Perry 2004.

Study characteristics
Methods	Trial name: Cafeteria Power Plus project Study design: cluster‐RCT Intervention duration: during 2 consecutive school years beginning in autumn 2000 Length of follow‐up from baseline: 2 years Differences in baseline characteristics: not reported Unit of allocation: schools Unit of analysis: canteen observations
Participants	School type: elementary schools Region: twin cities in metropolitan area of Minnesota, USA Demographic/socioeconomic characteristics: 26 schools had an enrolment that was 90% white and 21% of the school meals served were free or reduced price. Inclusion/exclusion criteria: not reported Number of schools allocated: 26 Numbers by trial group n (controls baseline) = 13 n (controls follow‐up) = 13 n (interventions baseline) = 13 n (interventions follow‐up) = 13 Recruitment: not reported Recruitment rate Schools: not reported Children: 91.7%. Of the 1820 students who were eligible to be observed in spring 2000, 1668 were observed and became the baseline sample. Of the 1820 students, 41 had moved, 7 parents and 44 students refused participation, and 60 students were absent.
Interventions	Number of experimental conditions: 2 (1 intervention, 1 control) Policies, practices or programmes targeted by the intervention ‐ Increasing the availability, appeal and encouragement of fruits and vegetables in the school lunch programme; emphasising changes in the lunch line; and, secondarily, the school snack cart. ‐ Increase the quality and quantity of fruits and vegetables served. ‐ Increase the choices of fruits and vegetables in the lunch line, to make them look more attractive (by putting them in small cups or arranging by colour), and to vary the type and preparation methods daily. ‐ Special events to promote fruits and vegetables. Implementation strategies EPOC: educational meetings ‐ Monthly meetings were held with the cook managers from each of the 13 intervention schools to discuss and share implementation issues and new ideas during the first school year. EPOC: educational outreach visits ‐ Intervention staff visited schools weekly, on average, and supported the activities for the kick‐off. EPOC: educational materials ‐ The "High 5 Flyers" that were hung in posters around the school cafeteria. EPOC: other ‐ Special events: sampling of fruit and vegetables, class challenges (to eat 3 serves of fruit and vegetables per day at lunch). EPOC: local consensus processes ‐ Monthly meetings were held with the cook managers from each of the 13 intervention schools to discuss and share implementation issues and new ideas during the first school year. Theoretical underpinning: Social Cognitive Theory Description of control: received training and materials at the end of the active study phase in autumn 2002.
Outcomes	Outcome relating to the implementation of school service policies, practices or programmes ‐ Verbal encouragement by food staff (mean % of observations) ‐ Number of fruits and vegetables on the snack cart (mean) ‐ Number of fruits and vegetables students can choose (mean) ‐ Fruit and vegetables rated as appealing (mean %) Data collection method: process measures for the study, collected in both the intervention and control schools, included direct observations of the canteen, lunch line, food cart and food service staff behaviour. Validity of measures used: not reported; however, the measure was objective. Outcome relating to cost: not reported Outcome relating to adverse consequences: not reported Outcome relating to child diet, PA or weight status: fruit and vegetable intake. Data collection method: trained observers watched the selected students from a distance in the cafeteria and recorded all items eaten at lunch and their portion. Validity of measures used: reported to be valid.
Notes	Research funding: grant from the National Cancer Institute (R01 CA59805). Conflicts of interest: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Schools were randomly allocated to experimental group. Random sequence generation procedure was not described.
Allocation concealment (selection bias)	Unclear risk	No information provided about allocation concealment and, therefore, it was unclear if allocation was concealed.
Blinding of participants and personnel (performance bias) Implementation outcome	High risk	Outcome group: observations: there was no blinding to group allocation of participants or personnel described and this was likely to influence performance.
Blinding of outcome assessment (detection bias) Implementation outcome	High risk	Outcome group: there was no mention that observers were blinded to group allocation and, therefore, the risk of detection bias was high.
Incomplete outcome data (attrition bias) Implementation outcome	Low risk	Outcome group: all 26 schools were retained in the study.
Selective reporting (reporting bias)	Unclear risk	No study protocol, therefore, it was unclear if there was selective outcome reporting.
Other bias	Unclear risk	Unclear if at risk of contamination. Did not appear at risk of other biases.
Recruitment to cluster	Low risk	Direct observations of school environment and food service staff.
Baseline imbalance	Low risk	There were no significant differences at baseline from the lunch observations for all the main outcome measures.
Loss of cluster	Low risk	All 26 schools were retained in the study.
Incorrect analysis	Unclear risk	Used mixed‐model regression procedures; however, it was unclear if adjusted for clustering.
Compatibility with individually randomised RCTs	Unclear risk	Unable to determine if a herd effect existed.
Overall risk of bias assessment	Unclear risk	Most domains were at unclear risk of bias.