A cluster-randomized controlled trial of an elementary school drinking water access and promotion intervention: Rationale, study design, and protocol

Abstract

Introduction:

Promoting water consumption among children in schools is a promising intervention to reduce sugar-sweetened beverage (SSB) intake and achieve healthful weight. To date, no studies in the United States have examined how a school-based water access and promotion intervention affects students’ beverage and food intake both in and out of school and weight gain over time. The Water First trial is intended to evaluate these interventions.

Methods:

Informed by the PRECEDE-PROCEED model and Social Cognitive Theory, the Water First intervention includes: 1) installation of lead-free water stations in cafeterias, physical activity spaces, and high-traffic common areas in lower-income public elementary schools, 2) provision of cups/reusable water bottles for students, and 3) a 6-month healthy beverage education campaign. A five year-long cluster randomized controlled trial of 26 low-income public elementary schools in the San Francisco Bay Area is examining how Water First impacts students’ consumption of water, caloric intake from foods and beverages, and BMI z-score and overweight/obesity prevalence, from baseline to 7 months and 15 months after the start of the study. Intervention impact on outcomes will be examined using a difference-in-differences approach with mixed-effects regression accounting for the clustering of students in schools and classrooms.

Discussion:

This paper describes the rationale, study design, and protocol for the Water First study. If the intervention is effective, findings will inform best practices for implementing school water policies, as well as the development of more expansive policies and programs to promote and improve access to drinking water in schools.

1. Introduction

Childhood obesity is a major public health concern that increases the risk of poor health outcomes in adulthood, including type 2 diabetes, high blood pressure, abnormal cholesterol levels, and fatty liver disease [1]. Sugar-sweetened beverages (SSBs), such as sodas and sports drinks, are prevalent sources of added sugars and a major contributor to calories in the diet [2-6]. Consumption of SSBs is associated with obesity and its comorbidities [2-6], particularly among children from low-income and minority backgrounds who consume more SSBs and have higher rates of obesity than other groups [7-9]. Replacing SSBs with water can lead to a significant reduction in total caloric intake [10] and can help prevent obesity [5]. In addition, increasing water intake may improve hydration and cognitive function among children [11,12].

Given that children spend substantial time in school and half of students come to school inadequately hydrated [12], schools are a prime location for increasing water intake to replace SSBs. Since 2010, California SB 1413 and the federal Healthy Hunger-Free Kids Act have required public schools to provide free drinking water during mealtimes [13]. These policies so far have not sufficiently increased students’ access to appealing drinking water as many schools only offer a single drinking fountain for hundreds of students [14-16]. Surveys suggest that the majority of schools provide drinking water through existing fountains and that few provide cups or reusable water bottles to promote student consumption of water [15]. Previous studies have shown very low rates of water fountain usage in schools, with safety concerns and water palatability cited as barriers to consumption [17-20]. Furthermore, most schools still offer juice and flavored milk in the cafeteria, tempting children to choose these sweet alternatives over water [21,22].

There is some evidence that interventions that increase access to appealing drinking water and that intentionally promote water in schools can increase student consumption of water [23], reduce intake of SSBs [24], and prevent obesity [25,26]. In a randomized controlled trial (RCT) of a school-based intervention in 32 low-income elementary schools in Germany where SSBs were unavailable, installation of appealing water sources, coupled with classroom-based education, led to a 31% reduction in children’s risk of being overweight [25]. Providing more appealing water sources, such as reusable water bottle filling stations or bottle-free water coolers (henceforth called “water stations”), along with drinking vessels, may increase water intake in schools and offer a cost-effective obesity prevention strategy [21,23-30]. However, no studies have examined how school-based drinking water access and promotion interventions affect students’ obesity and beverage intake both in and out of school. Although drinking water may be satiating [31,32], no studies have investigated how such interventions affect food intake.

The goal of the Water First study is to conduct a cluster RCT to examine how increased access to safe and appealing drinking water at school, coupled with robust education and promotion activities, impacts food and beverage intake (in and out of school) and obesity among elementary school students. This manuscript describes the rationale, design, and implementation of Water First, a five-year study funded by the National Heart, Lung, and Blood Institute at the National Institutes of Health.

