School Gardening Activities Aimed at Obesity Prevention Improve Body Mass Index and Waist Circumference Parameters in School-Aged Children: A Systematic Review and Meta-Analysis

Abstract

Background:

Childhood obesity, due to its increasing prevalence, is one of the most challenging problems in public health and prevention. Recent strategies have been implemented to tackle this trend, including school gardening projects. This systematic review analyzes the main elements of school gardening with a specific meta-analysis about its impact on anthropometric parameters.

Materials and Methods:

We searched studies in PubMed, EMBASE, and Cochrane Library on school gardening projects carried out until February 2019 and addressed to children aged 6 to 13 years.

Results:

Thirty-three studies passed the screening selection. Outcomes analyzed were fruits and vegetables consumption/knowledge, nutritional attitudes and behavior, anthropometric outcomes (i.e., BMI, waist circumference—WC) and other outcomes (i.e., blood pressure, science achievement, and physical activity). We performed descriptive analyses of these outcomes and found a general benefit from school gardening projects. We subsequently performed a meta-analysis on anthropometric outcomes, highlighting a significant reduction of both BMI percentile (%) (−1.37%) and WC (−1.30 cm).

Conclusions:

Although more high-quality studies are needed on this topic, along with a greater homogeneity of anthropometric measurements, the results of our article show a modest but positive impact of school gardening projects on anthropometric measures and more generally on children's health.

Introduction

Childhood obesity is a public health priority and its prevalence and economic burden have been steadily increasing worldwide.^1,2 The prevalence of childhood obesity ranges between 21% and 24%, affecting about 43 million children.^3,4 Evidence shows that childhood obesity is strongly associated with adult obesity and that it is the best predictor of morbidity and mortality in adulthood.^5,6 Furthermore, overweight and obese children can experience discrimination, alongside with higher levels of psychological and physical abuse by schoolmates, compared to other children.⁷

Public institutions have developed and implemented several strategies to prevent obesity in children, including programs aimed at promoting healthy eating habits through the increase of fruit and vegetable (F/V) intake, improvement of food education, and involvement in gardening activities.⁸ Studies have shown that children who learn healthy eating habits at a young age have a higher likelihood to maintain these habits through their life, compared to interventions carried out in adulthood.⁹ Other studies reported that family involvement is a key factor for children's eating habit improvement, and school can be a favorable environment to engage children and families.¹⁰ Recent evidence showed that multicomponent programs, including diet and educational activities, are more effective than programs exclusively focused on diet in changing children's eating attitudes. These approaches combine nutritional education programs, possibly with information on seasonality and local agricultural products, school gardening, and promotion of physical activity.^11–13

BMI, expressed as weight/height² (kg/m²), and waist circumference (WC), measured in centimeters, are commonly used to measure anthropometric characteristics and stratify the risk for overweight and obesity in adults and children.¹⁴ BMI and WC are easily obtainable and have proved to be strong predictors of metabolic syndrome, type 2 diabetes, and cardiovascular disease in adults and children.^15,16 Yet, the impact of school gardening interventions on anthropometric parameters, which have shown to predict morbidity and mortality in adulthood, has been poorly studied. Previous literature reviews about this topic focused on different outcomes related only to nutritional behavior, dietary intake, and academic achievement. On the contrary, this is the first study focused on covering the gap in knowledge in the scientific literature regarding changes in anthropometric outcomes related to school gardening programs.

We undertook a systematic review (SR) on the impact of school gardening projects on school-aged (6–13 years) children's data about F/V intake and knowledge; science achievement; physical activity levels; blood pressure; and markers in blood samples. We systematically reviewed and meta-analyzed the impact of school gardening projects on school-aged (6–13 years) children's anthropometric parameters, gathering information from experimental studies on BMI and WC changes in after-school gardening interventions. This work can provide clinicians and policy makers with robust evidence on the efficacy of tools aimed at increasing children's knowledge and engagement in practical/educational activities, and ultimately improving their knowledge about healthy habits and lifestyle, and health outcomes.

Materials and Methods

Search Strategy

Databases and search criteria

We performed a systematic search and review according to PRISMA guidelines.¹⁷ We searched primary (PubMed, EMBASE™) and secondary (Cochrane Library) electronic databases to identify relevant studies for the period between January 1946 and February 2019. We developed a specifically designed search string for this purpose: [(school garden* OR school gardening) AND (obes* or overweig* OR obesity OR overweight)]. We searched gray literature for other documents and performed a hand search of key journals, professional organizations' websites and guideline clearing houses. Moreover, we scanned reference lists of selected articles to find additional articles. We applied snowball (such as pursuing references of references and electronic citation tracking) and hand search techniques (such as scanning reference lists of selected reviews, primary articles and textbooks, and scanning for other documents in selected journals and guideline clearing houses) to the search strategy previously described.¹⁸

Patients, interventions, comparators, and outcomes question

The following patients, interventions, comparators, and outcomes (PICO) question addressed our primary search objective through the following search terms: (P) children attending school, aged 6–13 years, (I) school gardening projects for obesity prevention, (C) only nutritional education or no intervention, and (O) postintervention anthropometric measures, F/V knowledge, F/V consumption, and other outcomes such as blood pressure, science achievement, physical activity levels, and blood samples about metabolic syndrome markers.

Study Selection

Design

We included randomized controlled trials (RCTs), quasi-experimental, and observational studies whose primary intervention consisted in school gardening projects for primary school students (age range between 6 and 13 years). We excluded studies about preschool-aged children and adolescents. We included any outcome concerning anthropometric measurements, F/V consumption or knowledge and other outcomes such as blood pressure, science achievement, physical activity, and blood samples about metabolic syndrome markers. We excluded protocols as well as systematic or narrative reviews. We included in the meta-analysis only RCTs and quasi-experimental studies.

Data selection and extraction

Two reviewers (A.R. and D.G.) independently identified potentially eligible studies from title and abstract reading. The same authors assessed full texts of relevant studies included and consulted the references of the previously published primary and secondary articles to search for additional relevant articles. We resolved any disagreement through discussion and reaching a consensus between the two readers. When we did not reach any consensus, we requested an opinion of a third author (A.U.), which acted as an “in blind” final arbiter.

We used the same methodology for data extraction performed independently by two reviewers (A.R., D.G.), with the supervision of a third author (D.T.). We analyzed data using Microsoft Office Excel 2013 (Microsoft Corporation) and Stata (StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP).

Interventions

We included school gardening projects that involved gardening activities (with/without a nutritional education program). We excluded projects only based on theoretical lessons or on other activities that did not include school gardening. We considered interventions performed during school hours, after-school hours, or during summer school camps. We included any type of intervention addressed to scholars and those carried out in different school groups and with different components. We excluded interventions developed only in community gardens settings (i.e., addressed to the whole community) without a specific involvement of the schools.

Study Quality Assessment

We used the Cochrane Tool for Quality Assessment and the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) to assess study quality of the experimental and observational original studies included in this SR.

The Cochrane Tool allows analyzing seven bias groups: sequence generation and allocation concealment (both within the domain of selection bias or allocation bias), blinding of participants and personnel (performance bias), blinding of outcome assessors (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and an auxiliary domain, “other bias.”¹⁹ For each bias group, it is possible to assign a value of “high,” “low,” or “unclear” risk of bias when it is not specified if a specific bias is present or not. Every bias judgment helps to assign a global assessment to every RCT according to the Agency for Health care Research and Quality (AHRQ) standards (good, fair, and poor).²⁰

The STROBE statement is a 22 items tool specifically designed to evaluate observational studies quality.²¹ Items are associated with the different sections of an article, such as title and abstract (item 1), introduction (items 2 and 3), methods (items 4–12), results (items 13–17), discussion (items 18–21), and other information (item 22 on funding). Eighteen items are identical for the three different study designs, while four (specifically items 6, 12, 14, and 15) are differentially designed for each study type (i.e., cohort or case control).

STROBE does not provide a scoring stratification. As a general rule, the higher the score, the higher the quality of the study. We hence created three score thresholds corresponding to three levels of quality: 0–14 as poor quality, 15–25 as intermediate quality, and 26–33 as good quality of the study.

Statistical Analysis

We described the main descriptive characteristics of the studies in the SR, focusing on the following key-points: country, study design, type, duration and time of intervention, and outcomes. Subsequently, we performed a meta-analysis on the anthropometric measures that were analyzed by the included studies: BMI, BMI percentile (BMI%), BMI z-score, WC, and diastolic and systolic blood pressure. We decided not to analyze quantitatively F/V intake or knowledge or school/academic performances since we found very heterogeneous and inconsistent measurements. We compared values between intervention and control groups in those studies, which investigated these parameters both pre- and postintervention.

