Table 1.
Descriptive characteristics of the included studies stratified by health outcome (PA and mental health) and listed in alphabetical order and based on study design.
| Physical Activity | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Author, Year [Ref] | Study Period | Country | Study Design | Type of Greenspace | City | Sample Size | Statistical Analysis | Tool Used to Measure PA | Outcome Domain | Main Results | QS/9 |
| Observational Studies | |||||||||||
| Cerin E., 2017 [70] | 2002–2011 | BE, BR, CO, CZ, DK, HK, MX, NZ, UK, US | Cross-sectional | PUM | Ghent, Curitiba, Bogotá, Olomouc, Aarhus, Hong Kong, Cuernavaca, North Shore, Waitakere, Wellington, Christchurch, Stoke-On-Trent, Seattle, Baltimore | 6712 | Mixed-model regression measures | Accelerometer | PA regardless of the setting | MVPA in urban parks was lower in the late evening/night (1.2 ± 4.0 min/h) and higher in the afternoon (3.0 ± 4.0 min/h) of weekend days | 9 |
| Cohen D.A., 2014 [72] | 2006–2008 | US | Quasi-experimental post-only assessment | PUM | Los Angeles | n.a. | CEA | SOPARC | PA in greenspace only | Average visitor number: higher for pocket parks (n = 147) than larger UGS (n = 134). Total PA performed shows opposite trend: 324 vs 374 METs) | 8 |
| Cohen D.A., 2017 [74] | 2014 | US | Cross-sectional | PUM | 25 US cities > 100,000 residents each | n.a. | LRM | SOPARC | PA in greenspace only | Parks with walking loops attract 80% (95% CI: 42–139%) [p < 0.001] more visitors per hour and show increased levels of MVPA with 90% more MET-hours (95% CI: 49–145%) [p < 0.001] than unequipped counterparts | 8 |
| Copeland J.L., 2017 [75] | 2015 | CA | Cross-sectional | PUM | Lethbridge | 1646 | T-test | SOPARC | PA in greenspace only | Only 2.7% of adult visitors used fitness equipments for PA | 5 |
| Parra D.C., 2019 [83] | 2018 | US | Cross-sectional | RUGF | Wellston | 599 | Chi2 | SOPARC | PA in greenspace only | Children and middle-aged adults represented 41.1% and 50.3% of total park users, respectively. A total of 47% of them practised MVPA, 22% LPA and 30% was sedentary | 5 |
| Ramírez P.C., 2017 [86] | 2015 | CO | Cross-sectional | RUGF | Bucaramanga | 6722 | Chi2 | SOPARC | PA in greenspace only | Women more prone to use outdoor gyms than men (51.7% against 48.3%) and to practise intense PA levels (W = 53.5%; M = 46.5%) | 4 |
| Roemmich J.N., 2018 [88] | 2014 | US | Cross-sectional | PUM, UFAP | Grand Forks, ND and East Grand Forks, MN | 5486 | T-test | SOPARC | PA regardless of the setting | Rural parks dwellers display lower MPA prevalence than urban parks (34%, n = 240 against 48%, n = 1828) | 9 |
| Sallis J.F., 2016 [90] | 2002–2011 | BE, BR, CO, CZ, DK, HK, MX, NZ, UK, US | Cross-sectional | RUGF | Ghent, Curitiba, Bogotá, Olomouc, Aarhus, Hong Kong, Cuernavaca, North Shore, Waitakere, Wellington, Christchurch, Stoke-On-Trent, Seattle, Baltimore | 10,008 | SEV MEV GAMMs | Accelerometer | PA regardless of the setting | Positive correlation between PA and urban parks presence within 0.5 Km of the participants’ home in Ghent (exp[β] = 1.772; 95% CI: 1.177–2.669; p = 0.006) and Seattle (exp[β] = 2.064; 95% CI: 1.399–3.045; p < 0.001) | 8 |
| Spengler J.O., 2011 [93] | 2005 | US | Cross-sectional | PUM, SUG, RUGF | Tampa, Chicago | 3410 | Multilevel regression | SOPLAY | PA in greenspace only | Children perform MVPA most frequently (56.2% boys, 55.7% girls, p-value n.a.) in parks with playgrounds than in all other UGS | 6 |
| Suau L.J., 2012 [94] | 2005 | US | Cross-sectional | PUM, SUG, RUGF | Tampa, Chicago | 9454 | Multilevel regression | SOPLAY | PA in greenspace only | In Chicago’s parks, PA was greater in African American (F = 5.027; p < 0.01) and high-income neighborhoods (F = 5.027; p = 0.002) | 4 |
