Ermetici 2016.

Methods	Design: NRCT Timing: Prospective Allocation to group: "Three schools (262 adolescents) were chosen as the intervention schools because of previous collaboration with educational initiatives of our Institute, not related to nutrition and carried out with different pupils. Three other socio‐demographically matched schools (225 adolescents) were chosen as a comparison group" Number of clusters or sites: 6 schools Number of individuals: 462 Length of intervention: 2 years
Participants	General description of participants: Students in middle schools in Milan, Italy Age: Teenagers (age range 11 to 15 years, median age 12.5 years) Inclusion criteria: N/R Exclusion criteria: Quote: "Exclusion criteria were applied only for individuals with conditions interfering with anthropometric measurements, such as severe malformation" Recruitment: N/R Weight status at baseline: 24% of participants in the IG and CG were overweight at baseline, and 7% were obese SSB consumption at baseline: At baseline, mean SSB consumption (times per week) was 2.6 in the IG and 2.7 in the CG Equity considerations: Quote: "The schools were all in a limited area of about 20 square kilometers with a high degree of urbanization and about 10% of public green areas. (...) Most of the pupils attending the schools were Caucasian, with less than 5% of ethnic minorities including Asians, Arabs, and South Americans. The annual average income of the households was about 35,000 euros (about 38,000 dollars). The parental educational levels included mainly middle‐school and high‐school certificates." 49% of participants were female
Interventions	Intervention: Healthier vending machines (as part of a multicomponent intervention). (Quote: "In the intervention group schools, all traditional vending machines were replaced with machines containing healthy foods and beverages, including fresh fruit and vegetables, dried fruit, fruit juices, smoothies without added sugar, and drinkable yogurt, all from local farms, carefully evaluated by our expert nutritionist team. On the basis of a comparative analysis of nutritional facts, the products selected for the alternative healthy vending machines contained per 100 g an average of 60 less kilocalories, 14 g less added sugars, 0.1 g less salt, 1 g less saturated fats, and 1.4 g higher fiber, compared with the traditional vending machines. An agreement was reached with the vending machine supplier to keep prices as low as possible and to dedicate part of the proceeds to childhood health promotion initiatives. No other food and drink sources were available inside the schools, except for tap water. Students were free to bring food or drinks from home. Educational posters adapted from The Healthy Eating Plate conveying messages promoting healthy diet, water consumption, and daily exercise were posted in schools. The schools were asked to create more opportunities for exercise during breaks. Pupils were allowed to leave their classrooms and walk in the corridors or outdoor play areas, for a total of one additional hour a week of movement") Behavioural co‐intervention: Nutrition education (Quote: "Sixteen health‐promoting group lessons by expert nutritionists over two‐school‐years were included in the curriculum. Students were given an easy textbook developed by our nutritionist team as a support to school lessons and as a tool for involving parents. Automated text messages promoting a healthy diet and daily exercise were sent to the students and their parents three times a week throughout the two school years including school vacations. Text messages were sent close to meal times to encourage constructive debate in each family. Students were given a pedometer (PE320‐BL, Oregon Scientific Italia Srl) to encourage them to be physically active. A re‐usable BPA‐free TRITANtm water bottle was supplied to encourage water consumption") Control: No intervention
Outcomes	Measures of SSB intake: SSB intake (times/week), assessed with a "simple self‐completion questionnaire designed by the Italian National Institute of Health, adapted from a validated international standard questionnaire targeting adolescents" at baseline and 2‐year follow‐up Measures of intake of alternatives to SSB: N/R Anthropometric measures: Assessed in this study but not included in our review due to confounding by non‐beverage specific intervention components Adverse outcomes: The study does not report how data on adverse outcomes were collected, and if adverse outcomes were observed or not Other outcomes: Cost of the intervention per student (no details on how and when this outcome was assessed provided)
Context and implementation	Setting: Schools Sector: Education Country: Italy Year(s) when implemented: 2009 ‐ 2011 Mode of implementation: Pilot trial by researchers Level of implementation: Setting‐based intervention
Declarations	COI: "Disclosure: The authors declare no conflict of interest" Funding: "Funding agencies: This study was partially sustained by IRCCS Policlinico San Donato Ricerca Corrente Fund" Trial registration: N/R Protocol availability: Protocol mentioned in the report but not publicly available
Notes	None
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	This was a NRCT, and allocation was not random and not concealed
Allocation concealment (selection bias)	High risk	This was a NRCT, and allocation was not random and not concealed
Similarity of baseline outcome measurements (selection bias)	Low risk	Baseline outcome measurements were similar, and differences were tested for significance (all were non‐significant)
Similarity of other baseline characteristics (selection and performance bias)	Unclear risk	Study authors report that the IG and CG schools were similar with regard to a number of characteristics at baseline. However, the way schools were allocated to the IG and CG may have lead to unobserved differences. In particular,study authors hypothesise that motivation to participate differed between IG and CG. Quotes: "Although previous collaborations with the schools assigned to the intervention group in the present study were not inherent to nutritional education and were carried out with different students, the possibility of a different motivation between the intervention and control schools cannot be excluded. (...) The characteristics of the schools of the intervention and the control groups were sufficiently homogeneous, on the basis of previous reports of the Italian National Institute of Statistics. The schools were all in a limited area of about 20 square kilometers with a high degree of urbanization and about 10% of public green areas. The access to green areas as well as to fast food outlets and food stores was similar for all the schools"
Incomplete outcome data (attrition bias) All outcomes	Low risk	A per‐protocol analysis was done, but attrition was low (only 5 our of 487 students until year 2)
Blinding (performance and detection bias) Subjective outcomes	High risk	The outcome of interest to us (SSB intake) was self‐reported only, and the students' responses to the questionnaire are likely to have been influenced by the student's knowledge that they were part of the study, in particular due to the fact that an intensive behavioural co‐intervention was performed
Contamination (performance bias)	Unclear risk	The authors discuss contamination, but it is unclear how this might have occurred. Quote: "It was almost impossible to prevent cross‐contamination of the intervention between schools placed in a geographically limited area"
Selective reporting (reporting bias)	Unclear risk	It is not reported if outcomes were prespecified or not, or if the trial was registered or if a protocol was published. No data on a number of potentially relevant outcomes (e.g. prevalence of overweight and obesity at follow‐up, consumption of water and other alternatives to SSB, vending machine sales etc) are reported
Other bias	Low risk	No other concerns