We analyzed statistical heterogeneity to define if it was feasible to sum the results of the different studies considered eligible for the meta-analysis. We studied heterogeneity by the used graphic funnel plots and by calculating the I² statistic, which represents the percentage of the variance in effect estimates that is caused by heterogeneity rather than by sampling bias (chance). An I² statistic greater than 50% was considered significantly heterogeneous.

Where the number of studies was less than five or studies were substantially heterogeneous, we utilized a random-effects model in accordance with the Cochrane Handbook for Systematic Reviews of Interventions.²² We followed the method of DerSimonian and Laird to compute the random-effects estimates for the corresponding statistics.²³ If standard deviation was available for all the studies, we calculated the standard error and the variance for the meta-analysis. We chose to use the forest plots to show graphically effect estimates with 95% confidence intervals (CIs) for individual trials and pooled results. We carried out meta-analyses using Cochrane's RevMan (Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014).

Results

Search Findings

Our search yielded 203 studies (133 by databases browsing, 4 by hand search, and 66 from other SR bibliography), 33 of which were appropriate for review,^24–56 and 5 for meta-analysis.^{26,27,29–31} We excluded 140 studies, which did not meet the eligibility criteria or were duplicated. Figure 1 shows the flowchart with the study selection process.

Figure 1.

PRISMA flow diagram, which shows the process for studies selection. Flowchart designed according to: Moher et al.¹⁷

Characteristics of the Included Studies

The geographic origin of the studies was United States (n = 25, 75.8%),^{24–27,30–34,36–39,41,44–46,48–51,53–56} Australia (n = 3, 9.1%),^35,42,52 United Kingdom (n = 2, 6.1%),^28,40 Portugal (n = 1, 3%),²⁹ Netherlands (n = 1, 3%),⁴⁷ and Canada (n = 1, 3%).⁴³ Time interval for the school gardening program ranged between 2.5 months³⁵ and 2 school years.^31,53

The intervention sessions lasted approximately an hour each week, with the shortest length reported of 20 minutes^44,45 and the longest of 3 hours for week.³⁵ Sample sizes were heterogeneous and ranged between two classrooms (about 30 students)³⁸ and multiple campus schools (about 3000 students).³¹

The interventions took place in different settings and timings: school hours (n = 25; 75.8%),^{24–26,28,29,31–36,39,40,42,43,47–56} after-school (n = 5; 15%),^{27,30,37,38,45} summer school camps (n = 2; 6.1%),^44,46 and either during school hours or after-school (n = 1; 3.1%).⁴¹ Twenty-seven (81.8%)^{24–28,30–34,37,39–48,50,52–56} school interventions were designed for a single group of students, while in the remaining six studies (18.2%) ^{29,35,36,38,49,51} the students were divided in two groups: one implementing theoretical lessons only and one with an intervention including both theoretical lessons and practical gardening activities.

Four studies^26,31,41,53 described multicomponent interventions which integrated nutritional lessons and school gardening practice with other activities such as physical activity promotion, cooking class workshops, school menus implementation with healthy foods, and educational farms guided tours with the classes. In two studies, due to different schools facilities, not all students included in the educational programs were exposed to the intervention components in the same way.^41,53 We included in the SR: 6 RCTs (18.2%),^24–29 17 observational studies (51.5%),^40–56 and 10 quasi-experimental studies (30.3%).^30–39

Also of note, in two studies authors adopted the same study design, but reported different outcome results in each study.^24,25 Twenty-three studies (69.7%) had a control group,^{24–27,29–37,39–42,47,49,51,52,55,56} while 10 studies (30.3%) did not.^{28,38,43–46,48,50,53,54} In one study, the two intervention groups had the same components, but were carried out by different staff members (university staff vs. school teachers).²⁸

Only four studies monitored research outcomes after the intervention, with a follow-up period of 1 month,³⁵ 6 months,^28,36 and 1 year,⁴⁷ respectively. One quasi-experimental study, conducted by Hollar et al., consisted of two cohorts in two different school years.³¹ Therefore, we carried out the analyses considering school years separately. In the latter study, the authors investigated outcomes on the whole cohort study and a subgroup of low-income children accessing free/reduced price meals. Authors of a similar study analyzed subgroups containing overweight/obese children, and the pre–post intervention analyses were conducted both on the entire sample and on the subgroup.³⁰

Characteristics of the Outcomes Assessed

We found outcomes related to three main categories: F/V consumption/knowledge and nutritional behavior in 27 studies out of 33 (81.8%),^{24,26–28,30,32,34–54} anthropometric measures in 6 studies out of 33 (18.2%),^{26,27,29–31,40} and other outcome measures (i.e., physical activity, academic performance, blood pressure, and laboratory assessments like glucose, cholesterol, and triglycerides) in 11 studies out of 33 (33.3%).^{25–27,29–31,33,35,45,55,56} Out of 33 studies, 8 studies (24.2%) analyzed more than a single outcome category.

F/V-Related Outcomes

The outcomes related to F/V dietary habits were the following: F/V consumption, F/V knowledge, F/V dietary attitudes and behavior, and F/V home availability. In six study protocols, parents' collaboration was required to assist children to complete home food diary²⁸ or to fill out questionnaires about family demographic and family nutritional choices.^{26,34,42,44,53} Five studies focused on outcomes related to vegetables only (i.e., vegetable knowledge or intake) in school gardening implementation projects.^{32,39,45,47,54}

F/V Intake

Twenty studies analyzed F/V consumption through different methods: questionnaires (n = 12; 60%),^{26,27,30,34,37–43,45} food diary measuring food habits (compiled at home and/or at school: n = 5; 25%),^{28,35,48,49,53} and direct observation in the lunchrooms (n = 3; 15%).^32,51,54 In six schools, during the intervention, lunchroom menus were implemented with locally produced salads or vegetables or either vegetables harvested in school gardens.^{26,32,39,41,51,54} Five studies observed a positive impact of school gardening on combined F/V intake.^{34,40,41,49,53} Five studies measured an increased vegetable consumption,^{32,39,45,51,54} while one study observed a reduction of vegetable consumption in the intervention group although the reduction was stronger in the control group.²⁷ Three studies^27,30,49 showed an increase of fiber intake. Furthermore, one study found an increase of vitamin A and vitamin C intake through the results of F/V food diary.⁴⁹ Finally, in one study, the parental questionnaire analyses showed a positive significant relationship between the implementation of the intervention and the amount of F/V cooked at home.³⁴

F/V Knowledge

Seventeen studies examined F/V knowledge. Questionnaires assessed children's knowledge of basic nutritional principles^{26,28,34–38,41,43,46,47,50,51,53} or capacity for F/V identification^26,39,47,52 or capacity to describe food.⁴² Furthermore in two studies, questions concerned knowledge of locally grown F/V.^41,52 In 13 studies, the authors reported a knowledge increase^{26,28,35,36,39,41,43,46,47,50–53} in the intervention vs. the control group after the implementation of gardening projects, while the remaining studies did not show significant differences among groups.

F/V Dietary Attitudes and Behaviors

Twenty studies analyzed F/V dietary attitudes and behaviors related to intentions, attitudes, norms, and perceived behavioral control. Outcomes were related to general nutrition and F/V consumption,^{40,44,46,47,52,53} F/V preferences,^{26,35–39,41,43,44,46–48,50,51,53} “willingness to try new F/V,”^{28,34,35,39,42,51,52} and gardening self-efficacy.^37,41,44,52 The nutritional behavioral attitudes had a positive impact in two studies.^40,44 Four studies found a significant improvement in “willingness to try new F/V” outcomes in the intervention group^34,35,39,42 while other two studies showed a negative association, but only related to new type of fruit consumption.^28,52 F/V preference increased after intervention in two studies,^50,51 while preference for only vegetables increased in six studies^{35,36,39,44,47,48} and preference for only fruit in a single study.⁴³

In five studies, an improved self-reported preference for healthy F/V-based snacks instead of unhealthy foods was found.^{34,35,48,50,52} In one study, gardening self-efficacy was found significantly different in intervention vs. control group.⁴¹ Two studies researched F/V home availability,^24,44 but only one study observed an increase in F/V presence at home, especially for low-fat vegetables.²⁴

Anthropometric Outcomes

Studies that included analyzing anthropometric outcomes used nine different anthropometric measures: absolute BMI, BMI%, BMI z-score, weight z-score, weight, waist to height ratio, WC, body fat %, and height. Five out of six studies (83.3%) used multiple anthropometric measurements. A significant absolute BMI reduction was found in one study,³⁰ although the authors observed a change in measure in the overweight/obese subgroup only. Hollar et al.³¹ found a significant BMI% reduction in the second year of follow-up only, while at the first year, it was not significant. In the latter study, a significant BMI z-score and weight z-score reduction related to the low-income children subgroup was found. The results of other two studies showed a significant reduction in BMI z-score: in one study, the BMI z-score reduction was associated with a WC reduction,²⁷ while in the second study, the BMI z-score improvement was associated with a BMI% and waist to height ratio reduction.²⁶