| Author, year [Ref] | Study period | Country | Study design | Type of greenspace | City | Sample size | Statistical analysis | Tool used to measure PA | Outcome domain | Main results | QS/21 |
| Park S., 2018 [82] | 2013–2015 | US | Ecological | PUM | Los Angeles | 52,596 MPA, 5975 VPA | Chi2 | Accelerometer | PA in greenspace only | The proportion of park use time spent in MVPA (33.1%) was lower than the city-level average (35%) | 15/21 |
| Interventions | |||||||||||
| Author, year [Ref] | Study period | Country | Study design | Type of greenspace | City | Sample size | Statistical analysis | Tool used to measure PA | Outcome domain | Main results | QS |
| Andersen H.B., 2017 [64] | 2010; 2012 pre and post intervention | DK | Pre-post intervention | PUM, SUG | Copenhagen | 673 | Wilcoxon’s rank-sum test | Accelerometer, GPS, GIS | PA regardless of the setting | After intervention, 4.5 min/day increase in adolescents’ greenspace PA (95% CI: 1.8, 7.2; p < 0.001) | Fair |
| Cohen D.A., 2013 [71] | 2010–2011 | US | Randomized controlled trial | PUM | Albuquerfque, Chapel Hill, Columbus, Philadelphia | 36,000 | LRM | SOPARC | PA in greenspace only | Programmed activities (IRR: 1.79; p < 0.001) and the number of activity facilities (IRR: 1.13; p = 0.01) are associated with higher park use. Programmed activities (β = 192 ± 37; p < 0.001) and number of activity facilities (β = 28 ± 27; p = 0.30) are associated also with higher energy expended in the park too | Some |
| Cohen D.A., 2017 [73] | 2013–2015 | US | Randomized cluster trial | PUM | Los Angeles | 52,310 | DID models | SOPARC | PA in greenspace only | Free classes arm attracted more than twice park visits than the frequent user program. (p-value n.a.). (Among free classes arm it was show a 10% increase in total number of park users, more than twice the increasing percentage in frequent user program arm total number (p-value n.a.) | |
| Tester J., 2009 [95] | 2006–2007 | US | Pre-post intervention | SUG | San Francisco | 2041 | T-test | SOPARC | PA in greenspace only | Significant increase in visitors for PA among children (p < 0.05) and adults of both genders (p < 0.001) following parks’ renovations | |
| Mental Health | |||||||||||
| Author, year [Ref] | Study period | Country | Study design | Type of greenspace | City | Sample size | Statistical analysis | Tool used to measure MH | Outcome domain | Main results | QS/9 |
| Observational Studies | |||||||||||
| Andrusaityte S., et al., 2020 [65] | 2007–2009 | LT | Cross-sectional | PUM | Kaunas | 1489 | multivariate logistic regression | SDQ | Well-being/quality of life | Each increasing hour/week of park visits shows a non-significant association with mental difficulties: (aOR = 0.98 (0.96–1.01, [p < 0.05]) | 4 |
| Astell-Burt T., et al., 2019 [66] | 2006–2015 | AU | Cohort | Total greenspace | Sydney, Wollongong, and Newcastle | 4786 | multilevel logistic regression | K10 | Psychosocial stress | A 30% increase in total greenspace percentage is protective against both prevalent K10 psychological distress (aOR = 0.69 (0.47–1.02) [p = 0.03]) and incident K10 psychological distress (aOR = 0.46 (0.29–0.69) [p < 0.001]) | 8 |
| Balseviciene B., et al., 2014 [67] | 2007–2009 | LT | Cross-sectional | PUM | Kaunas | 1468 | LRM | SDQ | Well-being/quality of life | Proximity to city parks associated with increased mental difficulties in the lower maternal education subgroup (beta coefficient = 1.293, p < 0.05, R = 0.444). | 8 |
| Bixby H., et al., 2015 [68] | 2002–2009 | UK | Cross-sectional | PUM, SUG, RUGF, UFAP and BS | 50 largest cities in England | 5222 | Poisson linear regression | Mortality data: ICD-10 codes X60–84 | Suicide | Comparing quintiles 1 vs. 5 of greenspace coverage. RR of death from suicide was 1.02 (0.86–1.23) in men and 1.10 (0.77–1.57) in women [p < 0.05 for both]. | 5 |
