Americans can limit excess caloric intake and weight gain by drinking water and other beverages with few or no calories (1) and caloric intake could be reduced among children if sugar-sweetened beverages were replaced by plain water (2). However, more than 50% of US high school students drank 3 or fewer glasses or bottles of water per day in 2010 (3) and 19% drank 2 or more cans, glasses, or bottles of soda per day in 2011 (4). Adequate water consumption also prevents dehydration, which is associated with poor cognitive function in youth (5).

Providing greater access to free drinking water in schools is a strategy to support healthy beverage intake among youth (6,7). The Institute of Medicine recommended that free drinking water be available throughout the school day to students (8). The Healthy, Hunger-Free Kids Act of 2010 requires that schools participating in federal meal programs make free drinking water available to children during meals in the cafeteria (9). Outside of the cafeteria, access to free drinking water is regulated by state and local plumbing codes. These codes can specify the number of drinking fountains, the placement of fountains, and whether or not some or all of the plumbed fountains may be replaced by nonplumbed water dispensers. Plumbing codes are generally enforced during initial school construction or when schools undergo major renovation. Furthermore, it is unknown how plumbing codes might affect water availability over the life of a school as plumbing fixtures break down and temporary mobile classrooms are added to accommodate student enrollment beyond the original design capacity of school buildings. In 1999, 25% of public schools had plumbing in “less than adequate condition” (10), and in 2005, 33% of all public schools had temporary classroom facilities (11). Therefore, we investigated student-reported access to drinking fountains in schools and whether access differed according to individual and area-level characteristics, including state plumbing codes.

Overall, 57.3% of respondents reported that school drinking fountains or dispensers were widely available, 40.1% reported there were only a few, and 2.6% reported that there were no working fountains available in their schools. Statewide plumbing codes for school buildings are shown in Table 1. The minimum fountain to student ratio ranged from 1 fountain per 30 students (Arkansas) to 1 fountain per the first 150 students and 1 fountain per each additional 500 students (Washington). The most common ratio was 1 fountain per 100 students (31 states). Regarding nonplumbed water dispensers, 26 states allowed nonplumbed water coolers to be substituted for up to 50% of plumbed fountains, and 5 states allowed substitution of 100% of plumbed fountains.

Perceived school water access differed by race/ethnicity, region, fountain to student ratio specified in the plumbing code, and allowance of nonplumbed water dispensers (Table 2). Widely available fountains or dispensers were reported at the highest frequency by black youth (67.8%) and least frequently by Hispanic youth (50.4%). Regional prevalence of widely available school fountains ranged from 65.4% in the Midwest to 45.7% in the West. Widely available school drinking fountains were reported less frequently where state plumbing codes specified a fountain to student ratio of 1 fountain per more than 100 students (45.4% vs 60.1% for 1 per 100 students and 57.6% for 1 per fewer than 100 students). A greater percentage of students who lived in states where nonplumbed water sources could be substituted for all or half of drinking fountains reported widely available water access (61.0% for 100% substitution allowed and 61.1% for 50% substitution allowed vs 48.5% for no substitution allowed).

In multivariable analysis, black youth were more likely to report widely available fountains or dispensers compared with white youth (odds ratio [OR], 1.6; 95% confidence interval [CI], 1.1–2.3). Youth living in the Midwest (OR, 2.2; 95% CI 1.5–3.2) and South (OR, 1.6; 95% CI, 1.2–2.3) were significantly more likely to report widely available school drinking fountains than youth in the West. Youth living in states where plumbing codes required 1 fountain per more than 100 students were significantly less likely (OR, 0.6; 95% CI, 0.4–0.9) to report widely available drinking fountains than youth residing where plumbing codes specified 1 fountain per 100 students. Compared with youth residing where nonplumbed drinking fountains were not allowed, youth residing in states where 100% or 50% substitution of plumbed fountains with nonplumbed water dispensers was allowed were significantly more likely to report widely available drinking water fountains or dispensers (OR, 1.7; 95% CI, 1.1–2.6 for 100% substitution allowed; OR, 1.7; 95% CI, 1.3–2.2 for 50% substitution allowed) (Table 3).

We found that more than 40% of youth reported that there are few working drinking fountains or dispensers available in their school, and 1 in 25 youth reported there are no working fountains or dispensers. We also found regional differences in youth-reported access, with reported access lowest in the West and highest in the Midwest. These differences may be partially due to state plumbing codes, which differ greatly by region. To our knowledge, this is the only study to examine school drinking water access at a national level or to examine the association of state plumbing codes with water access in schools.

