014). We re‐classified judgements of unclear, low, or high risk of selection bias as specified in Appendix 4.
Blinding of participants and study personnel (performance bias due to knowledge of allocated interventions by participants and personnel during the trial)
  • Low risk of bias: blinding of participants and key study personnel was ensured, and it was unlikely that the blinding could have been broken; no blinding or incomplete blinding, but we judge that the outcome is unlikely to have been influenced by lack of blinding

  • Unclear risk of bias: insufficient information about blinding of participants and study personnel

  • High risk of bias: no blinding or incomplete blinding; outcome is likely to have been influenced by lack of blinding; blinding of trial participants and key personnel attempted, but likely that blinding could have been broken, and the outcome is likely to be influenced by lack of blinding


Blinding of outcome assessment (detection bias due to knowledge of allocated interventions by outcome assessment)
  • Low risk of bias: blinding of outcome assessment is ensured, and it was unlikely that blinding could have been broken; no blinding of outcome assessment, but we judge that the outcome measurement was unlikely to have been influenced by lack of blinding

  • Unclear risk of bias: insufficient information about blinding of outcome assessors

  • High risk of bias: no blinding of outcome assessment, and outcome measurement was likely to have been influenced by lack of blinding; blinding of outcome assessment, but likely that blinding could have been broken, and the outcome measurement was likely to be influenced by lack of blinding


Incomplete outcome data (attrition bias due to quantity, nature, or handling of incomplete outcome data)
For each included trial or each outcome, or both, we described the completeness of data, including attrition and exclusions from analyses. We stated whether the trial reported attrition and exclusions, and we reported the number of participants included in the analysis at each stage (compared with the number of randomised participants per intervention/comparator groups). We also noted if the trial reported reasons for attrition or exclusion, and whether missing data were balanced across groups or were related to outcomes. We considered the implications of missing outcome data per outcome such as high dropout rates (e.g. above 15%) or disparate attrition rates (e.g. difference of 10% or more between trial arms).
  • Low risk of bias: no missing outcome data; reasons for missing outcome data unlikely to be related to true outcomes (for survival data, censoring unlikely to introduce bias); missing outcome data balanced in numbers across intervention groups, with similar reasons for missing data across groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk was not enough to have a clinically relevant impact on the intervention effect estimate; for continuous outcome data, plausible effect size (mean difference or standardised mean difference) among missing outcomes was not enough to have a clinically relevant impact on observed effect size; appropriate methods, such as multiple imputations, were used to handle missing data

  • Unclear risk of bias: insufficient information to assess whether missing data in combination with the method used to handle missing data were likely to induce bias

  • High risk of bias: reason for missing outcome data was likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk enough to induce clinically relevant bias in the intervention effect estimate; for continuous outcome data, plausible effect size (mean difference or standardised mean difference) among missing outcomes enough to induce clinically relevant bias in observed effect size; 'as‐treated' or similar analysis done with substantial departure of the intervention received from that assigned at randomisation; potentially inappropriate application of simple imputation


Selective reporting (reporting bias due to selective outcome reporting)
We assessed outcome reporting bias by integrating results of the appendix 'Matrix of trial endpoints (publications and trial documents)' (Boutron 2014; Jones 2015a; Mathieu 2009), with those of the appendix 'High risk of outcome reporting bias according to the Outcome Reporting Bias In Trials (ORBIT) classification' (Kirkham 2010). This analysis formed the basis for the judgement of selective reporting.
  • Low risk of bias: the trial protocol was available and all the trial's pre‐specified (primary and secondary) outcomes that were of interest to this review were reported in the pre‐specified way; the study protocol was unavailable, but it is clear that published reports included all expected outcomes (ORBIT classification)

  • Unclear risk of bias: insufficient information about selective reporting

  • High risk of bias: not all the trial's pre‐specified primary outcomes were reported; 1 or more primary outcomes were reported using measurements, analysis methods, or subsets of the data (e.g. sub‐scales) that were not pre‐specified; 1 or more reported primary outcomes were not pre‐specified (unless clear justification for their reporting was provided, such as an unexpected adverse effect); 1 or more outcomes of interest in the Cochrane Review were reported incompletely, so that we could not enter them into a meta‐analysis; the trial report failed to include results for a key outcome that we would expect to have been reported for such a trial (ORBIT classification)


Specific to cluster‐RCTs: recruitment bias
For all cluster‐RCTs, we assessed recruitment bias by assessing whether individual participants were recruited to the trial, and individual‐level data were collected before or after clusters were randomised to an intervention or control group. This served as the basis for judgement of recruitment bias. Although the unit of randomisation is the cluster, bias could be introduced if individual participants knew whether the school or the classroom would receive the intervention or control condition prior to deciding whether or not to join the study.
  • Low risk of bias: investigators describe a procedure in which all participants were recruited and data were collected before randomising clusters to intervention or control groups, or if individual participants were not recruited at all but were identified prior to randomisation

  • Unclear risk of bias: insufficient information about the process to permit judgement

  • High risk of bias: clusters were randomised before recruitment to the trial was complete, and knowledge of whether each cluster was an ‘intervention’ or ‘control’ cluster could affect types of participants recruited


Specific to cluster‐RCTs: baseline imbalance
Often in cluster‐randomised trials, individuals within a cluster are more similar than participants across clusters. Particularly when the number of clusters randomised is small, there may be differences in baseline characteristics across study groups even if randomisation was successful.
  • Low risk of bias: investigators explored baseline imbalances and report that no imbalance was found or properly adjusted for any baseline balance in the analysis

  • Unclear risk of bias: investigators did not explore baseline imbalances, or it Is unclear whether any baseline imbalances exist

  • High risk of bias: baseline imbalances were observed between study groups and were not accounted for in the analysis


Specific to cluster‐RCTs: loss of clusters
  • Low risk of bias: no individual outcomes or clusters of missing data; reasons for clusters or outcomes missing unlikely to be related to intervention (for survival data, censoring unlikely to be introducing bias); missing outcome data balanced in numbers across clusters, with similar reasons for missing data across clusters; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk was not enough to have a clinically relevant impact on the intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardised difference in means) among missing outcomes not enough to have a clinically relevant impact on observed effect size; missing data have been imputed using appropriate methods

  • Unclear risk of bias: insufficient reporting of attrition/exclusions to permit judgement (e.g. number randomised not stated, no reason for missing data provided)

  • High risk of bias: reason for missing outcome data likely to be related to true outcome, with either imbalance in numbers or reasons for missing data across intervention groups; for dichotomous outcome data, the proportion of missing outcomes compared with observed event risk enough to induce clinically relevant bias in intervention effect estimate; for continuous outcome data, plausible effect size (difference in means or standardised difference in means) among missing outcomes enough to induce clinically relevant bias in observed effect size; 'as‐treated' analysis done with substantial departure of interventions received from that assigned at randomisation; potentially inappropriate application of simple imputation


Specific to cluster‐RCTs: incorrect analysis
Because participants in any cluster are often more similar than participants across clusters and tend to respond to an intervention in a similar manner, their data cannot be assumed to be independent. Failing to account for this is often referred to as 'unit‐of‐analysis error' because the unit of analysis is different from the unit of allocation (Whiting‐O'Keefe 1984), and many cluster‐randomised trials have been incorrectly analysed in this manner (Eldridge 2008).
  • Low risk of bias: investigators clearly describe consideration of the clustered nature of the data in their statistical analysis

  • Unclear risk of bias: insufficient reporting of statistical analysis procedures to permit judgement

  • High risk of bias: data are analysed with the unit of analysis at the individual level, and investigators do not consider the clustered nature of data in the analysis

Appendix 5. Health‐related quality of life: instruments

Instrument Dimensions (sub‐scales)
(no. of items) Validated
instrument Answer options Scores Minimum score
Maximum score
Weighting
of scores Direction
of
scales Minimal important difference
Strengths and Difficulties Questionnaire
Employed in
(Leahy 2019)
  • Prosocial behavior (5)

  • Emotional symptoms (5)

  • Conduct problems (5)

  • Hyperactivity (5)

  • Peer problems (5)

Yes (Goodman 1998 and Goodman 2001) 3‐point scale Item scores range from 0 to 2
Scale scores range from 0 to 10 0 to 40 Higher scores indicate more psychological difficulties
Perceived Stress Scale
Employed in
(Leahy 2019)
10 items Yes (Cohen 1983 and Chan 2013) 5‐point scale Item scores range from 0 to 4 0 to 40 Higher scores indicate a greater degree of subjective stress experienced by participants
KIDSCREEN
KIDSCREEN‐10
Employed in
(Jago 2019)
KIDSCREEN‐27
Employed in
Resaland 2016
KIDSCREEN‐10: 10 items
KIDSCREEN‐27:
  • Physical Well‐being (5)

  • Psychological Well‐being (7)

  • Autonomy & Parent Relations (7)

  • Social Support & Peers (4)

  • School Environment (4)

Yes
(Ravens‐Sieberer 2006, Ravens‐Sieberer 2010, and Ravens‐Sieberer 2014)
5‐point scale Items are scored as Rasch scales, then translated into T‐values with mean of 50 and standard deviation of 10 1 to 5 Higher T‐scores indicate higher HRQoL
Child Health Utility 9D (CHU9D)
Employed in Breheny 2020,
Jago 2019,
Harrington 2018 9 items Yes (Stevens 2011 and Stevens 2012) 5‐point scale Item levels range from 1 to 5 1 to 5 Yes Higher item scores indicate lower level of health
Pediatric quality of life inventory(PedsQL)
Employed in
Adab 2018
  • Physical Functioning (8)

  • Emotional Functioning (5)

  • Social Functioning (5)

  • School Functioning (5)

Yes (Varni 2003 and Varni 2017) 5‐point scale (aged 8 to 18)
3‐point scale (aged 5 to 7)
Item scores range from 0 to 4
Scale scores range from 0 to 100 0 to 100 No Higher scores indicate better HRQoL
  • Total score: 4.36

  • Physical Health: 6.66

  • Psychosocial Health: 5.30

  • Emotional Functioning: 8.94

  • Social Functioning: 8.36

  • School Functioning: 9.12

European Quality of Life‐5 Dimensions Youth Survey (EQ‐5D‐Y)
Employed in
Jago 2015 Descriptive system
  • Mobility (1)

  • Self‐care (1)

  • Everyday activities (1)

  • Pain/discomfort (1)

  • Happiness/worry/sadness (1)


Visual analogue scale: no sub‐scales
Yes (Ravens‐Sieberer 2010 and van Reenen 2014) Descriptive system: 3‐level scale
Visual analogue scale: vertical, continuous scale
Descriptive system: item levels range from 1 to 3
Visual analogue scale: 0 to 100
Descriptive system: 1 to 3
Visual analogue scale: 0 to 100 Yes Descriptive system: higher scores indicate worse health states
Visual analogue scale: higher scores indicate better perceived health state
 

Data and analyses

Comparison 1. PA programme vs no PA programme.

Outcome or subgroup title No. of studies No. of participants Statistical method Effect size
1.1 Physical activity participation: all data 5   Other data No numeric data
1.1.1 Children: before and after school programme 1   Other data No numeric data
1.1.2 Children: schooltime PA 1   Other data No numeric data
1.1.3 Children: multi‐component interventions 1   Other data No numeric data
1.1.4 Adolescents: multi‐component interventions 2   Other data No numeric data
1.2 Physical activity duration (minutes/d): meta‐analysis 33 20614 Mean Difference (IV, Random, 95% CI) 0.73 [0.16, 1.30]
1.2.1 Children 22 10715 Mean Difference (IV, Random, 95% CI) 1.01 [0.08, 1.93]
1.2.2 Adolescents 11 9899 Mean Difference (IV, Random, 95% CI) 1.84 [0.34, 3.35]
1.3 Physical activity duration by intervention type (minutes/d): meta‐analysis  33 20614 Mean Difference (IV, Random, 95% CI) 0.73 [0.16, 1.30]
1.3.1 Before and after school programmes 6 2571 Mean Difference (IV, Random, 95% CI) 0.77 [‐1.40, 2.94]
1.3.2 Enhanced PE 3 2050 Mean Difference (IV, Random, 95% CI) ‐0.23 [‐1.58, 1.11]
1.3.3 Multi‐component interventions 16 12135 Mean Difference (IV, Random, 95% CI) 2.42 [0.62, 4.22]
1.3.4 Schooltime PA 8 3858 Mean Difference (IV, Random, 95% CI) 5.30 [0.89, 9.72]
1.4 Physical activity duration: additional data 6   Other data No numeric data
1.4.1 Before and after school programmes 1   Other data No numeric data
1.4.2 Enhanced PE 2   Other data No numeric data
1.4.3 Multi‐component intervention 2   Other data No numeric data
1.4.4 Schooltime PA 2   Other data No numeric data
1.5 Sedentary time (minutes/d): meta‐analysis 16 11914 Mean Difference (IV, Random, 95% CI) ‐3.78 [‐7.80, 0.24]
1.5.1 Children 11 5766 Mean Difference (IV, Random, 95% CI) ‐3.35 [‐9.30, 2.60]
1.5.2 Adolescents 5 6148 Mean Difference (IV, Random, 95% CI) ‐5.67 [‐11.48, 0.14]
1.6 Sedentary time (minutes/d) by intervention type: meta‐analysis 16 11914 Mean Difference (IV, Random, 95% CI) ‐3.78 [‐7.80, 0.24]
1.6.1 Before and after school programmes 2 773 Mean Difference (IV, Random, 95% CI) 2.01 [‐15.28, 19.31]
1.6.2 Enhanced PE 1 540 Mean Difference (IV, Random, 95% CI) ‐11.18 [‐21.96, ‐0.40]
1.6.3 Multi‐component interventions 11 9164 Mean Difference (IV, Random, 95% CI) ‐4.60 [‐9.08, ‐0.12]
1.6.4 Schooltime PA 2 1437 Mean Difference (IV, Random, 95% CI) ‐3.26 [‐19.05, 12.52]
1.7 Sedentary time: additional data 4   Other data No numeric data
1.7.1 Before and after school programmes 1   Other data No numeric data
1.7.2 Enhanced PE 3   Other data No numeric data
1.7.5 Multi‐component intervention 1   Other data No numeric data
1.8 Physical fitness (mL/kg/min): meta‐analysis 13 3980 Mean Difference (IV, Random, 95% CI) 1.19 [0.57, 1.82]
1.8.1 Children 9 2215 Mean Difference (IV, Random, 95% CI) 1.47 [0.84, 2.09]
1.8.2 Adolescents 4 1765 Mean Difference (IV, Random, 95% CI) 0.58 [‐0.18, 1.35]
1.9 Physical fitness (mL/kg/min) by intervention type: meta‐analysis 13 3980 Mean Difference (IV, Random, 95% CI) 1.19 [0.57, 1.82]
1.9.1 Before and after school programmes 5 724 Mean Difference (IV, Random, 95% CI) 1.38 [0.34, 2.41]
1.9.2 Enhanced PE 4 1387 Mean Difference (IV, Random, 95% CI) 1.99 [0.76, 3.21]
1.9.3 Multi‐component interventions 3 1697 Mean Difference (IV, Random, 95% CI) ‐0.33 [‐0.73, 0.08]
1.9.4 Schooltime PA 1 172 Mean Difference (IV, Random, 95% CI) 2.70 [1.04, 4.36]
1.10 Physical fitness: additional data 29   Other data No numeric data
1.10.1 Before or after school programme 4   Other data No numeric data
1.10.2 Enhanced PE 5   Other data No numeric data
1.10.3 Multi‐component intervention 11   Other data No numeric data
1.10.4 Schooltime PA 10   Other data No numeric data
1.11 BMI: meta‐analysis [z‐scores] 21 22948 Mean Difference (IV, Random, 95% CI) ‐0.06 [‐0.09, ‐0.02]
1.11.1 Children 16 15732 Mean Difference (IV, Random, 95% CI) ‐0.06 [‐0.11, ‐0.01]
1.11.2 Adolescents 5 7216 Mean Difference (IV, Random, 95% CI) ‐0.03 [‐0.05, ‐0.00]
1.12 BMI by intervention type: meta‐analysis [z‐scores] 21 22948 Mean Difference (IV, Random, 95% CI) ‐0.06 [‐0.09, ‐0.02]
1.12.1 Before and after school programmes 2 1615 Mean Difference (IV, Random, 95% CI) ‐0.02 [‐0.05, 0.01]
1.12.2 Enhanced PE 1 174 Mean Difference (IV, Random, 95% CI) ‐0.08 [‐0.29, 0.13]
1.12.3 Multi‐component interventions 17 19489 Mean Difference (IV, Random, 95% CI) ‐0.06 [‐0.11, ‐0.01]
1.12.4 Schooltime PA 1 1670 Mean Difference (IV, Random, 95% CI) ‐0.03 [‐0.08, 0.02]
1.13 BMI: meta‐analysis [kg/m2] 50 34337 Mean Difference (IV, Random, 95% CI) ‐0.07 [‐0.15, 0.01]
1.13.1 Children 38 25447 Mean Difference (IV, Random, 95% CI) ‐0.11 [‐0.19, ‐0.02]
1.13.2 Adolescents 12 8890 Mean Difference (IV, Random, 95% CI) 0.05 [‐0.16, 0.25]
1.14 BMI by intervention type: meta‐analysis [kg/m2] 50 34337 Mean Difference (IV, Random, 95% CI) ‐0.07 [‐0.15, 0.01]
1.14.1 Before and after school programmes 9 2314 Mean Difference (IV, Random, 95% CI) ‐0.12 [‐0.25, 0.01]
1.14.2 Enhanced PE 10 3357 Mean Difference (IV, Random, 95% CI) ‐0.04 [‐0.32, 0.24]
1.14.3 Multi‐component interventions 20 24417 Mean Difference (IV, Random, 95% CI) ‐0.10 [‐0.24, 0.03]
1.14.4 Schooltime PA 11 4249 Mean Difference (IV, Random, 95% CI) ‐0.05 [‐0.14, 0.04]
1.15 BMI: additional data 9   Other data No numeric data
1.15.1 Before or after school programme 1   Other data No numeric data
1.15.2 Enhanced PE 1   Other data No numeric data
1.15.3 Multi‐component intervention 4   Other data No numeric data
1.15.4 Schooltime PA 3   Other data No numeric data
1.16 Health‐related quality of life: all data 7   Other data No numeric data
1.16.1 Children 5   Other data No numeric data
1.16.2 Adolescents 2   Other data No numeric data
1.17 Adverse events: all data 16   Other data No numeric data

1.15. Analysis.

Comparison 1: PA programme vs no PA programme, Outcome 15: BMI: additional data

BMI: additional data
Study Study population Intervention group Control Group Measurement period Overall effect Comment
Before or after school programme
Carlin 2018 Female students, aged 11 to 13 years old Brisk walking intervention Continued with normal PA habits 12 weeks No between group differences BMI values not reported
      6 months No between group differences  
Enhanced PE
Ardoy 2011 Students age 12 to 14, enrolled in first year of secondary school I1: 4 sessions/week of PE
I2: 4 sessions/week of PE with emphasis on increasing intensity
2 sessions/week of PE 16 weeks No between group differences BMI values not reported
           
Multi‐component intervention
Burke 1998 I1: standard PA and nutrition program including classroom lessons, fitness sessions daily, and nutrition program
I2: I1 plus a PA enrichment program for higher‐risk children
No program 9 months Significant difference between I1 and control group in boys only, P = 0.016
No significant difference between I1 and control in girls, or I2 and control
 
No BMI values reported
      15 months No significant differences between groups for boys or girls  
Ickovics 2019 Grades 5 and 6 students I1: PA school wellness policy
I2: PA + nutrition school wellness policy
C1: Nutrition school wellness policy
C2: delayed control
3 years No significant group by time interaction P = 0.94 No BMI values reported
           
Salmon 2008 Grade 5 students (approximately 10 to 11 years old) I1: behavioral modification group
I2: fundamental motor skills group
I3: combined behavioral modification and fundamental motor skills group
Usual classroom lessons 1 school year Adjusted BMI/sex‐age population median
I1: MD ‐0.40 (95% CI ‐1.11, 0.30) kg/m2
I2: MD ‐0.50 (95% CI ‐1.25, 0.25) kg/m2
I3: MD ‐1.30 (95% CI ‐2.29, ‐0.31) kg/m2
BMI/sex‐age population median not kg/m2 or z‐score
           
Suchert 2015 Students age 12 to 17 years Multilevel intervention targeting students, classrooms, schools, and parents No intervention 12 weeks BMI percentile, not kg/m2 or z‐score
      1 year BMI percentile: MD 1.09 (95% CI −0.64, 2.82), P = 0.215  
Schooltime PA
Donnelly 2017 Grades 2 and 3 students Academic Achievement and Physical Activity Across the Curriculum lessons, 160 min/week of MVPA Traditional classroom instruction and typical PE schedule 3 years BMI percentile (change from baseline): MD ‐2.3, 95% CI: ‐4.8, 0.2 BMI percentile, not kg/m2 or z‐score
           
