Table 4.

Studies addressing psychosocial outcomes included in this review

First author, year	Country, setting	Study design	Sample characteristics (inclusion criteria, number, age and sex)	Aims	Sampling methods	Intervention / Community garden program	Data collection Analysis (including adjustments)	Outcomes	Results
Brown et al. 2020 [25]	USA, Montana, Native American community	RCT (1) Group-based Community gardening program (2) control (no gardening)	Native Americans with prediabetes or diabetes N = 20 Age (y): 15/20 were 45–64 years, 5/25 25–44 years Male n(%): 4/20 (25%) CON n = 12 INT n = 8	Determine feasibility of a group gardening program and potential for collecting health outcomes	Convenience sample of participant expressing an interest in the gardening study at a diabetes clinic	Raised beds for gardening chosen for proximity to college and health centre. Plus 10 × 90-min structured sessions with hands-on gardening and food preparation activities Outcomes measured at 7 months after baseline	Outcomes were reported as medians and ranges. Change from baseline was compared between the groups using Wilcoxon rank sum tests. No adjustments Sample numbers reported for each outcome as there was missing information for some outcomes	Quality of life (QOL), CES Depression Scale Tension-anxiety, depression-dejection, anger-hostility, vigour-activity, fatigue-inertia, and confusion-bewilderment (from Profile of Mood States Inventory—POMS)	QOL – psychological INT (n = 7) 0 [-3.3, 2.0] CON (n = 11) 0.2 [-4.0, 4.5] P = 0.772 QOL – social INT (n = 7) 0 [0, 2.0] CON (n = 11) 0.2 [-6.7, 4.0] P = 0.430 QOL – environment INT (n = 7) -1.0 [-1.5, 0.5] CON (n = 11) 0 [-0.5, 4.0] P = 0.013 QOL – physical INT (n = 7) -0.6 [-1.7, 0.6] CON (n = 11) -0.6 [-5.1, 2.9] P = 0.707 POMS – total mood disturbance INT (n = 8) -2.0 [-16, 18] CON (n = 9) 9 [(-1.0, 30] P = 0.049 POMS – tension anxiety INT (n = 8) − 0.5 (− 4.0 to 6.0) CON (n = 10) 1.0 (− 2.0 to 12.) P = 0.062 POMS – depression-dejection INT (n = 8) − 0.5 (− 4.0 to 6.0) CON (n = 10) 3.5 (− 2.0 to 19) P = 0.105 POMS – anger-hostility INT (n = 8) 0 (− 8.0 to 10) CON (n = 9) 6.0 (− 6.0 to 20) P = 0.180 POMS – vigor-activity INT (n = 8) − 3.0 (− 13 to 13) CON (n = 9) 0 (− 7.0 to 7.0) P = 0.382 POMS – fatigue-inertia INT (n = 8) − 2.5 (− 16 to 7.0) CON (n = 10) 2.0 (− 10 to 21) P = 0.246 POMS – confusion-bewilderment INT (n = 8) 0 (− 4.0 to 6.0) CON (n = 9) 2.0 (0 to 17) P = 0.119
Gerber et al. 2017 [56]	USA	Cross-sectional survey	Bhutanese community leaders recruited participants & collected data Bhutanese refugees in the USA who self-select as community gardeners (n = 22) or non-gardeners (n = 28) Characteristics Female: 62% Age (mean ± SD): 45 ± 15 yrs	To explore differences in indicators of distress and social support among Bhutanese refugees that participate in community gardens compared with those who do not	Bhutanese community events, word-of-mouth	Waiting list for plots. Families typically garden on one or two plots	Descriptive statistics only. No comparisons, & not adjusted for anything	Symptoms of post-traumatic stress disease, anxiety & depression using the Refugee Health Screener (15-items, score > 12 refer to mental health service) Patient Health Questionnaire (15-items, 3-point scale; cut-points 5, 10 & 15 indicate low, medium and high somatic symptoms) Perceived social support (Medical Outcomes Study Social Support Survey; 19-item, 5-point scale)	On average, more Gardeners lived in a house, and had lower medical bills, compared with non-gardeners Refugee Health Screener referrals (mean ± SD) Gardeners: 11.6 ± 9.2 Non-gardeners: 11.0 ± 9.9 Comparisons “not statistically different” Scores > 5 on Patient Health Questionnaire Gardeners: 14/22 (64%) Non-gardeners: 13/28 (46%) Gardeners experienced more somatic symptoms Standardized effect size (d = 0.36 95% CI -0.21, 0.91) Social support (mean ± SD) Gardeners: 61.3 ± 13.2 Non-gardeners: 52.5 ± 12.1 Gardeners reported more social support Standardized effect size (d = 0.70 95% CI 0.12, 1.27)
