Daily fruit and vegetable consumption
The results for the primary outcome of the trial revealed that there was little difference in children’s mean change in fruit, vegetables or combined fruit and vegetable intake between the two groups. The Teacher-led group had slightly higher mean intakes for vegetables and combined fruit and vegetables than the Royal Horticultural Society (RHS)-led group; however there was no significant intervention effect after taking into consideration adjustment for confounders. There was also no intervention effect between the Teacher-led group and the RHS-led group for fruit, vegetables or combined fruit and vegetable intake at school only or at home only. Published studies have measured the relationship between children’s fruit and vegetable intake and a gardening intervention ,– although results are inconsistent. Out of the five studies, two reported a significant difference in fruit and vegetable intake ,; one study  found that boys had significantly higher consumption of fruit and vegetables compared to girls; one  reported a significant increase in vegetable consumption only and one  reported no differences in fruit or vegetable intake (measured separately). The quality of previous studies is variable with some using self-selection to determine which school , or which class  received the intervention. In this trial, the size of the gardening area or degree of existing activities were not requirements and all schools were randomly assigned to each intervention group.
The differences in key nutrients and foods intakes were explored to see if there was an effect of either intervention on mean intakes. Overall there was very little difference in values for key nutrients and foods. The only significant difference was for vitamin C intake although this could not be due to increases in fruit and vegetables consumption. Once the adjustments for confounders were made there was a 13 mg per day difference between the RHS-led and Teacher-led groups, with the Teacher-led group having a significantly higher intake of vitamin C. This may be due to higher fruit drink consumption although this was not tested. Two previous studies , also explored key nutrients, identifying a significant increase in dietary fiber in the gardening intervention group compared to the control group, with McAleese and Rankin  also reporting a significant increase in vitamins A and C. These studies had also identified improvements in fruit and vegetable consumption.
Intervention design, elements and geographic location
The fundamental aim of the RHS school gardening programme was to introduce children to basic gardening skills such as planting, watering, weeding, and harvesting. However, previous successful gardening interventions all involved additional elements in other settings as well as the gardening activities. Three interventions included cooking ,,; two interventions included nutrition education ,. Whereas, for both the RHS-led and the Teacher-led interventions in this study, gardening was only extended into additional curriculum lessons at the school’s discretion. The primary focus of the RHS teaching is to educate children in gardening. A more holistic approach with additional complementary components to gardening might be required to achieve a more pronounced change in children’s fruit and vegetable consumption. When schools integrate gardening activities throughout their curriculum it can have a positive impact on children’s fruit and vegetable intake. Schools need support such as lesson plans demonstrating how to implement nutrition and gardening into different areas within the existing curriculum. One of the additional classes for students in a previous study was an “add a veggie to lunch day” . These types of activities have shown positive results in improving children’s fruit and vegetable consumption . It should also be noted that all of the successful gardening interventions were implemented in countries with warmer climates than the UK - California, Minnesota, Alabama and Florida in the US and Newcastle in Australia where fruits and vegetables are grown all year round.
The interventions for this study were either run by the RHS regional advisor or teachers within each school. Of the five successful interventions in prior trials, three of them also used teachers to implement their intervention –. If the classroom teacher was passionate about gardening, then this could potentially assist with favourable implementation of the intervention and result in wide differences in success across schools. In some previous studies ,, the teachers not only taught the intervention but were also trained to complete the dietary assessment which could have introduced bias into the results. Only one published study  had an external company similar to the RHS, the Youth Farmers and Market Project, that implemented their intervention and therefore reduced the risk of bias. A systematic review of a range of school based interventions to increase fruit and vegetable intakes in primary school aged children found a moderate improvement in fruit intake but minimal impact on vegetable intake over the whole day . This suggests that any improvements in intake that may occur during school time are not necessarily maintained throughout the whole day.
Process measures evaluation
The process measures evaluations have provided some evidence to support previous research that an intense school gardening program can improve children’s fruit and vegetable intake. The Royal Horticultural Society has a benchmarking scheme (now called School Gardening Awards) which provides a list of evidence against which schools can rate their gardening activity . If schools are not doing any gardening activity this would be a level 0. The award levels 1 to 5 increase in detail and complexity including aspects of the school culture and ethos; the school garden itself; using gardening in teaching and learning and including the wider school and community in gardening. The results from this study demonstrated that whilst there was overall no significant difference in the primary outcome, when gardening in schools was implemented at an intense level and school gardening engagement substantially increased (measured by an improvement of at least 3 levels), it had a positive association on children’s fruit and vegetable intake. A limitation of the trial is that the RHS scheme involved all schools who wished to participate in the scheme, regardless of existing gardening level. There was no set minimum requirement for garden size or amount of fruit or vegetables grown. Whilst this is a positive for school gardening, it led to greater variability between the schools at baseline and potentially the level of implementation of the intervention.