Kobel 2014 Pupils at primary school, grades 1 and 2 Teacher training, PA education, and active breaks No intervention 1 year Adjusted BMI percentile (change from baseline): MD 0.5, 95% CI: ‐0.5, 1.5 BMI percentile, not kg/m2 or z‐score
           
Williamson 2007 Students in grades 2 to 6 Healthy Eating and Exercise program to increase PA during the school day and at home Alcohol/Drug/Tobacco abuse prevention program 2 years No between group differences, P = 0.5458 BMI values not reported
           

Characteristics of studies

Characteristics of included studies [author‐defined order]

Breheny 2020.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: UK schools with at least 20 pupils in school years 3 and 5
School exclusion criteria: none
Student inclusion criteria: students in Years 3 (aged 7 to 8 years) and 5 (aged 9 to 10 years)
Student exclusion criteria: pupils that had a disability preventing them from running or walking for 15 minutes and those who were unable to have their height and/or weight measured at baseline 
Setting: school
Age group: children
Gender distribution: females and males
Country where trial was performed: UK
Interventions Intervention: The Daily Mile. Each day, the teacher leads the class on a run or walk as fast as they can go in 15 minutes
Comparator: usual practice
Duration of intervention: 12 months
Duration of follow‐up: ‐
Number of schools: 40
Theoretical framework: ‐ 
Outcomes Fitness
BMI
Health‐related quality of life
Study registration ISRCTN 12698269
Publication details Language of publication: English
Funding: Birmingham City Council
Publication status: peer‐reviewed journal
Stated aim for study "The aim of this pragmatic cluster RCT is to assess the clinical and cost‐effectiveness of The Daily Mile in Birmingham primary schools for the purpose of improving health and well being"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "an independent statistician used a constrained randomisation based algorithm in a statistical package"
Allocation concealment (selection bias) Low risk Quote from publication: "an independent statistician used a constrained randomisation based algorithm in a statistical package"
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "it was not possible to mask school staff, children, family members and project staff"
Blinding of outcome assessment (detection bias)
All outcomes Low risk Quote from publication: "all research staff undertaking the physical measurements were blinded"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Quote from publication: "There was a large amount of missing data for the other secondary outcomes. For fitness and academic attainment this exceeded 56% at certain time points and therefore multiple imputation was performed and both complete case and imputed variables are reported"
Selective reporting (reporting bias) Low risk Comment: all applicable outcomes reported
Cluster RCT ‐ Recruitment bias Low risk Comment: schools enrolled and baseline data collected prior to randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: no baseline differences
Cluster RCT ‐ Loss of clusters High risk Comment: 1 cluster lost from control arm
Cluster RCT ‐ Incorrect analysis Low risk Comment: clustering accounted for in analysis 

Ketelhut 2020.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: within a pool of project schools from another study
School exclusion criteria: ‐ 
Student inclusion criteria: third grade, written parental/guardian consent
Student exclusion criteria: health conditions that did not allow unrestricted physical activity
Setting: school
Age group: children
Gender distribution: females and males
Country where trial was performed: Germany
Interventions Intervention: high‐intensity interval training incorporated into the first 20 minutes of regular PE classes (1 × 45 minutes and 1 × 90 minutes per week)
Comparator: regular PE classes (1 × 45 minutes and 1 × 90 minutes per week)
Duration of intervention: 12 weeks
Duration of follow‐up: ‐
Number of schools: 1
Theoretical framework: ‐
Outcomes Fitness
Study registration
Publication details Language of publication: English
Funding:
Publication status: peer‐reviewed journal
Stated aim for study "the aim of the present study was to examine the effects of a regular school‐based and child‐specific high intensity interval training intervention not only on aerobic fitness and peripheral blood pressure but moreover on different parameters of arterial stiffness"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "randomisation was carried out by the principal investigator using a computer‐generated random number table"
Allocation concealment (selection bias) Unclear risk Comment: not described
Blinding of participants and personnel (performance bias)
All outcomes Unclear risk Comment: not described; not likely done
Blinding of outcome assessment (detection bias)
All outcomes Unclear risk Comment: not described
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Unclear risk Comment: not described
Selective reporting (reporting bias) Unclear risk Comment: no protocol or registry document available.
Cluster RCT ‐ Recruitment bias Unclear risk Comment: not described
Cluster RCT ‐ Baseline imbalance Low risk Comment: no baseline differences
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis High risk Comment: analyses not adjusted for clustered nature of data

Belton 2019.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: (a) schools have a qualified PE teacher on staff, (b) first year students attending the school were time tabled for a minimum of 70 minutes of PE weekly, (c) schools were mixed gender and were situated in the greater area of a large Irish city
School exclusion criteria:
Student inclusion criteria: first year post primary students (12 to 13 years old) attending post primary education within a particular Irish geographical region
Student exclusion criteria:
Setting: school
Age group: adolescent
Gender distribution: females and males
Country/Countries where trial was performed: Ireland
Interventions Intervention: a whole‐school multi‐component intervention programme, aimed at reducing the age‐related decline of MVPA among adolescents. Key features include
1. PE component: PE teachers received 4 hours of Y‐PATH professional development including 6 targeted lesson plans focusing heavily on motivational climate, integrating health‐related activity core knowledge through fun and engaging practical lessons, with an emphasis on functional movement skill proficiency. Resource cards were used to prompt teachers to enable them to integrate a health‐related activity and fundamental movement skill focus within other core PE content areas. Students were given a PA journal to learn to track PA behaviours and identify ways to increase PA levels, and a PA directory containing information and contact details for local youth sport and PA clubs
2. Whole‐school teacher component: PA promotion workshops for teachers, and development and implementation of a school 'charter' for PA. Teachers were encouraged to be 'active role models'
3. Parent component: information evening for parents and information leaflets distributed through the school newsletter to highlight key strategies for promoting PA beyond the school environment
Comparator: usual care, consisting of regular delivery of the Irish Junior Cycle PE curriculum, and the broader school curricula
Duration of intervention: 2 years
Duration of follow‐up: 2 years
Number of schools: 20
Theoretical framework: social‐ecological framework, self‐determination theory
Outcomes PA duration
Study registration ISRCTN20495704
Publication details Language of publication: English
Funding: Dublin Local Sports Partnerships, Dublin City University Career Start grant
Publication status: peer‐reviewed journal
Stated aim for study "to investigate the effect of participation in the Y‐PATH intervention over a two‐year period on objectively measured MVPA levels of young people"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) High risk Quote from publication: "one school from each pair was then randomly allocated by the study principal investigator to the control group (and the other to the intervention group) using a manual number generator in blocks of 1:1, prior to the commencement of baseline testing"
Allocation concealment (selection bias) High risk Quote from publication: "one school from each pair was then randomly allocated by the study principal investigator to the control group (and the other to the intervention group) using a manual number generator in blocks of 1:1, prior to the commencement of baseline testing"
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "not possible given the nature of the intervention"
Blinding of outcome assessment (detection bias)
All outcomes High risk Quote from publication: "not possible given the nature of the intervention"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: BMI data missing > 10%
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: Large % missing data
Selective reporting (reporting bias) High risk Comment: BMI data not reported, stated in methods; secondary outcomes listed in clinical trials registry not reported
Cluster RCT ‐ Recruitment bias High risk Comment: baseline data collected after randomisation of schools
Cluster RCT ‐ Baseline imbalance Low risk Comment: groups balanced at baseline [author communication]
Cluster RCT ‐ Loss of clusters High risk Comment: 50% of clusters lost at 24 months
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "a three level multilevel structure was proposed with random intercepts, where time (Level one), pupils (Level two) and schools (Level three) served as the grouping variables, where time was treated as a fixed effect in the model but was also incorporated as a random slope effect (repeated measure) in the residual component"

Corepal 2019.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: post‐primary schools in Belfast that had previously participated in research projects with the university 
School exclusion criteria: ‐ 
Student inclusion criteria: Year 9 classes
Student exclusion criteria: advised by a general practitioner not to undertake MVPA, did not provide assent, or parents completed parental opt‐out consent form
Setting: school
Age group: adolescents
Gender distribution: females and males
Country where trial was performed: North Ireland
Interventions Intervention: StepSmart Challenge used gamification and self‐determination theory to encourage and support PA behaviour change. During phase 1, competitions were held between schools, between classes, and between students using material and social incentives. FitBit Zips were provided to track progress towards challenges. Phase 2 included a within‐student pedometer competition using the StepSmart Challenge website
Comparator: no intervention or incentives
Duration of intervention: 22 weeks
Duration of follow‐up: 52 weeks
Number of schools: 5
Theoretical framework: self‐determination theory
Outcomes PA duration
Sedentary time
Study registration NCT02455986
Publication details Language of publication: English
Funding: HSC R&D Enabling Research Award
Publication status: peer‐reviewed journal
Stated aim for study "This study investigated the feasibility of implementing and evaluating a school‐based gamified pedometer competition designed to promote physical activity among 12–14‐year‐olds, known as ‘The StepSmart Challenge’, which integrates core gamification strategies with self‐determination theory"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "stratified randomisation process (stratified by socioeconomic status, and whether schools were single‐sex or co‐educational) was undertaken by an independent statistician to assign schools to the intervention or control group using software available at http://www.randomization.com"
Allocation concealment (selection bias) Low risk Quote from publication: "stratified randomisation process (stratified by socioeconomic status, and whether schools were single‐sex or co‐educational) was undertaken by an independent statistician to assign schools to the intervention or control group using software available at http://www.randomization.com"
Blinding of participants and personnel (performance bias)
All outcomes Unclear risk Comment: not described
Blinding of outcome assessment (detection bias)
All outcomes Unclear risk Comment: not described
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: only 57.4% of returned accelerometers had data valid for analysis 
Selective reporting (reporting bias) Low risk Comment: all relevant outcomes reported
Cluster RCT ‐ Recruitment bias Unclear risk Comment: not described
Cluster RCT ‐ Baseline imbalance Unclear risk Comment: not described
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Comment: feasibility trial, thus only descriptive values reported

Ickovics 2019.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: K through 8 district schools New Haven, Connecticut
School exclusion criteria:
Student inclusion criteria: students enrolled in Grade 5 when the study began or began attending a target school in Grade 6
Student exclusion criteria:
Setting: school, urban
Age group: children
Gender distribution: females and males
Country where trial was performed: USA
Interventions Intervention 1: PA only
Intervention 2: Nutrition + PA
All schools received $500/y to establish a School Wellness team, focused on written policy implementation relevant to randomised condition. Research staff supported schools with 1 to 2 visits per month to provide workshops 
PA: high‐quality PE class at least 90 minutes/week to foster a lifelong appreciation for physical fitness and to participate in fitness activities; promotion of active transport; integration of PA into the classroom; fitness challenges. PA will not be used as a form of punishment (e.g. running laps, withholding recess), distribution of activity monitors, use of online tracking software, family‐targeted newsletters
Nutrition: appealing and attractive meals, clean and pleasant setting, no fried vegetables, only low‐fat, non‐flavoured milk, whole‐grain cereals. Schools will not use food or beverages as rewards or punishments, and will limit celebrations involving food to once per month and with only 1 food or beverage that does not meet school nutrition standards. Nutrition education provided through parent workshops, student materials, school menus, and bulletins. Schools asked to engage parents and students in thorough taste testing of new menu options and to assist with selecting food
Comparator 1: nutrition intervention only
Comparator 2: delayed control; schools received other health‐relevant training (e.g. oral health, cold or influenza prevention) during the study period, with obesity‐related materials delivered at study completion
Duration of intervention: 3 years
Duration of follow‐up: annually for 3 years
Number of schools: 12
Theoretical Framework: —
Outcomes BMI
Study registration NCT02043626 (retrospectively registered)
Publication details Language of publication: English 
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The objective of this cluster randomised trial is to assess whether implementation of specific nutrition and PA components of the written school wellness policies lead to healthier student outcomes, including BMI trajectories and behavioral correlates"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "randomisation was achieved using a computer‐generated sequence"
Allocation concealment (selection bias) Low risk Quote from publication: "to minimize selection bias, all schools were recruited before randomisation"
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "neither schools nor researchers could be blinded to study condition"
Blinding of outcome assessment (detection bias)
All outcomes High risk Quote from publication: "neither schools nor researchers could be blinded to study condition"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: most missing data due to students moving
Selective reporting (reporting bias) High risk Comment: mental health and standardised test scores not reported
Cluster RCT ‐ Recruitment bias Unclear risk Quote from publication: "schools were recruited before randomisation"
Comment: unclear when baseline data were collected
Cluster RCT ‐ Baseline imbalance Low risk Comment: there were no notable school‐level differences in size or relevant socioeconomic characteristics
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "multivariable analyses accounted for multilevel nature of data, adjusting for intra‐cluster correlation among repeated measures within students and schools, and allowed use of all study time points"

Jago 2019.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: primary schools from South Gloucestershire and North Somerset local authority
School exclusion criteria:
Student inclusion criteria: pupils in Year 3 or 4 (aged 7 to 9) at baseline
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country where trial was performed: UK
Interventions Intervention: Action 3:30R after‐school clubs, scheduled to run twice per week for 15 weeks and last 60 minutes per session. Sessions were designed to promote maximal participation, skill development, cooperation, problem‐solving, PA, and choice. Sessions began with fun warm‐up activities and moved through a series of small sided games and activities with a focus on fun and participation while improving fundamental movement skills such as running, catching, throwing, and using space in invasion games. Teaching assistants took part in 5 days (25 hours) of training to promote and foster aspects of motivation drawn within the club, with focus on creating a club climate that supported autonomy, relatedness, and competence. Teaching assistants completed a log book to indicate whether sessions were delivered fully, partially, or not at all and a register of attendance
Comparator:
Duration of intervention: 15 weeks
Duration of follow‐up: 15 weeks
Number of schools: 12
Theoretical framework: self‐determination theory
Outcomes PA duration
Sedentary time
BMI
Health‐related quality of life
Study registration ISRCTN34001941
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The aim of this study is to test the feasibility of the revised version of Action 3:30 which has been reworked to more successfully appeal to and engage girls and recruit less active children"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "computer generated"
Allocation concealment (selection bias) Low risk Quote from publication: "allocation will be performed (computer‐generated allocation) by an independent member of the Bristol Randomised Trials Collaboration who will be blind to the school identity"
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "randomisation will take place after baseline data collection has been completed. School is the unit of randomisation"
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: attempted to blind staff, but unsuccessful [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: low loss to follow‐up at student level
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Comment: low loss to follow‐up at student level
Selective reporting (reporting bias) Low risk Comment: all outcomes in protocol paper reported
Cluster RCT ‐ Recruitment bias Low risk Quote from publication: "all measures will be taken at baseline prior to randomisation"
Cluster RCT ‐ Baseline imbalance Low risk Comment: Control and intervention groups did not differ by demographics, PA, or psychosocial outcomes at baseline
Cluster RCT ‐ Loss of clusters High risk Comment: 2 schools dropped out after randomisation
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "multivariable mixed effects linear regression that took account of the clustering of pupils in schools was conducted"

Leahy 2019.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria:
School exclusion criteria:
Student inclusion criteria: students in Grade 11 at study schools who did not have an injury or illness that would preclude their participation in high‐intensity activity as outlined in the participant information and consent form (e.g. existing physical injury)
Student exclusion criteria:
Setting: school
Age group: adolescents
Gender distribution: females and males
Country where trial was performed: Australia
Interventions Intervention: Burn2Learn, a 14‐week multi‐component high‐intensity interval training intervention to improve older adolescents’ physical and mental health. Participants were prescribed 3 high‐intensity interval training sessions/week, for 14 weeks. Teachers were asked to offer at least 2 opportunities/week for high‐intensity interval training during class time, with the ultimate target of students performing 3 sessions/week (i.e. 1 self‐directed session outside of class). Any additional high‐intensity interval training sessions performed by participants were reported individually to the school champion to monitor session adherence. High‐intensity interval training sessions comprised a brief 2‐minute warm‐up, followed by 8 to 16 minutes of high‐intensity interval training, followed by a 2‐minute cool‐down (12 to 20 minutes total). Sessions were performed individually, in pairs, or in small groups. Participants were provided with pre‐designed high‐intensity interval training workouts that included a combination of aerobic‐based (e.g. shuttle runs) and resistance‐based (e.g. push‐ups) exercises designed to be performed using minimal space and equipment. Participants were able to select from the following high‐intensity interval training theme workouts: gym high‐intensity interval training, sport high‐intensity interval training, class high‐intensity interval training, dance high‐intensity interval training, combat high‐intensity interval training, and brain high‐intensity interval training. A full‐day professional learning workshop was provided for a school champion (i.e. teacher) and another member of staff to facilitate the Burn2Learn programme. The intervention included an introductory seminar for students, school‐based high‐intensity interval training sessions, parental videos, and an equipment and resource pack (including high‐intensity interval training task cards). Technique cards reinforcing correct technique were also provided to the intervention school and were used by the school champion during early weeks of the intervention. To encourage maintenance of the appropriate exercise intensity (i.e. > 85% maximum heart rate), participants were provided with heart rate monitors (Wahoo TICKR) during high‐intensity interval training sessions. Participants’ heart rates during sessions were viewed on smartphones utilising a commercially available group heart rate monitoring application (OnBeat)
Comparator: a wait‐list control group participated in usual school activities and received the intervention following the post‐test assessment period
Duration of intervention: 14 weeks
Duration of follow‐up: 14 weeks
Number of schools: 2
Theoretical framework: self‐determination theory
Outcomes Fitness
BMI
Health‐related quality of life
Study registration ACTRN12617000544370
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "the objective of this study was to evaluate the impact of a teacher‐facilitated high intensity interval training program for older adolescents, embedded within the school day in regard to 4 domains of feasibility (i.e. recruitment, retention, adherence, and program satisfaction). Preliminary efficacy was evaluated by testing the effect of the high intensity interval training program on cardiorespiratory fitness, muscular fitness, and psychological health"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "the 2 schools were randomised to the Burn2Learn intervention group, or a wait‐list control group using a coin flip by an independent researcher not involved in the project following baseline assessments"
Allocation concealment (selection bias) Low risk Quote from publication: "the 2 schools were randomised to the Burn2Learn intervention group, or a wait‐list control group using a coin flip by an independent researcher not involved in the project following baseline assessments"
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: not possible
Blinding of outcome assessment (detection bias)
All outcomes High risk Quote from publication: "trained research assistants, who were blinded to group allocation, conducted assessments for the primary outcome. Assessors responsible for the collection of secondary outcomes were not blinded to group allocation"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: reasons for dropout not reported
Selective reporting (reporting bias) High risk Comment: PA not reported
Cluster RCT ‐ Recruitment bias Low risk Quote from publication: "the 2 schools were randomised to the Burn2Learn intervention group, or a wait‐list control group using a coin flip by an independent researcher not involved in the project following baseline assessments"
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline demographics were compared, and all groups were similar [author communication]
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis High risk Comment: no clustering accounted for in analysis

Lonsdale 2019a.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: (1) school with students enrolled in Grades 8 and 9; (2) funded by the New Sout Wales Department of Education; (3) permission granted by the principal, the head PE teacher, and at least 1 Grade 8 PE teacher; (4) located in Western Sydney; (5) in a postal code with a mean decile rank that was below the median on the Australian Bureau of Statistics’ Index of Relative Socioeconomic Disadvantage
School exclusion criteria: —
Student inclusion criteria: students physically able to take part in Grade 8 PE
Student exclusion criteria: —
Setting: school
Age group: adolescent
Gender distribution: females and males
Country/Countries where trial was performed: Australia
Interventions Intervention: the 'Activity and Motivation in Physical Education’ (AMPED) intervention had 2 aims: (1) to help teachers deliver lessons that maximised opportunities for MVPA; and (2) to help teachers enhance their students’ motivation towards PE. Teachers’ learnt strategies that were categorised under 2 headings: (A) ‘Maximising Movement and Skill Development’ and (B) ‘Reducing Transition Time’. Strategies to enhance student motivation were organised under 2 further headings: (C) ‘Building Competence’ and (D) ‘Supporting Students’. Face‐to‐face workshops included brief presentations by the research team, but many of these teachers worked independently on the project’s website. This independent work was designed to help ensure teachers were comfortable working on the website, to facilitate later use. Throughout the entire intervention, teachers had access to online resources, a discussion forum, videos of good/poor practice, and the project’s mobile phone application, which included implementation and self‐reflection prompts
Comparator: standard teaching, wait‐list control 
Duration of intervention: 7 to 8 months
Duration of follow‐up: 14 to 15 months
Number of schools: 14
Theoretical framework: self‐determination theory, 
Outcomes PA duration
Sedentary time
Study registration ACTRN12614000184673
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The purpose of this study is to evaluate an intervention designed to increase the amount of health‐enhancing PA that secondary school students accumulate during their school‐based PE lessons"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "using a computer‐based randomisation plan generator...a researcher not associated with recruitment or data collection, and who will be blind to school identity, will carry out randomisation procedures"
Allocation concealment (selection bias) Low risk Comment: All 14 eligible schools randomised at 1 time point
Blinding of participants and personnel (performance bias)
All outcomes Low risk Quote from publication: "students participating in the study will also be blinded to hypotheses and school allocation. Teachers will be aware of their allocation to the intervention or control condition"
Comment: using only objective physical activity assessment, teacher's knowledge unlikely to bias results
Blinding of outcome assessment (detection bias)
All outcomes Low risk Quote from publication: "trained research assistants who will be blinded to school allocation will conduct baseline, post‐intervention and maintenance phase assessments"
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: 34% and 44% of physical activity data missing in intervention and control groups, respectively, at end of study
Selective reporting (reporting bias) Low risk Comment: all relevant outcomes reported
Cluster RCT ‐ Recruitment bias High risk Comment: teachers and students enrolled after schools randomised; teachers aware of allocation
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline characteristics balanced between 2 groups
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "we included four random intercept effects for: (1) lesson; (2) student; (3) teacher; and (4) class. When preliminary analyses suggested clustering at the school level, we included a fifth random intercept effect for this level"