2. Methods

2.1. Study overview

The Water First study is a cluster RCT of public elementary schools that is examining if increased access to fresh water and rigorous promotion of its intake will reduce student’s intake of caloric beverages (e. g. SSBs) and/or foods, thereby leading to lower BMI and obesity. The theory-based intervention, cultivated in prior developmental studies by the research team, promotes water consumption by: 1) installing lead-free water stations in school cafeterias, physical activity spaces and high-traffic common areas, 2) providing cups and reusable water bottles for students, and 3) conducting a 6-month healthy beverage education campaign that includes a kick-off assembly, class lessons, family engagement activities, signage, and rewards. Over a period of 5 years, 13 intervention schools receive the Water First intervention and 13 schools serve as controls that after study completion, receive a water station and access to education and promotion materials. A community advisory board consisting of school district officials, teachers, Department of Education staff, policy advisors, and parents reviewed all intervention materials and activities to ensure relevance for families and students. As the majority of students in California public schools are from Latinx backgrounds [33], we conducted focus groups with parents to inform the cultural relevance of intervention materials and activities for this population.

Student outcomes are measured at baseline, 7 months, and 15 months after the start of the study. The study time frame is designed to be realistic to fully implement a school-based intervention. From previous experience, we approximate 4 to 6 weeks for each round of data collection. This assumes completion of preparatory work (e.g., preparation for installation of stations, relationship building with 4th grade teachers) in study schools in the year prior to the intervention. The length of follow-up, including observations at 7 months and 15 months post-intervention, allow us to observe both immediate and long-term study impacts. Study outcomes include: 1) intake of water and other beverages (including direct observations, flowmeter measurements of water taken from water sources, beverage intake surveys) 2) total caloric intake from foods and beverages (from 24-h food and beverage diaries) and 3) BMI z-score and overweight/obesity prevalence.

2.2. Study design and randomization assignment

The study design is a parallel arm, cluster RCT in which 4th grade students are recruited from public elementary schools. The cluster RCT of 26 schools is staggered across different school districts over a four-year period. Even numbers of eligible schools within each district (n = 6 to 8 schools) are randomly assigned to the intervention or control condition in a 1:1 allocation ratio using a computerized random number function. Randomization occurs following meetings with schools and prior to baseline data collection. While random assignment of classes or individual students would be a more efficient design with greater power, the cluster RCT design was chosen because the intervention includes school-level environmental changes (e.g., installation of lead-free water stations).

2.3. Study setting and eligibility criteria

The study takes place in 26 public elementary schools in school districts located in the San Francisco Bay Area of California. Because low-income students are most at risk for obesity and poor nutrition and beverage intake habits [7-9], the intervention focuses on schools in which ≥50% of students are eligible for free or reduced-price meals (FRPM) through the National School Lunch program. Schools that are already engaging in efforts to promote water intake among students by providing appealing water sources, such as water stations and cups/reusable water bottles, are excluded.

To be eligible, schools must have at least 65 4th grade students enrolled, as classroom-based education and promotion activities and evaluation take place with 4th grade students. Fourth graders are included in the more intensive promotion and education for several reasons: 1) limiting the study to one grade reduces costs and increases feasibility, 2) older children consume SSBs at a higher rate than younger children, and can influence the behaviors of their younger peers [34,35], 3) older students are able to provide more reliable dietary and other survey information, and 4) it is easier to track 4th graders over the 15-month follow up period since they often remain in the same school, whereas 6th graders often move to different schools. If 4th graders are in combination classes (4th/5th or 3rd/4th grades in one classroom), students from the other grades are also included in the study. The 4th grade students in special education classes are allowed to participate in the intervention activities and evaluations to the best of their ability. Students who transfer to the school after baseline data collection are ineligible.

As there are undernourished children in the United States (U.S.), increased water consumption among underweight children holds a small but important risk of increasing satiety, thereby reducing food consumption. Researchers work with teachers and students’ families to develop individualized plans that allow students who have chronic diseases, conditions that require special dietary accommodations, or who are underweight (BMI-for-age < 5th percentile), to participate in activities that do not interfere with the dietary prescriptions of physicians or nutritionists. The weight of these high-risk students is also monitored closely for changes.

