Environmental Research 200 (2021) 111326 Contents lists available at ScienceDirect Environmental Research journal homepage: www.elsevier.com/locate/envres The relationship between residential proximity to outdoor play spaces and children’s mental and behavioral health: The importance of neighborhood socio-economic characteristics Carmen Pérez-del-Pulgar a, b, *, Isabelle Anguelovski a, b, c, Helen V.S. Cole a, b, Jeroen de Bont d, e, f, g, h, James Connolly a, b, m, Francesc Baró a, i, j, Yesika Díaz d, Mario Fontán-Vela k, l, Talita Duarte-Salles d, 1, Margarita Triguero-Mas a, b, n, 1 a Institute of Environmental Science and Technology (ICTA), Univesitat Autònoma de Barcelona (UAB), Barcelona, Spain b IMIM (Hospital del Mar Medical Research Institute), Barcelona, Spain c ICREA (Institució Catalana de Recerca i Estudis Avançats), Barcelona, Spain d Fundació Institut Universitari per a la recerca a l’Atenció Primària de Salut Jordi Gol i Gurina (IDIAPJGol), Barcelona, Spain e Universitat Autònoma de Barcelona (UAB), Bellaterra, Spain f ISGlobal, Barcelona, Spain g Spanish Consortium for Research on Epidemiology and Public Health (CIBERESP), Madrid, Spain h Universitat Pompeu Fabra (UPF), Barcelona, Spain i Vrije Universiteit Brussel (VUB), Geography Department, Pleinlaan 2, B-1050, Brussels, Belgium j Vrije Universiteit Brussel (VUB), Sociology Department, Pleinlaan 2, B-1050, Brussels, Belgium k Preventive Medicine Department, Infanta Leonor Universitary Hospital, Madrid, Spain l Public Health and Epidemiology Research Group, School of Medicine, University of Alcala, Madrid, Spain m School of Community and Regional Planning, University of British Columbia, Vancouver, Canada n Mariana Arcaya’s Research Lab, Department of Urban Studies and Planning, Massachusetts Institute of Technology, Cambridge, MA, USA A R T I C L E I N F O A B S T R A C T Keywords: Urban outdoor play spaces are reported to improve children’s health. However, there is little empirical evidence Childhood mental and behavioral health on the impact of outdoor play spaces on childhood mental and behavioral health. To fill this gap, we investigated Disorders of psychological development the associations between residential proximity to outdoor play spaces and the prevalence of diagnosed mental Urban environment and behavioral disorders. We explored whether these associations differ by individual and area-level socio- Green space economic status (SES). This cross-sectional study included 151 110 children who were 0–12 years old in 2014 Play space and were visited in public primary health care centers in Barcelona (Spain). Each child’s demographic and mental and behavioral disorders information was extracted for 2005–2014, including diagnoses on disorders of psychological development together with other four types of mental and behavioral disorders. The pediatrician diagnosed mental and behavioral disorders we explored in this study were: mood/affective; neurotic, stress- related and somatoform; psychological development; behavioral and emotional; and overall mental and behavioral disorders. We assessed 300 m network buffer residential proximity to overall outdoor play spaces (i.e., the overall sum of play spaces of any type), outdoor green play spaces, and to a diversity of outdoor play spaces. We used robust Poisson regression models to investigate the association between proximity to outdoor play spaces indicators and each health outcome. We tested interaction terms for indicators of proximity to outdoor play spaces and individual and area SES. For measures with significant interaction terms, we conducted stratified models. We found residential proximity to outdoor play spaces to be protective of disorders of psychological development. Proximity to overall outdoor play spaces, proximity to outdoor green play spaces and proximity to a greater diversity of outdoor play spaces were associated with a 4% (95% CI: 1,7), 4% (95% CI: 1,7) and 5% (95% CI: 2,9) lower prevalence rates of disorders of psychological development respectively. Most of the asso­ ciations were found to be in the same direction-although more pronounced-in low SES areas, but in the opposite direction for children living in high SES areas. No differences in these associations were found by individual SES. * Corresponding author. Universitat Autònoma de Barcelona, C/Doctor Aiguader, 88, 08003, Barcelona, Spain. E-mail address: carmenppulgar@gmail.com (C. Pérez-del-Pulgar). 1 Joint last authorship. https://doi.org/10.1016/j.envres.2021.111326 Received 7 January 2021; Received in revised form 16 April 2021; Accepted 10 May 2021 Available online 21 May 2021 0013-9351/© 2021 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). C. Pérez-del-Pulgar et al. Environmental Research 200 (2021) 111326 Residential proximity to outdoor play spaces is protective of children’s mental and behavioral health living in low SES areas. 1. Introduction 2. Materials and methods Today, cities in the Global North face the challenge of having un­ 2.1. Study design and population precedented high prevalence levels of childhood mental disorders (Amoly et al., 2014; Flies et al., 2019). A growing field of research at the We designed a semi-individual cross-sectional study in Barcelona intersection of public health and urban planning is enquiring the specific (Spain) using individual health data from the Information System for urban social and built-environment conditions that benefit or harm Research in Primary Care (SIDIAP; www.sidiap.org) in Catalonia, Spain human development in the first years of life (Derr et al., 2017; Flouri (Bolíbar et al., 2012). The Mediterranean coastal city of Barcelona had in et al., 2014; Malone, 2013). Increasing attention is also paid to envi­ 2014 a population of 1.6 million inhabitants, 200.890 of whom were ronmental (health) justice questions enquiring inequities in the distri­ children aged 0–12 (Barcelona City Council Statistical Yearbook, 2014). bution of urban social and built-environment