Study design

he study used a controlled, randomized trial with two groups. The first group received the educational intervention and the second served as a control. Schools were considered the units of the intervention, and randomization was performed by school. The intervention lasted one school year from September 2014 until July 2015. The sample size determination was performed based on the Muckelbauer and colleagues¹³ study to detect a difference in the plain water consumption of 220 ml between the intervention group (IG) and the control group (CG), with α = 0.05 and a power of 0.8. We estimated 102 children per school and thus required seven schools, 714 per group and a total of 1 428 children.

Participants

A community intervention study was performed in the Coyoacan Jurisdiction of Health including two health centers adjacent to the National Institute of Pediatrics. There exists a collaboration agreement with these two health centers. These centers have epidemiological oversight of 21 primary schools that were invited to participate. Eight schools accepted. These schools have an enrollment of 1 748 students in first, second and third grades of primary. Children whose parents accepted to participate and signed a privacy agreement were included. Children with a disease that, in the investigator’s opinion, could interfere with the study were excluded. Subjects that delivered a beverage consumption register (BCR) that were incomplete or incorrectly filled out were also excluded.

Intervention

At the beginning of the school year 2014-2015, the parents received a lecture by one of the investigators to inform them of the general characteristics of the study, clarify any doubts and receive instructions on how to fill out the beverage consumption register (BCR). Also, a nutritionist gave a session to sensitize the parents about the habits of consuming beverages to encourage them to drink plain water at home. The investigator also explained to the participating teachers the general characteristics of the study and clarified doubts. The teachers of the intervention groups were prepared by a nutritionist to use the educational strategies in a workshop that lasted six hours. At the beginning of the study, the teachers received a manual with the program and material necessary to give educational sessions.

The I prefer plain water education strategy was designed by Ogali (a nutrition consulting company [www.ogali.com]) and was validated in a previous study (unpublished). The educational support was based on three theories of knowledge: the social cognitive theory, social constructivism, and multiple intelligences. These have shown behavioral modification and the development of life abilities, permitting children to build permanent foundations to obtain a healthy life style (see “Educational Basis for the Strategy” [ www.prefieroaguasimple.com , web page under construction]). The strategy is designed for children in first, second and third grades of elementary school and adapted for each school year. It is based on games and a collaborative system in which the student, teacher, and family take part. It consists of four sessions of approximately 50 minutes each and is given every two weeks during class time (see “manual for teachers” [www.prefieroaguasimple., web page under construction]).

During the study, a group of nutritionists verified the presentation of the educational sessions and supported the teachers when asked to do so. There was a register of the date, duration, and the name of the teacher that gave the session.

A peemeter is a urine color chart that shows the hydration state according to the urine color and thus indicates whether the user needs to consume more water.¹⁴ Peemeters were placed in the lavatories of the IG schools. Additionally, the children received personal peemeters.

Instrumentation

The beverage consumption was evaluated using a diary BCR, which was completed only by the parents or caregivers at home during a 24-hour period for three separate days (two weekdays and one weekend day). The beverage intake during school hours was quantified when the student returned home after the parents could corroborate how much of the beverages that they sent were consumed.

Nine beverage categories were included and recorded seven times a day. The amounts were recorded in milliliters. This BCR has previously been used in large cross-sectional surveys among children and adolescents.¹⁵ The BCR was classified as implausible when a daily record in one category was > 3 500 ml.

The beverage categories were established based on a proposal by the National Health and Nutrition Examination Survey 2006.¹⁶

Process evaluation

During September 2014, each student’s clinical examination was performed. The beverage consumption baseline records were created in October 2014 (baseline or time point 1). Subsequently, the water dispensers were placed in each classroom of the eight schools. The four educational sessions were provided at the intervention schools during the November-December 2014 and January 2015 time periods. The second BCR was completed in February 2015 (four months - time point 2). From March to June, the water supply was maintained in both groups, and the third BCR was completed two weeks before the end of the school cycle in June 2015 (eight months - time point 3).