Other Outcomes

Other outcome measures considered in the studies were physical activity (n = 2; 18.2%), measured by questionnaires,^25,45 direct observation and use of accelerometers,²⁵ skills in school subjects (n = 5; 50%), including science,^26,33,55,56 math,^31,55 and reading,³¹ and other laboratory measures (n = 4; 40%),^27,29–31 that is, glucose, triglycerides, total cholesterol, high-density lipoprotein, and low-density lipoprotein, insulin (fasting blood samples), blood pressure, Homeostatic Model Assessment of Insulin Resistance (HOMA index), urinary sodium, and urinary creatinine. Finally, in one study²⁹ authors estimated salt intake from 24-hours urinary sodium excretion, while other authors measured the quality of school life, through a survey about student attitudes toward school, learning, and teachers and about other students' relationships.³⁵

Physical activity increased in the intervention groups in both studies that analyzed it.^25,45 Two studies found an increase in science achievement scores,^33,56 while math scores improved in another study,³¹ specifically in the low-income children subgroup. In one last study, math skills upgraded in the control group compared to intervention group.⁵⁵ Two studies observed a significant reduction of blood pressure,^30,31 while other laboratory measures (i.e., blood and urinary samples) showed no significant results in two studies.^27,29 Authors observed reduced salt intake after intervention,²⁹ while quality of school life showed no significant changes.³⁵ Table 1 shows the characteristics of the studies included in this SR, classified according to the three main outcomes explained in the text.

Table 1.

Characteristics of the Included Studies

First author (year)	Country	Study design	Age (N)	Intervention characteristics (duration – frequency)	Control	Outcomes	Results
F/V outcomes	Country	Study design	Age (N)	Intervention characteristics (duration – frequency)	Control	Outcomes	Results
Cotugna et al. (2012)³²	United States	Quasi-experimental	4th and 5th grade students (359)	School gardening activities combined with theoretical lessons about nutrition during school hours (n/a)	No intervention	V consumption	↑ V consumption (INT vs. CTR)
Davis et al. (2011)³⁰	United States	Quasi-experimental	9–11-year-old students (104)	School gardening activities combined with theoretical lessons about nutrition in after-school time (12 weeks – 90 minutes/week)	No intervention	F/V consumption	↑ Fiber consumption (INT vs. CTR)
Duncan et al. (2015)⁴⁰	United Kingdom	Observational	8–9-year-old students (77)	School gardening activities combined with theoretical lessons about nutrition during school hours (12 weeks – 30 minutes/week)	No intervention	F/V consumption and dietary attitudes and behaviors	↑ F/V consumption and ↑ dietary attitudes and behaviors (INT vs. CTR)
Evans et al. (2012)⁴¹	United States	Observational	6th and 7th grade students (214)	Multicomponent intervention: in-class lessons, after-school gardening program farm-to school, farmers' visits to schools, taste testing, field trips to farms. (5 months – n/a)	No intervention	F/V knowledge, consumption and dietary attitudes and behaviors	↑ F/V knowledge (INT vs. CTR) ↑ F/V knowledge, ↑ F/V consumption, ↑ self-efficacy and ↓ preference for unhealthy foods (INT with ≥2 components vs. INT with 1 component)
Gatto et al. (2015)²⁷	United States	RCT	7–11-year-old students (319)	School gardening activities combined with theoretical lessons about nutrition in after-school time (12 weeks – 90 minutes/week)	No intervention	F/V consumption	↑ Fiber consumption ↓ V consumption (INT vs. CTR)
Gibbs et al. (2013)⁴²	Australia	Observational	From 3rd to 6th grade students (592)	School gardening activities during school hours (gardening + kitchen class) (1 year – 45–60 minutes garden +90 minutes kitchen activities/week)	No intervention	F/V knowledge, consumption and dietary attitudes and behaviors	↑ willingness to try new foods (INT vs. CTR)
Hanbazaza et al.(2018)⁴³	Canada	Observational	From 1st to 6th grades students (66)	School gardening activities during school hours (2 school years – n/a)	No control group	F/V knowledge, consumption and dietary attitudes and behaviors	↑ F/V knowledge and ↑ F preference (baseline vs. post intervention)
Heim et al. (2009)⁴⁴	United States	Observational	From 4th to 6th grade students (93)	Gardening activities combined with theoretical lessons about nutrition at summer camps (12 weeks – 20/30 minutes/week)	No control group	Dietary attitudes and behaviors and F/V home availability	↑ V preference, ↑ F/V asking behavior at home and ↑ number of F/V ever eaten (baseline vs. post intervention)
Hermann et al. (2006)⁴⁵	United States	Observational	From 1st to 8th grade students (n/a)	School gardening activities combined with theoretical lessons about nutrition in after-school time (n/a – 90 minutes 5 days/week)	No control group	V consumption	↑ V consumption (baseline vs. post intervention)
Hutchinson et al. (2015)²⁸	United Kingdom	RCT	7–10-year-old students (1256)	2 intervention groups during school hours: (a) school gardening activities led by experts of horticulture institution combined with theoretical lessons about nutrition (b) same intervention activities led by teachers during school hours (1 school year – n/a)	No control group	F/V knowledge, consumption and dietary attitudes and behaviors	↑ V knowledge and ↓ F preference and ↓ willingness to try new F (INTa vs. INTb)
Koch et al. (2006)⁴⁶	United States	Observational	From 2nd to 5th grade students (56)	Gardening activities combined with theoretical lessons about nutrition at summer camps (variable, from 1 to 12 weeks – n/a)	No control group	F/V knowledge and dietary attitudes and behaviors	↑ F/V knowledge (baseline vs. post intervention)
Leuven et al. (2018)⁴⁷	Netherlands	Observational	10–12-year-old students (215)	School gardening activities combined with theoretical lessons (7 months a year for 2 school year – 2 hours/week)	No intervention	V knowledge and dietary attitudes and behaviors	↑ V knowledge and ↑ V preference (INT vs. CTR)
Lineberger and Zajicek (2000)⁴⁸	United States	Observational	3rd and 5th grade students (111)	School gardening activities combined with theoretical lessons about nutrition during school hours. (3 months a year for 2 school years – n/a)	No control group	F/V consumption and dietary attitudes and behaviors	↑ V preference and ↑ F/V snack preference (baseline vs. post intervention)
McAleese and Rankin LL (2007)⁴⁹	United States	Observational	10–13-year-old students (99)	2 intervention groups during school hours: a) only theoretical lessons on nutrition; b) theory and school gardening activities. (12 weeks – n/a)	No intervention	F/V consumption	↑ F/V, fiber, Vitamin A and Vitamin C consumption (INTb vs. INTa + CTR)
Meinen et al. (2012)³⁴	United States	Quasi-experimental	7–13-year-old students (801 + 995 parents)	School gardening activities during school hours (4 months – n/a)	No intervention	F/V knowledge, consumption and dietary attitudes and behaviors	↑ F/V consumption, ↑ willingness to try new F/V, ↑ F/V snack preference and ↑ F/V parent consumption (INT vs. CTR)
Morgan et al. (2010)³⁵	Australia	Quasi-experimental	11–12-year-old students (127)	2 intervention groups during school hours: a) only theoretical lessons on nutrition; b) theory and school gardening activities. (10 weeks – 3 lessons × 1 hour +180 minutes gardening activities/week)	No intervention	F/V knowledge, consumption and dietary attitudes and behaviors	↑ V knowledge, ↑ V preference,,↑ willingness to try new F/V and ↑ V snack preference (INTb vs. INTa)
Morris and Zidenberg-Cherr (2002)³⁶	United States	Quasi-experimental	9–10-year-old students (213)	2 intervention groups during school hours: a) only theoretical lessons on nutrition; b) theory and school gardening activities. (1 school year – n/a)	No intervention	F/V knowledge and dietary attitudes and behaviors	↑ F/V knowledge and ↑ V preference (INTb vs. INTa)
Nolan et al. (2012)⁵⁰	United States	Observational	From 2nd to 5th grade students (141)	School gardening activities combined with theoretical lessons about nutrition during school hours (n/a)	No control group	F/V knowledge and dietary attitudes and behaviors	↑ F/V knowledge, ↑ F/V preference and ↑ F/V snack preference (baseline vs. post intervention)
O'Brien and Shoemaker (2006)³⁷	United States	Quasi-experimental	4th grade students (38)	School gardening activities combined with theoretical lessons about nutrition in after-school time (10 weeks – 80 minutes/week)	No intervention	F/V knowledge, consumption and dietary attitudes and behaviors	No significant results (INT vs. CTR)
Parmer et al.(2009)⁵¹	United States	Observational	2nd grade students (115)	2 intervention groups during school hours: a) only theoretical lessons on nutrition; b) theory and school gardening activities. (6 months – theory lesson 1 hour/week + gardening activities 1 hour/week)	No intervention	F/V knowledge,V consumption and dietary attitudes and behaviors	↑ F/V knowledge and ↑ F/V preference (INTa + INTb vs. CTR) ↑ V consumption (INTb vs. INTa + CTR)
Poston et al. (2005)³⁸	United States	Quasi-experimental	From 3rd to 5th grade students (29)	2 intervention groups in after-school time: a) school gardening activities combined with theoretical lessons about nutrition in after-school time (3 months – 8 lessons 20–60 minutes) b) theoretical lessons on nutrition in after-school time (5 lessons 30–60 minutes)	No control group	F/V knowledge, consumption and dietary attitudes and behaviors	No significant results (INTa vs. INTb)
Ratcliffe et al. (2009)³⁹	United States	Quasi-experimental	11–13-year-old students (320)	School gardening activities combined with theoretical lessons about nutrition during school hours (4 months – 1 hour/week: 20 minutes theory lesson +40 minutes gardening activities)	No intervention	V knowledge, V consumption and dietary attitudes and behaviors	↑ V knowledge, ↑ V preference, ↑ willingness to try new V and ↑ V consumption (INT vs. CTR)
Scherr et al. (2017)²⁶	United States	RCT	9–10-year-old students (409)	Multicomponent intervention during school hours: (1) nutrition education and promotion, (2) family and community partnerships, (3) supporting regional agriculture, (4) foods available on the school campus, and (5) school wellness committees and policies. (1 school year – n/a)	No intervention	V knowledge, F/V consumption and dietary attitudes and behaviors	↑ V knowledge (INT vs. CTR)
Somerset and Markwell (2008)⁵²	Australia	Observational	8–13-year-old students (252)	School gardening activities during school hours (1 school year – n/a)	No intervention	F/V knowledge and dietary attitudes and behaviors	↑ F/V and locally grown products knowledge, ↑ F/V snack preference and ↓willingness to try new F (INT vs. CTR)
Wang et al. (2010)⁵³	United States	Observational	4th and 5th grade students (327)	Multicomponent intervention during school hours: gardening activities, healthy school lunch menus and education in nutrition, health, and the environment (organized in High intervention and Low intervention depending on the implementation degree at the schools) (2 school years – n/a)	No control group	F/V knowledge, consumption and dietary attitudes and behaviors	↑ F/V knowledge and ↑ F/V consumption (High degree INT vs. Medium and Low INT)
Wells et al. (2018)²⁴	United States	RCT	2th,4th and 5th grade students (2768)	School gardening activities combined with theoretical lessons about nutrition during school hours (2 years – n/a)	No intervention	F/V home availability	↑ V home availability (INT vs. CTR)
Wright and Rowell (2010)⁵⁴	United States	Observational	Students of an elementary school (n/a)	School gardening activities combined with theoretical lessons about nutrition during school hours (1 school year – n/a)	No control group	V consumption	↑ V consumption (baseline vs. post intervention)