2.4. Study recruitment

The study team recruited 6–8 schools per study year starting in the fall of 2016 to the fall of 2019. Demographics of schools and students who have completed the study at the time of this writing are shown in Table 1. First, the principal investigator (PI) reaches out to eligible school districts to assess interest in the study. In-person or phone meetings are held with interested district leaders including superintendents, board members, or directors of research and evaluation. During the meeting, the PI provides a short summary of the study, including the intervention elements, recruitment processes including consent and assent procedures, evaluation activities, and study incentives. Once school district approval is obtained and schools in the district are identified, the PI schedules similar meetings with the principal, 4th grade teachers, and other relevant school staff (food service, facilities managers, custodians). A memorandum of understanding between the school district and the academic collaborators is developed with each school district. For some districts, an institutional review board application, background check and fingerprinting, and proof of immunizations and tuberculosis testing for the research staff are also required. Study schools receive incentives of $500 and 4th grade teachers receive $200 for participating in the study. An administrative staff member (such as a principal or assistant principal) and a designated cafeteria staff member also receive $50 for completing surveys about school nutrition and physical activity policies and practices.

Table 1.

Sociodemographic Characteristics of Schools and Students for Schools in Study Years 1–2 (n = 12 schools).

Characteristics
School characteristics	n = 6681 overall students in study schools
Mean enrollment, mean (SD)	556.8 (148.3)
Free or reduced-price lunch eligible, n (%)	4373 (65.5)
Sex, n (%)
Male	3493 (52.3)
Female	3188 (47.7)
Racial/ethnic background, n (%)
American Indian or Alaska Native	27 (0.4)
Asian	1041(15.6)
Non-Hispanic Black or African American	211 (3.2)
Mexican American, Latino, or Hispanic	4513 (67.6)
Native Hawaiian or Other Pacific Islander	120 (1.8)
Non-Hispanic White	552 (8.3)
Two or more races	214 (3.2)
English learnersa, n (%)	2860 (42.8)
Mean percent, needs improvement – health risk body comp (SD)	29.5 (6.3)
Characteristics of students in study schools	n = 651 study students
Mean age in years (SD)	10.6 (0.3)
Sex, n (%)
Male	337 (51.8)
Female	314 (48.2)
Race/ethnicity, n (%)
American Indian/Alaska Native	3 (0.5)
Asian	100 (15.4)
Non-Hispanic Black/African American	22 (3.4)
Mexican Am, Latino, Hispanic	363 (55.8)
Native Hawaiian/Other PI	12 (1.8)
Non-Hispanic White	53 (8.1)
Two/more races	98 (15.1)
Primary language spoken at home, n (%)
English	211 (32.4)
Spanish	123 (18.9)
English/Other	314 (48.2)
Weight status, n (%)
Underweight	17 (2.6)
Normal	306 (47.0)
Overweight	137 (21.0)
Obese	191 (29.4)
Mean BMI Z-score (SD)	0.9 (1.1)

^aAccording to the CA Dept of Education, “students who do not speak, read, write or understand English well as a result of English not being their home language.”[36]

To recruit students and families, trained research staff provide presentations about the Water First study to 4th grade classes. Bilingual (Spanish/English) study information including consent forms and a short summary of the study are sent home to parents via students. The research team returns to classes to distribute new consent forms as needed for absent students or those who misplace forms. If families have questions about the study, the research team works with schools to contact parents via phone. Students receive small incentives (e.g., erasers, pencils) for returning consent forms. At intervention schools, all 4th grade students are allowed to participate in the school and classroom-based activities regardless of consent status. Students with parental consent are also asked to provide written assent prior to study participation. Only students with written parental consent and written personal assent can participate in the evaluation activities (e.g., surveys, food and beverage diaries, body mass index measurements). Students that complete all evaluation activities receive a movie theater voucher valid for one general admission. The study was approved by the University of California, San Francisco Committee on Human Research and the Stanford Institutional Review Board. The clinical trial was preregistered at clinicaltrials.gov (NCT03181971).

3. Intervention

3.1. Conceptual framework

The Water First intervention is based on the PRECEDE-PROCEED model and Social Cognitive Theory (SCT) (Table 2) [37,38]. The intervention incorporates environmental changes to promote healthy behaviors, along with lessons and school-wide promotion to influence students’ desire and perceived ability to increase water consumption. Intensive classroom lessons target the knowledge, attitude, and beliefs students have towards water and SSB consumption. Through these changes, the intervention aims to increase water intake, decrease SSB intake, and reduce overweight and obesity prevalence among students.

Table 2.