Müller 2019.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: quintile 3 schools selected based on geographic location, representation of various target communities, and commitment to support project activities
School exclusion criteria: < 100 learners in Grade 4
Student inclusion criteria 
  • Willingness to participate

  • Written informed consent by a parent or guardian

  • No participation in other clinical trials during the study period

  • Not suffering from medical conditions preventing participation in a maximum exercise test, as determined by qualified medical personnel


Student exclusion criteria:
Setting: school
Age group: children/adolescents
Gender distribution: females and males
Country where trial was performed: South Africa
Interventions Intervention 1: PA: a multi‐dimensional PA intervention during school time developed in collaboration with education authorities, teachers, and students. 4 key components included:
  • two 40‐minute PE lessons/week;

  • one 40‐minute moving‐to‐music lesson/week led by students from Nelson Mandela University;

  • in‐class activity breaks; and

  • low‐cost school environment adaptations (e.g. activity stations, painted games)


Intervention 2: PA + Health education: health education lessons were held to increase children’s awareness of intestinal parasite infections
Intervention 3: PA + Health education + Nutrition: a nutrition intervention consisting of classroom‐based lessons to help increase awareness of the importance of healthy nutrition were held
Comparator 1: Health education + Nutrition
Comparator 2: no intervention
Duration of intervention: 2 × 10‐week blocks
Duration of follow‐up: 15 months
Number of schools: 8
Theoretical framework:
Outcomes Fitness
BMI
Study registration ISRCTN68411960(retrospectively registered)
Publication details Language of publication: English 
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The overarching purpose of the Disease, Activity and Schoolchildren’s Health study was to investigate the dual disease burden (i.e. non‐communicable diseases and infectious diseases) among children in primary schools in disadvantaged neighbourhoods"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "generating the allocation sequence by a simple randomisation of the schools was carried out by the research team on the basis of a computer‐generated random number list"
Allocation concealment (selection bias) High risk Quote from publication: "research team allocated schools, concealment and blinding were not possible in our study design"
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "concealment and blinding were not possible in our study design"
Blinding of outcome assessment (detection bias)
All outcomes High risk Quote from publication: "concealment and blinding were not possible in our study design"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: 281 lost to follow‐up
Selective reporting (reporting bias) High risk Comment: different primary outcome reported in trial registry
Cluster RCT ‐ Recruitment bias Low risk Comment: individual learners were enrolled and baseline data collected before randomisation at the cluster level (of schools) [author communication]
Cluster RCT ‐ Baseline imbalance Low risk Quote from publication: "no significant differences in primary outcome measures, such as obesity, skin‐folds and cardiorespiratory fitness at baseline were detected, when comparing schools with and without PA intervention"
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "separate mixed linear regression models were employed with random intercepts for school classes, in order to adjust for cluster effects"

Ordóñez Dios 2019.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: bilingual public schools
School exclusion criteria:
Student inclusion criteria: children aged 11 to 12 years in their last year of primary school
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country where trial was performed: Spain
Interventions Intervention: two 45‐minute sessions of PE per week plus a daily run (starting at 250 m and progressing to 750 m)
Comparator: two 45‐minute sessions of PE per week as stipulated by law
Duration of intervention: 12 weeks
Duration of follow‐up: 12 weeks
Number of schools: 2
Theoretical framework:
Outcomes Fitness
BMI
Study registration
Publication details Language of publication: English/Spanish
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The objective of this study is to examine the possible effects of a daily physical activity intervention on physical fitness, coordination and attention"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Unclear risk Comment: not described
Allocation concealment (selection bias) Unclear risk Comment: not described
Blinding of participants and personnel (performance bias)
All outcomes Unclear risk Comment: not described
Blinding of outcome assessment (detection bias)
All outcomes Unclear risk Comment: not described
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Unclear risk Comment: number of participants enrolled or randomised not described; only number analysed
Selective reporting (reporting bias) Unclear risk Comment: no protocol published
Cluster RCT ‐ Recruitment bias Unclear risk Comment: unclear when baseline data were collected
Cluster RCT ‐ Baseline imbalance Unclear risk Comment: demographic characteristics not compared
Cluster RCT ‐ Loss of clusters Low risk Comment: no loss of clusters
Cluster RCT ‐ Incorrect analysis High risk Comment: clustered nature of data not accounted for in analysis

Seibert 2019.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: 40% or more of students qualifying for free or reduced lunch
School exclusion criteria:
Student inclusion criteria: Grades 3 to 8
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country where trial was performed: USA
Interventions Intervention: implementation of 3 Centre for Disease Control‐recommended evidence‐based strategies: (1) increasing the amount of time spent in MVPA during PE classes, (2) encouraging active classroom breaks, (3) providing organised PA opportunities during recess, and (4) providing organised PA opportunities before and after school
Comparator: continued with prior routine PA programming
Duration of intervention: 1 school year
Duration of follow‐up: 1 school year
Number of schools: 49
Theoretical framework:
Outcomes Fitness
Study registration NCT02411552 (retrospectively registered)
Publication details Language of publication: English
Funding: University of Wisconsin‐MadisonWisconsin Partnership Program and NIH T32 DK077586
Publication status: peer‐reviewed journal
Stated aim for study "In this study we evaluated the feasibility and effects of a large‐scale implementation program for CDC‐recommended school based physical activity strategies on cardiovascular fitness measured by the Progressive Aerobic Cardiovascular Endurance Run in low SES Wisconsin schools"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Unclear risk Comment: not described
Allocation concealment (selection bias) Unclear risk Comment: not described
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: principals and PE educators informed of intervention status
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: schools completed PACER as part of their regular PE class and submitted data to researchers
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Unclear risk Comment: not described
Selective reporting (reporting bias) Low risk Comment: all outcomes listed in clinical trials registry reported
Cluster RCT ‐ Recruitment bias High risk Comment: data collected after randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: between‐group differences at baseline adjusted for in analysis
Cluster RCT ‐ Loss of clusters Unclear risk Comment: not described
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "linear mixed effects modelling with school‐specific random effects were conducted to perform the comparisons of outcome measures at baseline between the 2 study arms"

Seljebotn 2019.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: all primary schools in the municipality of Stavanger, Norway
School exclusion criteria:
Student inclusion criteria: Grade 5 ‐ 9 to 10 years old
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country where trial was performed: Norway
Interventions Intervention: physically active lessons (45 minutes) 2 to 3 days/week on days without PE. Lessons were held mainly outdoors and included games, relays, and quizzes with curricular questions from theoretical subjects. Physically active lessons included at least 15 minutes of MVPA, were easily organised and adapted, included competitive and non‐competitive elements, and were enjoyable activities that included all children. Secondary components included physically active homework (10 minutes/d) and physically active recess (10 minutes/d). The intervention was intended to increase the amount of PA by 190 minutes/week, giving a total of 325 minutes/week of PA. To further improve the quality of the physically active lessons, a quality framework was stated at the back of the physically active lesson form and included tips of how differentiation, autonomy, collaboration, enjoyment, and high activity level could be ensured. To assist and support intervention teachers, 1 primary and 1 secondary contact person from the Active School project team was assigned to each intervention school. Contact persons attended meetings and regularly visited participating teachers and classes throughout the school year (1 to 4 visits/month, depending on requests from the schools). 1 pre‐intervention seminar and 1 midway seminar were arranged for the teachers to give information about the programme and to provide support. New physically active lessons were shared between intervention schools through a website
Comparator: control schools were asked to continue their normal routine, which included approximately 135 minutes/week of PA
Duration of intervention: 10 months
Duration of follow‐up: 10 months
Number of schools: 9
Theoretical framework:
Outcomes PA duration
Sedentary time
Fitness
BMI
Study registration NCT03436355
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "Based on current knowledge, the research question of this study is as follows: to what extent will increased PA in school affect children’s executive function and aerobic fitness?"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "the computer program 'Researcher Randomizer' was used to randomise the 2 groups into intervention and control groups"
Allocation concealment (selection bias) Unclear risk Comment: not stated
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants and study staff were not randomised [author communication]
Blinding of outcome assessment (detection bias)
All outcomes High risk  Comment: outcome assessors were not blinded [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: low loss to follow‐up; unrelated to intervention
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Comment: low loss to follow‐up; unrelated to intervention
Selective reporting (reporting bias) High risk Comment: composite measure of executive function reported rather than individual components listed as primary outcomes
Cluster RCT ‐ Recruitment bias High risk Comment: randomisation before participant enrolment
Cluster RCT ‐ Baseline imbalance Unclear risk Comment: not stated
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis High risk Quote from publication: "due to low variance between schools in these outcomes, multilevel analysis was not considered necessary"

Zhou 2019.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: middles schools from large, medium, and small metropolitan regions in China with similar student enrolment numbers, student‐teacher ratios, and outdoor facilities; having 80 to 100 students of both genders enrolled in seventh grade; located at least 5 kilometres apart from other study schools; agreed to randomisation of treatment; agreed to implement policy and curriculum modifications 
School exclusion criteria: −
Student inclusion criteria: junior high school healthy students; current enrolment, parental consent, and no physical disability
Student exclusion criteria: students who cannot exercise; members of varsity sports teams
Setting: school
Age group: adolescent
Gender distribution: females and males
Country where trial was performed: China
Interventions Intervention 1: school physical education ‐ minimum of 3 PE classes/week and daily 15‐minute recess, portable exercise equipment, redesign of PE curriculum, recess rhythmic aerobic routine, use of fitness and health handbook for knowledge and skills to be used on inclement weather days, bi‐weekly text messages to students
Intervention 2: after school programme ‐ bi‐weekly 45‐minute after school PA programme, portable exercise equipment, use of fitness and health handbook for knowledge and skills to be used on inclement weather days, bi‐weekly text messages to students
Intervention 3: school physical education + after school programme
Comparator: two 45‐minute PE classes per week
Duration of intervention: 32 weeks
Duration of follow‐up: ‐ 
Number of schools: 12
Theoretical framework: socioecological model, competence motivation theory
Outcomes PA duration
Sedentary time
Fitness
Study registration  
Publication details Language of publication: English
Funding: serving National Special Needs in Doctoral Talents Development Program—Performance Training and Health Promotion for Adolescents; the support programme for High‐level Teacher Team Development of BeijingMunicipal Institutions (IDHT20170515); Beijing Social Science Funding Project (No. 16YTB018); Scientific Research Project of Beijing Educational Committee (No. KM201710029002)
Publication status: peer‐reviewed journal
Stated aim for study "The purpose of the 4‐arm study was to examine the incremental effect of adding moderate physical activity and vigorous physical activity on the physical fitness among the children that were assigned to Arm 1‐school physical education intervention, Arm 2‐afterschool program intervention, Arm 3‐school physical education and after school program and Arm 4‐control condition"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Unclear risk Comment: not described
Allocation concealment (selection bias) Unclear risk Comment: not described
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students and staff not blinded
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: data collectors trained and blinded to allocation
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: very small quantity of missing data 
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Unclear risk Comment: only some participants wore accelerometers; no indication of completion rate
Selective reporting (reporting bias) Low risk Comment: all relevant outcomes reported
Cluster RCT ‐ Recruitment bias Unclear risk Comment: not described
Cluster RCT ‐ Baseline imbalance Low risk Comment: some baseline differences across groups, but adjusted for in analysis 
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis High risk Comment: not all analyses presented; unclear which tests were used

Adab 2018.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: all state primary schools in the West Midlands (UK), which included school years 1 to 5 (children aged 5 to 10 years) and were within a 35‐mile radius of the University of Birmingham
School exclusion criteria: schools with fewer than 17 pupils in the relevant year group (minimum cluster size) or those who were in special measures (status applied by the Office for Standards in Education when it considers that a school fails to supply an acceptable level of education and appears to lack the leadership capacity necessary to secure improvements)
Student inclusion criteria: all Year 1 pupils (aged 5 to 6 years) in participating schools
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country where trial was performed: UK
Interventions Intervention: 4 overlapping components
  • 30 minutes of additional MVPA on each school day—at least 15 minutes to be outside of break times

  • Term cooking workshops during school time, which parents were invited to attend to participate with their child and that were preceded by short classroom sessions for the children

  • 6‐week programme (Villa Vitality) developed to encourage healthy eating and increase PA and delivered by staff from an iconic sporting institution. School classes spent 2 days undertaking activities (indoor‐based movement routines, use of dance mats, ball skills session, interactive nutritional sessions, and opportunity to practise cooking skills) at an English premier league football club, separated by a 6‐week period during which teachers were asked to spend curriculum time working on a class project and involving children and their parents with weekly health challenges

  • Information sheets sign posting children and their families on ways to be active over the summer (identical for all schools) and PA opportunities in their local area (school‐specific sheets produced by the study team and checked before distribution by the school)


Comparator: continued with ongoing Year 2 health‐related activities plus citizenship education resources, excluding topics related to healthy eating and PA
Duration of intervention: 12 months
Duration of follow‐up: 15 and 18 months
Number of schools: 54
Theoretical framework: —
Outcomes PA participation
PA duration
Sedentary time
BMI
Health‐related quality of life
Study registration ISRCTN97000586
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal 
Stated aim for study "The main aim is to assess the clinical and cost‐effectiveness of the 12‐month childhood obesity prevention intervention programme, developed and refined in the Birmingham healthy Eating and Active lifestyle for CHildren Study, using usual practice in primary schools as the comparator"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "a blocked balancing algorithm was used to randomise participating schools to intervention or comparator arms. Schools were randomly allocated according to a randomisation scheme, which minimized imbalance on several characteristics"
Allocation concealment (selection bias) Low risk Quote from publication: "to ensure concealment of allocation we carried out randomisation after baseline measurements. Sessional researchers blind to arm allocation mainly undertook further data collection"
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded
Blinding of outcome assessment (detection bias)
All outcomes Low risk Quote from publication: "researchers blind to arm allocation mainly undertook further data collection"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: most loss to follow‐up due to children changing schools
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Comment: most loss to follow‐up due to children changing schools
Selective reporting (reporting bias) Low risk Comment: protocol published; all stated outcomes reported
Cluster RCT ‐ Recruitment bias Low risk Comment: schools and participants recruited before randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: clusters balanced at baseline
Cluster RCT ‐ Loss of clusters Low risk Quote from publication: 1 school lost at first follow‐up (N = 20 students) but retained in analysis
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "adjusted model included the baseline measurement and treatment arm as the independent variables, and to account for the clustered nature of the sample, school as the random effect"

Carlin 2018.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: a convenience sample of schools in Northern Ireland
School exclusion criteria:
Student inclusion criteria: all female pupils aged 11 to 13 years attending 6 post‐primary schools in Northern Ireland who were free from any medical condition that would limit their participation in a brisk walking intervention
Student exclusion criteria:
Setting: school
Age group: children/adolescents
Gender distribution: females 
Country where trial was performed: Northern Ireland
Interventions Intervention: brisk walking intervention consisting of structured 10– to 15‐minute walks spread across the school week before the first bell, at mid‐morning break, and at lunch time. Participants were instructed to attend at least 3 walking sessions/week and to increase the number of sessions that they attended to at least 5 walking sessions/week by Week 12 of the intervention. Walks were led by older pupils trained as walk leaders to ensure safety and intensity (i.e. at a pace sufficient to elicit moderate‐intensity PA). Walk leader training was facilitated by the research team at a lunchtime session and was informed by a PA coordinator from a local Health and Social Care Trust. Training was facilitated by a member of the research team and was delivered to walk leaders at a lunchtime session. Participants were provided with timetables of planned group walks, detailing start time and meeting location for each walk, and were given weekly verbal reminders by school staff and walk leaders to attend the walking sessions. Participants were provided with prompt cards from the research team containing general tips and advice in relation to brisk walking and information on setting goals. Schools were instructed to provide at least 2 walking sessions for participants to attend each day
Comparator: continued with normal PA habits. Following completion of the intervention, all control schools were provided study resources to implement their own school‐based brisk walking programme
Duration of intervention: 12 weeks
Duration of follow‐up: 6 months
Number of schools: 6
Theoretical framework: social cognitive theory
Outcomes Fitness
BMI
Study registration NCT02871830 (retrospectively registered)
Publication details Language of publication: English
Funding:
Publication status: peer‐reviewed journal 
Stated aim for study "The aim of this pilot study was to investigate the feasibility of peer‐led brisk Walking In ScHools intervention (the WISH study) and to investigate the impact of participating in a 12‐week school‐based walking programme on schooltime PA and sedentary behaviour post‐intervention (week 12) and at follow‐up (6 months). The secondary aim was to examine the effects of the intervention on a range of health‐related outcome measures"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "participants were randomly allocated by school, using a computer‐based random number generator to either receive the intervention or to act as controls"
Allocation concealment (selection bias) Low risk Comment: randomisation completed after baseline measurement
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "blinding of schools and participants was not possible"
Blinding of outcome assessment (detection bias)
All outcomes High risk Quote from publication: "researcher responsible for data collection and analysis was not blinded to group allocation"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: very few missing data due to student absence on measurement day
Selective reporting (reporting bias) High risk Comment: protocol specifies total weekly PA as the primary outcome; only schooltime PA was reported
Cluster RCT ‐ Recruitment bias Low risk Comment: schools and students recruited before randomisation
Cluster RCT ‐ Baseline imbalance High risk Comment: no statistical comparison between groups [author communication]
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis High risk Comment: clustering not accounted for in analysis

Harrington 2018.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: all state (government‐funded) secondary schools in Leicester City, Leicestershire, and Rutland, UK, with a Key Stage 3 (students age 11 to 14 years), and state schools that were geographically close to Leicester City, Leicestershire, and Rutland but in neighbouring counties
School exclusion criteria:
Student inclusion criteria: girls between the ages of 11 and 14 years and in Years 7, 8, and 9
Student exclusion criteria:
Setting: school
Age group: children/adolescents
Gender distribution: females
Country where trial was performed: UK
Interventions Intervention: Girls Active, a support framework for schools to review and change their PA, PE, and school sport culture and practices with support of the Youth Sport Trust and a hub school to develop a school action plan. Core components include submission of self‐review and action plans; attendance of lead teacher at initial training; use of package of resources or use of an alternative; engagement of young people as peer leaders; use of online, in‐person, or phone support of hub and/or development coach; lead teacher attendance at peer review day; and submission of mission analysis. Schools were provided with two £500 capacity funding instalments to coincide with action plan submission
Comparator: schools were not given any specific guidance or advice and were assumed to carry on with their usual practice of PE and sport provision
Duration of intervention: 14 months
Duration of follow‐up: 14 months
Number of schools: 20
Theoretical framework: social cognitive theory
Outcomes PA participation
PA duration
Sedentary time
BMI
Health‐related quality of life
Study registration ISRCTN10688342
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The aim of this study was to assess the effectiveness of the Girls Active PA programme in UK secondary schools"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) High risk Quote from publication: "sequentially numbered sections within a folder were used to implement the group allocations"
Allocation concealment (selection bias) Low risk Quote from publication: "following baseline measurements, schools were randomised by an independent statistician. The investigator team were not aware of the sequence until after randomisation"
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "the trial statistician was not blinded"
Blinding of outcome assessment (detection bias)
All outcomes Low risk Quote from publication: "measurement team members, except the team lead for the day, were blinded to group randomisation"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: only per protocol utilised; large missing data from control schools
Quote from publication: "the per protocol population included schools that engaged with 70% of the seven core components (as detailed above) of the programme over the 14 months and had complete data for the analysis concerned on ‘by analysis’ basis. In the control arm, the per protocol population included all schools/pupils randomised to that arm"
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: only per protocol utilised; large missing data from control schools
Quote from publication: "the per protocol population included schools that engaged with 70% of the seven core components (as detailed above) of the programme over the 14 months and had complete data for the analysis concerned on ‘by analysis’ basis. In the control arm, the per protocol population included all schools/pupils randomised to that arm"
Selective reporting (reporting bias) Low risk Comment: all outcomes in trial registry reported
Cluster RCT ‐ Recruitment bias Low risk Quote from publication: "randomisation will occur after baseline assessments and will be carried out by the Leicester Clinical Trials Unit"
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline differences at cluster level accounted for
Quote from publication: "generalized estimating equations, accounting for school level clustering, and adjusting for baseline MVPA, stratification factors of school size (< 850, ≥ 850) and percent of non‐White pupils (< 20%, ≥ 20%), percent free school meals and participant year group, were employed"
Cluster RCT ‐ Loss of clusters High risk Comment: 2 control schools lost to follow‐up
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "generalized estimating equations, accounting for school level clustering, and adjusting for baseline MVPA, stratification factors of school size (< 850, ≥ 850) and percent of non‐White pupils (< 20%, ≥ 20%), percent free school meals and participant year group, were employed"