First author (year)	Country	Study design	Age (N)	Intervention characteristics (duration)	Control	Outcomes	Results
Anthropometric outcomes	Country	Study design	Age (N)	Intervention characteristics (duration)	Control	Outcomes	Results
Cotter et al. (2013)²⁹	Portugal	RCT	10–12-year-old students (139)	2 intervention groups during school hours: a) regular lectures on the potential dangers of excessive salt intake; b) theory and school gardening activities (aromas plants). (6 months – 2 hours/week)	No intervention	Height, BMI, WC	No significant results (INT vs. CTR)
Davis et al. (2011)³⁰	United States	Quasi-experimental	9–11-year-old students (104)	School gardening activities combined with theoretical lessons about nutrition in after-school time (12 weeks – 90 minutes/week)	No intervention	Height, weight, BMI, BMI%, BMI z-score, WC, total fat %	↓ BMI (overweight/obese subgroup vs. CTR)
Duncan et al. (2015)⁴⁰	United Kingdom	Observational	8–9-year-old students (77)	School gardening activities combined with theoretical lessons about nutrition during school hours (12 weeks – 30 minutes/week)	No intervention	BMI	No significant results (INT vs. CTR)
Gatto et al. (2015)²⁷	United States	RCT	7–11-year-old students (319)	School gardening activities combined with theoretical lessons about nutrition during school hours (12 weeks – 30 minutes/week)	No intervention	BMI%, BMI z-score, WC, Body fat %	↓ BMI z-score and ↓ WC (INT vs. CTR)
Hollar et al. (2010)³¹	United States	Quasi-experimental	Elementary school students (3769)	Multicomponent intervention during school hours: healthy school lunch menus, theoretical lessons about nutrition + gardening activities, physical activity promotion and wellness projects (2 school years – n/a)	No intervention	BMI%, BMI z-score, weight z-score	↓ BMI% (INT vs. CTR in the second school year) ↓ BMI z-score and ↓ weight z-score (low-income subgroup vs. CTR)
Scherr et al. (2017)²⁶	United States	RCT	9–10-year-old students (409)	Multicomponent intervention during school hours: (1) nutrition education and promotion, (2) family and community partnerships, (3) supporting regional agriculture, (4) foods available on the school campus, and (5) school wellness committees and policies. (1 school year – n/a)	No intervention	BMI%, BMI z-score, waist to height ratios	↓ BMI%, ↓ BMI z-score and ↓ waist to height ratios (INT vs. CTR)

First author (year)	Country	Study design	Age (N)	Intervention characteristics (duration)	Control	Outcomes	Results
Other outcomes	Country	Study design	Age (N)	Intervention characteristics (duration)	Control	Outcomes	Results
Cotter et al. (2013)²⁹	Portugal	RCT	10–12-year-old students (139)	2 intervention groups during school hours: a) regular lectures on the potential dangers of excessive salt intake; b) theory and school gardening activities (aromas plants). (6 months – 2 hours/week)	No intervention	Blood pressure Urinary samples (sodium + creatinine) Estimated salt intake	↓ salt intake (INTb vs. INTa + CTR)
Davis et al. (2011)³⁰	United States	Quasi-experimental	9–11-year-old students (104)	School gardening activities combined with theoretical lessons about nutrition in after-school time (12 weeks – 90 minutes/week)	No intervention	Blood pressure	↓ Diastolic blood pressure (INT vs. CTR)
Gatto et al. (2015)²⁷	United States	RCT	7–11-year-old students (319)	School gardening activities combined with theoretical lessons about nutrition in after-school time (12 weeks – 90 minutes/week)	No intervention	Blood pressure Blood samples (cholesterol total and LDL + HDL, triglycerides, insulin and glucose)	No significant results
Hermann et al. (2006)⁴⁵	United States	Observational	From 1st to 8th grade students (n/a)	School gardening activities combined with theoretical lessons about nutrition in after-school time (n/a – 90 minutes 5 days/week)	No control group	Physical activity	↑ physical activity (baseline vs. postintervention)
Hollar et al. (2010)³¹	United States	Quasi-experimental	Elementary school students (3769)	Multicomponent intervention during school hours: healthy school lunch menus, theoretical lessons about nutrition + gardening activities, physical activity promotion and wellness projects (2 school years – n/a)	No intervention	Blood pressure Academic performance (math + reading)	↓ systolic and diastolic blood pressure (INT vs. CTR) ↑ math scores (low income subgroup vs. CTR)
Klemmer et al.(2005)³³	United States	Quasi-experimental	3rd, 4th and 5th grade students (647)	School gardening activities during school hours (1 school year – n/a)	No intervention	Academic performance (science)	↑ science achievement (INT vs. CTR)
Morgan et al. (2010)³⁵	Australia	Quasi-experimental	11–12-year-old students (127)	2 intervention groups during school hours: only theoretical lessons on nutrition; theory and school gardening activities. (10 weeks – 3 lessons × 1 hour +180 minutes gardening activities/week)	No intervention	Quality of school life	No significant results
Pigg et al.(2006)⁵⁵	United States	Observational	From 3rd to 5th grade students (196)	School gardening activities combined with theoretical lessons about nutrition during school hours (n/a)	No intervention	Academic performance (math + science)	↓ math scores (INT vs. CTR)
Scherr et al. (2017)²⁶	United States	RCT	9–10-year-old students (409)	Multicomponent intervention during school hours: (1) nutrition education and promotion, (2) family and community partnerships, (3) supporting regional agriculture, (4) foods available on the school campus, and (5) school wellness committees and policies. (1 school year – n/a)	No intervention	Academic performance (science)	No significant results
Smith and Motsenbocker (2005)⁵⁶	United States	Observational	Students of 3 primary schools (119)	School gardening activities combined with theoretical lessons about nutrition during school hours (14 weeks – 2 hours/week for school year (1.5 hours theory +30 minutes gardening))	No intervention	Academic performance (science)	↑ science scores (baseline vs. post intervention in INT group) No significant results (INT vs. CTR)
Wells et al. (2014)²⁵	United States	RCT	2th,4th and 5th grade students (2768)	School gardening activities combined with theoretical lessons about nutrition during school hours (2 years – n/a)	No intervention	Physical activity	↑ physical activity (INT vs. CTR)

BMI%, BMI percentile; CTR, control; F/V, fruits and vegetables; HDL, high-density lipoprotein; INT, intervention; LDL, low-density lipoprotein; n/a, not available; RCT, randomized controlled trial; WC, waist circumference.