Social Cognitive Theory and PRECEDE-PROCEED Constructs in the Water First Intervention^a.

Construct	Construct element	Element operationalization
Environment (b)	Improvements in water access at school and home to encourage consumption of water	1. Installation of safe and appealing water sources in cafeteria, physical activity space, and high traffic area of the school 2. Provision of disposable 3. Distribution of reusable water bottles for at school and at home use
Situation (a)	Educate students and families about the health importance of drinking water instead of sugary drinks	School assembly, water station signage, classroom lessons, multi-media education materials for students, family engagement activities
Behavioral capability (b)	Students and families learn how to increase intake of water instead of sugary drinks by improving: 1) Ability to locate water sources on campus 2) Capacity to identify beverages that are the highest in sugar 3) Capability to make water taste delicious without adding calories	Classroom lessons and family engagement activities
Expectations (a)	Students and families learn about the health consequences of drinking sugary drinks	School assembly, water station signage, classroom lessons, multi-media education materials for students (e.g., videos, graphic novel), family engagement activities
Expectancies (a)	Students will learn why drinking tap water has a lower environmental footprint than drinking sugary drinks	Water station signage, classroom lessons, multi-media educational materials for students, family engagement activities
Self-control (a)	Students will set goals to drink healthier beverages at home	Student beverage contracts with goals for beverage intake at school and at home
Observational learning (c)	Students learn about the importance of drinking water through teacher and peer modeling	Teachers and 4th grade students provided with reusable water bottles to encourage water intake at school and at home
Reinforcements (C)	Students receive reinforcements when observed drinking water; students provided with reinforcement tools to encourage water intake at school and at home	On one random day per week for 5 months, students receive prizes when they are observed drinking water at lunch; students receive a carabineer to make bringing water with them easy
Self-efficacy (c)	Students teach their parents about the benefits of drinking water instead of sugary drinks; students work with their families to complete family engagement activities	Classroom lessons and family engagement activities

^aPredisposing (a), enabling (b), and reinforcing (c) factors [37,38].

3.2. Environmental changes

The year-long intervention consists of school environment changes that increase the accessibility of safe and appealing drinking water to promote consumption (Table 2). Tap water is provided rather than individual-use bottled water because of its lower cost and environmental friendliness [39,40]. In addition, tap water, unlike most individual-use bottled water, contains fluoride, which can prevent dental caries [41]. It has been noted that some cultures and households with low educational attainment distrust tap water quality, which may reduce tap water consumption [42-45]. To address such concerns, water from the locations where new stations are installed are tested for lead by an EPA certified laboratory [46]. If lead levels exceed the American Academy of Pediatrics endorsed maximum level of 1 ppb, the school district assists with remediation. If remediation by the district is not feasible, an NSF® International-certified filtration device is installed to remove lead [47]. Follow-up testing is conducted after remediation to ensure lead removal. Results from the water testing are posted via signage near the water stations and disseminated to students, school staff and parents through lessons and family engagement activities (see below).

Installation of more appealing water sources can help address negative attitudes students may have about drinking fountains [18,19]. A Culligan® chilled bottle-free water cooler is installed in the cafeteria at each intervention school. In addition, each intervention school also receives durable water stations that are placed at two other locations: a gym or blacktop where students are physically active, and a high-traffic common area. The research team works with school principals and facilities staff to identify locations for new water stations based on existing water lines, electricity, and school need. As in previous successful studies [23,24], recyclable or compostable cups (based on school preference) are provided for use at lunch to increase accessibility to chilled water during mealtimes. During the first Water First lesson, 4th grade students receive two reusable water bottles – one for school use and one for home use. If students lose a water bottle for at school use, researchers replace it once. If they lose a bottle a second time, they are asked to bring in the study bottle given to them for at home use. The 4th grade teachers also receive reusable water bottles to model water intake at school. Teachers are asked to remind students to fill up their reusable water bottles at least once daily for school use, and are provided with supplies to sanitize and thoroughly air-dry bottles at least once weekly.

3.3. Education and promotion

Education and promotion activities used in Water First, including school assemblies and signage, classroom lessons, and parent engagement activities, have proven successful in previous studies [23,25,27].