Have 2018.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: schools in Danish municipalities that did not have a structured programme that incorporated PA in the classroom
School exclusion criteria:
Student inclusion criteria:
Student exclusion criteria: physical disability, no written parental consent
Setting: school
Age group: children
Gender distribution: females and males
Country where trial was performed: Denmark
Interventions Intervention: classroom‐based PA incorporated into math lessons for 1 school year. Subjects received an average of 6 math lessons of 45 minutes/week. Each lesson included at least 15 minutes of PA, with limited sedentary time. Teachers attended a series of workshops to provide them with the skills to implement task‐relevant physical activity into math teaching 
Comparator: children in control schools received regular classroom instruction, also with an average of 6 math lessons of 45 minutes/week. Math teachers in the control schools were asked not to make any changes to their usual teaching methods before study endpoint measurements
Duration of intervention: 10 months
Duration of follow‐up: 10 months
Number of schools: 12
Theoretical framework: theory of embodied cognition
Outcomes PA duration
BMI
Study registration NCT02488460 (retrospectively registered)
Publication details Language of publication: English 
Funding: non‐commercial funding (charitable trust)
Publication status: peer‐reviewed journal
Stated aim for study "We designed a randomised controlled trial with the primary objective of investigating how math achievement was affected by task‐relevant PA incorporated into math teaching for 7‐year‐old school children"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "randomisation was performed by random selection of sealed envelopes containing the intervention allocation stratified by municipality, in the presence of school leaders, municipality representatives and study researchers"
Allocation concealment (selection bias) Low risk Quote from publication: "randomisation was performed by random selection of sealed envelopes containing the intervention allocation stratified by municipality, in the presence of school leaders, municipality representatives and study researchers"
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: not possible
Blinding of outcome assessment (detection bias)
All outcomes Low risk Quote from publication: "the research assistants were blinded to the randomisation result for measurement of the outcomes and for data entry"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Quote from publication: "during the 9‐month intervention period, the dropout rate was 13.7% in the control group and 8.8% in the intervention group, which was not statistically significant. Dropouts were mainly attributed to subjects not present at follow‐up trials due to sickness or moving to a different city as well as subjects not being able to complete the test due to injury (e.g. the fitness test)"
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Quote from publication: "during the 9‐month intervention period, the dropout rate was 13.7% in the control group and 8.8% in the intervention group, which was not statistically significant. Dropouts were mainly attributed to subjects not present at follow‐up trials due to sickness or moving to a different city as well as subjects not being able to complete the test due to injury (e.g. the fitness test)"
Selective reporting (reporting bias) Low risk Comment: all main outcomes reported
Cluster RCT ‐ Recruitment bias Low risk Comment: all baseline data were measured prior to randomisation [author communication]
Cluster RCT ‐ Baseline imbalance Low risk Quote from publication: "there were no significant differences at baseline between intervention and control group in any descriptive characteristics except height"
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "because of the clustered nature of the data, schools were included as random effects in the analyses and the Kenward‐Roger degrees of freedom approximation was used"

Pablos 2018.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: public (government‐funded) schools in urban areas
School exclusion criteria:
Student inclusion criteria: in Grade 5 or 6 and not enrolled in any other research study
Student exclusion criteria:
Setting: school, urban
Age group: children
Gender distribution: females and males
Country where trial was performed: Spain
Interventions Intervention: Healthy Habits Program included free lunchtime extracurricular activities 2 times/week beginning with a brief 10‐minute talk about healthy habits (diet, PA, sleep, and hygiene) followed by a PE session consisting of a 15‐minute warm‐up, a 40‐minute main section (theme games for the first 22 sessions and modified sports for remaining sessions), and a 10‐minute calming down section involving another theme game led by a trained teacher. Total PA was 150 minutes/week, with intensity and duration increasing gradually throughout the intervention. Students also received a take‐home worksheet to reinforce session topics to be signed by parents or guardians. The programme was accompanied by three 45‐minute talks for parents and teachers about health habits for children
Comparator: continued with daily activities without participating in the Healthy Habits Program
Duration of intervention: 8 months
Duration of follow‐up: 8 months
Number of schools: 4
Theoretical framework: 
Outcomes Fitness
BMI
Study registration
Publication details Language of publication: English 
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "Its purpose is to bring health and education together within the school setting, with the involvement of the family, to achieve healthy lifestyle habits in the short and long term"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "each school was designated as a control or intervention group using 4 opaque envelopes containing the assigned treatment, which was handled by someone who was not involved in the study"
Allocation concealment (selection bias) Low risk Quote from publication: "each school was designated as a control or intervention group using 4 opaque envelopes containing the assigned treatment, which was handled by someone who was not involved in the study"
Blinding of participants and personnel (performance bias)
All outcomes Low risk Comment: participants were not blinded [author communication]
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: outcome assessors blinded [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Quote from publication: "30 children were excluded because of missing administrative data or absence from school when the measurements were taken. Complete data were collected for 158 of the 190 children"
Selective reporting (reporting bias) Unclear risk Comment: no protocol published or trial registry; cannot determine
Cluster RCT ‐ Recruitment bias High risk Comment: randomisation prior to student enrolment [author communication]
Cluster RCT ‐ Baseline imbalance Low risk Comment: BMI used to balance groups statistically at baseline [author communication]
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis High risk Comment: clustering not accounted for in analysis

Robbins 2018.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: located in urban community setting; enrolment greater than 100 girls in each school or more than double the number of girls needed for the study site (N = 50/school) in any combination of Grades 5, 6, 7, and/or 8; student body comprising at least 50% minority versus non‐minority race or ethnicity and including a similar percentage enrolled in the free and reduced lunch programme
School exclusion criteria: administrators not interested in participating, did not agree to random assignment, could not guarantee their availability at post‐intervention follow‐up
Student inclusion criteria: girls in Grades 5 through 7 (ages 10 to 14; 8th graders if needed in schools having only Grades 7 and 8); available and willing to participate in the PA club 3 days/week for 17 weeks; available for follow‐up (9 months after intervention ends); agree to school random assignment; able to read, understand, and speak English
Student exclusion criteria: involved in or planning to be involved in school or community sports or other organised PAs, such as dance lessons, that involve MVPA and require participation 3 or more days/week after school; a health condition precluding safe MVPA
Setting: school, urban
Age group: children/adolescents
Gender distribution: females 
Country where trial was performed: USA
Interventions Intervention 
  • An after‐school PA club 3 days/week at each school conducted by a club manager (1 of whom was male) and 3 to 4 female instructors, all of whom had recent experience conducting school‐ or community‐based PA programmes that involved girls whose ages were similar to those in this study

  • Two face‐to‐face 15‐ to 20‐minute motivational, individually tailored counselling sessions (1 at the beginning, the other at the end of the intervention) with a female health professional having experience with adolescents (e.g. registered or school nurse)

  • Interactive Internet‐based session via iPad (midpoint of intervention) set up by researchers at each school


Comparator: students did not receive any after‐school programming other than the programming currently offered by the school and community
Duration of intervention: 17 weeks
Duration of follow‐up: 13 months
Number of schools: 8
Theoretical framework: health promotion model and self‐determination theory 
Outcomes PA duration
Fitness
BMI
Study registration NCT01503333 (retrospectively registered)
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The primary aim was to evaluate the effect of a Girls on the Move school‐based intervention on minutes of MVPA among fifth‐ to eighth‐grade girls"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: a statistician used a computer programme to randomly assign the 2 schools in each pair to receive either intervention or control [author communication]
Allocation concealment (selection bias) Low risk Comment: schools were registered, paired, and randomised by an independent statistician [author communication]
Blinding of participants and personnel (performance bias)
All outcomes Low risk Quote from publication: "girls, parents or guardians, principals, nurses, and school staff are not told about hypotheses"
Blinding of outcome assessment (detection bias)
All outcomes Low risk Quote from publication: "the measurement and intervention teams function independently so as to blind members of the former group to each school’s randomisation status throughout the entire study"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Quote from publication: "we determined data to be missing at random, and multiple imputation was employed. Based on recommendations and the complexity of the process, 20 imputations were conducted at the individual level"
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Quote from publication: "we determined data to be missing at random, and multiple imputation was employed. Based on recommendations and the complexity of the process, 20 imputations were conducted at the individual level"
Selective reporting (reporting bias) Low risk Comment: outcomes listed in protocol appear to have been reported on
Cluster RCT ‐ Recruitment bias High risk Quote from publication: "prior to the start of the school year, the PI and project manager inform each principal regarding the randomisation status of his or her respective school to assist each principal with future planning. Principals are told that the randomisation information must remain confidential until completion of baseline data collection in the fall"
Cluster RCT ‐ Baseline imbalance Low risk Quote from publication: "both groups of girls (N = 1519) were similar for most baseline characteristics. However, the control group had a higher proportion of Black girls (P = .001) and higher BMI (P = .035) than the intervention group. The intervention group had a higher proportion of healthy weight girls, but a lower proportion of obese girls than the control group (P = .046)"
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost [author communication]
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "models included the group variable, cluster random effect of school, and the following fixed effects: age, BMI z‐score, race, socioeconomic status, ethnicity, pubertal stage, and study year"

Siegrist 2018.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: all secondary schools in the greater district of a city in Southern Germany that were willing to take part over 4 years and to be randomised to an intervention school with a lifestyle intervention programme or a control school
School exclusion criteria: outside of the study area (distance > 30 km)
Student inclusion criteria: Grade 5, parental consent
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country where trial was performed: Germany
Interventions Intervention: school prevention programme aimed at increasing PA inside and outside of school through regular exercise in sports lessons and additional PA in school (active breaks during lessons, active school breaks). Weekly lifestyle lessons were taught by school teachers and were reinforced by worksheets, homework, and practical instructions. The programme also intended to improve eating patterns and other health behaviours (reduction in media use and inactivity). Teachers took part in 4 to 6 annual training sessions. Parents received regular newsletters and were invited to a parental training programme 2 to 3 times per year, which included coaching parents to lead a more active lifestyle and providing nutritional counselling regarding family dinners. 
Comparator: control schools were asked to continue their usual activities
Duration of intervention: 18 months
Duration of follow‐up: 18 months
Number of schools: 15
Theoretical framework: social cognitive theory
Outcomes BMI
Study registration NCT00988754 (retrospectively registered)
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "we implemented a comprehensive cluster‐randomised school‐ and family based lifestyle‐intervention trial in secondary schools with aim to improve PA, physical fitness and cardio metabolic risk factors in children"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "randomisation of schools was conducted by sealed envelopes (1:1) in 8 intervention and 7 control schools"
Allocation concealment (selection bias) Low risk Quote from publication: "randomisation of schools was conducted by sealed envelopes (1:1) in 8 intervention and 7 control schools"
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "a limitation of our study is that the main coordinator of the study was not blinded to the group assignments of the schools"
Blinding of outcome assessment (detection bias)
All outcomes Low risk Quote from publication: "we tried to eliminate this bias by blinding medical examiners who were not aware of the group allocation of the participating children"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: most loss to follow‐up due to moving schools
Selective reporting (reporting bias) High risk Comment: quality of life and anthropometry not reported
Cluster RCT ‐ Recruitment bias High risk Comment: randomisation at school level before enrolment of students
Cluster RCT ‐ Baseline imbalance Low risk Comment: models adjusted for key baseline differences [author communication]
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "data were aggregated per cluster to account for the cluster‐randomised design in the statistical analysis"

Ten Hoor 2018.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: Dutch secondary schools
School exclusion criteria:
Student inclusion criteria:
Student exclusion criteria:
Setting: school
Age group: children/adolescents
Gender distribution: females and males
Country where trial was performed: The Netherlands
Interventions Intervention: strength training exercises during at least 30% of regular PE lessons, ~ 5 to 30 minutes/lesson. Teachers were instructed about the programme and specific strength exercises and safety guidelines, participated in workshops to improve their motivational speaking skills, were provided with materials (medicine balls, elastic bands, and free weights), and received a book with strength exercises and games. Once a month, a 1‐hour lesson based on motivational interviewing and facilitated by a trained mentor or PE teacher was used to increase motivation to be more physically active 
Comparator: continued with  usual curriculum
Duration of intervention: 1 year
Duration of follow‐up: 1 year
Number of schools: 9
Theoretical framework: theory of planned behaviour or reasoned action approach, self‐determination theory, social comparison theory, intervention mapping
Outcomes PA duration
Sedentary time
Study registration NTR5676 
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "In this cluster RCT, we investigated the 1‐year efficacy of incorporating strength exercises into gym classes, in combination with monthly motivational lessons (to engage in PAs after school) on the body composition and activity level of adolescents"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "randomised (stratified on education level; by flip of a coin by the first author under supervision of the fourth author) into an intervention condition (4 schools) or a standard curriculum control condition (5 schools)"
Allocation concealment (selection bias) Low risk Comment: allocation concealed from schools or directors; informed that they may or may not receive the intervention immediately [author communication]
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: not possible [author communication]
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: not possible [author communication]
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: 26% to 28% loss to follow‐up
Selective reporting (reporting bias) High risk Comment: strengths listed in clinical trials registry but not reported
Cluster RCT ‐ Recruitment bias High risk Comment: students were enrolled and baseline data collected after school randomisation [author communication]
Cluster RCT ‐ Baseline imbalance Low risk Quote from publication: "no baseline differences were found between the 2 conditions in age, height, weight, BMI (Z‐score), body composition, or PA outcomes (including wear time of the accelerometer)"
Cluster RCT ‐ Loss of clusters Low risk Comment: no loss of clusters
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "the random (variance) model part consisted of an unstructured covariance matrix for the within‐school variances and covariance of the 2 repeated measures plus a random intercept for the between‐school outcome variance"

Donnelly 2017.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: elementary schools within a 25‐mile radius of Lawrence, Kansas, including Grades 2 through 5, with at least 40 students in Grades 2 and 3, not participating in other classroom‐based PA interventions, and agreeing to be randomised
School exclusion criteria:
Student inclusion criteria: a random sample of Grade 2 and 3 students in each school from those who provided parental consent or child assent
Student exclusion criteria: students with physical or intellectual disabilities or learning disorders were part of the intervention or control group as a function of attending the school; however, some were ineligible to complete the outcome assessment due to their disability (i.e. blind, severe intellectual disability, etc.)
Setting: school
Age group: children
Gender distribution: females and males
Country where trial was performed: USA
Interventions Intervention: teacher‐delivered 10‐minute Academic Achievement and Physical Activity Across the Curriculum lessons twice per day (1 morning and 1 afternoon) 5 days/week plus 60 minutes of PE to total 160 minutes/week of MVPA
Comparator: teachers were asked to continue to use traditional classroom instruction and to continue with their typical PE schedule (2‐ to 30‐minute classes/week)
Duration of intervention: 3 years
Duration of follow‐up:
Number of schools: 17
Theoretical framework: —
Outcomes BMI
Fitness
Study registration NCT01699295 (retrospectively registered)
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The primary aim of this study is to assess differences in academic achievement in students who receive physically active lessons and students in control schools who receive regular academic lessons. Secondary aims include determining potential mediators of any association between Academic Achievement and Physical Activity Across the Curriculum and academic achievement, including changes in cognitive function, cardiovascular fitness, daily PA, BMI and attention‐to‐task. An extensive process analysis will also be performed to document the fidelity of the intervention"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "computer randomised by study statistician"
Allocation concealment (selection bias) Unclear risk Comment: not described
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students and teachers not blinded
Blinding of outcome assessment (detection bias)
All outcomes Low risk Quote from publication: "staff completing assessments, other than those obtained in the classroom, and staff performing data entry, were blinded to condition"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: large loss to follow‐up; no imputation for secondary outcomes
Selective reporting (reporting bias) High risk Comment: PA not reported
Cluster RCT ‐ Recruitment bias High risk Comment: students recruited after randomisation
Cluster RCT ‐ Baseline imbalance High risk Comment: baseline differences not controlled for
Cluster RCT ‐ Loss of clusters High risk Comment: loss of clusters from both groups
Cluster RCT ‐ Incorrect analysis Low risk Comment: clustering accounted for in analysis

Farmer 2017.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: state primary schools (Years 1 to 8 that are fully funded by the state and are co‐educational) with at least 150 pupils and a school decile ranking of 1 to 6 
School exclusion criteria:
Student inclusion criteria: children in school years 2 and 4
Student exclusion criteria: no exclusion criteria
Setting: school
Age group: children
Gender distribution: females and males
Country where trial was performed: New Zealand
Interventions Intervention: researchers, play worker, and school community worked together and received funding over the course of 1 year to develop a playground action plan tailored for each intervention school (e.g. addition of more interactive play equipment, alterations to school rules and policies). The majority of recommendations involved no to little cost, such as leaving trees that had been cut down in pieces or letting the grass grow long to encourage imaginative play, re‐purposing real estate signs for sledding down hills, purchasing raincoats and gumboots to allow outside play when wet, and using plastic piping and sand for water play
Comparator: asked not to change anything in school play spaces
Duration of intervention: 2 years
Duration of follow‐up: 2 years
Number of schools: 16
Theoretical framework: —
Outcomes PA duration
BMI
Study registration ACTRN12612000675820 (retrospectively registered)
Publication details Language of publication: English 
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The aim of our 2‐year cluster RCT (PLAY) was to determine whether providing greater opportunities for risk and challenge in primary schools increased PA and reduced relative body weight over the long term. A secondary aim considered the effect of the intervention on how children interacted with 1 another which forms the basis of a separate paper"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "pairs of schools were created by matching for region, school roll and decile ranking, and were randomly assigned to intervention or control by tossing a coin"
Allocation concealment (selection bias) Low risk Comment: coin toss used
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded
Blinding of outcome assessment (detection bias)
All outcomes Low risk Quote from publication: "measurements were obtained … by researchers blinded to group allocation"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Quote from publication: "the missing data were imputed using chained equations assuming that the data were missing at random"
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Quote from publication: "the missing data were imputed using chained equations assuming that the data were missing at random"
Selective reporting (reporting bias) High risk Comment: bullying, steps/d, nutrition not reported
Cluster RCT ‐ Recruitment bias High risk Comment: students enrolled after randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: clusters similar; models adjusted
Cluster RCT ‐ Loss of clusters Low risk Comment: no loss of clusters
Cluster RCT ‐ Incorrect analysis Low risk Comment: clustering accounted for in analysis

Sutherland 2017.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: government or Catholic elementary schools; located within Hunter New England Local Health District; socioeconomic score ≤ 5 (lower 50% of New South Wales) based on school post code; not participating in other PA studies
School exclusion criteria:
Student inclusion criteria: Grades 3 to 6
Student exclusion criteria: major physical or intellectual conditions impeding engagement in PA
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Australia
Interventions Intervention: a modified version of the Supporting Children’s Outcomes using Rewards, Exercise and Skills programme consisting of school committees and policy review, quality PE lessons, recess and lunchtime activity via student leadership, provision of equipment, and linkage with parents and community sporting organisations
Comparator: measurement components of the trial only; school PA practices according to curriculum
Duration of intervention: 6 months
Duration of follow‐up: 6 months
Number of schools: 46
Theoretical framework: social‐ecological theory and health promoting schools framework
Outcomes PA duration
Study registration ACTRN12615000437561 (retrospectively registered)
Publication details Language of publication: English 
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The primary aim of this paper is to report the effectiveness of an adapted version of an evidence based school PA program known as Supporting Children’s Outcomes using Rewards, Exercise and Skills on children’s MVPA. Secondary trial outcomes describe the impact on school implementation of practices including PE teaching quality and school PA practices"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "schools underwent stratified randomisation based on SES, allocated in a 1:1 ratio to intervention or control by an independent statistician using computerized random number function in Microsoft Excel"
Allocation concealment (selection bias) High risk Comment: random allocation of schools to group occurred post recruitment but before data collection
Blinding of participants and personnel (performance bias)
All outcomes Low risk Comment: personnel delivering intervention blinded; primary outcome assessment concealed from participants, so lack of blinding of participants unlikely to affect outcome
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: outcome assessors blinded
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Unclear risk Comment: no mention of missing data at end of study
Selective reporting (reporting bias) Low risk Comment: trial protocol referenced
Cluster RCT ‐ Recruitment bias Low risk Comment: participants recruited before cluster randomisation
Cluster RCT ‐ Baseline imbalance Unclear risk Comment: no analysis of clusters for similarity
Cluster RCT ‐ Loss of clusters High risk Comment: 1 cluster lost after randomisation; no reason given
Cluster RCT ‐ Incorrect analysis Low risk Comment: statistical analysis accounted for clustering

Torbeyns 2017.