Quality Assessment

As previously explained above, we analyzed for quality the two RCTs^24,25 which shared the same study design as one. We used the Cochrane Tool for Quality Assessment to assess 15 studies^24,26–39 (5 RCTs^24,26–29 and 10 quasi-experimental^30–39): 3 scored good quality,^24,26,27 while 12 scored fair quality (details are shown in Table 2).^28–39 About those rated as fair quality, one RCT had a high risk of bias due to the control group's features (identical activities into the intervention group, only carried out by different staff members),²⁸ while all the quasi-experimental studies had a high risk of bias due to absence of randomization.^30–39 Specifically in the latter studies, due to their study design, there was a potential risk of bias (marked in yellow under “other bias” in Table 2). In addition, in the study by Hollar et al.³¹ the outcome data related to blood pressure was not shown point-by-point and there was a graph without clear numerical references. In all the studies, there were no detailed information on allocation concealment; thus representing unclear risk of selection bias for all of the studies analyzed. Similarly, in four out of five RCTs analyzed,^24,27–29 random sequence generation details were not explained. For this reason, we assessed an unclear risk of selection bias. Finally, for one RCT,²⁹ we assessed a potential risk of bias because of the small sample size together with a difficulties in the achievement of 24 hours urine collection in a pediatric population.

Table 2.

Quality Assessment of Randomized Controlled Trials and Quasi-Experimental Studies

Using the STROBE tool, we rated 17 observational studies,^40–56 4 of which scored intermediate quality^45,46,48,54 and the remaining 13 scored good quality.^{40–44,47,49–53,55,56} The main critical issues assessed in these studies were small sample size, absence of eligibility criteria for study participants inclusion, unspecified/unclear follow-up period, unclear intervention type outcome, and lack of a control group.

Assessed Outcomes and Evidence Synthesis

Anthropometric measures

The meta-analysis of two studies^29,30 showed a 0.13 (95% CI: −0.94 to 1.20) nonsignificant increase (random model I² = 0%) of absolute BMI measure in the intervention group compared to control group (Fig. 2A). The meta-analysis of 4 studies^26,27,30,31 showed a 0.09 (95% CI: −0.19 to 0.01) nonsignificant decrease (random model I² = 78%) of BMI z-scores in the intervention group compared to control group (Fig. 2B). The meta-analysis of 4 studies^26,27,30,31 showed a 1.37 (95% CI: −2.38 to −0.37) significant decrease (random model I² = 41%) of BMI% in the intervention group compared to control group (Fig. 3A). The meta-analysis of three studies^27,29,30 showed a 1.30 cm (95% CI: −1.43 to −1.16) significant decrease (random model I² = 0%) of WC in the intervention group compared to control group (Fig. 3B).

Figure 2.

Impact of school gardening and control on BMI (A) and on BMI z-score (B).

Figure 3.

Impact of school gardening and control on BMI% (A) and on waist circumference (B). BMI%, BMI percentile.

Blood pressure

Between pre- and posttest comparisons, the meta-analysis of three studies^27,29,30 found a 1.91 mmHg (95% CI: −0.78 to 4.59) increase of diastolic values (random model I² = 73%) and a 0.58 mmHg (95% CI: −1.68 to 2.85) increase of systolic values (random model I² = 45%), both nonstatistically significant (Fig. 4A, B).

Figure 4.

Impact of school gardening and control on diastolic blood pressure (A) and of systolic blood pressure (B).

Discussion

This is the first SR and meta-analysis analyzing the impact of school gardening on anthropometric variables in school-aged children, along with other outcomes related to knowledge, attitudes, and other laboratory and clinical outcomes. Our study updates nonanthropometric outcomes, confirming a general improvement in outcomes related to the three main categories analyzed related to F/V, namely F/V consumption, knowledge, and nutritional behavior. With different degrees of evidence, we found a positive impact of school gardening projects on all of these outcomes. In particular, the selected articles demonstrate that children increased F/V daily/weekly intake^34,40,41,53 and improved their F/V knowledge.^{26,28,35,36,39,41,43,46,47,50–53} Moreover, the selected articles found a positive impact on nutritional behavioral attitudes, with a significant improvement in “willingness to try new fruits and vegetables” outcomes.^34,35,39,42 Other studies^{34,35,48,50,52} also showed that school gardening projects improved both F/V preference and self-reported preference for healthy F/V-based snacks. The abovementioned results are in accordance with previous secondary studies' results for these outcomes^57–64 and highlight their paramount importance in the obesity prevention field.

Gardening activities have the common purpose of improving children's knowledge about the natural world.⁶³ Developing this skill from childhood can establish an environmental responsibility and awareness about wellbeing and quality of life, as long-term effects also in adulthood.⁶³ Regarding F/V intake, some studies showed an increase of combined vegetables and fruit intake or of vegetables alone intake. There was no study which reported an increase in the fruit intake only. Similar results were found for willingness to try new F/V. We hypothesized that this has been caused by studies, which were designed to measure vegetable intake only^{32,39,45,47,54} or because school gardens often contain mostly vegetables (which are much more easy to cultivate). This peculiarity has been reported also in other secondary studies.^57–60,63

Among those, one study,⁶⁰ to explain the potential relationship among school garden, vegetable intake, and willingness to try new F/V, has assumed that gardening activities has the potential to bring the children closer to the vegetables and, at the same, to reduce their unwillingness to try new foods. Another review⁶⁴ demonstrated that an increased willingness to taste F/V in children attending kindergarten and first grade can result into a greater fruit and vegetables preference and intake also in older ages. This result is crucial for obesity prevention. School gardening activities, nutritional education, and healthy meals offered in the lunchrooms can give the opportunity for children to improve the quality of their diet and nutrient intake.⁵⁸ This last consideration could also explain the increased self-efficacy observed in one study⁴¹ we included in this SR.

As far as the organizational structure of the school garden programs are concerned, among the studies that reported two types of intervention^{29,35,36,38,49,51} (gardening + nutritional education vs. nutritional education only), four out of six showed significant better result in the group which undergo a double intervention compared to the control group.^35,36,49,51 These conclusions are in accordance with the literature which supports the strength of the school garden combined with nutrition education to achieve some outcomes related to obesity prevention (particularly vegetable intake).^60,65

In addition, we analyzed four different multicomponent programs,^26,31,41,53 and in two of them,^41,53 we found significant differences in terms of F/V outcomes between schools providing full-program components and other schools which offered partial programs. A secondary study⁵⁷ highlighted that multicomponent interventions have higher impact on change in nutrition-related behaviors than a single component intervention. Multicomponent interventions usually consist in the development of a series of interventions, in which families can find a strict involvement of the family (which has a crucial role in the obesity prevention⁵⁹), introduction of healthy foods in the school lunchrooms, which increases the quality of diet and different food options,⁵⁸ and finally, the implementation of food laboratories. Food workshops as well as active gardening activities (i.e., planting, cultivation and harvesting processes), are an example of “hands on” learning experience,⁵⁹ making children play an active role in the learning process, rather than being passively involved.⁵⁹

Other outcomes, such as blood pressure or laboratory outcomes, showed mixed and inconsistent results with the exception of academic achievement^31,33,56 and physical activity level.^25,45 Positive improvement for academic performance covered only science^33,56 and math³¹ subjects, while reading³¹ did not reach any statistical difference between intervention and control groups. In fact, other studies demonstrated a positive impact of school gardening on children's social development⁵⁸ and in other academic achievements.⁶⁴ This can be assumed by the demonstrated improvement of both observational skills of the children and interdisciplinary content of the school programs,⁵⁸ encouraging academic learning in the scientific area.^61,62 Furthermore, as demonstrated in the previous results, natural green areas have also the opportunity to increase the amount of physical activity carried out by children.⁶⁶

This is the first SR and meta-analysis studying the impact of school gardening on anthropometric variables in school-aged children. Our study shows a statistically significant reduction of WC and BMI% comparing pre- and postintervention. We found no significant differences in other anthropometric outcomes investigated, including blood pressure. The pooled analysis showed a 1.3 cm significant reduction of WC and a 1.37 reduction of BMI%. The use of WC and waist–hip ratio as anthropometric indicators has been analyzed in a previous WHO report.⁶⁷ The working group states that anthropometric parameters (i.e., BMI, WC, and waist–hip ratio) can well estimate the risk of chronic disease. Therefore, WC and waist–hip ratio, regardless of cutoff points developed, could be utilized alone or in association with BMI. In addition, experts recommend, during the cutoff points development for WC and waist–hip ratio, to adopt a statistical approach based on policy relevance methodology with a final clear explanation of the decision rules.⁶⁷ This confirms that there is no consensus in the scientific community on cutoffs defining a clinically significant WC decrease, particularly for the pediatric population.