conditions and their The city was divided in 1061 census tracts in 2014 with a median size of associated health benefits or harms by neighborhood disadvantage, 3.6 ha and average population per tract of 1511 residents. Outdoor play racial and ethnic characteristics (Anguelovski, 2013; Byrne and Wolch, spaces are fundamental for children’s social life in a city with an urban 2009; Hughey et al., 2016; Kamel et al., 2014; Pérez del Pulgar et al., form as dense and compact as Barcelona (Busquets, 2006). The SIDIAP 2020; Rigolon and Flohr, 2014; Strife and Downey, 2009; Vaughan database is a large pseudo-anonymized database of electronic health et al., 2013). records for all visits in primary care centers managed by the Catalan Neighborhood social conditions related with the concentration of Health Institute since 2005. It has a catchment of 5.5 M people, disadvantage, such as poverty, crime and unemployment rates have approximately 74% of the population living in Catalonia. SIDIAP in­ been associated with mental, behavioral and cognitive problems cludes demographic data, clinical variables, immunizations, specialists’ (Christian et al., 2015). In terms of the relation of the available outdoor referrals, prescriptions and dispensation of medications, history of sick play spaces – which are a fundamental part of children’s social life in leave and any acute and chronic health problems registered during a dense cities (Busquets, 2006)- with children’s mental and behavioral primary care visit. health, links have been found for different types of residential green The selection of children with ages from 0 to 12 registered in the space availability (Christian et al., 2015; McCormick, 2017; Markevych SIDIAP database and living in the municipality of Barcelona in 2014 et al., 2018; Vanaken and Danckaerts, 2018; Alderton et al., 2019; resulted in the extraction of data of 151 110 individuals corresponding Engemann et al., 2019; Madzia et al., 2019; Mnich et al., 2019; Bijnens to 75.22% of the age group in the city of Barcelona in 2014. We chose 12 et al., 2020; Thygesen et al., 2020). Specifically, green space availability years old as a cut-off age defining the end of childhood and the onset of has been associated with better emotional and behavioral outcomes adolescence (Britannica, 2020). . For the selected individuals their (Amoly et al., 2014; Balseviciene et al., 2014; Flouri et al., 2014; Mar­ complete data history on demographic data (date of birth, gender and kevych et al., 2014; Sobko et al., 2018), better mood indicators for nationality) and mental and behavioral health diagnoses from 2005 depression and anxiety (Maas et al., 2009), stress reduction and atten­ until 2014 was retrieved. This study was approved by the ethics com­ tion restoration (Huynh et al., 2013; Taylor and Kuo, 2011) and mittee of the Jordi Gol i Gurina Institute for Research in Primary Care self-discipline (Taylor et al., 2002). The diversity of outdoor play spaces (IDIAPJGol, 20/163). (i.e. diversity of types of outdoor play spaces such as green, sports-oriented, socially exposed, quiet, or with different sets of play 2.2. Outdoor play space indicators equipment) to which a child is exposed to has been found to provide opportunities for different types of play experiences and meet a broader We collected comprehensive spatial data on public outdoor play array of needs of children of different genders, physical abilities, ages spaces in the city of Barcelona either specifically planned for children or and developmental stages (Dyment and O’Connell, 2013; Luken et al., where children frequently play in the city (Lynch, 1977). Data was ob­ 2011; Stanley, 2011) and potentially associated health benefits. The tained from the Urban Ecology Department. We identified 1665 play congestion of green spaces has been identified as an access barrier -as it spaces, including playgrounds, plazas, parks, gardens, urban forests and may decrease the attractiveness of the space - (Biernacka and Kronen­ recreational sports fields for 2014. berg, 2018) which can reduce its use and the associated health benefits. We applied a geospatial analysis of residential proximity to these To our knowledge, no study has accounted for the differential asso­ play spaces based on the intersection between 300 m-network buffers ciations between the residential proximity to green and non-green out­ around play spaces and the census tracts (Graph 1). The 300 m threshold door play spaces – such as parks, plazas, sport fields or playgrounds – as was chosen for being the average walking independent mobility stan­ well as the proximity to a diversity of play opportunities and children’s dard for children defined by UNICEF (United Nations, 2018). Network mental and behavioral health. Furthermore, few studies have explored buffers were calculated based on the centroid of the play spaces, except differences in the distribution of the benefits of outdoor play space for fenced parks and urban forests, for which the main access point (i.e., proximity for children’s mental health by individual or area-level socio- main entrance) was used as the center of the network buffer. Based on demographic characteristics (for exceptions see Flouri et al., 2014; this analysis, for each census tract we calculated the following in­ Wells, 2000). In response to those research gaps, the aim of our study is dicators: 1) total number of 300 m network buffers around play spaces to investigate the association between residential proximity to outdoor intersecting each Barcelona census tract; 2) number of 300 m network play spaces and the prevalence of diagnosed mental and behavioral buffers around outdoor green play spaces intersecting each Barcelona disorders in children. census tract; 3) number of 300 m network buffers around outdoor We further incorporate some insights from environmental justice by playgrounds intersecting the census tract; 4) number of 300 m network exploring potential inequities in the distribution of the mental