Grier et al. 2015 [57]	USA; Dan River, Virginia	Pre-/post	n = 43 Characteristics Ethnicity: 97.7% African American Age (mean) 8.7y Male: 46.5% Weight status: 34.1% overweight 18.2% obese Inclusion criteria Age: 5–17 y Child AND parent reside in housing authority full-time	To report on feasibility (demand, acceptability, implementation and limited-effectiveness) of a CG and nutrition education program	Two public housing authority sites – active members of the Dan River Partnership for a Healthy Community Adult site leaders knew families and youth; distributed recruitment material	Junior Master Gardener curriculum with nutrition focused lessons (informed by SCT). Weekly gardening sessions or gardening + nutrition education with site leaders	Interviewer administered survey Repeated measures ANOVA (ITT and complete case; ITT presented); effect sizes calculated	Psychosocial factors related to F&V consumption (not actual consumption) and nutrition knowledge	Increased self-efficacy for asking for F&V (ES: 0.39; p = 0.013) No change in willingness to try F&V (ES = 0.10; p = 0.310), self-efficacy for eating F&V (ES = 0.21; p = 0.119) or nutrition knowledge (ES = 0.10; p = 0.583)
Hawkins et al. 2011 [49]	UK, Wales, Cardiff	Cross-sectional study (1) Indoor exercise group (2) Walkers (3) Allotment gardeners (4) Home gardeners	Characteristics (1) n = 23 Age (y): 72.9 ± 6.9 Male: 3 (13%) (2) n = 25 Age (y): 62.4 ± 6.8 Male: 8 (32%) (3) n = 25 Age (y): 65.7 ± 9.1 Male: 17 (68%) (4) n = 21 Age (y): 69.5 ± 7.7 Male: 2 (10%) Inclusion criteria  ≥ 50 y attending various local activity groups	Measure health status and perceived stress of allotment gardeners compared to other activity groups (indoor exercisers, walkers, home gardeners)	Recruited via leaflets, posters and visits to groups from researcher Response rate 87.8%	Compared leisure activity groups to members of allotment gardening group No intervention	Perceived stress: Cohen & Williamson 1988 Health-related Quality of Life, (Mental health component) using the SF-36v2 Social provisions Stress level btw groups adjusted for area-level SEP	Perceived stress	Significantly lower perceived stress among allotment gardeners than other activity groups Perceived stress (mean ± SD) (1) 15.8 ± 6.1 (2) 13.6 ± 5.4 (3) 9.8 ± 5.8 (4) 12.0 ± 4.8 QOL Mental health (median, IQR) (1) 50 (46.6–54.4) (2) 56.1 (51.7–58.4) (3) 55.3 (50.1–58.6) (4) 55.8 (50.7–58.8)
Hartwig and Mason 2016 [33]	USA, MN, Twin Cities	Cross-sectional surveys	n = 97 Characteristics Female: 65% English: 18% good/fluent Age (mean): 39y (16–80 y) Ethnicity: 67% Karen (Burmese)	To evaluate church CGs serving refugee and immigrant populations, reporting primary health and social benefits	All gardeners at 8 gardens invited (Response rate = 44–45%) Gardens purposively sampled based on: - 2 yrs participation - # gardeners - primary language of gardeners	8 church gardens serving refugees and immigrants	Measured early and late season harvest (Jul-Sept) Descriptive stats used: change in mean/% early and late season No adjustments	Depression Gardening alone/with others (social interaction)	Change in depression risk not reported (12% reached cut-off for additional screening) Frequency of social interactions declined from early to late season harvest