Due to the preventive nature of the intervention, our study included both children with normal and abnormal BMI. Therefore, we considered pre–post intervention nonincreasing or slightly reduced BMI values as a positive result. Furthermore, for more clarity, anthropometric variables should be stratified according to baseline values, as described by Davis et al.³⁰ Although it is still not clear if gardening activities have a strong impact on children's health, using a conservative approach, we may consider any reduction in WC or BMI% as a positive impact, particularly if addressed to overweight and obese children. This is also particularly important in the contemporary population, in which anthropometric measures, including WC and BMI%, show increasing trends in the pediatric population.^68–70 Finally, literature shows an impact, along with all the other anthropometric variables considered, also on metabolic syndrome of the children undergoing obesity prevention multicomponent programs.⁷¹ Previous studies have shown that there are two main reasons to target childhood anthropometric outcomes. First, overweight and obese children and teens are, compared to normal BMI children, much more likely to develop overweight and obesity in adult age. Second, as adults it is more challenging to lose the excess weight, compared to childhood or teen age, once they become overweight or obese. Moreover, children who learn healthy eating habits at a young age have a higher likelihood to maintain these habits through their life, compared to interventions carried out in adulthood.⁷² Furthermore, obese children today are experiencing diseases and health problems previously observed only in adults. The resulting complications of these diseases (i.e., type 2 diabetes, cardiovascular disease, hypertension, osteoporosis, and some carcinomas) can be a leading cause to a shorter life.⁷³

A possible criticism of similar interventions, as investigated by a recent SR⁷⁴ on interventions for preventing obesity in children, may be the potential impact of adverse or unintended consequences of the interventions. The study found no evidence of adverse outcomes, such as underweight, unhealthy eating practices, teasing, stigmatization, body image perceptions, satisfaction, or self-worth, although it acknowledged that only a minority of the included studies assessed adverse effects. Literature has previously shown that multicomponent interventions are effective in reducing the analyzed anthropometric variables.^75–77 Interventions for childhood obesity prevention should include both diet and physical activity and be preferentially targeted toward school-aged children.^65,74

In particular, for school-aged children, there is evidence of a positive effect of the interventions on anthropometric variables (i.e., BMI%, BMI z-score) in school settings. As highlighted by Katz et al., the 6 hours that school-aged children spend in school every day, for more than half the year, constitute a substantial part of their time and their lives.^78,79 Therefore, it is important to consider school as one of the major drivers and elective settings for children education on healthy lifestyles.^80,81 Recently, the programs for childhood obesity prevention have started to include activities and lessons focused on the importance of physical activity and on the dangers of sedentary lifestyle. In this context, some studies have shown a positive impact of the school and community gardens on outcomes such as blood pressure, BMI, and mental well-being. These studies showed also that outdoor activities improve the promotion of physical activity.^82,83 In addition, strategies for overweight/obesity prevention, including gardening activities starting from childhood, could contribute to reduce health inequalities. Indeed, recent evidence highlights the association between low socioeconomic status and higher BMI in children aged 6–12 years.^84,85 Three studies included in this review confirmed this evidence, showing that school gardening projects can improve health outcomes of low-income children who belong to families that may be unable to afford the cost of healthy food by increasing their access to healthy food in the school lunchrooms.^27,30,31

This study has important strengths. First, we followed a robust methodology, based on the Cochrane guidelines and PRISMA methodology for SRs and meta-analysis. Second, we included in our qualitative analysis non-RCT studies to provide comprehensive data from all the available literature. Third, we assessed the risk of bias for the various studies. The results of this study must be also interpreted in light of important limitations. First, although the number of studies on school gardening has steadily increased in the last 10 years, it is still too low to draw conclusions on this topic.

We tried to address this lack of data by being as comprehensive as possible, carrying out hand-search and trying to find additional references and/or data in the gray literature. Second, the studies we analyzed had considerable heterogeneity, with particular regard to outcomes and measures. Unfortunately, anthropometric outcomes are measured in many different ways and use different units of measure. We tried to address this problem by considering all the potential outcomes (i.e., absolute BMI, BMI%, BMI z-score), and, whenever possible (i.e., more than one study with comparable outcome), we summed up the different data in different meta-analyses. Third, we observed that only a few studies (recent RCTs or quasi-experimental studies) analyzed anthropometric outcomes. School gardening has become an important topic in childhood obesity prevention only recently. We hypothesize that researchers have adopted anthropometric outcomes as more objective measures compared to traditional ones (i.e., F/V knowledge or F/V consumption). Fourth, we generally found mid/low-quality studies, particularly quasi-experimental and observational studies. More specifically, the majority of studies did not include long follow-up periods, therefore registering only short-term changes. Furthermore, F/V outcomes have been achieved mostly through self-reported measures (i.e., questionnaires and/or in a school/pediatric environment), which could have led to a selection bias.⁸⁶ This factor may have affected the overall strength of our conclusions. However, it also reveals the state of the art and can encourage researchers to carry out further high-quality studies on this important topic.

Finally, it may appear unrealistic and unreasonable to expect that school-based gardening interventions alone are capable of appreciably influencing anthropometric variables, or other health outcomes, of the pediatric population. Obesogenic environment is, in fact, part of our modern societies. Children, like adults, have easy access to highly palatable, energy-dense foods, and at the same time, they are living into an ever-proliferating world of labor-saving technologies and electronic amusement. These last factors are particularly important outside of school hours. In particular, numerous studies showed the effect of the family lifestyle on the children's dietary habits, therefore, family involvement will be a primary target for obesity prevention programs in the near future.^87,88

Conclusions

This is the first SR, conducted in association with a meta-analysis, measuring the impact of school gardening projects on school-aged children's anthropometric outcomes. Our study shows a moderate degree of descriptive evidence for improvement in consumption of healthy foods in school-aged children undergoing intervention in school and after-school contexts, along with their knowledge of fruits, vegetables, and their healthy properties. Moreover, we found a modest but clinically significant reduction of WC and BMI% in children undergoing gardening project activities during the school hours.

It is key to implement global initiatives aimed to prevent childhood overweight and obesity, which has become a major public health issue globally and is the main driver for adult obesity, metabolic syndrome, and early mortality. Secondarily, since the global financial restraints require a targeted funding allocation, it is important to provide evidence-based, cost-effective, preventive interventions for childhood overweight and obesity.⁸⁹ New research on individual factors and personal lifestyles (i.e., diet or physical activity), as well as on the obesogenic environment is a priority to ensure high-quality obesity prevention programs.

Finally, we strongly advocate for the adoption of high-quality study designs in preventive medicine, such as RCTs or quasi-experimental studies, and the use of anthropometric measures, according to the last guidelines on childhood obesity.^90,91 This will pave the way for direct comparisons and meta-analyses to guide decisions for the implementation of school-based preventive interventions (including gardening activities) targeted to prevent and fight childhood overweight and obesity.

Footnotes

Acknowledgments

A.R., D.T., and M.P.F. conceived and designed the study. A.R., A.U. carried out the review, acquired, analyzed, and interpreted the data. D.T. checked data extractions. A.R. and D.G. drafted the article, which was critically revised for important intellectual content by all authors. A.R. and D.G. carried out the statistical analysis. D.G. supervised the study. All authors have read and approved the final article. D.G. is guarantor and had full access to all the data and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Funding Information

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Author Disclosure Statement

No competing financial interests exist.

