Schwartz 2016.

Methods	Design: CBA Timing: Retrospective Allocation to group: The study used a stepped‐wedge design in which schools received water jets (i.e. the intervention) at different time points during a 5‐year period, and crossed over from the CG to the IG at the moment when they received the water jet. The timing of water jet installation in each school depended on factors outside the control of the investigators. Quote: "Water jets were not randomly assigned, but spread across a large number of schools as part of a district initiative. (...) Interviews with district‐ and school‐level personnel suggest the specific timing of water jet adoption by a school reflects the interaction of opportunity, convenience, and happenstance, rather than systematic targeting" Number of clusters or sites: 1227 schools Number of individuals: 1,065,562 Length of intervention: 5 years
Participants	General description of participants: Students at public elementary and middle schools in New York City, US Age: Children and teenagers (the exact age range is not reported) Inclusion criteria: Students attending public elementary and middle schools in New York City participating in the New York FITNESSGRAM survey, an routinely administered annual school health examination. Quote: "New York schools have conducted the New York FITNESSGRAM annually for all students in kindergarten through 12th grade since the 2005‐2006 academic year as part of the standards‐based physical education program" Exclusion criteria: Quote: "We excluded charter and special education schools from our sample. We also excluded schools (in a given year) where less than 50% of students participated in the New York FITNESSGRAM (only 1.6% of kindergarten through eighth‐grade student observations) and students with missing New York FITNESSGRAM data (4.1 % of students in schools not dropped owing to low New York FITNESSGRAM coverage)" Recruitment: See inclusion criteria Weight status at baseline: At baseline, 39% of students in the IG and CG were overweight or obese, and 21% of students in the IG and CG were obese SSB consumption at baseline: N/R Equity considerations: Both sexes were equally represented among the participants in the IG, as were a variety of ethnicities (37% black, 37% Hispanic, 14% white, 12% Asian). 11.3% were foreign‐born, 12% had limited English proficiency, 12.1% were on special education, and 85% received free or reduced‐price school lunches (i.e. were from lower‐income families)
Interventions	Intervention: Improved access to drinking water at school (Quote: "In 2009, New York’s Department of Health and Mental Hygiene and Department of Education launched an intervention to increase access to drinking water at lunchtime by placing ‘water jets’ in school cafeterias. Water jets are electrically cooled, large, clear jugs with a push lever for fast dispensing that give students access to clean New York tap water. Water jet machines both chill the water and oxygenate it to keep it tasting fresh and cost approximately $1000 per machine. Plastic disposable cups are also provided by schools for use with the water jet machines") Behavioural co‐intervention: None reported Control: No intervention
Outcomes	Measures of SSB intake: Number of half‐pints of fat‐free chocolate milk purchased per student per year, assessed continuously during the study period with data from an administrative data base maintained by the New York City Department of Education Office of School Food Measures of intake of alternatives to SSB: Number of half‐pints of low‐fat and skim white milk purchased per student per year, assessed continuously during the study period with data from an administrative data base maintained by the New York City Department of Education Office of School Food Anthropometric measures: Mean z BMI, percentage of overweight, and percentage of obese, assessed annually with data from the New York FITNESSGRAM data base. Data were collected by physical education teachers trained in measurements protocols Adverse outcomes: The study does not report how data on adverse outcomes were collected, and if adverse outcomes were observed or not, but notes that a reduction in total milk intake was observed (see above) Other outcomes: None included in this review
Context and implementation	Setting: Schools Sector: Education Country: USA Year(s) when implemented: 2008 ‐ 2013 Mode of implementation: Government regulation and public investment Level of implementation: Policy‐level intervention
Declarations	COI: "Conflict of Interest Disclosures: None reported" Funding: "This project was supported by award 1R01HD070739 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development. (...) The sponsor played no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication" Trial registration: N/R Protocol availability: N/R
Notes	None
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	This was a CBA, and allocation was not random and not concealed
Allocation concealment (selection bias)	High risk	This was a CBA, and allocation was not random and not concealed
Similarity of baseline outcome measurements (selection bias)	Low risk	Quote: "In the 2007‐2008 academic year, water jet schools had similar percentages of overweight students (38.9% vs 39.2%) and obese students (21.1% vs 21.4%) than non–water jet schools" (baseline differences for milk consumption are not reported)
Similarity of other baseline characteristics (selection and performance bias)	Low risk	Quote: "Overall, schools with water jets (n = 374) were not systematically different than those without (n = 482) on baseline characteristics. (...) However, water jet schools did have different percentages of Asian, special education, and foreign born students than non–water jets schools (P < .05). Differences were small in magnitude"
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Body weight outcome observations were excluded for schools which had a FITNESSGRAM coverage of less than 50% in the respective year, i.e. which assessed the body weight of fewer than 50% of their students. The study does not report if average FITNESSGRAM coverage differed between IG and CG
Blinding (performance and detection bias) Objective outcomes	Low risk	Outcomes were objective
Contamination (performance bias)	Low risk	Allocation was by institution (school), and it seems unlikely that a substantial number of schools in the CG installed water jets independently from the programme run by the New York Department of Education (i.e. before the scheduled date, and before they crossed over to the IG)
Selective reporting (reporting bias)	Low risk	All outcomes mentioned in the Methods section are reported, and the choice of outcomes is well‐justified, making it unlikely that outcomes with positive results were selected, and those with negative results dropped from the analysis in a post hoc manner
Other bias	Unclear risk	This study used a complex quasi‐experimental stepped‐wedge cross‐over design. Results may have been biased by underlying temporal trends