Heilmayr and Friedman, 2020 [34]	USA, CA	RCT with 5 INT groups: (1) Community gardening (2) moderate indoor exercise (3) Exposure to nature (4) Social club (watching films) (5) Indoor container gardening	University students Baseline data reported in combination (not by group allocation) Age (y): 20.6 ± 3.3 Male: 31.2% (1) n = 21 (2) n = 21 (3) n = 23 (4) n = 22 (5) n = 23	To compare community gardening with four theoretically driven comparison groups to understand possible causal mechanisms around how community gardens have improved outcomes	Convenience sample recruited via flyers, emails and the Psychology Subject Pool	4 week INT; assigned an activity for 2–3 h/wk	Data were analyzed by ANOVA with pre-/post-test values to assess how groups changed from baseline and a group by time interaction	Emotional wellbeing (Comprised of: Perceived stress, Happiness, Self-efficacy, Positive and Negative Affect) Social relationships (Comprised of: Companionship, Social integration)	Emotional wellbeing (post-test only; mean ± SD) (1) 65.4 ± 14.8 (2) 66.6 ± 15.5 (3) 66.1 ± 13.3 (4) 63.6 ± 15.5 (5) 67.1 ± 14.4 Social relationships (post-test only; mean ± SD) (1) 62.3 ± 10.8 (2) 63.5 ± 11.9 (3) 59.3 ± 14.1 (4) 59.3 ± 16.6 (5) 60.7 ± 11.4
Hopkins and Holben 2018 [35]	USA, OH, rural Appalachia (Athens)	Cross-sectional study	-n = 50 Inclusion criteria: CG plot in Athens Characteristics Ethnicity: 81.6% white Female: 67.4% Education:46.9% college educated	To examine relationships among food security, produce intake and behaviors, health and social capital among community gardeners	All community gardeners (n = 120) in Athens	No INT Individuals with CG plots	Survey distributed via email (Response rate = 42%) Descriptive stats reported, no adjustment	Social capital (made new friends)	Social cohesion 74% have made new friends due to CG No association of food security with social capital
Koay et al. 2020 [58]	Singapore	Cross- sectional survey	(1) Community gardener N = 45 Male n (%): 25(56%) Age (mean ± SD): 60.2y (± 13.3) Ethnicity: 40 (89%) Chinese Education: 18(40%) Tertiary (2) Home gardener N = 38 Male n(%) 6(84%) Age (mean ± SD): 43.8 ± 13.0 Ethnicity: Chinese 35(92%) Education: Tertiary 33(87%) (3) Non-gardening control N = 28 Male n (%) 12(43%) Age (mean ± SD): 55.5 ± 11.6 Ethnicity: Chinese 23(82%) Education: Tertiary 13(46%)	Study relationship between community gardening and mental health benefits	Snowball recruitment from gardens and outdoor activity groups	Community in Bloom program of government supported	Multivariate ANCOVA with adjustment for age and connection to nature	Perceived stress scale (10-item, 5-point scale) Personal Wellbeing Index (7-item, 11 point scale) Brief resilience scale (6-item, 5-point scale)	Perceived stress scale (mean ± SD) (1) 11.4 ± 6.4 (2) 15.5 ± 6.1 (3) 7.0 ± 0.8 Personal wellbeing index (mean ± SD) (1) 8.2 ± 1.1 (2) 7.0 ± 1.2 (3) 7.0 ± 0.8 Brief resilience scale (mean ± SD) (1) 3.7 ± 0.7 (2) 3.5 ± 0.6 (3) 3.0 ± 0.8
Litt et al. 2015 [38]	USA; Denver, Colorado	Cross-sectional survey	n = 469 Characteristics Age (mean): 46.1y (± 15.9) Female: 67.4% Education: 57.4% college educated Identified as gardeners: 59.3% Inclusion criteria English or Spanish speaking, ≥ 18yrs	To examine the direct and indirect pathways by which gardening influence self-rated health	Area-based sample of general population n = 1154 randomly drawn from 40 block groups 13 gardens identified; List-based census of community gardeners n = 300	No intervention Individuals participating in CGs compared with non-gardeners	Surveys interviewer administered Path analysis controlling for age, education, years in neighborhood, % college education in neighborhood, observed incivilities	Social involvement Collective efficacy Neighborhood attachment	Path model results: Data fit model adequately, accounting for 22% variance in self-rated health and 4% in F&V intake Gardening predicted social involvement (β = 0.36; p < 0.001) Social involvement (β = 0.11, p < 0.05) and aesthetics (β = 0.46, p < 0.001) predicted Collective efficacy Collective efficacy predicted neighborhood attachment (β = 0.29, p < 0.001)