References

World Health Organization (WHO). GLOBAL STATUS REPORT on Noncommunicable Diseases 2014 “Attaining the Nine Global Noncommunicable Diseases Targets; a Shared Responsibility.” Geneva, Switzerland, 2014. https://apps.who.int/iris/bitstream/handle/10665/148114/9789241564854_eng.pdf?sequence=1 (last accessed February 19, 2020).

Tremmel

, Gerdtham

U-G

, Nilsson

, et al. Economic burden of obesity: A systematic literature review. Int J Environ Res Public Health, 2017; 14:435.

, Xue

. Pediatric obesity: Causes, symptoms, prevention and treatment (review). Exp Ther Med, 2016; 11:15–20.

Kelishadi

, Azizi-Soleiman

. Controlling childhood obesity: A systematic review on strategies and challenges. J Res Med Sci, 2014; 19:993–1008.

Dietz

, Gortmaker

. Preventing obesity in children and adolescents. Annu Rev Public Health, 2001; 22:337–353.

Story

, Sallis

, Orleans

. Adolescent obesity: Towards evidence-based policy and environmental solutions. J Adolesc Health, 2009; 45(3 Suppl):1–5.

Bacchini

, Licenziati

, Garrasi

, et al. Bullying and victimization in overweight and obese outpatient children and adolescents: An Italian Multicentric Study. PLoS One, 2015; 10:1–13.

Story

, Nanney

, Schwartz

. Schools and obesity prevention: Creating school environments and policies to promote healthy eating and physical activity. Milbank Q, 2009; 87:71–100.

Savage

, Fisher

, Birch

. Parental influences on eating behavior: Conception to adolescence. J Law Med Ethics, 2007; 35:22–34.

10.

Budd

, Hayman

. Addressing the childhood obesity crisis: A call to action. MCN Am J Matern Nurs, 2008; 33:111–118.

11.

Zota

, Dalma

, Petralias

, et al. Promotion of healthy nutrition among students participating in a school food aid program: A randomized trial. Int J Public Health, 2016; 61:583–592.

12.

Morano

, Rutigliano

, Rago

, et al. A multicomponent, school-initiated obesity intervention to promote healthy lifestyles in children. Nutrition, 2016; 32:1075–1080.

13.

Scherr

, Linnell

, Smith

, et al. The shaping healthy choices program: Design and implementation methodologies for a multicomponent, school-based nutrition education intervention. J Nutr Educ Behav, 2014; 46:e13–e21.

14.

Messiah

, Arheart

, Lipshultz

, et al. Body mass index, waist circumference, and cardiovascular risk factors in adolescents. J Pediatr, 2008; 153:845–850.

15.

Messiah

, Arheart

, Natale

, et al. BMI, waist circumference, and selected cardiovascular disease risk factors among preschool-age children. Obesity, 2012; 20:1942–1949.

16.

Magnussen

, Koskinen

, Chen

, et al. Pediatric metabolic syndrome predicts adulthood metabolic syndrome, subclinical atherosclerosis, and type 2 diabetes mellitus but is no better than body mass index alone: The Bogalusa Heart Study and the Cardiovascular Risk in Young Finns Study. Circulation, 2010; 122:1604–1611.

17.

Moher

, Liberati

, Tetzlaff

, et al. The PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. PLoS Med, 2009; 6:e1000097.

18.

Greenhalgh

, Peacock

. Effectiveness and efficiency of search methods in systematic reviews of complex evidence: Audit of primary sources. Br Med J, 2005; 331:1064–1065.

19.

Higgins

JPT

, Altman

, Gotzsche

, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ, 2011; 343:1–9.

20.

Higgins

JPT

, Green

. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 (updated March 2011) The Cochrane Collaboration, 2011. www.handbook.cochrane.org

21.

von Elm

, Altman

, Egger

, et al. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: Guidelines for reporting observational studies. Int J Surg, 2014; 12:1495–1499.

22.

Higgins

JPT

, Thompson

, Deeks

, et al. Measuring inconsistency in meta-analyses. BMJ Br Med J, 2003; 327:557–560.

23.

DerSimonian

, Laird

. Meta-analysis in clinical trials. Control Clin Trials, 1986; 7:177–188.

24.

Wells

, Meyers

, Todd

, et al. The carry-over effects of school gardens on fruit and vegetable availability at home: A randomized controlled trial with low-income elementary schools. Prev Med, 2018; 112:152–159.

25.

Wells

, Myers

, Henderson

. School gardens and physical activity: A randomized controlled trial of low-income elementary schools. Prev Med, 2014; 69(S):S27–S33.

26.

Scherr

, Linnell

, Dharmar

, et al. A multicomponent, school-based intervention, the shaping healthy choices program, improves nutrition-related outcomes. J Nutr Educ Behav, 2017; 49:368–379.e1.

27.

Gatto

, Martinez

, Spruijt-Metz

, et al. LA sprouts randomized controlled nutrition, cooking and gardening programme reduces obesity and metabolic risk in Hispanic/Latino youth. Pediatr Obes, 2017; 12:28–37.

28.

Hutchinson

, Christian

, Evans

CEL

, et al. Evaluation of the impact of school gardening interventions on children's knowledge of and attitudes towards fruit and vegetables. A cluster randomised controlled trial. Appetite, 2015; 91:405–414.

29.

Cotter

, Cotter

, Oliveira

, et al. Salt intake in children 10–12 years old and its modification by active working practices in a school garden. J Hypertens, 2013; 31:1966–1971.

30.

Davis

, Ventura

, Cook

, et al. LA sprouts: A gardening, nutrition, and cooking intervention for Latino youth improves diet and reduces obesity. J Am Diet Assoc, 2011; 111:1224–1230.

31.

Hollar

, Lombardo

, Lopez-Mitnik

, et al. Eective multi-level, multi-sector, school-based obesity prevention programming improves weight, blood pressure, and academic performance, especially among low-income, minority children. J Health Care Poor Underserved, 2010; 21:93–108.

32.

Cotugna

, Manning

, Didomenico

. Impact of the use of produce grown in an elementary school garden on consumption of vegetables at school lunch. J Hunger Environ Nutr, 2012; 7:11–19.

33.

Klemmer

, Waliczek

, Zajicek

. Growing minds: The effect of a school gardening program on the science achievement of elementary students. Horttechnology, 2005; 15:448–452.

34.

Meinen

, Friese

, Wright

, et al. Youth gardens increase healthy behaviors in young children. J Hunger Environ Nutr, 2012; 7:192–204.

35.

Morgan

, Warren

, Lubans

, et al. The impact of nutrition education with and without a school garden on knowledge, vegetable intake and preferences and quality of school life among primary-school students. Public Health Nutr, 2010; 13:1931–1940.

36.

Morris

, Zidenberg-Cherr

. Garden-enhanced nutrition curriculum improves fourth-grade school children's knowledge of nutrition and preferences for some vegetables. J Am Diet Assoc, 2002; 102:91–93.

37.

O'Brien

, Shoemaker

. An after-school gardening club to promote fruit and vegetable consumption among fourth grade students: The assessment of social cognitive theory constructs. Horttechnology, 2006; 16:24–29.

38.

Poston

, Shoemaker

, Dzewaltowski

. A comparison of a gardening and nutrition program with a standard nutrition program in an out-of-school setting. Horttechnology, 2005; 15:463–467.

39.

Ratcliffe

, Merrigan

, Rogers

, et al. The effects of school garden experiences on middle school-aged students' knowledge, attitudes, and behaviors associated with vegetable consumption. Health Promot Pract, 2011; 12:36–43.

40.

Duncan

, Eyre

, Bryant

, et al. The impact of a school-based gardening intervention on intentions and behaviour related to fruit and vegetable consumption in children. J Health Psychol, 2015; 20:765–773.

41.

Evans

, Ranjit

, Rutledge

, et al. Exposure to multiple components of a garden-based intervention for middle school students increases fruit and vegetable consumption. Health Promot Pract, 2012; 13:608–616.

42.

Gibbs

, Staiger

, Johnson

, et al. Expanding children's food experiences: The impact of a school-based kitchen garden program. J Nutr Educ Behav, 2013; 45:137–146.

43.

Hanbazaza

, Triador

, Ball

GDC

, et al. The impact of school gardening on cree children's knowledge and attitudes toward vegetables and fruit. Can J Diet Pract Res, 2015; 76:133–139.

44.

Heim

, Stang

, Ireland

. A garden pilot project enhances fruit and vegetable consumption among children. J Am Diet Assoc, 2009; 109:1220–1226.

45.

Hermann

, Parker

, Brown

, et al. After-school gardening improves children's reported vegetable intake and physical activity. J Nutr Educ Behav, 2006; 38:201–202.

46.

Koch

, Waliczek

, Zajicek

. The effect of a summer garden program on the nutritional knowledge, attitudes, and behaviors of children. Horttechnology, 2006; 16:620–625.