and buffers around sport-oriented outdoor play spaces intersecting the behavioral health benefits of outdoor play space (Anguelovski et al., census tract; and 5) number of 300 m network buffers around com­ 2020), by examining whether play space-health associations differ by munity outdoor public play spaces intersecting the census tract. individual and area-level socio-demographic characteristics. Whereas indicator 1 counted each play space buffer once, indicators 2–5 were sometimes overlapping with one outdoor play space possibly 2 C. Pérez-del-Pulgar et al. Environmental Research 200 (2021) 111326 falling into several categories. (e.g., a play space with play equipment health diagnoses that were not associated with physiological distur­ located in a park was considered both a playground and a green play bances, physical factors or psychoactive substance use. These health space). Indicators 2–5 were meant to capture the type of play opportu­ diagnoses are registered by pediatricians following the International nities and experiences offered by the outdoor play space (Gibson, 1979; Statistical Classification of Diseases and Related Health Problems-10 Hart, 1979; Frost, 1992). Parks, gardens, urban forests and playgrounds (ICD-10) clinical evaluation and coding criteria. Each child can have and recreational sports fields located inside parks were counted as several entries for the same and/or different diagnoses over the years. outdoor green play spaces (Indicator 2). Outdoor play spaces with Accordingly, the mental and behavioral health indicators we included in traditional play equipment were counted as playgrounds (Indicator 3). our study were: Sports fields were counted as sport-oriented outdoor play spaces (Indi­ Mood/affective disorders: containing disorders in which the funda­ cator 4) and plazas were considered as outdoor community play spaces mental disturbance is a change in affect or mood to depression or to (indicator 5) (Graph 1). elation, ICD-10 codes: F30–F39. Based on these indicators, we computed our three main exposure Neurotic, stress-related and somatoform disorders: including several variables: a) Residential proximity to overall play spaces, defined as the phobias, anxiety and severe stress, ICD-10: F40–F48. total number of outdoor play spaces whose 300 m network buffer in­ Disorders of psychological development: which in most cases refer to tersects the census tract divided by the number of children living in the disorders in language, visuo-spatial skills, and motor coordination, ICD- census tract; b) Residential proximity to green play spaces, defined as 10: F80–F89. the total number of green outdoor play spaces whose 300 m network Behavioral and emotional disorders: with onset usually occurring in buffer intersects the census tract divided by the number of children childhood and adolescence and characterized by a lack of persistence in living in the census tract and c) Residential proximity to a diversity of activities that require cognitive involvement, ICD-10: F90–F98. play spaces, calculated as an index of the diversity of play types offered Overall mental and behavioral disorders: indicating the presence of by play spaces whose 300 m network buffer intersected the census tract any of the previously mentioned disorders (ICD-10: F30–F39, F40–48, (Graph 1). F80–F89, F90–F98) and/or unspecified mental disorder (ICD-10: F99). For exposure variables a) proximity to overall outdoor play spaces To assess prevalence for each mental and behavioral indicator, we and b) proximity to outdoor green spaces, we divided the number of created dichotomous variables for each of the previously cited mental outdoor play spaces by the number of children living in each census tract and behavioral disorder groups where 1 indicated that the child had one to account for the potential pressure of use of each outdoor play space. or more diagnoses of the considered indicator in the period 2005–2014. This is based on the assumption that a high competition for outdoor play For all health indicators, the date of the diagnosis was defined as the first space (i.e. low number of proximate outdoor play spaces per child) can recorded date at which the disorder appeared in their record. be considered an entry barrier (Biernacka and Kronenberg, 2018) and be associated to lower mental health benefits. The variable c) proximity to a diversity of play spaces index was 2.4. Individual and area-level socio-economic status covariates calculated using the Shannon index of diversity (Spellerberg and Fedor, 2003)) considering the types of outdoor play space defined in indicators Individual variables on gender (dichotomous variable), age in 2014 2–5: green play spaces; playgrounds; sport-oriented play spaces; and (as a continuous variable), and nationality (dichotomous variable community play spaces, as follows: defined as either nationality from Global South – i.e., Latin-America and the Caribbean, Africa and Asia/Middle East- or Global North – i.e. ∑5 ( ( )) H= − ni * log ni Europe and Anglo-Saxon America-were retrieved from the SIDIAP i=2 N N database. Nationality was used as a proxy of racial and ethnic charac­ teristics, as used in other recent studies in the context of Barcelona where, n is the total amount of outdoor play spaces of each type, N the (Anguelovski et al., 2018). total amount of play spaces in the census tract and i the indicator. To estimate area-level socio-economic status (SES), we developed a census tract SES index linked to the children’s census tract of residence. 2.3. Mental and behavioral disorders To do so, first we selected relevant variables a priori using information from similar indicators developed in the Spanish context to theory- We extracted information on children’s mental and behavioral inform our selection (Anguelovski et al., 2018; Domínguez-Berjón Graph 1. Illustration of the process of composition of the exposure variables. 