Increasing free drinking water access in schools is gaining greater attention as a means of encouraging healthy beverage intake among children, with recent regulatory requirements implemented at both the state and federal level (9,14). Data on the availability of free drinking water in schools is limited. A 2011 national survey of school administrators found that approximately two-thirds of middle and high school students had drinking fountains available in the school cafeteria, approximately 80% had fountains available in gymnasium or locker room areas, approximately 98% had fountains available in hallways near classrooms, and approximately 40% had fountains in other school locations (15). On a local level, Patel et al reported in 2011 that only half of schools surveyed in the Bay Area of California had water available in food service areas before enactment of California Senate Bill 1413 and the national 2010 Healthy, Hunger-Free Kids Act, which both require schools to provide access to free drinking water during mealtimes in school food service areas (16). Regarding school wellness policies, Cradock et al found that school wellness policies in Massachusetts rarely covered access to drinking water (17). Although school administrators reported in 2011 that approximately half of middle and high school students were covered by written policies regarding availability of free drinking water (15), examination of actual district wellness policy documents found that approximately 85% did not include a written policy regarding availability of free drinking water throughout the day (18). Regarding water offered for sale in schools, data from the US Department of Agriculture’s School Nutrition Dietary Assessment IV suggest that bottled water is made available for sale through vending and a la cart sales during lunch in most middle and high schools (19).

Our study suggests that water access may be greater where plumbing codes require at least 1 drinking fountain per 100 students and where codes allow the substitution of nonplumbed water dispensers for plumbed drinking fountains. However, the clustering of these 2 attributes in state plumbing codes because of widespread adoption of published model codes makes it difficult to separate the effects of fountain to student ratio and allowing nonplumbed water fountains to be substituted for plumbed water fountains. Specifically, all state plumbing codes that specify fountain to student ratios of 1 fountain per more than 100 students also do not allow any substitution for plumbed drinking fountains. Although 3 non-Western states specify fountain to student ratios of 1 fountain per more than 100 students (Maine, Nebraska, and South Dakota), the small samples from these states limited our ability to separate the effect of plumbing codes on school water availability from that of region.

Regional factors other than plumbing codes could explain regional differences. In general, the arid climate of many Western states can create challenges in ensuring adequate drinking water supplies for their populations and may thus encourage an atmosphere of water conservation that overrides access concerns (20). Water quality problems may also affect the availability of tap water in schools, and regional differences exist in perceptions of tap water safety (21). The lowest level of access reported in our study occurred in the West and Northeast. Although systematic data on water quality at the point of use in schools are not available, problems such as agricultural contamination of drinking water in California’s Central Valley (22) and point-of-use drinking water contamination from old plumbing in Philadelphia schools (23) have been reported. However, undocumented problems with water quality in schools could exist in other places as well.

This study has several limitations. First, we could not separate the effects of plumbing codes from those of region. Second, we used youth-reported drinking fountain access, from a survey offered only in English, rather than objective measurement. Future research is needed on how student-reported water access corresponds to actual access and to validate water access survey items. Third, we did not have information on local municipal plumbing codes or school wellness policies, which may go beyond state plumbing codes in specifying water access requirements in schools. Fourth, we did not have data on the age of the school facilities. Age of school facilities may be a factor in how plumbing codes associate with actual water access, because codes may change over time and broken fountains may come out of service and not be replaced. Furthermore, water access may change over time as the number of students attending a school changes and schools undergo additions or renovations that either add or remove fountains. Finally, because of selection bias associated with the use of a convenience sample from a mail panel survey and a low response rate, our findings may not be generalizable nationally. However, the study population included participants from 47 states and the District of Columbia who were sampled proportionately to state populations, and participants were weighted to match the US Current Population Survey (12). A previous study has shown that prevalence estimates of certain items from HealthStyles (eg, health conditions and behaviors) are comparable to findings in the Behavioral Risk Factor Surveillance System, which uses a probability sampling technique (24).

We found that 4 in 10 youth report that there are only a few working drinking fountains or dispensers in their schools, and 1 in 25 indicated that they had no working fountains or dispensers in their school. Additionally, student-reported drinking fountain availability, after controlling for other factors, was significantly different by race/ethnicity, region, the fountain to student ratio specified in state plumbing codes, and whether the plumbing code allowed nonplumbed water sources. Further research regarding how drinking fountain requirements in plumbing codes relate to drinking water availability in schools and other places would be useful for states and communities that wish to increase drinking water access to support healthy beverage intake among their residents.