It is evident that barriers to implementing a school garden program do exist. School gardens require a long term commitment , and a supportive team involved in maintaining the garden over the summer months when the schools are closed is crucial. The intensity as well as the sustainability of the gardening intervention could also affect the success of long term change in children’s fruit and vegetable intake . All of these issues make improving children’s fruit and vegetable consumption a challenging task. The main barriers for increasing children’s fruit and vegetable intake have been identified to be availability, accessibility, convenience, taste preferences, peer pressure, parental/school support and knowledge . The main barrier that teachers’ cite for not implementing school based interventions, is preparation time . In this study the teacher’s continual willingness to engage with the intervention and their own beliefs in the importance of the garden, as well as daily contact with the children, could explain the current findings that, whilst not significant, the Teacher-led intervention tended to have a higher increase in fruit and vegetable consumption compared to the specialist RHS-led intervention.
Limitations and strengths
As with many RCTs evaluating changes in dietary behaviour, there were limitations to this research. One of the disadvantages of the research design was the lack of a comparison group that received no intervention. A second trial linked to this study consisting of 1475 children was conducted with schools from London boroughs adjacent to those in this trial which were randomised to receive either the Teacher-led (n = 756) or the comparison group (n = 719). In that trial the comparison group was a delayed intervention, so that the RHS did not provide any gardening advice to these schools during the course of the trial. The results revealed that the Teacher-led group consumed on average 15 g (95% CI: −36, 148) more fruit and vegetables than the comparison group, however this difference was not statistically significant .
One of the main limitations of previous literature in this area is study design and the use of convenience sampling ,, and therefore a strength of this study is that it is the first cluster RCT to evaluate the effectiveness of a school gardening intervention on children’s diets. However, there are a number of potential sources of bias. The majority of the schools initially contacted did not volunteer to take part in the trial, perhaps because they are offered a range of different school programmes to choose from. This selective response may have led to differences in schools that took part compared with those that did not. In addition, many of the schools in London have a high proportion of children changing schools each academic year, which would have contributed to the high dropout rate (~30%) between baseline and follow-up.
Difficulties in delivery of the intervention and a lack of consistency of delivery may have also led to problems with analysing the effectiveness of the gardening program. Although the RHS follow an established program and aim to spend half a day every six weeks in each school, sowing, growing and harvesting the same fruits and vegetables for every school involved in the program, there may be reasons why this doesn’t happen such as the plants don’t get watered or the weather was inclement. Although efforts were made to hide the intervention group from the fieldworkers the trial was not double blind. Furthermore, schools that dramatically improved their gardening engagement during the programme may be very different in other ways not assessed in this study, Schools that initially had no garden but improved substantially may be different from schools that started with a garden but also improved their gardening engagement.
The results were analysed using a robust statistical methodology, namely multilevel analysis , which has the benefit that the means and confidence intervals for the different foods and nutrients are more accurate. The original plan was to measure the difference in follow-up intake of fruit and vegetables, adjusting for baseline intake. However, due to the skewed distribution of the residuals for fruit and vegetable intake, a change score was calculated .
A further limitation was that the sample size at baseline was lower than planned. Small sample sizes reduce the power to detect a statistical difference between groups and can lead to an overestimation of the standard errors. Nevertheless, this trial is the largest trial to evaluate school gardening to date. Furthermore, the current trial involved a highly diverse population in terms of ethnicity and socio-economic groups. Interventions targeting children living in more deprived areas have the potential to reduce inequalities in health, however the results from this trial which included many children in low socio-economic groups was not more effective than programmes with a more general appeal.
The dietary data was collected using CADET; a validated 24-hour food tick list for children aged 3–11 years old ,. The strength of the CADET diary is that it uses age and gender specific food portion sizes to calculate food and nutrient intake. A one-day tick list is a less burdensome  and effective way of gathering nutrient information from children. However, the disadvantage of using a 24-hour tick list questionnaire is that it uses pre-allocated portion sizes for each food item which are based on average weighed intakes from UK children . This method may not reflect true nutrient intake in the longer term. This study attempted to improve the quality of the dietary data by providing parents and children with an instruction DVD to help explain how to complete the CADET Home Food Diary.
Recommendations for future research
Despite the lack of evidence of a quantitative impact of school gardening on children’s dietary behaviour reported here, the literature often describes positive attributes of school gardening identified through qualitative methods. When a school garden is successfully integrated into the school environment, it can provide a link between the community and the school. The RHS believes that school gardening can provide vital links to members of the community who otherwise have little involvement with their child’s education . This is supported in academic literature , but was not assessed here.
The results from this study suggest that gardening alone, delivered at a low level of intensity, will not increase children’s fruit and vegetable consumption; however, an intense gardening programme, using a holistic approach and incorporating additional related activities with parental involvement has the potential to improve children’s fruit and vegetable intake. The authors suggest that engaging, high quality gardening interventions that also incorporate additional components such as educational activities, visits to farms and cooking programmes should be introduced into schools. The benefits of gardening programmes may be far reaching and therefore not evident for many years making evaluation difficult. Further research evaluating gardening schemes should plan follow up of more than one year to also take into account problems with weather that may reduce harvesting and productivity. Programmes could also extend to community gardens in the school vicinity that parents can get involved with. This would ensure that interventions tackle individual intake, family intake, the community as well as the school environment . The World Health Organisation and Food and Agriculture Organisation believe that school based interventions play a fundamental role in improving the population’s fruit and vegetable consumption . However, to improve children’s fruit and vegetable intake schools need support from the food industry and government to improve access and cost of fruit and vegetables in all settings in which children spend time.