Study characteristics
Methods Study design: RCT
Participants Student inclusion criteria: Grades 3 and 4 of high school in Ninove, Belgium
Student exclusion criteria:
Setting: school
Age group: adolescents
Gender distribution: females and males
Country/Countries where trial was performed: Belgium
Interventions Intervention: students were instructed to cycle on a height‐adjustable cycling desk (LifeSpan C3‐DT5 Bike Desk) for 4 class hours/week (4 × 50 min). Participants were free to adjust the cycling intensity to their preference. All participants were asked to not change their lifestyle during the study (e.g. PA levels outside the classroom)
Comparator: asked to not change their lifestyle during the study (e.g. PA levels outside the classroom)
Duration of intervention: 22 weeks
Duration of follow‐up: 22 weeks
Number of schools: 1
Theoretical framework: —
Outcomes Fitness
BMI
Study registration
Publication details Language of publication: English
Funding: other funding (no sources of funding received)
Publication status: peer‐reviewed journal
Stated aim for study "Thus, the purpose of this study was to examine the impact of providing cycling desks in a classroom for 5 months on energy expenditure, physical health parameters, cognitive performance, brain functioning and educational measures in an adolescent population"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Unclear risk Comment: not stated
Allocation concealment (selection bias) Unclear risk Comment: not stated
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded
Blinding of outcome assessment (detection bias)
All outcomes Unclear risk Comment: not stated
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Quote from publication: "7 more participants (4 intervention, 3 control) were excluded because they were absent for more than 1 week during the intervention period"
Selective reporting (reporting bias) Unclear risk Comment: no protocol referenced

Daly 2016.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: suburb primary schools in Canberra, where the average household income approximates the mean for Australian city dwellers
School exclusion criteria:
Student inclusion criteria: all Grade 2 children in good health, able to participate freely in PE, willing to undertake a series of venous blood collections
Student exclusion criteria:
Setting: school 
Age group: children/adolescents
Gender distribution: females and males
Country/Countries where trial was performed: Australia
Interventions Intervention: specialist PE teacher to replace general classroom teacher programme. An all‐inclusive, enjoyable, challenging yet non‐threatening environment for PA that encouraged students to discover answers to a range of physical movement problems and game strategies themselves, through experimentation and self‐discovery. Lesson plans included game play, fitness activities, skill practice, and core movements
Comparator: continued with usual PE programme (traditional fitness and stretching exercise, including running and walking around the oval) conducted by the generalist classroom teachers, none of whom were formally trained in PE. Teachers rarely participated in activities
Duration of intervention: 4 years
Duration of follow‐up: 4 years
Number of schools: 29
Theoretical framework: —
Outcomes PA duration 
Sedentary time
Study registration ACTRN12612000027819
Publication details Language of publication: English
Funding: other funding (charitable trust)
Publication status: peer‐reviewed journal 
Stated aim for study "The primary aim of the Lifestyle of Our Kids project is to investigate relationships of (a) PA in general, and (b) an externally provided specialist PE program in schools, with physiological and psychological health and development in young children. This study is intended to provide a range of integrated scientific evidence upon which conclusions may be drawn in regard to optimising childhood health and development"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "after acceptance, the allocation of schools to the intervention or control groups was determined randomly, 2 members of the research team drawing from a shuffled set of 29 envelopes, each of which contained a school name"
Allocation concealment (selection bias) Low risk Comment: allocation concealed until after enrolment
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: research staff blinded to group only for some measures
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Comment: large loss to follow‐up but unrelated to intervention (i.e. relocation)
Selective reporting (reporting bias) High risk Comment: academic performance and hand‐eye coordination not reported
Cluster RCT ‐ Recruitment bias Low risk Comment: students enrolled after randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: no baseline imbalance
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "school, year, and subject were included in the model as random effects to account for the sample design and hence possible dependence structure in the data"

de Greeff 2016.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: primary schools in the Northern part of The Netherlands
School exclusion criteria:
Student inclusion criteria: Grade 2 or 3; all children from that class participated in the intervention programme
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: The Netherlands
Interventions Intervention: physically active mathematics and language lessons that were taught in the classroom. During each lesson, the children stood behind or beside their school desk. In each lesson, 10 to 15 minutes was spent on mathematics and 10 to 15 minutes on language. Physical exercises were relatively easy to perform and were aimed at exercising at moderate to vigorous intensity level. During the lessons, all children performed basic exercises and specific exercises simultaneously. Specific exercises were performed when children solved an academic task. For example, the word ‘dog’ must be spelled by jumping in place for every mentioned letter, or children had to jump 6 times to solve the multiplication ‘2 × 3’. Basic exercises (marching, jogging, or hopping in place) were performed during the remaining part of the lesson (e.g. when children were thinking about a sum)
Comparator:
Duration of intervention: 2 × 22 weeks
Duration of follow‐up: 3 years
Number of schools: 12
Theoretical framework: —
Outcomes Fitness
BMI
Study registration ISRCTN17021806
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "This study is part of the project ‘‘Fit en Vaardig op school’’ (fit and academically proficient at school), which is a RCT including a school‐based intervention program for primary school children with the primary aim to improve academic performance"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Unclear risk Comment: not described
Allocation concealment (selection bias) Low risk Comment: performed by the National Bureau for Economic Policy Analysis ‐ not involved with the study
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students not blinded
Blinding of outcome assessment (detection bias)
All outcomes Unclear risk Comment: not described
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Unclear risk Comment: unclear where different sample sizes come from across papers
Selective reporting (reporting bias) Low risk Comment: all outcomes from clinical trials registry reported on
Cluster RCT ‐ Recruitment bias Unclear risk Comment: not described
Cluster RCT ‐ Baseline imbalance Low risk Comment: adjusted for in analysis
Cluster RCT ‐ Loss of clusters High risk Comment: 2 schools did not start the second intervention period
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "a random intercept was considered for each child (level 1) and for each school (level 2), to account for the common experience the children share within each school"

Drummy 2016.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: convenience sample of primary schools
School exclusion criteria:
Student inclusion criteria: students aged 9 and 10
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Northern Ireland
Interventions Intervention: teachers in the intervention group were asked to lead a 5‐minute activity break 3 times/d for 12 weeks. The activity break began with gentle jogging on the spot as a warm‐up for less than 1 minute, followed by moderate to vigorous intensity exercises such as hopping, jumping, and running on the spot, scissor kicks, etc. Teachers could select which exercises to include in each activity break. They were encouraged to vary activities each day. Children participated in the activity break in the classroom beside their desks
Comparator: control groups continued with their normal daily routine throughout the 12‐week period
Duration of intervention: 12 weeks
Duration of follow‐up: 12 weeks
Number of schools: 7
Theoretical framework: —
Outcomes PA duration
BMI
Study registration
Publication details Language of publication: English
Funding:
Publication status: peer‐reviewed journal
Stated aim for study "Thus, the present study examined the effect of a classroom‐based activity break on accelerometer‐determined MVPA and anthropometric variables in a sample of children attending primary schools in Northern Ireland"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) High risk Comment: stratified by rural/urban location and given alternate numbers by researcher [author communication]
Allocation concealment (selection bias) Unclear risk Comment: concealed from participants but investigator could have foreseen assignment [author communication]
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants could not be blinded
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: no blinding [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: > 10% loss to follow‐up; reasons not collected [author communication]
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: > 10% loss to follow‐up; reasons not collected [author communication]
Selective reporting (reporting bias) Unclear risk Comment: no protocol; appears all outcomes reported
Cluster RCT ‐ Recruitment bias High risk Comment: baseline data collected after randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: no baseline differences
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis High risk Comment: clustering not accounted for in analysis

Jarani 2016.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: elementary schools in Tirana that were not engaged in any PA programme or intervention, had gym space larger than 180 m, and a typical indoor air investigation, which included observation of conditions in the area of the gym and comfort parameters such as humidity and temperature
School exclusion criteria:
Student inclusion criteria: Grades 1 and 4
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Albania
Interventions Intervention 1: exercise group intervention programme emphasised PA exercises (e.g. gait exercises to improve speed). PE lessons were organised to allow maximum participation of children using station or circuit teaching framework to provide opportunities for continuous practice on different exercises at the same time. The exercise group programme was structured in 4 different modules: movement awareness, object or manipulative skills, rhythm, tumbling or gymnastics for Grade 1; and throwing or catching, rhythm, fitness, tumbling or gymnastics for Grade 4, respectively. PE structures station or circuit activities were orientated on an individual school child
Intervention 2: games group intervention programme was focused on fun games. PE lessons were organised to allow maximum participation of children using station or circuit teaching framework to provide opportunities for continuous practice on different exercises or games at the same time. The programme was structured in 4 different modules: movement awareness, object or manipulative skills, rhythm, tumbling or gymnastics for grade 1 and throwing or catching, rhythm, fitness, tumbling or gymnastics for grade 4, respectively. The games group‐based PE focused on involving more than 3 or 4 children in a game organised to address each of the 4 modules
Comparator: traditional PE school curriculum given by classroom teachers using traditional PE programmes including mainly typical sports
Duration of intervention: 5 months
Duration of follow‐up: 5 months
Number of schools: 4
Theoretical framework: —
Outcomes Fitness
BMI
Study registration
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The aim of this study was to examine the effectiveness of 2 new PE programmes on skill‐ and health‐related fitness among elementary school children aged 7 and 10 years in Tirana, Albania, without changing the traditional frequency and duration of the current PE in Albanian schools. Another aim of the study was to compare 2 PE programme approaches to improve children’s physical fitness: 1 with the emphasis on exercises and the other on games using station circuit teaching framework during PE lessons"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: random numbers table
Allocation concealment (selection bias) Low risk Comment: allocation was concealed [author communication]
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: blinding of students not possible
Blinding of outcome assessment (detection bias)
All outcomes High risk Quote from publication: "...not able to conduct the collection of data blinded to what group the children were randomly allocated"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: high completion rates; no differences between groups; no differences between completers and non‐completers
Selective reporting (reporting bias) Unclear risk Comment: no protocol published
Cluster RCT ‐ Recruitment bias High risk Comment: classes were randomised before students enrolled [author communication]
Cluster RCT ‐ Baseline imbalance Low risk Comment: accounted for in analysis
Cluster RCT ‐ Loss of clusters Low risk Comment: no loss of classes or schools
Cluster RCT ‐ Incorrect analysis Low risk Comment: school and class as random effects in model

Kocken 2016.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria:
School exclusion criteria:
Student inclusion criteria: 9 to 11 years
Student exclusion criteria:
Setting: school, mix
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: The Netherlands
Interventions Intervention: the intervention Extra Fit! comprised a variety of theory and practical lessons on nutrition and PA to provide an attractive programme for children. The intervention was focused on the main behavioural changes: decreasing consumption of high‐energy or high‐fat foods and sugar‐sweetened drinks; promoting a healthy breakfast; increasing consumption of fruits and vegetables; reducing television viewing and computer gaming or browsing; and increasing PAs at school and outside school hours. Behavioural determinants of the Theory of Planned Behaviour that were targeted were knowledge (theory lessons and practical assignments), attitude (group discussions and food diaries), social norm (group discussions and homework assignments), and perceived behavioural control (modelling through assignments, e.g. preparing a healthy meal and PA games)
Comparator: regular school programme or curriculum on nutrition and PA
Duration of intervention: 2 × 16 weeks
Duration of follow‐up: 24 months
Number of schools: 45 
Theoretical framework: theory of planned behaviour
Outcomes PA duration
Sedentary time
BMI
Study registration
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The aim of Extra Fit! was to improve dietary habits, PA and inactivity behavior in order to prevent overweightness"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Unclear risk Comment: not described
Allocation concealment (selection bias) Unclear risk Comment: not described
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants could not be blinded to intervention
Blinding of outcome assessment (detection bias)
All outcomes Unclear risk Comment: interviewers were blinded with respect to group status of the child’s school (intervention or control) for nutrition interviews, not other measures
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: large loss to follow‐up
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: large loss to follow‐up
Selective reporting (reporting bias) Unclear risk Comment: no protocol
Cluster RCT ‐ Recruitment bias Low risk Comment: all students within randomised schools enrolled
Cluster RCT ‐ Baseline imbalance Low risk Comment: hip‐waist ratio was significantly higher in the control group at time 0 (t‐test, P < 0.01). There were no other significant differences in baseline characteristics between intervention and control groups
Cluster RCT ‐ Loss of clusters High risk Comment: 20/65 schools dropped out after randomisation; 7 schools lost to follow‐up after intervention began
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "outcome analyses were conducted using multilevel regression models, with schools included as a random effect"

Lau 2016.

Study characteristics
Methods Study design: RCT
Participants Student inclusion criteria: all students in Grade 4
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: China
Interventions Intervention: children participated in two 60‐minute Xbox 260 Kinect gaming sessions/week after school for 12 school weeks. Children were free to choose games from the 12 offered sports in Season 1 or Season 2 within a play session. This approach was chosen to encourage children’s autonomy and to enhance attractiveness and the challenge of game play. Children and their partners with consensus of opinion had their own choice on the order of games, what they wanted to play, and the duration of each game play. Participants could get awarded based on degree and speed of movement and level of difficulty
Comparator: adopted regular PA and PE class and received no additional intervention
Duration of intervention: 12 weeks
Duration of follow‐up: 12 weeks
Number of schools: 1
Theoretical framework: —
Outcomes PA duration
Fitness
BMI
Study registration
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "Thus, besides the effect of an active video games intervention on children’s aerobic fitness and PA level, this study also sought to explore the active video games impact on players’ psychological correlates, including PA task efficacy, barrier efficacy, and enjoyment"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: random numbers table [author communication]
Allocation concealment (selection bias) High risk Comment: allocation not concealed from investigators [author communication]
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: no blinding of participants
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: outcome assessors were blinded [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: no loss to follow up
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Comment: no loss to follow up
Selective reporting (reporting bias) Unclear risk Comment: no protocol published

Resaland 2016.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: schools had ≥ 7 pupils in Grade 5
School exclusion criteria:
Student inclusion criteria: children were healthy (with no serious or chronic illness) and able to participate in daily PA and PE. Participants had to be able to complete standard academic performance tests (our primary outcome)
Student exclusion criteria:
Setting: school, mix
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Norway
Interventions Intervention: comprised 3 components aimed at providing children with the opportunity to engage in 165 minutes of PA/week more than the control group
  • Physically active lessons for 90 minutes/week, conducted in the playground; physically active educational lessons were delivered in 3 core subjects – Norwegian, mathematics, and English

  • PA breaks (5 minutes/d) implemented in the classroom during academic lessons

  • PA homework (10 minutes/d) prepared by teachers


In addition, pupils participated in the curriculum‐prescribed 90 minutes/week of PE and the curriculum‐prescribed 45 minutes/week of PA. Thus, PA (165 minutes/week) and PE or PA (135 minutes/week) components provided children opportunities to engage in school‐based PAs 300 minutes/week. The intervention was designed so activities could be varied and enjoyable for the children. Teachers were encouraged to motivate children during active lessons, to stimulate their positive feelings and attitudes towards PA. The intervention was designed so approximately 25% of daily PA was of vigorous intensity, defined as “children sweating and being out of breath.” Teachers achieved the vigorous‐PA‐intensity component through selecting a variety of high‐intensity activities such as running, relay racing, obstacle courses, and various forms of high‐activity play
Comparator: curriculum‐prescribed 90 minutes/week of PE and 45 minutes/week of PA (total 135 minutes/week). It was specified to control schools that they should carry out the amount of PA and PE that they would have done regardless of the study
Duration of intervention: 7 months
Duration of follow‐up: 7 months
Number of schools: 60
Theoretical framework: socioecological conceptual framework
Outcomes PA duration
Sedentary time
Fitness
Health‐related quality of life
Study registration NCT02132494
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "Our primary objective is to investigate the effect of a 1‐year school‐based PA intervention (Active Smarter Kids; ASK) on academic performance on a sample of 10‐year‐old boys and girls attending elementary school in Norway"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: a neutral third party (Centre for Clinical Research, Haukeland University Hospital, Norway) performed the randomisation
Allocation concealment (selection bias) Low risk Quote from publication: a neutral third party (Centre for Clinical Research, Haukeland University Hospital, Norway) performed the randomisation
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: groups not blinded to intervention
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: only data manager and statisticians blinded to group allocation
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: low loss to follow‐up
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Comment: low loss to follow‐up
Selective reporting (reporting bias) Low risk Comment: outcomes listed in protocol appear to have been reported
Cluster RCT ‐ Recruitment bias Low risk Comment: > 97% recruitment
Cluster RCT ‐ Baseline imbalance Low risk Comment: no differences
Cluster RCT ‐ Loss of clusters High risk Comment: N = 3 clusters withdrew after randomisation
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "analyses were performed using a mixed‐effect model with school as a random effect"

Sutherland 2016.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: government and Catholic schools; schools with post codes ranked in the bottom 50% of New South Wales post codes based on the Socio‐Economic Indexes for Australia; between 120 and 200 Year 7 students (to meet sample size requirements); and not participating in other major PA or health intervention studies
School exclusion criteria:
Student inclusion criteria: all Year 7 students in participating schools will be eligible to participate in the study measurement
Student exclusion criteria: classes catering for students with severe physical and mental disabilities will be excluded
Setting: school
Age group: adolescents
Gender distribution: females and males
Country/Countries where trial was performed: Australia
Interventions Intervention: intervention involved implementation of 7 PA intervention strategies and 6 strategies to support implementation of the intervention. PA intervention strategies included:
  • teaching strategies to maximise students’ PA in health and PE lessons;

  • development and monitoring of student PA plans within PE lessons;

  • enhanced school sport program;

  • development or modification of school policies;

  • PA programmes during school break;

  • promotion of community PA providers; and

  • parent engagement


The intervention implementation strategies included
  • in‐school PA consultant (change agent)

  • establishing leadership and support

  • teacher training

  • school resources

  • teacher prompts

  • intervention implementation performance feedback


Comparator: schools allocated to control group participated in measurement components of the trial only and delivered PA teaching and promotion practices according to PE curriculum and school‐based initiatives
Duration of intervention: 24 months
Duration of follow‐up: 24 months
Number of schools: 10
Theoretical framework: socioecological theory and health‐promoting schools framework
Outcomes PA duration
BMI
Study registration ACTRN12612000382875
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The aim of this study is to determine whether a multi‐component PA intervention implemented in disadvantaged secondary schools can reduce the decline in PA associated with adolescence"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: block randomisation using random numbers function
Allocation concealment (selection bias) Low risk Comment: independent statistician conducted randomisation after data were collected
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: data were collected by trained research assistants blind to group allocation
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Quote from publication: "2 sensitivity analyses were conducted, first using only those that provided complete adiposity outcomes at all 3 time points (complete cases), and second using multiple imputation to fill in the missing data. The multiple imputation model used the method of chained regression equations, including variables that were prognostic of missing data and additional demographic and outcome data to create 5 imputed data sets. The results from fitting the model were pooled over the 5 data sets using Rubin’s method"
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Quote from publication:  "PA outcome data were analysed assuming data were “missing at random.” Sensitivity analyses were undertaken for the primary outcome, initially adjusting for any variables on which students with and without 24‐month follow‐up accelerometer data were significantly different, and secondly, using multiple imputation"
Selective reporting (reporting bias) Low risk Comment: all outcomes in protocol reported
Cluster RCT ‐ Recruitment bias Low risk Comment: randomisation occurred after baseline data collected
Cluster RCT ‐ Baseline imbalance Unclear risk Comment: did not assess
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "a 3‐level hierarchical model was used to capture correlations in the data with random intercepts for repeated measures on individuals clustered within schools"