Machida 2019 [39]	Japan	Cross-sectional survey	Web-based survey limited to age 60–69 y, professional farmers excluded (1) Community gardeners n = 129 Male n (%): 87(67%) Age (y): 64.1 ± 2.6 (2) Home gardeners (HG) n = 371 Male n(%):280(76% Age (mean ± SD): 63.9y ± 2.7 (3) Non-gardeners n = 500 Male n(%): 327 (65%) Age (mean ± SD): 63.3y ± 2.5	To study the relationship between community or home gardening and health status or a healthy lifestyle	The survey was conducted by a marketing company with 4.2 million people registered across all 47 prefectures in Japan	NA	Odds Ratios adjusted for sex, age, family structure and employment status (not described)	Happiness (single item, 11-point scale) dichotomized to ≤ 6 vs ≥ 7 Psychological distress using 4 items of the K6 (4-point scale) dichotomized at ≤ 8 vs ≥ 9	Happiness (1) CG: 1.60 (1.18, 2.16) (2) HG: 0.89 (0.59, 1.34) Distress (1) CG: 0.85 (0.57, 1.27) (2) HG: 0.72 (0.38, 1.36)
Mourao et al. 2019 [54]	Portugal	Cross-sectional survey	Invitation from the Urban Allotment Garden office, sent to 30 gardeners per session. Six sessions performed, resulting in 65 validated responses Lived in urban areas of the council, 90.8% Characteristics Male: 56.9% Age: 26–45 yrs: 36.9% 46–65 yrs: 47.7%  > 65 yrs: 15.4%	To evaluate the happiness and well-being of the Portugal population, based on the urban organic allotment gardens	Self-administered questionnaires	Permanent resident, garden a family plot	Pearson correlation. No adjustment	Personal wellbeing scale Subjective happiness scale	Gardening frequency: Once a week: 10.5% Few days a week: 47.7% Daily: 41.5% Degree of life satisfaction (personal well-being index): Mean 74.5% (0–100%) Greater frequency to gardens was associated with higher perspective of subjective happiness and compared to their peers
Soga et al. 2017 [50]	Japan, Tokyo, Nerima district in central Tokyo	Cross-sectional survey	Gardeners (n = 165) vs non-gardeners (n = 167) Characteristics Gardeners: Male: 68.1% Age (mean ± SD): 62 ± 17y Non-gardeners: Male: 42% Age (mean ± SD): 61 ± 16y	To quantify effects of allotment gardening on physical, psychological and social health	Gardeners located by face-to-face recruitment at allotment gardens (90% response rate). Non-gardeners recruited via a letter sent to 1000 Nerima households (20% response rate)	No INT	Adjusted for sex, age, household income, employment, smoking, drinking, vegetable intake and PA (days per week of > 30 min/day of moderate activity)	Mental health using 12-item General Health questionnaire (scores 0–12)	Mental health Compared with non-gardeners, mean mental health scores for gardeners (± SE) was -0.91 (0.42) higher (P < 0.05), indicating improvements in mental health
Swami 2020 [59]	UK, London	Cross-sectional survey	English-speaking adults. One participant per allotment (1) allotment gardeners (n = 84) (2) non-gardeners (n = 81) Full sample Male (%): 40% Age (mean ± SD): 44.7y ± 18.2	To examine the effect of allotment gardening on state body image	Gardeners recruited from 12 allotment sites via direct approach. Non-gardeners recruited from supermarkets closest to the allotment sites. They could no “do anything in the garden”	Not described	Non-gardeners ‘matched’ to gardeners but no description of matching process or characteristic. Analysis by unpaired t-test with no adjustments (including no adjustment of matching criteria) Bonferroni correction of p values	State body image using a 10 cm visual analogue scale Body Appreciation Scale-2 (10-items, 5-point scale) Functionality Appreciation Scale (7-item, 5-point scale) Authentic Pride subscale of the Body and Appearance Self-Conscious Emotions Scale (6-items, 5-point scale)	Body appreciation (mean ± SD) (1) 3.5 ± 0.8 (2) 3.1 ± 0.8 Functionality appreciation (mean ± SD) (1) 3.5 ± 0.7 (2)3.2 ± 0.8 Body pride (mean ± SD) (1) 3.1 ± 0.9 (2) 2.6 ± 0.9