3.4. Kickoff assembly

The Water First program is launched at each intervention school through an interactive school-wide assembly. The assembly, which typically includes a live performance by a local singer/songwriter, aims to educate students about the importance of drinking water in a fun and engaging manner [48]. The school principal and research team also introduce the Water First program and establish rules about the orderly use of the water stations and the proper disposal of the cups in recycling/composting bins at lunch.

3.5. Signage, emoticons, and prizes

Signs communicating reasons why it is beneficial to drink water are posted next to the newly installed water stations. These signs are changed weekly to reinforce each week’s lesson content taught in the 4th grade classrooms. In addition, signage displaying “go”, “slow”, “whoa” concepts are placed in close proximity to beverages in the school cafeteria and classroom. Green emphasizes consumption of “go” drinks, such as water and plain milk, that are nutritious and can be consumed frequently. Yellow signifies “slow” drinks, which provide some nutritional value but should be consumed in moderation (e.g.,100% fruit juice, chocolate milk). “Whoa” drinks, which have lower nutritional value and have considerable added sugar (e.g., sodas, sports drinks), are symbolized as red to urge students to consume them less often. To reinforce water intake during lunchtime, prizes are awarded weekly, on a day chosen at random, to students observed drinking water and disposing of cups in appropriate bins. In a previous healthy beverage intervention in elementary schools, use of prize reinforcement increased students’ selection of plain milk instead of chocolate milk [49,50].

3.6. Lessons

Trained research staff lead eight weekly, 15-min interactive lessons in 4th grade classrooms about the health, fiscal, and environmental benefits of drinking water. These lessons were developed, piloted, and refined by the research team in previous studies in schools in the San Francisco Bay Area [23,27]. Lessons include both didactic and interactive components (Table 3). Students are asked to develop individual or group projects (e.g., posters, skits, puppet show, poem, video) that communicate some aspect of lesson content. During the final class session, students present their project to their teacher, their peers, and any family members who are available to attend. Students receive a certificate and button for completing the Water First lessons.

Table 3.

Water first lessons, family engagement activities, and reinforcement tools.

Lesson Theme/ Objective	Lesson Activities	Family Engagement Activitiesa/Reinforcement Tools
1. Introduction to Water First program and importance of drinking water for health	Overview of Water First Program Watch video about the importance of drinking water for health and discuss key points Distribute reusable water bottles to students/teacher Introduction of final student presentations	Family engagement activity: Watch Share The Love, Share the Water video (DVD English/Spanish) with family and write down what family learned Reinforcement tool: Reusable water bottles (2 per student, 1 for school and 1 for home use)
2. Water is Safe Review origin, treatment, and safety of tap water	Explain water quality, including where local water originates and how it is treated, and provide results of lead testing of drinking water at school	Family engagement activity: Teach family about origin of tap water, its treatment, and results of lead testing of water at school Reinforcement tool: Graphic novel handout
3. Water is Healthy Discuss added sugar content in beverages and why water is a healthy alternative	Explain how to read a nutrition label Calculate added sugar in beverages Visually display the teaspoons of added sugar in beverages Overview of “Go, Slow, Whoa” concept for beverages	Family engagement activity: Calculate the added sugar in drinks at home; identify beverages as go, slow, or whoa drinks Reinforcement tool: Graphic novel handout
4. Water is Easy Examine locations in and out of school where tap water is available	Brainstorm locations to access tap water in the community School scavenger hunt to locate drinking water sources at school	Family engagement activity: Work with family to complete contract to identify at least one healthy beverage goal Reinforcement tool: Carabineer for easy water access; graphic novel handout
5. Water is Tasty Consider ways to make water taste delicious without adding sugar	Overview of ways to make water tasty without adding sugar Watch short video about how to create fruit and vegetable infused waters Taste test of infused waters; guess ingredients used	Family engagement activity: Show family fruit and veggie water video (DVD) and design recipe Reinforcement tool: Graphic novel handout and recipe card
6. Water is Green Discuss the environmental footprint of beverages	Review environmental impact of drinking tap water vs. single-use bottled water vs. sugary drinks Rank order beverages by the amount of water that is needed for production Discuss plastic waste generated by single-use bottles	Family engagement activity: Work on final project and discuss lesson content with family Reinforcement tool: Ice cube molds to encourage tap water intake; graphic novel handout
7. Water makes $ense Overview of how drinking tap water is cost-saving and can be beneficial for oral health	Rank order commonly consumed beverages by cost per ounce Calculate annual cost-savings of drinking tap water instead of bottled water and sugary drinks Watch video about oral health benefits of drinking fluoridated tap water	Family engagement activity: Work on final project Reinforcement tool: Crazy straw; graphic novel handout
8. Final project and presentation	Review of lesson content Student presentations for classmates, teachers, and parents	Certificate of completion “Share the Love, Share the Water” button

^aStudents who complete family engagement activities receive prizes such as temporary tattoos, stickers, pencils, and erasers.