| Camargo D.M., et al., 2017 [69] | 2015 | CO | Cross-sectional | PUM and SUG | Bucaramanga | 1392 | Multiple regression | EQ5D-5L | Well-being/quality of life | Positive associations between quality of life and: tree conditions status -> aPR = 1.20 (1.07–1.34), perceived safety -> aPR = 1.22 (1.04–1.44) [p < 0.05 for both] | 8 |
| Feda D.M., et al., 2015 [77] | 2008–2010 | US | Cross-sectional | PUM, SUG and RUGF | New York and Buffalo | 68 | Multiple regression analysis | PSS | Psychosocial stress | Percentage of park area predicted perceived stress β = −62.573, [p < 0.03] | 8 |
| Guite H.F., et al., 2006 [78] | n.a. | UK | Cross-sectional | Not specified | Greenwich (London) | 2696 | mutivariate logistic regression | SF-36v2 | Well-being/quality of life | Dissatisfaction with open UGS access significantly associated with lowest quartile for well being and quality of lifeOR = 1.69 (1.05–2.74) | 8 |
| Ihlebæk C., et al., 2018 [79] | 2000–2001 | NO | Cross-sectional | PUM, RUGF, UFAP, BS | Oslo | 8638 | Logistic regression | Not validated questionnaire | General mental health | With enhanced exposure to UGS, significant drop in MH disorders prevalence in women (−6% p = 0.049) but not in men (−2.5% p = 0.129) | 6 |
| Lee H.J., et al., 2019 [80] | 2015 | KR | Cross-sectional | PUM, SUG, UFAP and BS | 7 metropolitan areas in Korea | 11,408 | Binary logistic regression analysis | Not validated questionnaire | Depression and Psychosocial stress | Inverse relationship between stress levels, depressive symptoms and urban green area ratio (p < 0.005) | 7 |
| Pope, D., et al., 2018 [84] | 2009–2013 | UK | Cross-sectional | PUM | Sandwell | 1680 | Multivariable logistic regression | GHQ-12 | Psychological stress | Wider greenspace accessibility associated with reduced PD [OR = 0.13 (0.42, 0.94)] | 6 |
| Reklaitiene, R., et al., 2014 [87] | 2006–2008 | LT | Cross-sectional | PUM | Kaunas | 6944 | Multiple logistic regression | CES-D10 | Depressive symptoms | Living >300 m away from UGS and using them ≥4 h/week showed higher odds 1.92 (1.11–3.3) and 1.68 (0.81–3.48) of depressive symptoms | 6 |
| Ruijsbroek, A., et al., 2017 [89] | 2013 | ES, NL, LT, UK | Cross-sectional | NGS | Barcelona, Doetinchem, Kaunas, Stoke-on-Trent | 3771 | Multilevel regression analyses | MHI-5 | Nervous and feelings of depression in the past month | Only in Barcelona, NGS quantity was associated with better MH status (1.437 ± 0.71) p < 0.05 | 9 |
| Van Dillen, S.M., et al., 2012 [96] | 2007 | NL | Cross-sectional | SUG | 80 Dutch urban neighborhoods | 1641 | Multilevel regression | MHI-5 | General mental health | Perceived general health and green areas, had a significant interaction with the following parameters: quantity = 0.27 (0.013), quality = 0.126 (0.066), interaction term = 0.084 (0.040) | 5 |
| Zhang, Y., et al., 2015 [97] | 2014 | NL | Cross-sectional | PUM; SUG | Groningen | 223 | Multivarite ANOVA | MHI-5 | General mental health | Differences in neighborhood have a positive and significant influenceon mental health, β = 0.15, t(245) = 2.10, p < 0.05 | 5 |
| Author, year [Ref] | Study period | Country | Study design | Type of greenspace | City | Sample size | Statistical analysis | Tool used to measure MH | Outcome domain | Main results | QS/21 |
| Nutsford, D., et al., 2013 [81] | 2008–2009 | NZ | Ecological | PUM, SUG, RUGF, UFAP | Auckland City | 319,521, of which 7552 cases | Negative binomial regression models | Record linkage (treatment) | Mood state and general anxiety | Better access UGS access, and decreased distance (less than 3km) reduced the risk of anxiety/mood disorders treatment by 4% and 3% respectively (p < 0.01) | 12/21 |
| Interventions | |||||||||||