Human subjects approval statement

This study was conducted according to the guidelines established in the Declaration of Helsinki and the “International Ethical Guidelines for Health-related Research Involving Humans” (CIOMS). All procedures involving human subjects/patients were approved by Institutional Review Board of the National Institute of Pediatrics (IRB 8064, CONBIOETICA-09-CEI-025-20161215, approval number INP 013/2014). A consent form was received from the parents and directors of each school. One of the parents signed a privacy agreement. All data were recorded anonymously.

Statistical analysis

The analysis was performed using the IBM SPSS 22.0 statistical package. We compared the baseline consumption of the different beverage categories between the two groups using the Mann-Whitney U test. To compare the differences in beverage consumption at the three time points for both groups, we performed a Friedman test of related samples. The educational strategy was considered effective if the student’s consumption of plain water increased by ≥ 220 ml. We compared the percentages of children who at the end of the study increased water consumption by 220 ml using the χ2 test. We considered p < 0.05 to indicate statistical significance.

Baseline consumption

Of the 777 children, 408 were girls and 369 boys, with a median age of 7.1 years and an age range from 5.6 to 9.4 years. The average baseline consumption of total liquids was 1 337 ml/day (SD 546). No significant differences were identified in the baseline consumption of total liquids and categories between the IG and CG (the results are not shown). The baseline results are shown in table I.

Beverage consumption

A total of 413/777 subjects were assigned to the IG, including 214 girls and 199 boys with a median age of 7.1 years, a minimum age of 5.6 years and a maximum age of 8.8 years. A total of 364/777 subjects were assigned to the CG, including 194 girls and 170 boys with a median age of 7.2 years, a minimum age of 5.6 years and a maximum age of 9.6 years. The median baseline of the total liquid consumption in the population was 1 231 ml/day (mean 1 337 ml/day, SD 546). The beverage consumption at the three evaluation time points is shown in table II.

In the IG, the increase in the median quantity of plain water consumption was 109 ml at time point 2 (four months) and 167 ml at time point 3 (eight months). In the CG, no increase was found during the first period. During the final period, we found an increase of 37 ml. During the study, significant changes occurred between the groups. At the end of the school year, an increase of 220 ml was achieved in plain water consumption in 95/264 children in the CG and in 166/413 children in the IG (X2 p < 0.001).

The total liquid intake in the IG presented a median increase of 80 ml from time point 1 to time point 2 and 41 ml from time point 1 to time point 3. The CG presented median increases of 33 ml from time point 1 to time point 2 and 38 ml from time point 1 to time point 3. The changes were not significantly different between the groups; however, the changes in the IG measurements were significantly different between time point 1 and time point 2 (figure 2).

^*Friedman test

Figure 2 Change in beverage consumption at the three time points(intervention group). Mexico City, 2015 

When comparing the consumption of other beverages between the groups, we found significant changes in sugar-sweetened beverages and sodas (figure 2). Except milk and milk beverages, the consumption of hot beverages, light beverages and functional beverages was low, and no significant changes were observed throughout the study.

Milk and milk beverages represented the second most commonly consumed liquids after plain water, with a similar median in the two groups throughout the study. No significant differences were found. Figure 3 shows the percentage differences between the baseline (time point 1) and final consumption (time point 3) medians of plain water, sugar-sweetened beverages, sodas and total liquids.

Figure 3 Percentage variation between consumption medians at time 1 and time 3 points (intervention group). Mexico City, 2015 

Limitations

The present study provided the BSR, a validated research instrument,¹⁵ to the parents to gather key research information. Some parents were not committed to tracking fluid consumption and others were confused about how it should be completed causing the removal of records from the analysis. Also, during school hours, only the drinks that the parents sent to be consumed during recess were quantified reducing the proportion of children who increased their water consumption by 220 ml/day.

In conclusion, when incorporated into the educational program for each school grade of early primary school and associated with free, convenient access to plain water inside the classrooms, the I prefer plain water educational strategy is effective in increasing the consumption of plain water and decreasing the consumption of sugar-sweetened drinks. This strategy did not modify the consumption of total liquids, milk and milk beverages or other beverages