Tarp 2016.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: all schools participating in a primary school project about PA and health
School exclusion criteria:
Student inclusion criteria: all students in Grade 6 and 7 classes
Student exclusion criteria:
Setting: school, mix
Age group: adolescents
Gender distribution: females and males
Country/Countries where trial was performed: Denmark
Interventions Intervention: 60 minutes of PA during school time on all school days including scheduled activities during recess were initiated by teachers and volunteer students. PA homework consisted of a booklet containing suggestions for various daily activities of 5 to 10 minutes, and students were instructed to perform at least 1 activity on all days. It was emphasised that these activities were in addition to usual daily activities. A 2‐week cycling campaign was launched in the middle of the intervention to facilitate active transportation by cycling. A custom‐made “activity watch” was used as a shared tool by teachers and students to serve as motivation and to sum up the amount of time the class had engaged in PA during academic subjects and scheduled recess activities
Comparator: schools were asked to continue with their normal practice
Duration of intervention: 20 weeks
Duration of follow‐up: 20 weeks
Number of schools: 14
Theoretical framework: social cognitive theory, socioecological theory, health promoting schools framework from WHO
Outcomes PA duration
Fitness
BMI
Study registration NCT02012881
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The primary aim of the present paper is to describe the effectiveness of a school‐based PA intervention in enhancing executive functions and academic performance in adolescents. Furthermore, the effect of the intervention on PA levels, cardiorespiratory fitness and adiposity were assessed as these are potential explanatory variables"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "the randomisation process was conducted by the principal investigator as draws of folded paper with school names from a bowl in the presence of other senior researchers"
Allocation concealment (selection bias) Low risk Quote from publication: "the randomisation process was conducted by the principal investigator as draws of folded paper with school names from a bowl in the presence of other senior researchers"
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "baseline measures were carried out after the randomisation with schools aware of the randomisation results. This was also known to the staff conducting the tests"
Blinding of outcome assessment (detection bias)
All outcomes Unclear risk Quote from publication: "baseline measures were carried out after the randomisation with schools aware of the randomisation results. This was also known to the staff conducting the tests"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: large loss to follow‐up on some outcome measures. Significant differences at baseline between students and those lost to follow‐up included in analyses
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: large loss to follow‐up on some outcome measures. Significant differences at baseline between students and those lost to follow‐up included in analyses
Selective reporting (reporting bias) Low risk Comment: all outcomes listed in protocol paper reported
Cluster RCT ‐ Recruitment bias High risk Comment: schools randomised before participants recruited
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline imbalance accounted for in analysis
Cluster RCT ‐ Loss of clusters High risk Quote from publication: "2 intervention schools withdrew the acceptance before the start of the study, but after the randomisation. Additionally, 3 schools not originally enrolled in the study were invited from local networks and randomised"
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "as the units of randomisation were schools a “random effect” was added by using mixed models (maximum likelihood based) to accommodate the clustering of students within these units"

Cohen 2015.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: 16 government primary schools located within 30 minutes' drive from the University of Newcastle, with a Socio‐Economic Indexes for Areas ≤ 5 (lowest 50%) 
School exclusion criteria:
Student inclusion criteria: all students in Grades 3 and 4 (stage 2) at study schools
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Australia
Interventions Intervention: implemented in 3 phases. Phase 1 focused on teacher professional learning, student leadership workshops, and PA promotion tasks to achieve awards. Examples of tasks included acting as equipment monitor, organising games during recess and lunch, and writing a PA promotion article for the school newsletter. Equipment was provided to the school during this phase, and the school committee was established. In phase 2, schools were encouraged to implement 6 PA policies to support the promotion of PA and fundamental movement skill competency within the school. A member of the research team met with the principal at the intervention schools to explain the policies. The member of the research team then conducted a meeting with all staff members to explain the policies and to provide strategies for implementation of the policies. In addition, the research team used a range of strategies targeting the home environment (newsletters, parent evening, and fundamental movement skill homework) to engage parents and encourage them to support their children’s PA. Phase 3 addressed strategies to improve school–community links (e.g. inviting local sporting organisations to assist with school sport programmes)
Comparator: control group followed the usual PE and school sport programmes. The New South Wales Department of Education and Communities requires by policy that all schools provide students with 120 minutes/week of planned PA. In government primary schools, sports programmes are similar among schools
Duration of intervention: 12 months
Duration of follow‐up: 12 months
Number of schools: 8
Theoretical framework: socioecological model
Outcomes PA duration
Fitness
Study registration ACTRN12611001080910
Publication details Language of publication: English 
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "Supporting Children's Outcomes using Rewards, Exercise, and Skills is a multi‐component school‐based intervention that combines a range of evidence‐based behavior change strategies to promote PA and fundamental movement skills competency among primary school aged children from low‐income communities"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "randomly allocated to the intervention or control group using a computer‐based random number producing algorithm by a researcher not involved in the current study"
Allocation concealment (selection bias) Low risk Comment: randomisation occurred after baseline assessment [author communication]
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded to intervention
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: assessors were blind to treatment allocation at baseline but not at follow‐up assessments
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: large loss to follow‐up; different by group
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: large loss to follow‐up; different by group
Selective reporting (reporting bias) Low risk Comment: all outcomes in protocol reported
Cluster RCT ‐ Recruitment bias Low risk Comment: baseline assessments were conducted prior to randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline imbalance adjusted for statistically
Cluster RCT ‐ Loss of clusters Low risk Comment: no loss of clusters reported
Cluster RCT ‐ Incorrect analysis Low risk Comment: analysis accounted for clustered design

Jago 2015.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: state secondary schools operating within 3 local authorities: Bristol City Council, North Somerset Council, Bath and North East Somerset Council
School exclusion criteria: excluding special educational needs, dance academies, and privately or independently funded schools
Student inclusion criteria: Year 7 girls who are able to engage in PE class
Student exclusion criteria:
Setting: school
Age group: children/adolescents
Gender distribution: females
Country/Countries where trial was performed: UK
Interventions Intervention: intervention consisted of up to 40, 75‐minute dance sessions provided twice per week after school. Session plans included guidance on how to reinforce the underpinning self‐determination theory principles as well as advice on activities, group work, and dance skill development. To reflect a ‘normal’ dance session, instructors were able to decide on the genre of dance used, after consultation with the girls at their school
Comparator: provided data only
Duration of intervention: 20 weeks
Duration of follow‐up: 52 weeks
Number of schools: 18
Theoretical framework: self‐determination theory
Outcomes PA participation
PA duration
Sedentary time
Health‐related quality of life
Study registration ISRCTN52882523(retrospectively registered)
Publication details Language of publication: English 
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "Determine the effectiveness of the Bristol Girls Dance Project intervention to improve the objectively‐assessed (accelerometer) mean weekday min of MVPA accumulated by Year 7 girls 1 year after the baseline measurement (Time 2 = time 0 + 52 weeks)"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: using Stata to balance arms by minimisation
Allocation concealment (selection bias) Low risk Comment: 1 investigator conducted randomisation; 1 notified schools of allocation after randomisation completed
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: trained field workers; blinded to intervention allocation; collected data
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Comment: minimal missing data
Selective reporting (reporting bias) Low risk Comment: all outcomes in protocol reported in results
Cluster RCT ‐ Recruitment bias Low risk Comment: randomisation after baseline data collection
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline imbalances adjusted for
Cluster RCT ‐ Loss of clusters Low risk Comment: no loss of clusters
Cluster RCT ‐ Incorrect analysis Low risk Comment: all models adjusted for clustering of girls in schools

Madsen 2015.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: > 50% of students eligible for free or reduced‐price lunch; average of at least 60 students/grade; no exposure to Playworks in the past 5 years
School exclusion criteria:
Student inclusion criteria: all Grade 3, 4, and 5 students from the 6 participating schools
Student exclusion criteria:
Setting: school, urban
Age group:
Gender distribution: females and males
Country/Countries where trial was performed: USA
Interventions Intervention: each intervention school received 1 part‐time registered dietician coach and 1 full‐time Playworks coach for 2 school years. Each year, the registered dietician coach delivered a 12‐week nutrition and energy balance education curriculum that included food tasting, PA games to reinforce nutrition messages, and strategies to help students meet their nutrition and PA goals. Registered dietician coaches also worked with a team of school staff and parents to implement classroom wellness policies and to make improvements in school food, including increased offerings of fruits and vegetables to meet the Bronze‐level Healthier USA School Challenge criteria. The Playworks coach structured recess activities before and during school hours to encourage active participation by all students. The Playworks coach also led a PA session with individual classes every other week. Classroom teachers were trained to implement Playworks games and classroom management strategies in their PE sessions with students. Last, Playworks coaches led after‐school sports leagues throughout each year
Comparator:
Duration of intervention: 2 years
Duration of follow‐up: 2 years
Number of schools: 6
Theoretical framework: social cognitive theory
Outcomes Fitness
BMI
Study registration
Publication details Language of publication: English
Funding: non‐commercial funding (charitable trust)
Publication status: peer‐reviewed journal
Stated aim for study "The current study was designed to test the impact of Energy Balance 4 Kids with Play on students’ nutrition and PA knowledge and behaviours, fitness, and BMI z‐score over a 2‐year period beginning in the fall of 2011"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: random number generator [author communication]
Allocation concealment (selection bias) High risk Comment: 1 school knowingly assigned to intervention group
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students not blinded
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: no blinding [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: missing data explained; similar between groups
Selective reporting (reporting bias) Low risk Comment: all outcomes reported in protocol paper
Cluster RCT ‐ Recruitment bias High risk Comment: baseline data collected after randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: differences adjusted for in analyses
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters were lost [author communication]
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "mixed‐effects linear regression models were used ...[with] school as a random effect to account for clustering"

Muros 2015.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: rural environment and similar socioeconomic status
School exclusion criteria:
Student inclusion criteria: 10 to 11 years old
Student exclusion criteria:
Setting: school, rural
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Spain
Interventions Intervention 1: intervention consisted of 60‐minute sessions of PA held twice per week. PA was controlled by means of heart rate monitoring. The aim of the training sessions was to improve aerobic capacity using PAs specifically targeted for health gains such as motor skills, games, and sports. Play was used in all activities to motivate students and achieve the desired level of PA. All games and tasks were designed and developed by a group of experts in education and sports science and were directed by the same supervisor
Intervention 2: PA and nutritional educational interventions combined
Intervention 3: the same intervention as PA and nutritional educational intervention group and replaced the oil that they normally consumed with extra virgin olive oil during the final month of the intervention
Compartor 1: nutritional educational sessions informed participants about the benefits of a Mediterranean diet (high fruit, vegetables, legumes, fish, cereals, and unsaturated‐to‐saturated fat ratio, and low meat, meat products, and dairy products) and lifestyle. Nutritional education involved both parents and students. One or both parents could attend the sessions. It was compulsory for pupils to attend both nutrition sessions held during school time
Comparator 2: continued with their usual activities
Duration of intervention: 6 months
Duration of follow‐up: 6 months
Number of schools: 5
Theoretical framework:  —
Outcomes Fitness
BMI
Study registration
Publication details Language of publication: English
Funding: non‐commercial funding (charitable trust)
Publication status: peer‐reviewed journal
Stated aim for study "The aim of our study was to investigate the effects of 4 experimental conditions and 1 no intervention control group on health‐related parameters, such as the lipid, physiological and anthropometric profiles of children"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: computer‐generated random numbers [author communication]
Allocation concealment (selection bias) Low risk Comment: researchers could not know the intervention assignment [author communication]
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: for all outcomes, investigators were blinded to grouping
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: only 56% of participants have end of study measures; reasons not reported
Selective reporting (reporting bias) Unclear risk Comment: no protocol published
Cluster RCT ‐ Recruitment bias Low risk Comment: students and parents enrolled before randomisation
Cluster RCT ‐ Baseline imbalance High risk Comment: groups imbalanced at baseline; not controlled for
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost (only 1/group)
Cluster RCT ‐ Incorrect analysis High risk Comment: clustering not accounted for in analysis

Suchert 2015.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: schools selected from a complete list of all secondary schools in Schleswig Holstein obtained from the Ministry of Education
School exclusion criteria: schools for disabled students
Student inclusion criteria: all students in participating classes
Student exclusion criteria:
Setting: school
Age group: adolescents
Gender distribution: females and males
Country/Countries where trial was performed: Germany
Interventions Intervention: the intervention operates at 4 levels: individual, class, school, and parents
  • Student: pedometer use, interactive user account on the "läuft" homepage

  • Class: competitions, educational and practical resources

  • School: informational material, introductory seminar for participating teachers

  • Parent: informational material, parent‐teacher conference


Comparator: no further intervention
Duration of intervention: 12 weeks
Duration of follow‐up: 1 year
Number of schools: 29
Theoretical framework: —
Outcomes Fitness
BMI
Study registration ISRCTN49482118
Publication details Language of publication: English 
Funding: non‐commercial funding (charitable trust)
Publication status: peer‐reviewed journal
Stated aim for study "The aim of the study is to evaluate the implementation and effectiveness of the “läuft” PA program among adolescents in grade 8"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "a stratified randomisation was carried out on the school level (according to type of school and number of participating classes) with the computer program Randomization In Treatment Arms"
Allocation concealment (selection bias) Low risk Comment: randomisation conducted using a computer programme
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded to intervention
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: assessors were not blinded [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Quote from publication: "students in the study sample engaged in less out‐of‐school sports activities at baseline than students lost to post‐assessment (P = 0.006). Attrition analyses revealed no further differences. There was no selective attrition between groups"
Selective reporting (reporting bias) High risk Comment: reporting of medical testing listed in protocol missing
Cluster RCT ‐ Recruitment bias High risk Comment: participants recruited after clusters randomised
Cluster RCT ‐ Baseline imbalance High risk Comment: clusters not balanced at baseline
Cluster RCT ‐ Loss of clusters Low risk Comment: there was no dropout on school or class level
Cluster RCT ‐ Incorrect analysis Low risk Comment: analysis accounted for clustered design

Andrade 2014.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: > 90 students in Grades 8 and 9 and located in the urban area of Cuenca, Ecuador
Eligible schools were paired according to 4 criteria
  • Total number of students of the school

  • Monthly school fee (as proxy for socioeconomic status of the school)

  • Gender (male or female only or mixed gender).

  • Time schedule of classes (morning: 7:00 to 13:00, or afternoon: 12:00 to 18:00)


School exclusion criteria: schools with no matching pair
Student inclusion criteria: 2 Grade 8 and 2 Grade 9 classes were randomly selected; all students in those grades were invited to participate
Student exclusion criteria: pregnant adolescents and those with muscle or bone injury or a concomitant disease
Setting: school, urban
Age group: adolescents
Gender distribution: females and males
Country/Countries where trial was performed: Ecuador
Interventions Intervention: ACTIVITAL intervention for PA objectives were to decrease daily screen time (1 hour to 2 hours/d), to increase daily PA levels to reach 60 minutes/d, and to have the school offer more opportunities to be active. Individual‐based strategies included delivery of an educational package organised at the classroom level. Persons in charge of delivering the educational package received an introduction to the intervention objectives and a basic workshop on healthy eating and PA. The PA environmental strategy included workshops with parents that were organised at the same time as classes with adolescents and covered similar topics; organisation of social events at school such as an interactive session with famous young athletes; and environmental modification ‐ a walking trail was drawn on the school playground in the second year of the intervention. There was no minimum dose for activities for each of the intervention strategies
Comparator: standard curriculum as determined by the Ecuadorian government, geared at increasing sport skills and includes and a mandatory 80 minutes of PE/week
Duration of intervention: 28 months
Duration of follow‐up: 28 months
Number of schools: 20
Theoretical framework: social cognitive theory, information‐motivation behavioural skills model, control theory, theory of planned behaviour
Outcomes PA participation
PA duration
Sedentary time
Fitness
BMI
Study registration NCT01004367
Publication details Language of publication: English 
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "We implemented a school‐based health promotion intervention ACTIVITAL that aimed at improving diet and PA. ACTIVITAL was developed using participatory approaches and tailored to the Ecuadorian school context"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "using a random number generation with random allocation of the intervention within each pair"
Allocation concealment (selection bias) Unclear risk Comment: no description of allocation concealment
Blinding of participants and personnel (performance bias)
All outcomes Low risk Quote from publication: "adolescents were not informed about the existence of a counterfactual school"
Blinding of outcome assessment (detection bias)
All outcomes Unclear risk Comment: not described
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: loss to follow‐up greater in control school, but data imputed
Quote from publication: "age, BMI z‐score, gender, physical activity knowledge and socioeconomic status were used as predictors in models to impute data in the pairs"
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Comment: loss to follow‐up greater in control school, but data imputed
Quote from publication: "age, BMI z‐score, gender, physical activity knowledge and socioeconomic status were used as predictors in models to impute data in the pairs"
Selective reporting (reporting bias) Low risk Comment: all published outcomes reported
Cluster RCT ‐ Recruitment bias High risk Comment: students enrolled after randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: adjusted for in models
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Comment: mixed models used to account for effect of clustering

Jago 2014.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: primary schools in the Greater Bristol area
School exclusion criteria:
Student inclusion criteria: all Year 5 and 6 children who are physically able to engage in PE classes
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: UK
Interventions Intervention: Action 3:30 school PA sessions. The focus of the sessions was to promote children’s perceptions of autonomy, belonging, and competence. Amongst a range of techniques, to promote autonomy, teaching assistants were encouraged to provide children with choices within activities, such as leading warm‐ups and adapting games, and there were child‐led sessions in which children chose the activities. Teaching assistants supported competence by setting progressive activities targeting quick successes balanced with providing optimal challenge and providing specific praise for attempts as well as outcomes. Relatedness was supported through empathic teaching assistant‐child interactions, with teaching assistants showing interest in the children’s lives outside the intervention and encouraging teamwork
Comparator: schools provided data at Time 0, Time 1, and Time 2 only; no other contact was made by the research team
Duration of intervention: 20 weeks
Duration of follow‐up: 9 months
Number of schools: 20
Theoretical framework: self‐determination theory
Outcomes PA duration
Study registration ISRCTN58502739 (retrospectively registered)
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The main research question for a future definitive trial is, ‘Is Action 3:30, an after‐school PA intervention that is based on behaviour‐change theory and delivered by teaching assistants, effective in improving the PA, attitudes and confidence of Year 5 and 6 children?’ Specific aims: 1) Estimate the likely recruitment, attendance, and retention rates of pupils to the Action 3:30 after school PA intervention. 2) Estimate the likely impact on PA while the club was still running and 4 months after contact sessions had ended. 3) Develop a reliable costing tool and assess the feasibility of obtaining programme cost data. 4) Estimate the sample size for an adequately powered evaluation of the Action 3:30 intervention"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "randomisation was conducted by an independent statistician in the trials unit with no other involvement in the project"
Allocation concealment (selection bias) Low risk Comment: conducted by clinical trials unit not involved in the study
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students and staff not blinded
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: attempted to blind data collectors, but group allocation was often revealed by students or staff [author communication]
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Comment: small loss to follow‐up; reasons provided
Selective reporting (reporting bias) Low risk Comment: all outcomes in protocol paper reported
Cluster RCT ‐ Recruitment bias Low risk Quote from publication: "schools were randomly assigned to intervention or control arms once baseline data had been processed"
Cluster RCT ‐ Baseline imbalance Low risk Comment: models adjusted for baseline imbalance
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "models [used]... robust standard errors used to take account of the cluster randomised design"

Kipping 2014.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: all state primary and junior schools with children in Years 4 to 6 (aged 8 to 11) in the area covered by Bristol City Council and North Somerset Council
School exclusion criteria: special schools (i.e. those for children whose additional needs cannot be met in a mainstream setting) because they are unlikely to be teaching the standard national curriculum and children may not be able to take part in all measurements
Student inclusion criteria: children in Year 4 (age 8 to 9) at the recruitment stage
Student exclusion criteria:
Setting: school, mix
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: UK
Interventions Intervention: Active for Life Year 5 is a school‐based intervention that aims to increase children’s self‐efficacy and knowledge, together with motivating parents, to increase children’s levels of PA, reduce sedentary behaviour, and increase consumption of fruits and vegetables; a secondary aim is to improve other aspects of healthy activity and diet. The 5 components of the intervention are:
  • training for Year 5 classroom teachers and learning support assistants;

  • provision of 16 lesson plans and teaching materials;

  • provision of 10 parent‐child interactive homework activities;

  • provisions of written information for school newsletters; and

  • written information for parents


Comparator: continued standard education provision for the school year, and any involvement in additional health‐promoting activities, but no access to intervention teacher training and no known access to teaching materials
Duration of intervention: 1 school year
Duration of follow‐up: 1 school year
Number of schools: 60
Theoretical framework: social cognitive theory
Outcomes PA duration
Sedentary time
BMI
Study registration ISRCTN50133740
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The aims of Active for Life Year 5 are to determine the effectiveness and cost‐effectiveness of the intervention in children aged 9 to 10 years to improve the following primary outcomes: 1) Daily time spent in, and amount of, PA. 2) Daily time spent in sedentary behaviour. 3) Portions of fruit and vegetables consumed/d. And secondary outcomes: 1) Time spent screen‐viewing/d. 2) Portions of: snacks, high fat foods, and high energy drinks consumed/d. 3) BMI. 4) Waist circumference. 5) Whether overweight or obese. The aim is to determine whether the intervention affects these outcomes in the short‐term (i.e. immediately at the end of the intervention) and in the longer term (i.e. 12 months after the end of the intervention)"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "schools were grouped and randomly allocated them to control or intervention. 1 author (DAL) who was unaware of any characteristics of the schools did the randomisation (identification numbers were used to relate schools to the 2 stratifying variables, and DAL had no knowledge of which schools these numbers linked to)"
Allocation concealment (selection bias) Low risk Quote from publication: "randomisation was concealed by using the Bristol Randomised Trials Collaboration’s automated (remote) system"
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students not blinded
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: field workers collecting data were blind to allocation
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: similar missing data at baseline and at end of study
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Comment: similar missing data at baseline and at end of study
Selective reporting (reporting bias) Low risk Comment: all outcomes in protocol paper reported
Cluster RCT ‐ Recruitment bias Low risk Comment: randomisation after enrolment
Cluster RCT ‐ Baseline imbalance Low risk Comment: analyses adjusted for baseline values
Cluster RCT ‐ Loss of clusters Low risk Comment: no loss of clusters
Cluster RCT ‐ Incorrect analysis Low risk Comment: multi‐level regression models accounted for clustering within school

