Hager 2018.

Study characteristics
Methods	Trial name: Wellness Champions for Change (WCC) Study design: randomised, controlled school‐level pilot study, i.e. cluster‐RCT Intervention duration: 1 year Length of follow‐up from baseline: 1 year Differences in baseline characteristics: no significant differences in school demographics by intervention group. The schools reported having wellness teams was not different between groups. Unit of allocation: clusters – schools Unit of analysis ‐ Implementation outcomes: schools ‐ Behavioural/health outcomes: –
Participants	School type: elementary, middle or high schools Region: Maryland school districts (mid‐Atlantic state), USA Demographics/socioeconomic characteristics ‐ Most schools (79.4%) had a FARMS rate > 40% ‐ Schools were located in a range of geographic locations, with nearly a third each in rural/town, suburban and urban areas. Inclusion/exclusion criteria Inclusion:schools were nominated based on 3 guidelines: ‐ A specified number of schools (range 6–15 schools/district, depending on the district size (the number of schools per district ranged from > 100 schools to < 10 schools)) ‐ Preference was given to schools with higher FARMS eligibility rates (> 40%) (per funder request, not a mandatory inclusion criterion) ‐ Only standard schools (e.g. no part‐day high schools or alternative schools) Exclusion:none. Number of schools allocated: Schools:63 schools (20 – WCC training + TA; 23 – WCC training; 20 – wait control) Students:– Numbers by trial group n (controls baseline) = 20 schools n (controls follow‐up) = 17 schools n (interventions WWC training plus TA baseline) = 20 schools n (interventions follow‐up WWC + TA) = 19 schools n (interventions WWC training baseline) = 23 schools n (interventions follow‐up WWC) = 19 schools Recruitment Districts:6 districts approached based on results from 2012–2013 state‐wide survey, indicating that the districts had a low proportion of schools (< 40%) with wellness teams. After a meeting with district leaders to explain the study, the leaders nominated schools within their district to participate. Schools:principals of nominated schools were emailed by the study team and asked to participate. Students:– Recruitment rate Districts: 83.3% Schools:100% Students:–
Interventions	Number of experimental conditions: 3 randomised within district to: ‐ WCC training (6‐hour, single‐day teacher training) ‐ WCC training plus TA ‐ delayed training (control) Policies, practices or programmes targeted by the intervention The formation of a wellness team, whether wellness teams were active, and whether implementation of local wellness policy components occurred. Implementation strategies EPOC: tailored interventions ‐ Wellness teams tailored for individual school needs. EPOC: educational meetings ‐ The WCC training took place in August–September, before or at the beginning of the school year. In addition to providing TA, the wellness specialists led the WCC trainings. Wellness specialists attended a full‐day training on how to lead the WCC training session how to complete the School Health Index via the Alliance for a Healthier Generation website and TA provision. EPOC: external funding ‐ All participating districts were offered financial resources (about USD1000/school), with guidance that the funds could be used to purchase items to support Smarter Lunchroom initiatives, purchase supplies that supported wellness activities, pay for substitute time for trainings, pay a wellness champion stipend, or a combination of these. EPOC: educational materials ‐ Including a tool kit with detailed training notes and background readings. EPOC: local opinion leaders ‐ Schools nominated an individual to be the school wellness champion Theoretical underpinning ‐ Of the evidence‐based intervention/policy/practice or programme: the WCC training was to be developed based on Social Cognitive Theory. A single‐day training was developed for wellness team leaders based in both Social Cognitive Theory and Social Ecological Theory. Specifically, for Social Cognitive Theory, reciprocal determinism was a major thread throughout the training. ‐ Of the implementation strategy: formative research findings indicated the importance of 1. forming a wellness team; 2. having buy‐in and support from key stakeholders including teachers, principals, parents and district administrators; 3. offering resources (e.g. a list of healthy snacks or sample letters sent to parents); 4. building partnerships (e.g. partnering with parent, teacher and community groups to provide additional resources, creating school‐level clubs. Description of control: delayed control group that received training the following summer.
Outcomes	Outcome relating to the implementation of school policies, practices or programmes Primary outcome variables: the formation of a wellness team, whether wellness teams were active, and whether implementation of local wellness policy components occurred. Data collection method: surveys were sent via e‐mail to a school administrator or the head of an ongoing wellness team, relying on a single respondent per school. Formation of wellness team: single survey item. The presence of a wellness team was determined by a single item, "My school had a school‐level wellness team responsible for implementing local wellness policies in place during the [2014‐2015 (baseline) or 2015‐2016 (follow‐up)] school year." Also, to determine whether wellness teams were active: an 8‐item active wellness team sum score was generated. Additionally, the occurrence of implementation of local wellness policy components was assessed by the Maryland Wellness Policies and Practices Project School Survey II. Validity of measures used ‐ The researchers measured local wellness policy implementation using a scale was adapted from the original Maryland Wellness Policies and Practices Project School Survey. The original scale (17 items) demonstrated test–retest reliability (Spearman correlation = 0.70; P < 0.001; item‐by‐item percent agreement = 75.6%) and high internal consistency (a = 0.923). Items in version II (29 items) were added based on the updated School Health Index and language in the local wellness policy proposed rule stemming from the HHFKA. The version II local wellness policy implementation scale also demonstrated high internal consistency (a = 0.933). Outcome relating to cost: not reported Outcome relating to adverse consequences: not reported Outcome relating to child diet, PA or weight status: not reported Data collection method: not reported Validity of measures used: not reported
Notes	Research funding: USDA Team Nutrition Training Grant, Centers for Disease Control and Prevention Cooperative Agreement No.2B01OT009025 through the Maryland Department of Health, and the Summer Program in Obesity, Diabetes, and Nutrition Research Training Grant under National Institutes of Health Award No.T35DK095737. Conflicts of interest: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The schools themselves were randomised; however, the school district leaders nominated the schools initially. Additionally, during the intervention 2 schools randomised to the intervention arm were switched to the control and 3 schools randomised to the control were moved to the intervention group.
Allocation concealment (selection bias)	Unclear risk	Unclear, no statement regarding this.
Blinding of participants and personnel (performance bias) Implementation outcome	High risk	Outcome group: all outcomes. The nature of the study precluded blinding.
Blinding of outcome assessment (detection bias) Implementation outcome	High risk	Outcome group: all outcomes were assessed by self‐report. Self‐report is considered high risk as the participant is able to determine intervention outcomes.
Incomplete outcome data (attrition bias) Implementation outcome	Unclear risk	It was unclear from the information provided if attrition occurred (or to what extent) and if it was adjusted for during analysis.
Selective reporting (reporting bias)	Unclear risk	Unclear as no trial registration or protocol available to assess reporting bias.
Other bias	Unclear risk	Unclear if at risk of contamination. Did not appear at risk of other biases.
Recruitment to cluster	Low risk	Recruitment and baseline information collected prior to randomisation.
Baseline imbalance	Low risk	No major baseline imbalances.
Loss of cluster	Low risk	No loss of cluster.
Incorrect analysis	High risk	Clustering of schools within school district was not accounted for in the models because randomisation occurred within districts.
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.