Tharrey et al. 2020 [46]	France, Montpellier	Longitudinal cohort study Data collected at baseline and 1 year later	Characteristics (1) Community gardeners (n = 66) Male n(%): 16(24.2) Age (y): 44.0 ± 14.0 (2) Non-gardeners (n = 66) Male n (%): 16(24.2) Age (y): 44.9 ± 13.7 Inclusion criteria Starting gardening in a CG; residents of Montpelier; ability to read French	To assess the impact or urban community garden participation the adoption of sustainable lifestyles	Gardeners recruited when new to the gardening community Non-gardeners recruited via volunteers for a population-based survey on food supply behaviors	Community gardens plots used collectively or individually	Analyzed with mixed-effects models with group by time interaction Adjustments for education, BMI, meals consumed outside the home, social desirability where appropriate	Warwick-Edinburgh Mental Wellbeing Scale (WEMWBS; 14-item, 5-point scale) Loneliness scale v3 (20-items 4-point scale)	Wellbeing at 1 year (mean ± SD) (1) 51.5 ± 6.9 (2) 51.5 ± 5.7 Loneliness at 1 year (mean ± SD) (1) 40.1 ± 10.9 (2) 40.5 ± 9.5
van den Berg et al. 2010 [51]	The Netherlands, “large cities”	Cross-sectional survey	Gardeners (n = 121) from 12 allotment gardens Non-gardener (n = 63) Characteristics Gardeners: Male: 53% Age (mean ± SD): 62 ± 12 y Non-gardeners: Male: 41% Age (mean ± SD): 56 ± 14 y	To directly compare the health, wellbeing and physical activity of allotment gardeners to that of controls without an allotment garden	Gardeners sent invitations to their home addresses Non-gardeners were responders living next to the home address of allotment gardeners	Ranged from residential parks, day-recreational parks and food production parks	Adjusted for age, sex, education, income, access to a garden at home, PA in winter and stressful life events, and included an age by gardening interaction term. Results separated by age. For all outcomes	Stress in past month (2-items, 6-point scale), Life Satisfaction Index (8-item, 3-point response) Loneliness (2-items, 0–1 responses) Social contacts (2-items, scores range 1–12)	All meanadjusted ± SE Stress  < 62 yrs Gardeners 3.2 ± 0.1 Non-gardeners 2.9 ± 0.2  ≥ 62 yrs Gardeners 2.1 ± 0.1 Non-gardeners 2.5 ± 0.2 Life satisfaction  < 62 yrs Gardeners 2.2 ± 0.1 Non-gardeners 2.2 ± 0.1  ≥ 62 yrs Gardeners 2.3 ± 0.1 Non-gardeners 2.0 ± 0.1 Loneliness  < 62 yrs Gardeners 0.7 ± 0.1 Non-gardeners 0.6 ± 0.1  ≥ 62 yrs Gardeners 0.3 ± 0.1 Non-gardeners 0.8 ± 0.2 Social contacts  < 62 yrs Gardeners 6.1 ± 0.4 Non-gardeners 7.0 ± 0.5  ≥ 62 yrs Gardeners 8.1 ± 0.4 Non-gardeners 6.2 ± 0.7
Young et al. 2020 [60]	Switzerland, Zurich	Cross-sectional survey	Materials provided in 4 languages used locally. Limited to one person per allotment (1) Allotment gardeners (n = 108) Male (%): 52% Age (y): 59 (SD NR) (2) Domestic gardeners (n = 193) Male (%): 33% Age (y): 54 (SD NR)	To identify whether gardening is a source of stress (i.e. stress as a result of the garden)	Allotment gardeners drawn in a two-stage probabilistic sampling strategy (response rate 48%.) Domestic gardeners drawn from a random sample of individuals living in Zurich (response rate 27%)	Allotments typically 100–200 m2, with rules to prohibit invasive species and construction on site. Domestic gardens are available to householders who can afford to buy/rent a residence with a garden (~ 10% of population)	Independent t-test Structural equation model (SEM) with robust standard errors, full information maximum-likelihood for missing data and adjustment for age, gender, employment, job type and biodiversity preference	Single question “I often feel under pressure when I think of the tasks that need doing in my garden” (5-point response)	Garden-related stress (mean ± SD) (1) 2.2 ± 1.2 (2) 2.5 ± 1.1 Allotment gardeners reported lower stress than domestic gardeners (β = -0.167, p = 0.013) when controlling for socioeconomic variables in SEM

Abbreviations: CG Community garden, CI Confidence interval, COM Comparison group, CON Control group, ES Effect size, F&V Fruit and vegetable, INT Intervention group, ITT Intention-to-treat, NR Not reported, OR Odds ratio, PA Physical activity, RCT Randomized controlled trial, SD Standard deviation, SE Standard error, SEP Socioeconomic position