3 C. Pérez-del-Pulgar et al. Environmental Research 200 (2021) 111326 et al., 2008). Accordingly, we used the following area-level data: index 2.6. Sensitivity analyses of family income (household disposable income); percentage of popu­ lation (i) with university degree or higher, (ii) with non-western na­ We conducted several sensitivity analyses to evaluate the robustness tionality (i.e. all African Countries, Philippines, Pakistan, Peru, Bolivia, of our findings. First, to explore whether differences between the area- Ecuador, Colombia, Dominican Republic, those being the most level SES indicators influenced our findings we performed models numerous and identifiable Global South nationalities in Barcelona’s substituting the adjustment of our area-level SES index by the dis­ census tract available statistical data), (iii) aged 65 or more, (iv) un­ aggregated variables for area-level average monthly rent per census employed; percentage of mono-parental households; percentage of tract, income (household disposable income per neighborhood), edu­ housing stock (i) privately owned, (ii) occupied by the owner, (iii) in cation (percentage of population with university degree or higher per deficient condition; and average monthly rent (€/m2). All the variables census tract) and nationality (percentage of population with non- used to develop the area level SES index were available at the Barcelona Western nationality per census tract), which are the variables included city Statistics Department, were at the census tract level (apart from in the area level SES index, plus, an indicator of the percentage of family income which was at neighborhood level) and referred to year population with non-Western nationality per census tract which did not 2014 (apart from percentage of housing stock occupied by the owner or show to be correlated with other socio-economic variables in the PCA in deficient condition and percentage of population unemployed which but is theoretically relevant according to the literature (Domí­ was available from 2011). Second, following the same methodology as nguez-Berjón et al., 2008; Anguelovski et al., 2018). the one used to develop the previously-used MEDEA indicator (Domí­ Second, to evaluate the robustness of our findings to variations in the nguez-Berjón et al., 2008), we performed a principal component anal­ individual nationality indicator we performed: 1)models categorizing ysis (PCA) to identify clustered variables whose variation could be individual nationality as Spanish and non-Spanish, given the high share explained by one index. For that purpose, we used an orthogonally of children with Spanish nationality (87.4%) and the potential con­ rotated principal component analysis (PCA) according to the Varimax founding effect it can have on the Global South/Global North individual method and selected the extraction of components with Eigenvalues nationality categorization and 2) models excluding children with Asian greater than 1. The analysis indicated a first component that explained nationality and from Central and Eastern Europe from the analysis, 25.94% of the total variance of the initial ten variables. We considered considering that these nationalities included an especially heteroge­ that indicators were highly correlated with the first component if their neous mix of social, economic and ethnic statuses in the context of component loadings were greater or equal to 0.70 (Domínguez-Berjón Barcelona. Models excluding children with Asian nationality from the et al., 2008), which resulted in the selection of three variables related to Global South category and children with Central and Eastern European the economic and educational capital of the census tract (percentage of nationality from the Global North category reduced the sample to 145 population with university degree or higher, family income and average 026 children. monthly rent), ruling out the rest of variables. We then conducted a Third, since we use the outdoor play space proximity indicators second orthogonally rotated PCA including only the three selected in­ associated to the children’s census tract of residence in year 2014 but dicators and adopted the extraction of one component. We used the first some children might have changed census tract of residence in the (and only) component factor, which explained 84.05% of the total period 2005–2014, we performed models excluding children who variance of the three indicators (percentage of population with univer­ changed census tract of residence in the period 2011–2014 (data on sity degree or higher component loading 0.93, family income compo­ change of residency is not available prior to 2011). Models excluding nent loading 0.93 and average monthly rent component loading 0.82), children who moved in the studied period reduced the sample to 114 as our census tract SES index. Using this index, Barcelona census tracts 460 children. were classified alongside four equal groups according to the distribution Last, we conducted sensitivity analyses with the indicators of chil­ of de value of the index across the city’s census tracts. dren’s residential proximity to outdoor play spaces as a categorical versus continuous variable to assess the assumption of a linear rela­ 2.5. Statistical analysis tionship between proximity to outdoor play spaces and mental and behavioral disorder outcomes in the robust Poisson model. We conducted descriptive analyses of the prevalence of our mental and behavioral health indicators by gender, age, nationality and area- 3. Results level SES index. Then, we estimated the relationship between outdoor play space indicators and prevalence of mental and behavioral disorders 3.1. Descriptive statistics outcomes. For each outcome/exposure combination, we developed robust Poisson regression models to estimate the change in the preva­ Descriptive statistics for the characteristics of study participants, the lence of each health outcome associated with an interquartile range prevalence of our investigated disorder outcomes and the play space (IQR) increase of each play