Nathan 2020.

Study characteristics
Methods	Trial name: Pilot of the Physically Active Children in Education (PACE) Study design: 2 × 2 factorial cluster‐RCT Intervention duration: 9 months (January 2017 to September 2017) Length of follow‐up from baseline: students' baseline data were collected between February and March 2017. Follow‐up data collection occurred approximately 9 months postbaseline (October–November 2017). Teacher surveys were conducted at follow‐up only. Differences in baseline characteristics: as shown in Table 2, there were large differences in the proportion of rural vs urban schools intervention vs comparison Unit of allocation: school Unit of analysis Implementation outcomes: cluster (school) Behavioural/health outcomes: individual (student)
Participants	School type: 12 Catholic elementary schools Region: Hunter region of NSW Australia Demographics/socioeconomic characteristics: almost 2% of intervention schools were rural, compared to 46% of comparison schools Inclusion/exclusion criteria Inclusion ‐Catholic schools from the study region Exclusion ‐ Schools participating in another PA intervention ‐ Catered for both elementary and secondary students ‐ Exclusively for children requiring specialist care ‐ Did not use the school communications application ("Skoolbag"), as this was required for the lunchbox treatment group. Number of schools allocated Schools ‐ PA group = 6 ‐ No PA support (groups 2 and 4) = 6 Students ‐ PA group = 1323 ‐ No PA support (groups 2 and 4) = 825 Numbers by trial group n (controls baseline) = 1323 n (controls follow‐up) = 940 n (interventions baseline) = 825 n (interventions follow‐up) = 562 Recruitment Schools:Catholic schools from the study region were stratified by school size (small or large) and placed in a random order and invited to participate. School principals provided written informed consent. Recruitment continued until 12 schools consented to participate. Students:all students (ages 5–12 years) attending participating schools were invited to participate in the data collection component of the trial, which required active parent and child consent. Recruitment rate Schools:60% Students:56.9%
Interventions	Number of experimental conditions: 4: 1. PA support only, 2. lunchbox support only, 3. both PA and lunchbox support, 4. waitlist comparison. Outcome data comparing the PA outcomes of students from schools that received the PA support (groups 1 and 3) to those who did not (groups 2 and 4). Policies, practices or programmes targeted by the intervention PA policy: the NSW Department of Education Sport and Physical Activity Policy, required teachers to implement ≥ 150 minutes (up from 120 minutes) of planned moderate, with some vigorous, PA across the school week for students in kindergarten to grade 10. Schools were also required to develop a local school policy that stated their commitment to the ongoing implementation of the 150‐minute policy. Implementation strategies EPOC: educational materials ‐ In‐school champions received an "intervention manual" and classroom teachers received various printed and electronic instructional materials. EPOC: educational outreach visits or academic detailing ‐ Conduct educational outreach visits. EPOC: local opinion leaders ‐ Nominated ≥ 2 in‐school champions. EPOC: tailored interventions ‐ Support officers provided in‐school champions with support remotely, i.e. via telephone or e‐mail twice per term to support implementing the intervention. ‐ Centralised technical support. EPOC: other ‐ Mandate change. ‐ Schools were offered teacher relief funding to cover in‐school champions’ attendance at the workshop. Theoretical underpinning ‐ Of the evidence‐based intervention/policy/practice or programme: none ‐ Of the implementation strategy: Theoretical Domains Framework Description of control: waitlist comparison and lunchbox support. Comparison schools were asked to continue their usual PA practices. However, during the trial period, teachers from either intervention or comparison group schools were able to access NSW government‐run programmes directed at supporting school promotion of healthy eating and PA, generally.
Outcomes	Outcome relating to the implementation of school policies, practices or programmes Teacher implementation of PA policy across a school week. Data collection method: teacher logbooks were reviewed to determine the mean minute's teachers implemented PA. The daily minutes of PA implemented were then summed for each teacher to obtain the total number of minutes of PA that were delivered across that school week. Validity of measures used: the use of teacher logbooks is frequently used in classroom‐based obesity prevention interventions, with high response rates (i.e. > 80%) and established reliability and validity. Outcome relating to cost: – Outcome relating to adverse consequences: – Outcome relating to child diet, PA or weight status ‐ Student daily activity counts per minute. ‐ Student mean daily MVPA and sedentary behaviour Data collection method: accelerometer data was collected Monday to Friday for the whole school day (i.e. 9 A.M. to 3 P.M.). Counts per minute was calculated by dividing the total accelerometer counts by the minutes of wear time. Accelerometer counts were classified as sedentary, light‐intensity PA, and MVPA using the vertical axis wrist cut points developed by Chandler 2016. Validity of measures used: –
Notes	Research funding: Hunter Medical Research Institute (HMRI), Hunter Children's Research Foundation (HCRF), and Hunter New England Population Health. The authors thank the participants for their involvement in this study. N.K.N. is supported by an NHMRC TRIP Fellowship (APP1132450) and a Hunter New England Clinical Research Fellowship; L.W. is supported by an NHMRC Career Development Fellowship (APP1128348), Heart Foundation Future Leader Fellowship (101175), and a Hunter New England Clinical Research Fellowship; R.L.S. is supported by an NHMRC TRIP Fellowship (APP1150661). None of the funding bodies had a role in the design, data collection, analysis, interpretation of data, and dissemination of findings. All authors declare that they have no financial disclosures. Conflicts of interest: all authors declared that they had no financial disclosures.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Schools were randomly allocated in a 1:1:1:1 ratio to 1 of the 4 arms by an independent investigator using a computerised random number generator.
Allocation concealment (selection bias)	Low risk	Random allocation using independent investigator so unlikely to foresee allocation.
Blinding of participants and personnel (performance bias) Implementation outcome	High risk	Outcome group: implementation. The nature of the trial precluded blinding. Outcome: child PA accelerometer data. Low: objective measure.
Blinding of outcome assessment (detection bias) Implementation outcome	High risk	Teacher log‐book used to assess implementation outcome. Outcome: child PA accelerometer data. Low: objective measure.
Incomplete outcome data (attrition bias) Implementation outcome	Unclear risk	Outcome group: teacher self‐reported implementation. Conducted follow‐ups with teachers only. Did not report if some teacher's logbooks were unavailable or not. Outcome: child PA Unclear: there were some missing from analysis due to invalid wear time, and some students did not wear accelerometer at both time points.
Selective reporting (reporting bias)	High risk	The outcome assessing the mean minutes teachers implemented PA was not prospectively registered a priori in ANZCTR registry. Child PA: low: prospectively registered.
Other bias	Unclear risk	Unclear if at risk of contamination. Appeared free of other biases.
Recruitment to cluster	Unclear risk	Unclear at what time point students were recruited (before or after randomisation of schools).
Baseline imbalance	Low risk	There was imbalance between student baseline socioeconomic location and remoteness classification between the experimental groups. These variables were also controlled for by including them as fixed effects in all PA outcome models.
Loss of cluster	Low risk	No loss of clusters.
Incorrect analysis	Low risk	Clustering was taken into account. Analyses of study outcomes were performed under an ITT framework, with participants (students and teachers) analysed according to the group their school was randomised into. 2 random intercepts were included in the model to account for the clustered design of the trial, and to account for the repeated measurements taken on children.
Compatibility with individually randomised RCTs	Unclear risk	No statement regarding this.
Overall risk of bias assessment	Unclear risk	Most domains were at unclear risk of bias.