Kobel 2014.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: all primary schools of the state of Baden‐Württemberg
School exclusion criteria: teachers who already took part in the programme in the academic year 2009/2010 were not included in the study
Student inclusion criteria: pupils at primary school participating in the Baden‐Württemberg study, Grades 1 and 2
Student exclusion criteria: none
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Germany
Interventions Intervention: a progressive, teacher‐led intervention from Grades 1 to 4 using a spiral curriculum. Each teacher takes part in a 3‐part training course led by a colleague or other teacher (not an external expert) to enhance programme acceptance and facilitate translation into the school environment for sustainability. The intervention consists of 20 units/school year of lessons on “beverages”, “PAs”, and “recreational activities”. These units are spread over the whole academic year. Furthermore, the intervention consists of 2 PA exercises that are performed every school day ("active breaks”, each exercise takes between 5 and 7 minutes). Additionally, “family homework” exercises are given, which are small tasks related to the lesson’s topics to involve parents. Further, samples for parents’ evenings and templates for letters to parents in 3 languages (i.e. German, Turkish, and Russian) are included
Comparator: in the academic year 2010/2011, there was no intervention in the control group; although interested, class teachers belonging to the control group received no local training and no materials for the intervention; they were registered for participation in the academic year 2011/2012. In the academic year 2011/2012, these class teachers started with the 3‐part local training course
Duration of intervention: 1 year
Duration of follow‐up: 1 year
Number of schools: 91
Theoretical framework: social cognitive theory, saluto‐genetic competence, action‐oriented approach
Outcomes PA participation
PA duration
BMI
Study registration DRKS00000494
Publication details Language of publication: English
Funding:
Publication status: peer‐reviewed journal
Stated aim for study "The purpose of this study, therefore, is to investigate the children’s behaviours after a 1‐year intervention in respect of the programme’s key aspects: an increase of PA, a decrease in time spent with screen media as well as more regular breakfast, and a reduction of the consumption of sugar‐sweetened beverages"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: the randomisation list was generated by an independent person in the Institute of Epidemiology and Medical Biometry Ulm University, using a validated system, which involves a pseudo‐random number generator to ensure that the resulting treatment sequence will be both reproducible and non‐predictable [author communication]
Allocation concealment (selection bias) Low risk Comment: allocation was concealed by using an independent statistician
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students and teachers knew group assignment
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: outcome assessors at schools were blinded [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: > 10% loss to follow‐up; no explanation given
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: > 10% loss to follow‐up; no explanation given
Selective reporting (reporting bias) High risk Comment: skin‐fold thickness reported in protocol but not in article
Cluster RCT ‐ Recruitment bias High risk Comment: students enrolled after randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline demographics similar
Cluster RCT ‐ Loss of clusters High risk Comment: N = 6 classes withdrew because of randomisation to control group
Cluster RCT ‐ Incorrect analysis High risk Comment: statistical analysis does not account for clustering

Martinez‐Vizcaino 2014.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: schools in different municipalities in the province of Cuenca, Spain
School exclusion criteria:
Student inclusion criteria: literate in Castilian Spanish, no physical or mental disorder identified by parents or teachers that would prevent student from doing PA, no chronic disease that paediatrician or family doctor considered would prevent student from participating in MOVI‐2, collaboration of a family member who would respond to questionnaires on lifestyle
Student exclusion criteria:
Setting: school, mix
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Spain
Interventions Intervention: MOVI‐2 consisted of an extracurricular play‐based and non‐competitive PA programme. The primary objective of MOVI‐2 was to increase weekly PA while improving health‐related fitness. MOVI‐2 included basic sports games, traditional games, and other outdoor activities such as cycling or gymkhanas. The programme included two 90‐minute PA sessions during weekdays in the evening from 4:00 to 5:30 PM and one 150‐minute session on Saturday morning each week
Comparator: standard PE curriculum (2 hours/week of PA at low to moderate intensity)
Duration of intervention: 8 months
Duration of follow‐up: 8 months
Number of schools: 20
Theoretical framework: socioecological model
Outcomes BMI
Study registration NCT01277224 (retrospectively registered)
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "Our study assessed the impact of a standardized PA program on adiposity and cardio metabolic risk in grades 4 and 5 schoolchildren. The program consisted of noncompetitive recreational activities focused on developing aerobic and muscular fitness"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: computer‐generated procedure
Allocation concealment (selection bias) Low risk Comment: opaque envelopes
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "it was impossible to blind parents, children, and teachers to the existence of the intervention group program"
Blinding of outcome assessment (detection bias)
All outcomes High risk Quote from publication: "although it was not possible to blind the investigators who measured other study variables at baseline and at the conclusion of the study as to trial group allocation, the analysts who processed and analysed the study data were blinded in this respect"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: > 10% loss to follow‐up; reason not provided
Selective reporting (reporting bias) Low risk Comment: all outcomes reported in protocol described
Cluster RCT ‐ Recruitment bias High risk Comment: students recruited after randomisation
Cluster RCT ‐ Baseline imbalance Low risk Quote from publication: "there were no statistically significant differences between intervention and control participants in any baseline characteristics"
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost [author communication]
Cluster RCT ‐ Incorrect analysis Low risk Comment: clustering accounted for in analysis

Nogueira 2014.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: 2 local independent primary schools (Gold Coast, Australia) of essentially identical size and demographic (ethnicity and socioeconomic profile), with comparable school fees, school hours, curricula, and time devoted to PE and other PAs 
School exclusion criteria:
Student inclusion criteria: students who were of sound general health, fully ambulatory, and gave their consent to participate
Student exclusion criteria: students taking medications known to affect bone, muscle, or metabolism; recovering from a limb fracture or other immobilising injury in the past 6 months; affected by any condition not compatible with PA; parents declined to consent
Setting: school
Age group: children
Gender distribution: females and males 
Country/Countries where trial was performed: Australia
Interventions Intervention: instructor‐led exercise bouts comprising 10 minutes of continuous high‐intensity movement intended to improve musculoskeletal and metabolic health. Programme was largely based on capoeira, a Brazilian sport that combines martial arts with dance, and a broad range of continuous movements of medium to high impact, applied at varying speeds and directions to increase heart rate and to load a variety of muscle groups and skeletal regions in upper and lower body. Occasional small prizes (e.g. balls, game vouchers) were provided to reward participation and improvement
Comparator: control school children continued to undertake usual school activities over the course of the intervention year
Duration of intervention: 9 months
Duration of follow‐up: 21 months
Number of schools: 2
Theoretical framework: —
Outcomes BMI
Fitness
Study registration
Publication details Language of publication: English
Funding:
Publication status: peer‐reviewed journal
Stated aim for study "The aim of the current study then, was to test the efficacy of a brief, novel, and enjoyable bone‐ and fat‐targeted exercise program on parameters of bone, muscle and fat in healthy pre‐and peri‐pubertal boys over the course of a school year. The aim of the CAPO Kids intervention trial then was to determine the effect of a brief, simple, enjoyable, musculoskeletal‐ and fat‐targeted exercise programme on quantitative‐ultrasound‐derived bone quality, fat and metabolic health in pre‐ and early‐pubertal girls over the course of a school year"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: coin toss used
Allocation concealment (selection bias) Low risk Comment: group allocation concealed from participants and investigators prior to randomisation [author communication]
Blinding of participants and personnel (performance bias)
All outcomes Low risk Quote from publication: "control school participants were aware of neither the intervention activity nor the overall purpose of the study"
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: testers were not blinded to intervention
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: overall loss to follow‐up was 9% and was related mainly to student relocation or absence from school on the days of testing
Selective reporting (reporting bias) Unclear risk Comment: no protocol published
Cluster RCT ‐ Recruitment bias Low risk Comment: students enrolled after randomisation, but almost complete enrolment
Cluster RCT ‐ Baseline imbalance Low risk Comment: adjusted for in analysis
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis High risk Comment: clustering not accounted for in analysis

Santos 2014.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: within the provincial jurisdiction, minimum of 200 students/school, and offering Grades 1 through 6
School exclusion criteria:
Student inclusion criteria:
Student exclusion criteria: consent not received or condition that limited participation in PA
Setting: school, mix
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Canada
Interventions Intervention: an older class was paired with a younger class. Each week, older students received a 45‐minute healthy living lesson from their classroom teacher. Later that week, older students acted as peer mentors, teaching a 30‐minute lesson to their younger “buddies”. The “Go Move!” aspect included two 30‐minute structured aerobic fitness sessions/week, called fitness loops, with student pairs. Students were encouraged to complete the fitness loops at a vigorous intensity using perceived exertion scales. The “Go Fuel!” component included lessons about distinguishing nutritious from unhealthy (nutrient poor, energy‐rich) foods and beverages. As part of the “Go Feel Good!” component, students were taught to value themselves and classmates based on individual traits rather than on peer influence. The peer‐led model facilitated social skills, self‐esteem, and social responsibility
Comparator: standard curriculum
Duration of intervention: 7 months
Duration of follow‐up: 7 months
Number of schools: 20
Theoretical framework: —
Outcomes BMI
Study registration NCT01979978 (retrospectively registered)
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The present study was designed to overcome these limitations by using a cluster‐randomised effectiveness trial to test the hypothesis that a school‐based, peer‐led healthy living program would reduce adiposity and increase PA among children"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "schools were randomised in a computer‐generated random selection process and blocked to ensure equal representation from rural and First Nations (i.e. Indigenous) schools in both intervention and control arms"
Allocation concealment (selection bias) Unclear risk Comment: randomisation completed by investigator not involved in data collection; concealment unclear
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students and teachers not blinded
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: research assistants blinded to study assignment
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: low loss to follow‐up; no differences in outcome measures were noted between completers and non‐completers
Selective reporting (reporting bias) Low risk Comment: outcomes match clinical trial registry
Cluster RCT ‐ Recruitment bias High risk Comment: students enrolled after randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline differences controlled for
Cluster RCT ‐ Loss of clusters High risk Comment: 1 intervention school withdrew from the study
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "all analyses were adjusted for random effect of student and within‐school clustering using a compound symmetry structure for the within‐student correlations"

Toftager 2014.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: all municipalities in the Region of Southern Denmark were invited; 5 municipalities (Esbjerg, Nordfyn, Varde, Vejle, and Sønderborg) accepted the invitation and were asked to enrol public schools that contained Grade 8
School exclusion criteria: schools that were placed in the countryside and had more than 50% of all students living farther than 2 km Euclidian distance from the school, had a majority of students that were non‐native Danish
Student inclusion criteria:
Student exclusion criteria:
Setting: school, mix
Age group: children/adolescents
Gender distribution: females and males
Country/Countries where trial was performed: Denmark
Interventions Intervention: intervention consisted of 11 intervention components changing the physical and organisational environment of schools. The multi‐component intervention was developed according to socioecological models of behavioural change and was constructed in accordance with existing knowledge‐based research and practical experiences from Danish school settings. A detailed written description of intervention components was delivered to all participating schools and included 4 physical environment changes and 7 organisational environment changes. Required physical environment changes included the following components:
  • upgrade existing outdoor areas at the school for PA, including unfixed equipment;

  • develop and build playgrounds specially designed for adolescents: play spots;

  • improve safety for active transport to and from school; and

  • establish an after‐school fitness programme


Organisational environment changes included:
  • formulate and implement school PA policy;

  • educate teachers as ‘‘kick‐starters’’, who facilitate and motivate PA during recess;

  • establish school play patrol: older students trained to initiate play and games for minors during school recess;

  • mandatory outdoor recess and/or free access to gym or sports hall;

  • school traffic patrol: older students helping minors cross the streets near the school;

  • educate and train students in safe cycling; and 

  • present school project or theme week once per year focused on learning about and doing PA during school lessons


http://www.forebyggelsescenter.dk
Comparator:
Duration of intervention: 2 years
Duration of follow‐up: 2 years
Number of schools: 14
Theoretical framework: socioecological model of behavioural change
Outcomes PA duration
Sedentary time
Fitness
Study registration ISRCTN79122411
Publication details Language of publication: English
Funding: other funding (funded by a non‐profit organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The aim of the School site, Play Spot, Active transport, Club fitness and Environment Study was to develop, document, and assess a comprehensive intervention in local school districts that promote everyday PA among 11‐ to 15‐year‐old adolescents"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: drew lots [author communication]
Allocation concealment (selection bias) High risk Comment: allocation not concealed [author communication]
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded to intervention group
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: outcome assessors not blinded [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Quote from publication: "out of the participating adolescents at baseline 13% (N = 162) had moved to another school at follow‐up, and 2% (N = 27) withdrew consent"
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Quote from publication: "out of the participating adolescents at baseline 13% (N = 162) had moved to another school at follow‐up, and 2% (N = 27) withdrew consent"
Selective reporting (reporting bias) Low risk Comment: all main outcomes in protocol reported
Cluster RCT ‐ Recruitment bias High risk Comment: individuals recruited and baseline measures taken after randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: no baseline differences between groups
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Comment: clustering of students within schools was accounted for by including schools as a random effect in analyses

Fairclough 2013.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: within pre‐defined geographical units known as Neighbourhood Management Areas, 1 high and 1 low socioeconomic status school (defined as percentage of students per school eligible to receive free school meals) were randomly selected
School exclusion criteria:
Student inclusion criteria: all children within Year 6 (10 to 11 years old)
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: UK
Interventions Intervention: Children’s Health, Activity and Nutrition: Get Educated! curriculum consisted of 20 weekly lesson plans, worksheets, homework tasks, lesson resources, and a CD‐ROM. Lessons provided an opportunity for children to discuss, explore, and understand the meaning and practicalities of PA and nutrition as key elements of healthy lifestyles. The core message of the PA and sedentary behaviour components was “move more, sit less” with no specific prescription given as to what forms of PA the children should do. Nutrition components focused on topics such as energy balance, macronutrients, and eating behaviours. Homework tasks involved the whole family because formative work emphasised the importance of family support
Comparator: classes in the comparison schools received normal instruction. This did not involve a specific unit of Personal, Social, and Health Education focused on healthy eating and PA, but concepts related to these areas may have been touched on informally during other lessons
Duration of intervention: 20 weeks
Duration of follow‐up: 8 months
Number of schools: 12
Theoretical framework: social cognitive theory
Outcomes PA duration
Sedentary time
Fitness
BMI
Study registration ISRCTN03863885 (retrospectively registered)
Publication details Language of publication: English 
Funding:
Publication status: peer‐reviewed journal
Stated aim for study "The Children’s Health, Activity and Nutrition: Get Educated! intervention was designed to promote healthy weight in primary school children through a teacher‐delivered curriculum‐based intervention with family involvement, focused on PA and dietary behaviour. The aim of this pragmatic evaluation was to assess the effectiveness of the Children’s Health, Activity and Nutrition: Get Educated! intervention on measures of body size, PA, and food intake"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: random number generator used
Allocation concealment (selection bias) High risk Comment: not described
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students not blinded
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: outcome assessors not blinded [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: large loss to follow‐up, bigger in intervention group
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: large loss to follow‐up, bigger in intervention group
Selective reporting (reporting bias) Low risk Comment: outcomes match with trial registry
Cluster RCT ‐ Recruitment bias High risk Comment: randomisation conducted prior to baseline measures
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline differences adjusted for in analyses
Cluster RCT ‐ Loss of clusters High risk Comment: 1 school lost from intervention group
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "multilevel models can analyse the hierarchical nature of non‐independent, nested data by taking into account the dependency of observations. Children were defined as the first level unit of analysis, and school was the second level unit of analysis"

Ford 2013.

Study characteristics
Methods Study design: RCT
Participants Student inclusion criteria: aged 5 to 11 years, from 2 primary schools located within the southeast of England
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: UK
Interventions Intervention: accumulated brisk walking programming during school time
Comparator: normal school lessons during walking sessions, which involved seated literacy work
Duration of intervention: 15 weeks
Duration of follow‐up: 15 weeks
Number of schools: 2
Theoretical framework: —
Outcomes PA duration
BMI
Study registration
Publication details Language of publication: English
Funding:
Publication status: peer‐reviewed journal
Stated aim for study "The purpose of this study was to determine whether a 15‐week accumulated brisk walking programme, performed within a primary school setting, is effective in eliciting changes in body composition in 5‐ to 11‐year‐olds"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "participants were divided into 2 groups using the random number generation function in SPSS"
Allocation concealment (selection bias) Unclear risk Comment: not described
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded
Blinding of outcome assessment (detection bias)
All outcomes Unclear risk Comment: not described
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Unclear risk Comment: 22 dropped out; study authors do not indicate which group they were from
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Unclear risk Comment: 22 dropped out; study authors do not indicate which group they were from
Selective reporting (reporting bias) Unclear risk Comment: no protocol published

Grydeland 2013.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: more than 40 pupils in Grade 6 and located in the 3 or 4 largest towns or municipalities in the 7 counties surrounding the county of Oslo
School exclusion criteria:
Student inclusion criteria: all Grade 6 students in 37 included schools
Student exclusion criteria:
Setting: school
Age group: children/adolescents
Gender distribution: females and males
Country/Countries where trial was performed: Norway
Interventions Intervention: collaboration with school principals, teachers, school health services, and parent committees to increase students' PA during school hours and leisure time, and to reduce screen time. Teachers held 1 structured lecture on energy balance; initiated 10‐minute PA breaks during class at least once/week; hung posters in classrooms; launched active commuting campaigns; distributed fact sheets to parents once per month; and implemented a computer‐tailored programme for students. Each school received an “Activity box” with sports equipment and toys to promote PA during recess
Comparator:
Duration of intervention: 20 months
Duration of follow‐up: 32 months
Number of schools: 37
Theoretical framework: social cognitive theory and socioecological framework
Outcomes PA duration
Sedentary time
BMI
Study registration
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The overall goal of the HEalth In Adolescents study was to design, implement, and evaluate a comprehensive, intervention programme to promote healthy weight development among young adolescent schoolchildren (11 to 13 year olds)"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "schools were randomised by simple drawing"
Allocation concealment (selection bias) Low risk Quote from publication: "allocation could not be predetermined"
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "neither participants nor investigators were blinded for condition"
Blinding of outcome assessment (detection bias)
All outcomes High risk Quote from publication: "neither participants nor investigators were blinded for condition"
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: < 10% loss to follow‐up
Incomplete outcome data (attrition bias)
Physical activity and sedentary time Low risk Comment: < 10% loss to follow‐up
Selective reporting (reporting bias) Low risk Comment: all outcomes in protocol paper reported
Cluster RCT ‐ Recruitment bias High risk Comment: schools were randomised prior to baseline data collection [author communication]
Cluster RCT ‐ Baseline imbalance Low risk Comment: no baseline differences
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Comment: clustering accounted for in analysis

Melnyk 2013.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: 11 high schools from 2 school districts in the southwestern USA. The choice of schools was designed to provide diversity across race or ethnicity as well as socioeconomic status
School exclusion criteria:
Student inclusion criteria: teens 14 to 16 years of age enrolled in a health class at 1 of 11 participating high schools, assented to participation, had a custodial parent who consented to the teen's participation in the study and optionally for himself or herself, could speak and read in English, parents could speak and read either Spanish or English
Student exclusion criteria: a medical condition that would prevent participation in the PA component of the programme
Setting: school
Age group: adolescents
Gender distribution: females and males
Country/Countries where trial was performed: USA
Interventions Intervention: goal‐setting to promote engagement in healthy lifestyle behaviours and problem‐solving for typical adolescent challenges; educational content to increase teens' knowledge of how to lead a healthy lifestyle; homework to reinforce skills learned in the classroom; 20 minutes of PA within each of the 15 Creating Opportunities for Personal Empowerment components. Teachers chose types of physical activities, which commonly included movement within the classroom, brisk walking, dodge ball, kickball, obstacle courses, "Tank" (a game suggested by the research team), and basketball
Comparator: Healthy Teens attention control curriculum was intended to promote knowledge of common adolescent health topics and health literacy. Content included pertinent health information for teens
Duration of intervention: 15 weeks
Duration of follow‐up: 10 months
Number of schools: 11
Theoretical framework: cognitive‐behavioural theory
Outcomes BMI
Study registration NCT01704768
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The purpose of this study is to test the short and more long‐term efficacy of the Creating Opportunities for Personal Empowerment Healthy Lifestyles Thinking, Emotions, Exercise, Nutrition intervention, versus an attention control program (Healthy Teens) on the healthy lifestyle behaviours, BMI and BMI percentile, social skills, depressive or anxiety symptoms and academic performance of 779 culturally diverse high school teens enrolled in the southwest region of the USA for the ultimate purpose of preventing overweight or obesity, mental health disorders and poor academic functioning"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: drawing names from a hat
Allocation concealment (selection bias) Low risk Comment: allocation could not be predetermined
Blinding of participants and personnel (performance bias)
All outcomes Low risk Comment: students and teachers were blinded to intervention
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: outcome assessors were blinded [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: large loss to follow‐up; no description of handling missing data
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: large loss to follow‐up, no description of handling missing data
Selective reporting (reporting bias) Low risk Comment: all specified outcomes reported on
Cluster RCT ‐ Recruitment bias Low risk Comment: students blinded to group allocation
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline differences adjusted for
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost to follow‐up [author communication]
Cluster RCT ‐ Incorrect analysis Low risk Comment: clustering was accounted for in analysis as per protocol paper [author communication]