space indicator. We used Poisson regression indicators are presented in Table 1. Maps of our outdoor play space with robust variance for understanding that it provides correct estimates indicators, and area-level SES index are presented in Supplemental and is a better alternative for the analysis of cross-sectional studies with Material, Figures S1–S3. binary outcomes than a logistic regression, since the prevalence ratio Our sample had a balanced presence of both genders (51.43% of (PR) is more interpretable and easier to communicate than the odds boys). Most children had a Global North nationality (91.2%), mainly ratios (OR) (Barros and Hirakata, 2003). We adjusted for gender, age, Spanish (87.41%) but also South European (1.43%), Central and Eastern nationality and area level SES index and calculated prevalence ratios European (1.57%), Northern European (0.79%) and Anglo-Saxon and 95% Wald confidence intervals in all models. Then, to evaluate the American (0.15%) (data not shown). For the children that had a effect modification of individual and area-level SES variables, we Global South nationality, most were from Latin-America and the included interaction terms between IQR increases in outdoor play space Caribbean (4.35%), and the rest were: 2.45% Asia and Middle Eastern, indicators and SES variables in the robust Poisson regression models 1.44% Northern African, and 0.39% Central and Southern African. The (one interaction by each SES variable in each outcome/exposure com­ study population had an even distribution of children in different ages bination model). For measures with significant interaction terms, we 0–12 (median 6 and IQR = 6). then conducted stratified models. We used SPSS, version 26 and set In our study population, bit more than 10% of the children - more statistical significance at p-value<0.05. frequently boys (62.51%) than girls (data not shown) - were diagnosed in the period 2005–2014 with any mental and behavioral disorder, at a median age of 6 (IQR = 4). From these, the disorder outcome most 4 C. Pérez-del-Pulgar et al. Environmental Research 200 (2021) 111326 Table 1 mental and behavioral disorder decreased to over 6% among the study Descriptive statistics of sample sociodemographic characteristics, health out­ population living in census tracts with the highest area-level SES index. comes and outdoor play space proximity variables (n = 151 110). There was a big variability within all indicators of proximity to Variable n (%)/median outdoor play spaces from the census tracts of residence. The minimum (IQR) number of overall play spaces per 1000 children within 300 m from Sociodemographic characteristics children’s census tracts of residence was 0.37 and the maximum 77.53 (data not shown). Similarly, the minimum number of green play spaces Gender, Girls [n (%)] 73403 (48.57) Nationality, Global South [n (%)] 13073 (8.65) per 1000 children within 300 m from children’s census tracts of resi­ Age [years: median (IQR1)] 6 (6) dence was 0 and the maximum 60.67. Children of our study were Area-level SES index [index: median (IQR)] − 0.39 (1.34) exposed to a diversity of outdoor play spaces between 0 and 1.28. The Area-level SES index 1st quartile, lowest SES [n (%)] 49650 (32.85) indicator of proximity to overall outdoor play spaces was strongly Mood/affective disorders, Diagnosed [n (%)] 107 (0.22) Neurotic, stress-related and somatoform disorders, Diagnosed 1027 (2.07) correlated with the indicator of proximity to green play spaces (Spear­ [n (%)] man’s correlation coefficient r = 0.82). Contrary, the correlation be­ Disorders of psychological development, Diagnosed [n (%)] 1998 (4.02) tween overall play spaces and diversity of play spaces and between Behavioral and emotional disorders Diagnosed [n (%)] 3747 (7.55) green play spaces and diversity of play spaces was very weak (r = 0.05 Overall mental and behavioral disorders, Diagnosed [n (%)] 6101 (12.29) and r = 0.16 respectively). Area-level SES index 2nd quartile [n (%)] 45089 (29.84) Mood/affective disorders, Diagnosed [n (%)] 51 (0.11) Neurotic, stress-related and somatoform disorders, Diagnosed 800 (1.77) 3.2. Main results [n (%)] Disorders of psychological development, Diagnosed [n (%)] 1494 (3.31) Lower prevalences of disorders of psychological development were Behavioral and emotional disorders Diagnosed [n (%)] 2699 (5.99) Overall mental and behavioral disorders, Diagnosed [n (%)] 4527 (10.04) associated with higher values of the indicators of outdoor play space Area-level SES index 3rd quartile [n (%)] 29953 (19.82) (Table 2). For all play space indicators, an IQR increase of the play space Mood/affective disorders, Diagnosed [n (%)] 52 (0.17) indicator was significantly associated with lower prevalence of disorders Neurotic, stress-related and somatoform disorders, Diagnosed 593 (1.98) of psychological development. That is, an increase in 64 overall play [n (%)] spaces per 1000 children (i.e., an IQR increase in the overall play spaces Disorders of psychological development, Diagnosed [n (%)] 870 (2.90) Behavioral and emotional disorders Diagnosed [n (%)] 1673 (5.59) indicator) within 300 m from children’s census tracts of residence was Overall mental and behavioral disorders, Diagnosed [n (%)] 2826 (9.43) associated with a 4% (95% CI: 1,7) decrease in the prevalence of dis­ Area-level SES index 4th quartile, highest SES [n (%)] 26418 (17.48) orders of psychological development. Each IQR (36) increase in the Mood/affective disorders, Diagnosed [n (%)] 48 (0.18) number of green play spaces per 1000 children within 300 m from the Neurotic, stress-related and somatoform disorders, Diagnosed 333 (1.26) [n (%)] census tracts of residence was associated with a 4% (95% CI 1,7) lower Disorders of psychological development, Diagnosed [n (%)] 482 (1.82) prevalence of disorders of psychological development. And an increase Behavioral and emotional disorders Diagnosed [n (%)] 1097 (4.15) in 0.29 units of the Shannon index within 300 m from children’s census