3.7. Family engagement activities

As the majority of SSB intake by elementary students occurs in the home environment [51], the educational curriculum also includes interactive bilingual (English/Spanish) family engagement activities to help increase water intake and decrease intake of SSBs in out-of-school environments. Students who complete the activities with their family and return the engagement activity worksheet with their parent or guardian’s signature receive small prizes such as an eraser, temporary tattoo, or a pencil.

3.8. Control schools

Following the conclusion of the intervention and final 15-month data collection for all study measures, the control schools receive a water station, a $500 incentive, and access to the intervention education and promotion materials.

4. Study outcomes

4.1. Primary Outcomes: Body Mass Index (BMI) Z-Score and BMI Percentile

At baseline, 7 and 15 months after the start of the study, trained research staff measure each participating student’s height and weight in private using a calibrated scale and stadiometer as outlined in the National Health and Nutrition Examination Survey Anthropometry Procedures Manual [52]. Height and weight measurements along with students’ sex and date of birth are used to obtain BMI z-scores and age and sex-adjusted BMI percentiles using a statistical program from the Centers for Disease Control and Prevention [53].

4.2. Secondary outcomes: Caloric intake from foods and beverages

At baseline and 7 months after the start of the study, students are asked to complete 24-h food and beverage diaries. The food and beverage diary protocol has been used in numerous school-based studies and was adapted from a previously validated instrument [54,55]. Trained research staff conduct a 45-min classroom session in which they introduce the diary protocol and instruct students on how to use food and beverage assessment measures [54]. The measures include food and beverage description guides, cups and spoons to measure the portion sizes of foods and beverages consumed at home, and diary pages for students to write down information about the type, portion size, time, and location of foods and beverages consumed. Students are encouraged to complete the diary in consultation with parents or other adults who care for them, particularly for items that describe different food or beverage components. Trained dietary interviewers review the responses with the students within 1–2 days of diary completion. Interviewers use a multiple-pass method that is commonly used for assessing dietary intake in population-based studies, in which interviewers ask multiple questions about the previous 24 h using probes to help individuals remember and describe foods and beverages consumed. To further improve the validity of the dietary data, the multiple-pass method is informed by a school food and beverage scan that is completed by the research team [55]. Here, prior to reviewing diaries with students, the research team conducts observations, interviews food services staff, and collects school menus to record school food and beverage offerings.

Diary data are analyzed using USDA’s Food and Nutrient Database for Dietary Studies (FNDDS) [56]. The FNDDS is used in the National Health and Nutrition Examination Survey and contains descriptions for 6940 foods and beverages including portion sizes and weights, preparation method, energy content, and nutrients. Foods and beverages consumed by elementary school students are matched to food and beverage codes in FNDDS by a registered dietitian. For combination foods and/or beverages (e.g., smoothies) we construct a combination code using individual food and beverage codes from FNDDS. Energy content or calories for each food/beverage from FNDDS are the outcomes of analysis. As the team encounters new foods or beverages, the database is adapted to include their nutrient information, including calories from food and beverage company sources available online.

4.3. Secondary outcomes: Beverage intake frequency

Due to budgetary constraints, the food and beverage diaries are administered only at baseline and 7 months following the start of the study. To capture changes in beverage intake over the course of the 15-month study, students also complete a beverage frequency questionnaire at the baseline, 7-, and 15-month time points. The beverage frequency questionnaire was adapted from the Beverage and Snack Questionnaire [57] and asks students to report if they drank different beverages while they were at school and out of school. Response options are never, 1 per week, 2–4 per week, 5–6 per week, 1 per day, 2–3 per day, and 4+ per day.

