This study evaluated the ‘water campaign’ programme. We found an effect on SSB on the basis of two of the three sources of information that were used to assess SSB consumption (i.e. ‘parent report’ and ‘observation report’).
Although the intervention had no effect on whether or not children consumed SSB on a daily basis, their average SSB consumption did change: after one year of intervention, on the basis of information gathered using the ‘parent report’, both average SSB consumption and average SSB servings were lower for children in the intervention group than for children in the control group. On the basis of information gathered using the ‘child report’, no significant differences in average SSB intake (in litres or servings) were found between children in the intervention and control group. An explanation for this discrepancy is lacking, but the lack of effect seen with the ‘child report’ can most likely be attributed to the fact that children are still too young to properly estimate their behaviour. Children’s inability to conceptualize – not only SSB but also the concepts of frequency and averaging – make it debatable whether these young children provide valid responses to food questionnaires that have items covering periods greater than one day –. In addition, research has shown that parents are more prone to reporting socially desirable answers compared to children . This could also partly explain the fact that SSB consumption reported by children was higher than that reported by parents. On the basis of the ‘parent report’, no differences in intervention effect were found between the younger children (grades 2 to 4) and the older children (grades 5 to 7) (p > 0.05, data not shown). The parent-reported SSB consumption is probably more reliable and is supported by similar findings in the observations.
After one year of intervention, the number of children bringing SSB to school was lower in the intervention condition than in the control condition. Although the observations did not measure total daily SSB consumption, merely what children brought along to school for break time, they were the most objective measure of SSB consumption in our study. Furthermore, what children bring along to school is most probably largely dependent on their parents’ decisions.
The stratified analyses performed on the basis of the ‘observation report’ demonstrated that intervention effects are limited to subgroups. Differences in intervention effect were found between the two school pairs. Replication of the study with more clusters is recommended to confirm or reject our findings. Also, the effect of the intervention differed according to caregiver’s educational level in a manner that contradicted our expectations. Because the intervention schools are located in socially more deprived neighbourhoods, we expected to see an intervention effect among children of caregiver’s who have lower levels of education. This contradictory finding could be due to some degree of response bias: we may have had higher responses from caregivers with a higher level of education. It could also be explained by the large group of caregivers with an ‘unknown’ educational level that we found in the ‘observation report’.
A number of studies have been published on interventions that aimed to reduce SSB consumption by promoting water. These studies found similar but smaller intervention effects: for example, Tate et al. found a 80.7 ml decrease in SSB intake after a 6-month intervention and Sichieri et al. found a 55.0 ml SSB decrease after a one-year intervention ,. The study of Muckelbauer et al. found a significant increase in water consumption, but no effects on the consumption of juice or soft drinks were observed after adjustment for ethnic background and baseline intake . Compared with these other studies the intervention effects in our study are thus encouraging.
Although the intervention was aimed at reducing the intake of children’s SSB consumption by promoting the intake of water, water consumption was not an outcome measure of our study. Despite this, we did explore the average intake of water, measured in litres, as reported in the parent and child questionnaires. On the basis of the ‘parent report’, there was a significant overall increase in water intake over time in both the intervention and control groups (respectively p < 0.001 and p = 0.015). However, on the basis of the ‘parent report’ and the ‘child report’ we found the intervention to have no effect on children’s water consumption (p > 0.05; see Additional file 1: Table S2). When we also explored whether the decrease in SSB consumption could be explained by an increase in water intake, we found that children with reduced SSB consumption did not differ in their water consumption at follow-up (p > 0.05; data not shown). These findings correspond with those of Veitch et al. . However, since the mechanisms underlying the decrease in SSB consumption still remain unclear, further research is required.
The fact that we found an effect on SSB consumption does not necessarily imply a decrease in total energy intake or weight gain. However, a number of studies have indicated that a reduction in SSB consumption can have beneficial effects on total energy intake and weight status/BMI. For instance, Daniels and Popkin demonstrated that replacing SSB with water reduced total energy intake, implying less weight gain which may well contribute to preventing overweight . In addition, the study by De Ruyter et al. demonstrated that replacing SSB with sugar-free alternatives resulted in reduced weight gain . We explored the effects of the intervention on child’s BMI and weight status which are shown in Additional file 1: Table S3. Children in the intervention group had a significant higher increase of BMI compared to children in the control group (0.26BMI, 95% CI 0.11;0.40, p = 0.001). According the effect size criteria by Cohen, this can be regarded as a negligible effect (d = 0.03) .
The intervention in our study was a school- and community-targeted intervention, developed using social marketing. Our results suggest that a combined school and community approach may be beneficial for children to successfully develop healthier intake of drinks, supporting Bleich's et al. findings . Furthermore, the use of social marketing meant that it was also possible to aim the intervention at a specific population (i.e. Turkish and Moroccan families) within a specific setting (i.e. socially more deprived neighbourhoods). However, when we explored whether such tailoring of the water campaign specifically to these minorities improved the effects seen among these children, we were unable to detect significant differences in intervention effect between children of Turkish and Moroccan background and children from other ethnic backgrounds (p > 0.05 in all three data sets; data not shown). However, the fact that the intervention had similar effects among all ethnic groups could be an indication that the reach and participation among this hard-to-reach target audience has improved, possibly due to the application of social marketing. We recommend that future studies should include a larger sample to increase the power for detecting behavioural changes within such a varied population.
Strengths and limitations
The main strengths of this study are the setting and the duration (i.e. activities in daily practice at primary schools and in neighbourhoods for over a year). The study’s pragmatic setting means that the effects can be generalized to similar settings. A further strength of this study is that we used observations as well as questionnaires to determine the children’s SSB consumption.
A limitation of this study is the fact that randomization on the individual level was not possible. A further limitation is the small number of clusters (i.e. four), which inhibited multi-level analyses but was countered by adding the ‘school pair’ variable in the analyses. Since the use of self-report questionnaires to assess behavioural change is subject to limitations (e.g. misreporting of behaviour and providing socially desirable answers), we used different methods (i.e. observations and questionnaires) and assessed questionnaires from both parents and children. The non-response of parents to the parent questionnaire (complete case analyses only possible for 35%) is another limitation of this study. Our study included a diverse group of children with different ethnic backgrounds; between the three data reports the child’s ethnic backgrounds differed in distribution. Although no intervention effect of ethnic background and intervention condition was found, the intervention effects should be interpreted and generalized with caution (especially our findings based on the ‘parent report’). We assessed SSB intake ‘on average a day’ with the parent and child questionnaires and observed SSB consumption ‘on a random school day’. Further research is recommended to gain insight into different patterns of the child’s SSB consumption (e.g. on week-days vs. weekend-days). It may be debatable whether some beverages should be in- or excluded from the definition ‘SSB’. We recognize that some beverages may have additional nutritional benefits for children; however, we defined SSB in this study based on the amount of sugar within the beverages. A next step in altering the child’s consumption intake could be to give attention to and differentiate even more between SSB’s with and without nutritional value for the child’s diet. Finally, the water campaign consists of several components that promote water consumption. However, when applying such a multi-component intervention, it remains unclear which intervention activities are essential for obtaining the observed effects. We were unable to gather detailed implementation information as it was impossible to register the delivery of components at an individual level. Further research is therefore needed to understand the pathways of the behaviour changes that seem to have occurred.