Overall mental and behavioral disorders, Diagnosed [n (%)] 1807 (6.84) tracts of residence was associated with a 5% (95% CI 2,9) lower prev­ Mental and behavioral disorders alence of disorders of psychological development. Mood/affective disorders, Diagnosed [n (%)] 258 (0.17) Neurotic, stress-related and somatoform disorders, Diagnosed 2753 (1.82) An increase in 64 overall play spaces per 1000 children (i.e., an IQR [n (%)] increase) within 300 m from children’s census tracts of residence was Disorders of psychological development, Diagnosed [n (%)] 4844 (3.21) also associated with a 2% (95% CI: 1,3) decrease in the prevalence of Behavioral and emotional disorders Diagnosed [n (%)] 9216 (6.10) overall mental and behavioral health disorders. The rest of mental and Overall mental and behavioral disorders, Diagnosed [n (%)] 15261 (10.10) Outdoor play space indicators behavioral health outcomes were not statistically significantly associ­ Overall play spaces per 1000 children [number of play spaces: 67 (64) ated with any of the play space indicators. median (IQR)] We did not find any suggestion of differences on the associations Green play spaces per 1000 children [number of play spaces: 34 (36) between outdoor play space indicators and individual nationality (p- median (IQR)] value>0.08 for the interaction terms). However, we found indications of Diversity of play spaces [index: median (IQR)] 0.9 (0.29) differences by area-level SES index on the associations between preva­ lence of disorders of psychological development with overall play spaces diagnosed was behavioral and emotional disorders, again mostly boys and also with diversity of play spaces (p-value = 0.02 and p-value = 0.01 (64.59%) and at the same median age than for the overall diagnosed. for the interaction terms, respectively). We also found some marginal Less than 5% of our study population were diagnosed from each of the indication of differences by area-level SES index on the associations remaining disorder outcomes. From these remaining disorders, children between prevalence of disorders of psychological development with in our sample were diagnosed of disorders of psychological development green play spaces (p-value = 0.07), which we considered worth at a median age of 5 (IQR = 4), with the prevalence being higher among exploring given the significance of the interaction terms of all other boys (65.28%). Similarly, children (particularly boys, 54.09% of the exposures with this outcome. Moreover, we found indications of dif­ diagnosed in our sample) were diagnosed of neurotic, stress-related and ferences by area-level SES index on the association of overall mental and somatoform disorders at a median age of 7 (IQR = 4). Last, the average behavioral disorders with overall play spaces (p-value<0.01 for the age of mood/affective disorders was slightly higher, 8 years old (IQR = interaction term). 4), despite those more frequently diagnosed were also boys (51.16%) Stratified models by area-level SES index revealed a general trend of The distribution of our study population across the four quartiles of proximity to overall outdoor play spaces and diversity of play spaces area-level SES index followed a decreasing trend, with over 30%of the being protective of disorders of psychological development for children study population living in census tracts with the lowest area-level SES living in low SES census tracts but a risk factor for children living in high index and around 18% of the study population living in census tracts SES census tracts (see Table 3). An IQR (64) increase in the number of with the highest area-level SES index. The prevalence of all mental and overall proximate outdoor play spaces per 1000 children within 300 m behavioral disorder outcomes also followed a decreasing trend as the from children’s census tract of residence was associated with an 8% area-level SES index increased. Over 12% of the study population living lower prevalence of disorders of psychological development for children in census tracts with the lowest area-level SES index were diagnosed living in the first and second lowest area-level SES index census tracts (i. with any mental and behavioral disorder, whereas the prevalence of any e. area-level index first quartile with 95% CI: 2,13 and area-level index second quartile with 95% CI: 2,14). Similarly, an IQR (0.29) increases in 5 C. Pérez-del-Pulgar et al. Environmental Research 200 (2021) 111326 Table 2 Adjusted associations between prevalence of mental and behavioral disorders and residential proximity to outdoor play space indicators derived from robust Poisson regression models. Associations reported for 1-IQR increase in outdoor play space indicators within 300 m from children’s census tracts of residence. Overall play spaces Green play spaces Diversity of play spaces Prevalence Ratio (95% CI) p-value Prevalence Ratio (95% CI) p-value Prevalence Ratio (95% CI) p-value Mood/affective disorders 0.95 (0.83, 1.09) 0.48 0.96 (0.85, 1.08) 0.51 1.04 (0.88, 1.23) 0.68 Neurotic, stress-related and somatoform disorders 0.98 (0.94, 1.02) 0.37 0.99 (0.95, 1.02) 0.46 1.02 (0.97, 1.09) 0.41 Disorders of psychological development 0.96 (0.93, 0.99) 0.02 0.96 (0.93, 0.99) <0.01 0.95 (0.91, 0.98) <0.01 Behavioral and emotional disorders 0.98 (0.96, 1.01) 0.15 0.99 (0.97, 1.01) 0.31 0.99 (0.96, 1.01) 0.32 Overall mental and behavioral disorders 0.98 (0.97, 0.99) 0.05 0.99 (0.97, 1.00) 0.09 0.98 (0.96, 1.01) 0.14 Note: All models include individual gender, nationality, age and area-level SES index as covariates. Table 3 Adjusted associations between prevalence of mental and behavioral disorders and residential proximity to outdoor play space indicators derived from robust Poisson regression models. Associations reported for 1-IQR increase in outdoor play space indicators within 300 m from children’s census tracts of residence. Models stratified by area-level SES index. Overall play spaces Green play spaces Diversity of play spaces Prevalence Ratio (95% CI) p-value Prevalence Ratio (95% CI) p-value Prevalence Ratio (95% CI) p-value Disorders of psychological development Area-level SES index 1st quartile (lowest SES) 0.92 (0.87, 0.98) <0.01 0.94 (0.89, 0.99) 0.02 0.92 (0.86, 0.97) <0.01 Area-level SES index 2nd quartile 0.92 (0.86, 0.98) <0.01 0.92 (0.87, 0.97) <0.01 0.97 (0.90, 1.05) 0.44 Area-level SES index 3rd quartile 1.09 (1.01, 1.18) 0.04 1.04 (0.96, 1.13) 0.35 0.93 (0.85, 1.01) 0.06 Area-level SES index 4th quartile (highest SES) 1.02 (0.90, 1.16) 0.71 1.02 (0.90, 1.15) 0.80 1.16 (1.03, 1.32) 0.02 Overall mental and behavioral disorders Area-level SES index 1st quartile (lowest SES) 0.94 (0.92, 0.97) <0.01 (not estimated)2 (not estimated) Area-level SES index 2nd quartile 0.98 (0.95, 1.01) 0.23 (not estimated) (not estimated) Area-level SES index 3rd quartile 1.04 (0.99, 1.08) 0.10 (not estimated) (not estimated) Area-level SES index 4th quartile (highest SES) 1.08 (1.02, 1.14) 0.01 (not estimated) (not estimated) Note: All models include individual gender, nationality and age as covariates. the diversity of proximate play spaces was associated with 8% (95% categorizing nationality as Spanish and non-Spanish, when excluding CI:3,14) lower prevalence of disorders of psychological development for children with Asian nationality and from Central and Eastern Europe, or children living in the lowest SES census tracts (i.e., area-level index first when excluding children who changed census tract of residence from the quartile). Also, an increase in 36 green play spaces per 1000 children analysis - were generally consistent with those of the main analyses in within 300 m from children’s census tracts of residence was associated terms of direction and statistical significance (Supplemental material, with a 6% (95%CI: 1, 11) lower prevalence of disorders of psychological Tables S1–S4). Only some minor differences were found when development for children living in the lowest area-level SES index substituting the adjustment of our area-level socio economic indicator census tracts, and with an 8% (95%CI: 3, 13) lower prevalence for those by the disaggregated variables of income, education, nationality, and living in the second lowest area-level SES index census tracts. However, monthly rent, were we found negative associations between proximity for children living in the highest SES census tracts these associations to overall outdoor play spaces and green outdoor play spaces and reversed their direction: Each IQR (64) increase in the proximate overall behavioral and emotional disorders, at the expense of the associations outdoor play spaces was associated with 9% (95%CI: 1,18) higher with disorders of psychological development (Supplemental Material prevalence of disorders of psychological development for children living Table S1). in the second highest area-level SES index census tracts (third quartile). IQR increases in the diversity of proximate outdoor play spaces were 4. Discussion associated with a 16% (95% CI:3,32) higher prevalence of disorders of psychological development for children living in the highest area-level In this study including 151 110 children living in Barcelona, we SES index census tracts (i.e., area-level index forth quartile) (see found that lower prevalence of disorders of psychological development Table 3). was consistently associated to increases in the residential proximity to Similar results were found for proximity to overall outdoor play overall and green outdoor play space as well as to a greater diversity of spaces in its associations with overall mental and behavioral disorders: play opportunities. Meanwhile, lower prevalence of overall mental and An IQR (64) increase in the overall proximate outdoor play spaces was a behavioral disorders was also found to be associated with increases in protective factor for children living in low SES census tracts but a risk the proximity to overall outdoor play spaces. Our findings also indicate factor for children living in high SES census tracts (see Table 3). That is, that these associations are not equal across area-level SES characteris­ each IQR (64) increase in the proximate overall outdoor play spaces was tics. We found that, for those children living in the lowest SES census associated with 6% (95%CI: 3,8) lower prevalence of disorders of psy­ tracts, the indicators of residential proximity to different indictors of chological development for children living in the lowest SES census outdoor play space had a protective role for their mental and behavioral tracts (i.e., area-level index first quartile) and an 8% (95%CI: 2,14) health. However, residential proximity to overall play spaces and to a higher prevalence of disorders of psychological development for chil­ diversity of proximate play spaces were risk factors for those children dren living in the highest SES census tracts (i.e. area-level index forth living in higher SES census tracts. quartile). The association we found between proximity to overall outdoor play spaces with lower prevalence of disorders of psychological development 3.3. Sensitivity analyses is novel. We hypothesize that our findings could be explained by the activities that outdoor play spaces have been shown to offer. For Our findings - when substituting the adjustment of our area-level example, previous research has indicated that a higher presence of socio economic indicator by the disaggregated variables; when outdoor play spaces is associated with greater overall physical activity 6 C. Pérez-del-Pulgar et al. Environmental Research 200 (2021) 111326 (Dunton et al., 2014; Timperio et al., 2008). Outdoor play spaces have SES areas. Higher-income families also have greater access to larger also been linked to other precursors of better mental health such as backyards and private gardens, second homes -which is especially active participation and interaction (WHO, 2001), negotiation of one’s prevalent in Spain (Módenes and López-Colás, 2007) -or are financially identity and sense