Sacchetti 2013.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria:
School exclusion criteria:
Student inclusion criteria: Grade 3
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Italy
Interventions Intervention: PA consisted of at least 30 minutes of physical exercise/d, divided between schoolyard (vigorous activity) and classroom (moderate activity). Twice weekly, a further 50 minutes of PE was spent in the gym, according to the standard curriculum of PE. On average, then, during school hours, children were engaged for around 45 minutes in specific PA, which was moderate
Comparator: control group followed the standard programme of PE involving 2 lessons of around 50 minutes/week in the gym, taught by the ordinary classroom teacher
Duration of intervention: 2 years
Duration of follow‐up: 2 years
Number of schools: 26
Theoretical framework: —
Outcomes BMI
Study registration
Publication details Language of publication: English 
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The aim of this study was to assess whether a school‐based intervention of PA education was effective in improving physical abilities and influencing physical behavior in a representative group of primary school children. As a second aspect, the possible effect on body weight was considered. In the study, we compared the PA habits, the physical performances, and BMI measurements in an intervention group and a control group of children at baseline (age: 8 to 9 years) and after a 2‐year follow‐up (age: 10 to 11 years)"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: computerised random number generator [author communication]
Allocation concealment (selection bias) Low risk Comment: allocation was concealed from participants and/or study personnel prior to randomisation [author communication]
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students not blinded
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: outcome assessors were not blinded [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Quote from publication: "any loss to follow‐up was due to children who moved to other schools (14.2% and 13.9% respectively in intervention and control groups)"
Selective reporting (reporting bias) Unclear risk Comment: no protocol published
Cluster RCT ‐ Recruitment bias High risk Comment: clusters randomised before individuals recruited
Cluster RCT ‐ Baseline imbalance Low risk Quote from publication: "both in boys and girls, no significant differences were found between the intervention and control groups in age, BMI, and frequency and duration of the practice of extra scholastic sports"
Cluster RCT ‐ Loss of clusters Low risk Comment: no loss of clusters reported
Cluster RCT ‐ Incorrect analysis High risk Comment: clustering not accounted for in analysis

Siegrist 2013.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: primary schools throughout Bavaria, Germany
School exclusion criteria: none
Student inclusion criteria: attendance in Grade 2 or 3 and written consent from parents
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Germany
Interventions Intervention: the focus of the multi‐faceted JuvenTUM intervention was on directly educating and encouraging children, teachers, and parents to live active and healthy lifestyles. Additionally, school environmental settings (e.g. physical environment, organisation of school breaks, playing during school time, sports facilities) were altered to promote more PA. http://www.juventum.med.tum.de/
Comparator: principals were instructed to continue with school activities as usual, without changing policies related to PA or nutrition during the study period
Duration of intervention: 1 year
Duration of follow‐up: 1 year
Number of schools: 8
Theoretical framework: —
Outcomes BMI
Study registration
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The aim of the present project was to evaluate a simple and ubiquitously applicable school‐based educational program to increase PA, fitness, and lifestyle awareness and to improve health obesity measures"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: schools randomised by drawing a lot [author communication]
Allocation concealment (selection bias) Low risk Comment: allocation was concealed, and recruitment was based on willingness to be randomised to either group [author communication]
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: participants not blinded to intervention
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: outcome assessors were not aware of group allocation [author communication]
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: children with missing data were ill or were absent from school or had left school
Selective reporting (reporting bias) Unclear risk Comment: no protocol available or trial registered
Cluster RCT ‐ Recruitment bias High risk Comment: individual students enrolled after randomisation at school level [author communication]
Cluster RCT ‐ Baseline imbalance Low risk Comment: intervention and control schools were comparable with regard to socioeconomic status of the population and recreational environments
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis High risk Comment: analysis did not account for cluster design

Aburto 2011.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: located in the south of Mexico City, classified by the Secretary of Public Education as low socioeconomic status, received benefits from the Federal School Breakfast Program, demonstrated minimum facilities necessary for execution of the intervention, possessed the standard Secretary of Public Education–issued set of sports equipment, enrolled at least 350 students, and consisted of 2 or more classrooms/grades
School exclusion criteria:
Student inclusion criteria: all students in Grades 4 and 5
Student exclusion criteria:
Setting: school, urban
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Mexico
Interventions Intervention 1 ‐ Basic: environmental and policy changes at the school level meant to foster an environment conducive to increased PA. These changes were complemented with an educational campaign to increase students’ and school staff’s awareness of the importance of PA for health
Intervention 2 ‐ Plus: all components of the basic intervention plus additional changes requiring more resources such as an added daily exercise session held before classes began, in which all students participated
Comparator: control group experienced no change to the standard practices of public schools in Mexico City related to PE (once/week) and recess (non‐existent)
Duration of intervention: 6 months
Duration of follow‐up: 6 months
Number of schools: 27
Theoretical framework: reciprocal determinism
Outcomes BMI
Fitness
Study registration
Publication details Language of publication: English
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The objective of this investigation was to test the effect of a school‐based environmental intervention on PA and physical fitness of students attending public primary schools in Mexico City"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "using a statistical program to draw a random sample from the complete list of eligible schools, 27 schools were randomly chosen for inclusion. Using the same program to draw 3 samples from the 27, the 3 intervention groups were randomly generated"
Allocation concealment (selection bias) Unclear risk Comment: not described
Blinding of participants and personnel (performance bias)
All outcomes High risk Quote from publication: "the nature of the environmental intervention precluded blinding of the participants or field staff"
Blinding of outcome assessment (detection bias)
All outcomes High risk Quote from publication: "data analysts and researchers were blinded to the meaning of all numeric codes until data analyses were complete"
Comment: outcome assessors not blinded
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: low loss to follow‐up for outcome measures
Selective reporting (reporting bias) High risk Comment: anthropometry not reported; only P value for fitness
Cluster RCT ‐ Recruitment bias High risk Comment: data collected after randomisation
Cluster RCT ‐ Baseline imbalance Low risk Comment: accounted for in analysis
Cluster RCT ‐ Loss of clusters High risk Comment: 3 clusters from control group lost to follow up due to time constraints
Cluster RCT ‐ Incorrect analysis Low risk Quote from publication: "robust standard error accounting for the design effect of school were calculated"

Ardoy 2011.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria:
School exclusion criteria:
Student inclusion criteria: age 12 to 14, enrolled in first year of secondary school
Student exclusion criteria: partial injury or illness or chronic disease that prevented involvement in PE classes
Setting: school
Age group: adolescents
Gender distribution: females and males
Country/Countries where trial was performed: Spain
Interventions Intervention 1: experimental group 1 doubled the academic load stipulated for this subject (4 sessions/week). Experimental group 1 was taught in the same sessions as the control group, doubling the volume of booster sessions with the same objectives, content, and methods. For practical matters and questions of viability, extra sessions were carried out during the afternoon, under the same conditions, and in the same facilities as the usual sessions (held during the morning). Sessions were carried out in accordance with the established curriculum, with approval of the educational institution and the parents
Intervention 2: experimental group 2 also received 4 sessions/week in which there was special emphasis on increasing the intensity of sessions. Experimental group 2 objectives and content were the same, but with high intensity across all sessions. For practical matters and questions of viability, extra sessions were carried out during the afternoon, under the same conditions, and in the same facilities as the usual sessions (held during the morning). Sessions were carried out in accordance with the established curriculum, with approval of the educational institution and the parents
Comparator: control group received 2 sessions of PE/week as established by regulations currently in force in Spain. For all groups, content included the same teaching units (subjects) and sessions on physical fitness and health, games and sports, personal driving qualities, movement, and environmental activities. This intervention was designed to adhere closely to characteristics and context of educational content to obtain results of great application and transfer to national education policies
Duration of intervention: 16 weeks
Duration of follow‐up: 16 weeks
Number of schools:
Theoretical framework: social cognitive theory, information‐motivation behavioural skills model, control theory, theory of planned behaviour
Outcomes BMI
Fitness
Study registration NCT01098968 (retrospectively registered)
Publication details Language of publication: Spanish
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The purpose of the present study is to examine the effect on physical fitness and body composition of: a) doubling the number of sessions of PE/week; b) doubling the number of sessions of PE/week plus increasing their intensity; and c) increasing the intensity of the sessions, while maintaining the same number/week"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Unclear risk Comment: not described
Allocation concealment (selection bias) Unclear risk Comment: not described
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students could not be blinded
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: randomisation was blinded for those who performed the outcome assessment
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: minimal missing data and very few dropouts in each group
Selective reporting (reporting bias) Low risk Comment: all outcomes specified in protocol paper reported on
Cluster RCT ‐ Recruitment bias Unclear risk Comment: unclear when baseline data were collected
Cluster RCT ‐ Baseline imbalance Low risk Comment: secondary analysis adjusted for age
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis High risk Comment: clustering within classes not accounted for in analysis.

de Heer 2011.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: schools in El Paso, Texas, were selected according to school location (for logistical purposes, half of those chosen were located within 5 miles of the University of Texas at El Paso campus), size, socioeconomic status, and percentage of children with limited English proficiency
School exclusion criteria:
Student inclusion criteria: enrolled in 1 of the target grades (3 to 5) and had no condition that would endanger their own or others’ safety
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: USA
Interventions Intervention: a 20‐ to 30‐minute health education component followed by 45 to 60 minutes of PA after school in the schoolyard or in the multi‐purpose room using a bilingual health education curriculum, Bienestar (well‐being), that is, culturally targeted to Mexican Americans. The curriculum included modules on healthy eating, exercise, diabetes, and self‐esteem. Activities for the programme emphasised cardiovascular activity and aerobic recreational games
Comparator 1: students received Grade 4 health workbooks and incentives at pretest and follow‐up measurements, but they did not attend after‐school sessions
Comparator 2 (spillover control group): students were in an intervention classroom but declined an invitation to participate in the after‐school programme. Students received Grade 4 health workbooks and incentives at pretest and follow‐up measurements, but they did not attend after‐school sessions
Duration of intervention: 12 weeks
Duration of follow‐up: 4 months
Number of schools: 6
Theoretical framework: social cognitive theory
Outcomes BMI
Fitness
Study registration
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "We developed, implemented, and evaluated a culturally tailored health education and PA after‐school program for a population of predominantly Hispanic elementary school children"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: random number generator in excel [author communication]
Allocation concealment (selection bias) Low risk Comment: enrolment and baseline assessment prior to generation of randomisation sequence [author communication]
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: students could not be blinded
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: project staff collected these measurements in collaboration with each school’s PE teachers during PE classes (not blinded)
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: differential loss to follow up; did not present imputed analysis
Selective reporting (reporting bias) Unclear risk Comment: no protocol paper; appears all outcomes are reported
Cluster RCT ‐ Recruitment bias Low risk Comment: enrolment and baseline assessment prior to generation of randomisation sequence [author communication]
Cluster RCT ‐ Baseline imbalance Low risk Comment: models adjusted for baseline imbalance
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost [author communication]
Cluster RCT ‐ Incorrect analysis Low risk Comment: analysis accounted for clustering of students within classrooms and classrooms within schools

Jago 2011.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: student body at least 50% minority (African American, Hispanic or Latino, and/or American Indian) and/or greater than 50% eligible for free or reduced lunch; annual student attrition from all causes is ≤ 25%; expected cohort size at end of study is at least 50 per school; school authorities willing to accept randomisation, permit grade‐wide data collection, and assist with mass mailings to students' homes; possess Federal Wide Assurance to conduct research and agree to adhere to the protocol; schools must also have at least 1 play area that satisfied intervention requirements
School exclusion criteria:
Student inclusion criteria: enrolled in Grade 6 in fall 2006, able to participate in school’s standard PE programme, not previously diagnosed with diabetes, providing parent's or guardian’s informed consent and minor child informed assent to participate in data collection and evaluation procedures
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: USA
Interventions Intervention: intervention had 4 integrated components. The first component was change in the total school food environment. The second component was a programme of peer‐led, teacher‐facilitated learning activities. The third component was a social marketing campaign that had a different theme for each semester of the intervention. The fourth element was a revised, more active PE curriculum. Schools also received around $10,000 of equipment and a teacher assistant to facilitate small‐group activities that were intended to increase activity time during sessions
Comparator: control group activities were limited to recruitment and data collection only
Duration of intervention: 2.5 years
Duration of follow‐up: 2.5 years
Number of schools: 42
Theoretical framework: developmental learning frameworks
Outcomes Fitness
BMI
Study registration NCT00458029 (retrospectively registered)
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "HEALTHY was a primary prevention trial with a public health objective of preventing the development of risk factors for type 2 diabetes in adolescents"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Comment: coordinating centre developed randomisation scheme
Allocation concealment (selection bias) Low risk Comment: randomisation occurred at a coordinating centre
Blinding of participants and personnel (performance bias)
All outcomes Low risk Comment: study took measures to mask intervention to both students and parents
Blinding of outcome assessment (detection bias)
All outcomes Low risk Comment: separate study staff not aware of treatment assignment administered data collection protocols
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: loss to follow‐up similar between groups; most often due to transfer to another school
Selective reporting (reporting bias) Low risk Comment: all outcomes reported on
Cluster RCT ‐ Recruitment bias Low risk Comment: students enrolled after randomisation, but investigators attempted to mask students and parents to school assignment
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline balance; models adjusted for confounders
Cluster RCT ‐ Loss of clusters Low risk Comment: no clusters lost
Cluster RCT ‐ Incorrect analysis Low risk Comment: analyses accounted for clustered design

Jansen 2011.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: schools in more deprived inner‐city areas with high proportions of immigrant children in the city of Rotterdam
School exclusion criteria: schools that could not be paired based on size, proportion of migrants, and neighbourhood
Student inclusion criteria: Grades 3 to 8
Student exclusion criteria:
Setting: school, urban
Age group: children/adolescents
Gender distribution: females and males
Country/Countries where trial was performed: The Netherlands
Interventions Intervention: 4 components: (1) implementation of 3 PE sessions/week by a PE teacher; (2) organisation of additional voluntary sport and play activities outside school hours; (3) classroom education with 3 main lessons on healthy nutrition, active living, and healthy lifestyle choices adapted for each grade. Each lesson finishes with joint goal‐setting, and individual counselling by the school nurse is offered if needed; (4) health promotion gathering at the beginning of the school year for parents and involvement of local sport clubs
Comparator: continued with usual curriculum: 2 PE sessions/week by classroom teacher or a PE teacher, dependent on school policy
Duration of intervention: 1 school year
Duration of follow‐up: 1 school year
Number of schools: 20
Theoretical framework: theory of planned behaviour and ecological model
Outcomes BMI
Fitness
Study registration ISRCTN84383524 (retrospectively registered)
Publication details Language of publication: English
Funding:
Publication status: peer‐reviewed journal
Stated aim for study "The purpose of this study was to evaluate the effect of a school‐based intervention program to reduce overweight and improve fitness in primary school children"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Low risk Quote from publication: "randomisation took place within each pair with the toss of a coin by an officer of the municipal education service"
Allocation concealment (selection bias) Low risk Comment: randomisation was done with a coin toss; participants could not foresee assignment
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: blinding was not feasible based on nature of intervention
Blinding of outcome assessment (detection bias)
All outcomes High risk Comment: blinding was not feasible due to the presence of a PE teacher during PE class in intervention schools
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness Low risk Comment: very low dropout; missing data not described but were imputed in analysis
Selective reporting (reporting bias) Low risk Comment: all specified outcomes reported
Cluster RCT ‐ Recruitment bias High risk Comment: schools were randomised prior to baseline data collection [author communication]
Cluster RCT ‐ Baseline imbalance Low risk Comment: baseline differences adjusted for
Cluster RCT ‐ Loss of clusters High risk Comment: 6 clusters lost after randomisation
Cluster RCT ‐ Incorrect analysis Low risk Comment: multi‐level analyses were used to allow for clustering of observations within schools

Magnusson 2011.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: 3 pairs of schools in the city of Reykjavik were selected and matched for size (i.e. number of students and total number of grades, with at least 30 students entering Grade 2 in 2006)
School exclusion criteria:
Student inclusion criteria: all children attending Grade 2
Student exclusion criteria:
Setting: school
Age group: children
Gender distribution: females and males
Country/Countries where trial was performed: Iceland
Interventions Intervention: students had opportunities to engage in PA during PE lessons, during recess, and during classes where PA was to be integrated into various subjects of the general curriculum. Teachers at intervention schools were provided access to PA equipment intended to be used during regular school lessons. Teaching materials promoting PA, such as books and DVDs on classroom workouts and cooperative activity games, etc., were provided. After the first year of intervention, an additional PE lesson was introduced at the intervention schools. PE teachers at each of the intervention schools carried out this additional lesson, which was specifically tailored to suit all children while maintaining a high level of intensity
Comparator: followed general PA curriculum, compulsory on a national level, consisting of two 40‐minute PE sessions/week, in addition to 2 swimming lessons/week, taught over the course of a 6‐week period any time during the school year
Duration of intervention: 2 years
Duration of follow‐up: 2 years
Number of schools: 6
Theoretical framework: social cognitive theory
Outcomes BMI
Fitness
PA duration
Study registration
Publication details Language of publication: English
Funding: non‐commercial funding (research funding body)
Publication status: peer‐reviewed journal
Stated aim for study "The objectives of this study were to compare changes in volume and intensity of PA among the group of intervention children to PA levels of children who only received general curriculum‐based PA (controls) and further, to assess whether the intervention effect on PA was modified by gender or BMI"
Notes  
Risk of bias
Bias Authors' judgement Support for judgement
Random sequence generation (selection bias) Unclear risk Comment: not described
Allocation concealment (selection bias) Unclear risk Comment: not described
Blinding of participants and personnel (performance bias)
All outcomes High risk Comment: blinding not possible
Blinding of outcome assessment (detection bias)
All outcomes Unclear risk Comment: not described
Incomplete outcome data (attrition bias)
Anthropometrics, Fitness High risk Comment: large quantity of missing data; reasons not provided
Incomplete outcome data (attrition bias)
Physical activity and sedentary time High risk Comment: large quantity of missing data; reasons not provided
Selective reporting (reporting bias) Unclear risk Comment: protocol not published
Cluster RCT ‐ Recruitment bias High risk Comment: schools recruited and randomised 8 months before baseline measurements
Cluster RCT ‐ Baseline imbalance Unclear risk Comment: adjusted for BMI; many others not considered (e.g. only half of participants had data on socioeconomic status)
Cluster RCT ‐ Loss of clusters Unclear risk Comment: not described
Cluster RCT ‐ Incorrect analysis Low risk Comment: clustering accounted for in analysis

Okely 2011.

Study characteristics
Methods Study design: cluster‐RCT
Participants School inclusion criteria: secondary schools in New South Wales that submitted an expression of interest and completed a profile used to pair‐match schools
School exclusion criteria:
Student inclusion criteria: girls, formally enrolled in Grade 8 within participating schools, provided written consent from themselves and their parent(s) or guardian(s)
Student exclusion criteria:
Setting: school
Age group: adolescents
Gender distribution: females 
Country/Countries where trial was performed: Australia
Interventions Intervention: schools developed and implemented unique 18‐month action plans with a member of the research team. Intervention strategies were designed to prevent a decline in participation in MVPA levels among girls over the course of the intervention. Each school followed an identical process in developing the intervention, which involved developing an action learning team and the school‐specific action plan. The action plan addressed formal curriculum, school environment, and links with the community. During intervention, schools participated in monthly meetings with a member of the research team to share their progress towards study outcomes
Comparator: continuation of their usual programmes
Duration of intervention: 18 months
Duration of follow‐up: 18 months
Number of schools: 25
Theoretical framework: health‐promoting schools framework
Outcomes PA duration
Sedentary time
Study registration ACTRN12610001077055
Publication details Language of publication: English 
Funding: non‐commercial funding (governmental organisation)
Publication status: peer‐reviewed journal
Stated aim for study "The primary aim of the study was to test if an 18‐month school‐based intervention targeting school sport and PE (through the formal curriculum), school ethos (including policies and school breaks such as lunchtime), and links with the local community, could prevent the decline in objectively meas