4.4. Secondary outcome: Water intake

4.4.1. Proportion of students drinking from water sources in key school locations

Trained researchers observe the number of students that use water sources in 1) the cafeteria at lunch, 2) at recess, and 3) during PE class, at both intervention and control schools, at baseline, monthly during the intervention period, at 7 months, and at 15 months after the start of the study. To calculate the proportion of students who use water sources during observational periods, researchers divide the number of individuals who use the water source by the total number of students in attendance in the area [17]. As outdoor temperature could affect water intake [58], during observations researchers also document ambient temperature using a phone-based weather application.

4.4.2. Volume of water taken from water stations

School facilities staff install flowmeters on the water line for reusable water bottle-filling stations and bottle-free coolers at intervention schools to track the volume of water taken from water sources. Trained researchers record flowmeter measurements before and after lunch once a week to estimate the volume of water taken from water sources during lunchtime. Researchers calculate the average volume (e.g., ounces) of water taken per student by dividing the volume of water taken from a source during a given time (e.g., lunch) by the number of students in attendance in the area.

4.5. Covariates

4.5.1. Students’ physical activity, screen time, and demographics

Students also complete surveys at all time points that include questions about their physical activity, screen time, and demographics (age, gender, race/ethnicity, and language spoken at home) [59,60].

4.5.2. School nutrition and physical activity practices and policies

At baseline and the 15-month follow-up, each school principal is asked to complete a survey about the current practices and policies related to nutrition and physical activity at their school [60]. A member of the food service team at each school is also asked to complete a survey regarding school nutrition practices and policies [60].

4.5.3. School beverage access

Researchers also use an adapted Harvard Water audit tool to document the condition of the water sources at the schools at baseline, and the 7- and 15-month time points, as the appeal of the sources may influence water consumption [61]. Information about access to beverages during the school day (within a quarter-mile radius from the school), including the location, type of beverage, venue where it is sold (e.g., vending machine, ala carte, carts or stalls), along with information about marketing of beverages (e.g., product, brand, location, size and type of advertisement), is also noted.

4.6. Fidelity measures

4.6.1. Fidelity visits

At intervention schools, researchers conduct random biweekly visits to examine the fidelity of the intervention. These visits include checking that cups are available to students in cafeterias, signage is posted near new water stations and in classrooms, reusable water bottles are being used by the 4th grade and 4th grade combination classes, and that teachers are role-modeling water intake.

4.6.2. School logs

On a weekly log, food service staff mark whether they put out compostable/recyclable cups for students to use next to the new water station in the cafeteria. Teachers also complete a weekly log in which they indicate whether or not they help students to fill intervention water bottles with water daily and to clean bottles once a week. Researchers also track the number of students in each classroom who return family engagement activity forms.

4.7. Statistical analysis and power calculation

4.7.1. Data analysis

Stata SE 16.1 (StataCorp LP, College Station, Texas) or SAS version 9.4 (SAS Institute Inc. Cary, NC, 2013) will be used for analyses. We will use range checks and examine variable distributions to establish, as needed, appropriate cut-points for measures that are also informed by the literature.

Analyses will test our hypothesis that students in schools receiving the Water First intervention will experience significant increases in their consumption of water, and declines in their caloric intake from beverages and BMI z-score as compared to students in control schools. Mixed-effects logistic and linear regression models will be used for our primary and secondary dependent variables to assess the change in outcomes during follow-up. The primary independent variable will be the interaction of intervention status and time, while adjusting for baseline differences and secular trends. The model will also account for clustering of students within schools and classes and adjust for student race/ethnicity, gender, and age.

4.7.2. Power calculations

Our sample size was determined by requiring 80% power to detect a difference of 5% in the change in our primary outcome, BMI percentile, between baseline and 15 months [25]. We assumed a standard deviation (SD) for change in BMI percentile of 25.6% based on previous studies. These calculations account for clustering by adjusting for the design effect, assuming an intra-class correlation (ICC) of 0.005 (between vs. in school differences in BMI percentile change). Enrolling 13 schools in each intervention arm with 50 students per school (n = 1300 total) was found to be sufficient to guarantee 80% power [25,62].

Given this sample size, we also anticipate having 80% power to detect a difference of 20 kcal from total daily beverages consumed by students in intervention schools compared to control schools (two-sided alpha of 0.05). This is based upon a SD for change in energy intake of 125 kcal based on our previous pilot study [62].