47.

Leuven

JRFW

, Rutenfrans

AHM

, Dolfing

, et al. School gardening increases knowledge of primary school children on edible plants and preference for vegetables. Food Sci Nutr, 2018; 6:1960–1967.

48.

Lineberger

, Zajicek

. School gardens: Can a hands-on teaching tool affect students' attitudes and behaviors regarding fruit and vegetables?. Horttechnology, 2000; 10:593–597.

49.

McAleese

, Rankin

. Garden-based nutrition education affects fruit and vegetable consumption in sixth-grade adolescents. J Am Diet Assoc, 2007; 107:662–665.

50.

Nolan

, Mcfarland

, Zajicek

, et al. The effects of nutrition education and gardening on attitudes, preferences, and knowledge of minority second to fifth graders in the Rio grande valley toward fruit and vegetables. Horttechnology, 2012; 22:299–304.

51.

Parmer

, Salisbury-Glennon

, Shannon

, et al. School gardens: An experiential learning approach for a nutrition education program to increase fruit and vegetable knowledge, preference, and consumption among second-grade students. J Nutr Educ Behav, 2009; 41:212–217.

52.

Somerset

, Markwell

. Impact of a school-based food garden on attitudes and identification skills regarding vegetables and fruit: A 12-month intervention trial. Public Health Nutr, 2009; 12:214–221.

53.

Wang

, Rauzon

, Studer

, et al. Exposure to a comprehensive school intervention increases vegetable consumption. J Adolesc Health, 2010; 47:74–82.

54.

Wright

, Rowell

. Examining the effect of gardening on vegetable consumption among youth in kindergarten through fifth grade. WMJ, 2010; 109:125–129.

55.

Pigg

, Waliczek

, Zajicek

. Effects of a gardening program on the academic progress of third, fourth, and fifth grade math and science students. Horttechnology, 2006; 16:262–264.

56.

Smith

, Motsenbocker

. Impact of hands-on science through school gardening in Louisiana public elementary schools. Horttechnology, 2005; 15:439–443.

57.

Savoie-Roskos

, Wengreen

, Durward

. Increasing fruit and vegetable intake among children and youth through gardening-based interventions: A systematic review. J Acad Nutr Diet, 2017; 117:240–250.

58.

Berezowitz

, Bontrager Yoder

, et al. School gardens enhance academic performance and dietary outcomes in children. J Sch Health, 2015; 85:508–518.

59.

Davis

, Spaniol

, Somerset

. Sustenance and sustainability: Maximizing the impact of school gardens on health outcomes. Public Health Nutr, 2015; 18:2358–2367.

60.

Langellotto

, Gupta

. Gardening increases vegetable consumption in School-aged children: A meta-analytical Synthesis. Horttechnology, 2012; 22:430–445.

61.

Blair

. The child in the garden: An evaluative review of the benefits of school gardening. J Environ Educ, 2009; 40:15–38.

62.

Williams

, Dixon

. Impact of garden-based learning on academic outcomes in schools: Synthesis of research between 1990 and 2010. Rev Educ Res, 2013; 83:211–235.

63.

Ohly

, Gentry

, Wigglesworth

, et al. A systematic review of the health and well-being impacts of school gardening: Synthesis of quantitative and qualitative evidence. BMC Public Health, 2016; 16:286.

64.

Robinson-O'Brien

, Story

, Heim

. Impact of garden-based youth nutrition intervention programs: A review. J Am Diet Assoc, 2009; 109:273–280.

65.

Gori

, Guaraldi

, Cinocca

, et al. Effectiveness of educational and lifestyle interventions to prevent paediatric obesity: Systematic review and meta-analyses of randomized and non-randomized controlled trials. Obes Sci Pract, 2017; 3:235–248.

66.

Wells

, Ashdown

, Davies

EHS

, et al. Environment, design, and obesity. Environ Behav, 2007; 39:6–33.

67.

Nishida

, Ko

, Kumanyika

. Body fat distribution and noncommunicable diseases in populations: Overview of the 2008 WHO Expert Consultation on Waist Circumference and Waist-Hip Ratio. Eur J Clin Nutr, 2010; 64:2–5.

68.

Visscher

TLS

, Heitmann

, Rissanen

, et al. A break in the obesity epidemic? Explained by biases or misinterpretation of the data?. Int J Obes, 2015; 39:189–198.

69.

Sigmund

, Badura

, Sigmundová

, et al. Trends and correlates of overweight/obesity in Czech adolescents in relation to family socioeconomic status over a 12-year study period (2002–2014). BMC Public Health, 2018; 18:1–11.

70.

Hamřík

, Sigmundová

, Pavelka

, et al. Trends in overweight and obesity in Czech school children from 1998 to 2014. Cent Eur J Public Health, 2017; 25(Supplement 1):S10–S14.

71.

Rothberg

, McEwen

, Kraftson

, et al. Impact of weight loss on waist circumference and the components of the metabolic syndrome. BMJ Open Diabetes Res Care, 2017; 5:3–8.

72.

Pandita

, Sharma

, Pandita

, et al. Childhood obesity: Prevention is better than cure. Diabetes Metab Syndr Obes Targets Ther, 2016; 9:83–89.

73.

Dietz

. Overweight in childhood and adolescence. N Engl J Med, 2004; 350:855–857.

74.

Waters

, De Silva-Sanigorski

, Burford

, et al. Interventions for preventing obesity in children. Sao Paulo Med J, 2014; 132:128.

75.

Blanchette

, Brug

. Determinants of fruit and vegetable consumption among 6-12-year-old children and effective interventions to increase consumption. J Hum Nutr Diet, 2005; 18:431–443.

76.

Van Cauwenberghe

, Maes

, Spittaels

, et al. Effectiveness of school-based interventions in Europe to promote healthy nutrition in children and adolescents: Systematic review of published and grey literature. Br J Nutr, 2010; 103:781–797.

77.

Evans

CEL

, Christian

, Cleghorn

, et al. Systematic review and meta-analysis of school-based interventions to improve daily fruit and vegetable intake in children aged 5 to 12 y. Am J Clin Nutr, 2012; 96:889–901.

78.

Katz

, O'Connell

, Njike

, et al. Strategies for the prevention and control of obesity in the school setting: Systematic review and meta-analysis. Int J Obes, 2008; 32:1780–1789.

79.

Katz

. School-based interventions for health promotion and weight control: Not just waiting on the world to change. Annu Rev Public Health, 2009; 30:253–272.

80.

Moens

, Braet

, Bosmans

, et al. Unfavourable family characteristics and their associations with childhood obesity: A cross-sectional study. Eur Eat Disord Rev, 2009; 17:315–323.

81.

Karnik

, Kanekar

. Childhood obesity: A global public health crisis. Int J Prev Med, 2012; 3:1–7.

82.

Rees-Punia

, Holloway

, Knauft

, et al. Effects of school gardening lessons on elementary school children's physical activity and sedentary time. J Phys Act Health, 2017; 14:1–22.

83.

Soga

, Gaston

, Yamaura

. Gardening is beneficial for health: A meta-analysis. Prev Med Rep, 2017; 5:92–99.

84.

Lazzeri

, Pammolli

, Pilato

, et al. Relationship between 8/9-yr-old school children BMI, parents' BMI and educational level: A cross sectional survey. Nutr J, 2011; 10:76.

85.

Wang

, Lim

. The global childhood obesity epidemic and the association between socio-economic status and childhood obesity. Int Rev Psychiatry, 2012; 24:176–188.

86.

Miller

, Abdel-Maksoud

, Crane

, et al. Effects of social approval bias on self-reported fruit and vegetable consumption: A randomized controlled trial. Nutr J, 2008; 7:1–7.

87.

Gray

, Hernandez Alava

, Kelly

, et al. Family lifestyle dynamics and childhood obesity: Evidence from the millennium cohort study. BMC Public Health, 2018; 18:1–15.

88.

Campbell

. Biological, environmental, and social influences on childhood obesity. Pediatr Res, 2016; 79:205–211.

89.

Cecchini

, Sassi

. Preventing obesity in the USA: Impact on health service utilization and costs. Pharmacoeconomics, 2015; 33:765–776.

90.

Faggiano

, Concina

, Molinar

, et al. Primary prevention of cancer and cardiovascular diseases: An overview of scientific literature. Epidemiol Prev, 2014; 38:19–22.

91.

Evans

, Ranjit

, Hoelscher

, et al. Impact of school-based vegetable garden and physical activity coordinated health interventions on weight status and weight-related behaviors of ethnically diverse, low-income students: Study design and baseline data of the Texas, Grow! Eat! Go! (TGEG) clu. BMC Public Health, 2016; 16:973.