| Author, year [Ref] | Study period | Country | Study design | Type of greenspace | City | Sample size | Statistical analysis | Tool used to measure MH | Outcome domain | Main results | QS |
| Coventry P.A., et al., 2019 [76] | 2017 | UK | Pre-post intervention | PUM | York | 45 | One-way ANOVA + Bonferroni correction for multiple comparisons | SWEMWBS, UWIST-MACL | Affective/general and well-being/quality of life/ stress and (physical) arousal | UWIST-MACL mean difference (pre-post intervention stress levels across all participants at all locations) of −3.53 (4.79–2.28) [p < 0.001] | Fair |
| Pratiwi, P.I., et al., 2019 [85] | 2019 | JP | Pre-post intervention | PUM | Matsudo | 24 | Wilcoxon’s rank-sum test | POMS-STAI | Mood state and general anxiety | POMS scores: 0.71 in spring and 0.896 in summer. STAI score 0.896 and 0.933 respectively | Fair |
| Song, C., et al., 2015 [91] | 2014 | JP | Pre-post intervention | SUG | Kashiwa City | 20 | Wilcoxon’s rank-sum test | STAI | Anxiety and mood state | STAI score was 19.3% significantly lower after the urban park walk than after the city area walk (urban park: 39.0 ± 6.3; city area: 48.4 ± 7.5; p < 0.01) | Fair |
| South, E.C., et al., 2018 [92] | 2011–2014 | US | Randomized cluster trial | PUM | Philadelphia | 149 | Pairwise comparison using time serious regression | K6 | General mental health and depression | ITT analysis of the greening intervention demonstrated a non-significant reduction in overall self-reported poor MH with respect to non-intervention (−62.8%; 95% CI, −86.2% to 0.4%; p = 0.051) but a significant reduction in depressive symptoms (−41.5%; 95%CI, −63.6% to −5.9%; p = 0.03) | Low |
AU: Australia; BE: Belgium; BR: Brazil; BS: “Blue” spaces; CA: Canada; CEA: Cost-effectiveness analysis; CES-D10: Center for the Epidemiological Studies of Depression Short Form 10-items; CI: Confidence Interval; CO: Colombia; CZ: Czech Republic; DID: Difference-in-differences; DK: Denmark; EQ5D-5L: EuroQol 5 Dimensions-5 Levels; ES: Spain; F: Fisher’s F-test distribution; GAMM: Generalized Additive Mixed Models; GHQ-12: General Health Questionnaire-12; GIS: Geographic Information Systems; GPS: Global Positioning Systems; Exp: Expected; HK: Hong Kong; ICD-10: International Statistical Classification of Diseases and Related Health Problems 10; IRR: Incidence Rate Ratio; ITT: intention-to-treat JP: Japan; K6: Kessler-6-Psychological Distress Scale; K10: Kessler Psychological Distress Scale; KR: Korea; LRM: Linear regression model; LT: Lithuania; M: Men; METS: Metabolic Equivalents; MEV: Multiple Environmental Variable; MH: mental health; MHI-5: The Revised Mental Health Inventory-5; MN: Minnesota; MPA: Moderate-intensity Physical Activity; MVPA: Moderate/Vigorous Physical Activity; MX: Mexico; N: Number; ND: North Dakota; NL: Netherlands; NZ: New Zealand; OR: Odds Ratio; PA: Physical Activity; POMS-STAI: Profile of Mood States—State Trait Anxiety Inventory; PSS: Perceived Stress Scale; PUM: Parks and urban meadows; QS: Quality Score; RR: Relative Risk; RUGF: Recreational and urban gardening facilities; SDQ: Strengths and Difficulties Questionnaire; SEV: Single Environmental variable; SF-36v2: SF36 subscales for mental health; SOPARC: System of Observing Play and Recreation in Communities; SOPLAY: System for Observing Play and Leisure Activity in Youth; STAI: State-Trait Anxiety Inventory; SUG: “small” urban greenspaces; SWEMWBS: Short Warwick–Edinburgh Mental-Wellbeing Scale; UFAP: Urban forests and agricultural parks; UGS: urban greenspace; UK: United Kingdom; US: United States; UWIST-MACL: Measured by the University of Wales Institute of Science and Technology -Mood Adjective Checklist; VPA: Vigorous Physical Activity; W: Women; aPR: adjusted Prevalence Ratio; aOR: adjusted Odds Ratio; n.a.: not available; β: β coefficient.