Lastly, we anticipate having 80% power to detect a difference of 54 daily food calories consumed by students in intervention schools compared to control schools (two-sided alpha of 0.05). Calculations assume a SD for change in daily food calories consumed of 714 and an ICC of 0.05 based on prior evaluations of 4th and 5th grade California students’ dietary intake [63].

These calculations are also adjusted for the cluster design assuming an ICC of 0.05 (between vs. within school differences in change in caloric intake) based on prior studies of nutrition behaviors of students from prior studies [62,63].

5. Discussion

Obesity is one of the most common chronic conditions in children, one that increases the risk of significant health complications in adulthood, making this health concern an important target for health interventions [1]. Intake of SSBs is a known contributor to obesity [2-6]. Encouraging children to drink water as an alternative to SSBs is an important obesity prevention strategy [10], and can also prevent dental caries [64] and may help improve child cognition [11,12].

Previous studies suggest that providing appealing drinking water in schools with drinking vessels such as cups or reusable water bottles can increase intake of water, reduce intake of SSBs, and prevent obesity [23-26]. In these studies [21,23-30], however, it was not known whether decreases seen in obesity occurred through reductions in the intake of caloric beverages or food. Nor was it clear whether such interventions promoted healthier choices only within the confines of the school, or if they also extended to outside the school. Moreover, most of the previous studies have not tested an approach that couples widespread access to appealing drinking water with comprehensive activities to promote drinking water not only at school, but also in the home environment.

Through a five year-long cluster RCT in 26 low-income, public elementary schools in the San Francisco Bay Area, the Water First study is testing how increased access to appealing and safe drinking water in key school locations and activities to promote water intake among students affects students’ weight status and what they are drinking and eating.

The Water First study will provide new evidence about how school water access and promotion impact beverage and food intake and obesity in real-world settings where schools may or may not offer high-calorie beverages. If the intervention is effective, findings will not only inform best practices for implementing current school water policies, but they will guide the development of new, more expansive policies and programs to promote and improve access to drinking water in schools.

A strength of the study is the rigorous study design - a cluster RCT with parallel arms for comparison. In addition, the research team has built strong partnerships with school district staff at all levels that facilitates the recruitment of schools and study implementation. Another strength is that the study targets public elementary schools serving students from low-income and predominantly Latinx backgrounds who are known to consume more SSBs and are at greatest risk for obesity than other groups [7-9]. As noted, all intervention materials were adapted to be culturally relevant to this population. For example, we included examples of culturally relevant beverages such as aguas frescas and horchatas in intervention materials and student beverage intake surveys. Furthermore, all of the intervention materials are available in English and Spanish, capturing two of the most common languages spoken by families of students in the U.S., thereby increasing the accessibility and comprehension of information by the majority of students and families. Lastly, the team also incorporated extensive process data collection about intervention implementation and fidelity that will inform interpretation of study results.

This study also has limitations. Methods require parental consent for student participation in the evaluation activities. While the study team uses multiple strategies to maximize participation (e.g., providing incentives when students’ return forms, redistributing forms on multiple occasions, working with school staff to call families to answer questions about the study), students who participate may have a greater level of family engagement than those who do not. Also, while this California-based study may not be generalizable to states with different demographics, because 1 in 8 children in the U.S. who attend a public school, attend a public school in California, this study is relevant to a large population of U.S. children [33]. Another limitation pertains to our ability to gather high-quality dietary data. Although our collection of multiple days of dietary data is superior to capturing only a single day, we use one 24-h food and beverage diary at each time point to reduce participant burden and to decrease study cost. A single day of intake is routinely used to compare group intakes in population-based studies [65]. Furthermore, although half of students may be underhydrated [12], we do not assess how the Water First intervention affects students’ hydration status. Future studies should consider assessing changes in hydration by using non-invasive urine samples to measure urine specific gravity. Finally, even though use of flowmeters on water stations to measure water consumption provides an objective way to monitor water intake at school, this approach fails to capture water that is spilled or discarded rather than consumed.

5.1. Implications

The Water First study will provide a rigorous evaluation of an intervention that aims to increase water consumption, decrease SSB intake, and prevent obesity among elementary school students through water promotion and access to more appealing and safe water sources in school settings. Study findings will contribute to evidence on the role that drinking water access and promotion programs in schools play in fostering healthy beverage intake habits and preventing obesity among all children, including those from the most vulnerable racial/ethnic groups.