of purpose (Cederborg, 2020; Compton-Lilly et al., able to travel further distances and provide alternative access to outdoor 2017), increased sense of community (Anguelovski, 2014; play spaces, meaning that municipal outdoor play spaces might not be Perez-del-Pulgar et al., 2020), awareness of one’s self and others (Mayer the primary outdoor play spaces and the determinants of high SES et al., 2009), experience of independent mobility (Schoeppe et al., children’s good mental health. In contrast, children from urban low SES 2016), free play and exploratory thinking (Holt et al., 2015), stress areas are those for whom urban outdoor play and contact with nature mitigation (Ulrich et al., 1991), and attention restoration (Taylor and tend to be mostly facilitated by formal municipal play spaces rather than Kuo, 2011). private play spaces, to which they tend to have more limited access, Our findings with regards to the protective role of the proximity to hence the protecting effects our findings indicate. In addition, residents outdoor green play spaces are in line with previous studies reporting of Barcelona’s working-class neighborhoods have long mobilized for better children’s development (Alderton et al., 2019; Christian et al., public play spaces and value their construction, design, access, and use. 2015; Vaden-Kiernan et al., 2010; Wells, 2000; Wu et al., 2014) and Under these conditions outdoor play spaces tend to become a source of general mental health outcomes (Sobko et al., 2018; Tillmann et al., pride for the community (Pérez del Pulgar, 2020) which potentially gets 2018) associated with higher exposure to green spaces. However, our transmitted to children (Putra et al., 2021) having an impact on how findings differ from previous studies with regards to the particular children use, experience and care for these spaces. More regular uses of mental and behavioral health disorders significantly associated with these spaces and feelings of safety and attachment might explain some of exposure to green spaces. Whereas previous studies present exposure to the positive health outcomes. Last, we might hypothesize that the ten­ green being protective of mood disorders (Maas et al., 2009), neurotic dency of high SES families with children to move to areas with better and stress related disorders (Huynh et al., 2013) or behavioral and children’s facilities, including outdoor play spaces (Lilius, 2019) is emotional disorders (Amoly et al., 2014; Balseviciene et al., 2014; Flouri enhanced amongst high SES families with children with mental health et al., 2014; Markevych et al., 2014; Mårtensson et al., 2009; Taylor disorders. In any case, more research on the pathways through which et al., 2002; Taylor and Kuo, 2011), our results show mainly a protection proximity to outdoor play spaces is associated to higher risks of disor­ to disorders of psychological development. We argue that these differ­ ders of psychological development in children living in the higher SES ences are secondary and could point to the difficulty to clearly distin­ areas. guish and separate each category of mental disorders given the porosity Our study faced some limitations. First, our study considered the 300 between the different mental health disorders and the heterogeneity of m network distance from the outdoor play spaces to the census tract their clinical presentations (American Psychiatric Association, 2013). boundaries as a proxy for residential proximity to the play space because We are unaware of previous studies on the association of the prox­ we did not have children’s individual residential addresses. This expo­ imity to a greater diversity of play spaces and the prevalence of chil­ sure introduced an ecological bias and, in some cases the actual walking dren’s mental and behavioral health outcomes overall or disorders of distance from the child’s home to the outdoor play space might be psychological development in particular. In this respect our results are higher than 300 m. Second, the study’s cross-sectional design, limits its novel. However, our results are coherent with previous research draw­ ability to determine causality. Third, mental and behavioral health is ing attention to the importance of the built environment (Derr and associated with a wide range of individual child and parental factors (e. Tarantini, 2016; Malone, 2013; Moore-Cherry, 2014; Pérez del Pulgar g. physiological conditions of the child, family members with mental et al., 2020; Woolcock et al., 2010), its richness, diversity of purposes and behavioral health disorders or specific traumatic events in the and play experiences (Dyment and O’Connell, 2013; Luken et al., 2011; child’s life) that, for data limitations, cannot be included in the present Stanley, 2011) for children’s wellbeing. study as confounder variables. Fourth, individual race and ethnicity Our results point to a modification of the association between resi­ have been suggested to influence the association between the exposure dential proximity to outdoor play space and mental and behavioral to outdoor play spaces and health outcomes, but that information was health by area-level socio-demographic characteristics in a manner unavailable to us for this study. The most similar available data was on consistent with previous literature, suggesting that outdoor play spaces individual nationality, aggregated in mixed categories that included may mitigate the negative influences of other aspects of the physical children with very different types of advantages and disadvantage, ac­ environment, such as poor housing, deteriorated neighborhoods, over­ cess to material and cultural resources, education, incomes, language crowded schools, or the prevalence of crime and violence (Engemann etc. We ran a sensitivity analysis to test the effect of this factor, but we et al., 2019). Thus, the role of residential proximity to overall outdoor cannot rule out that other (not available) data may have been better play spaces, green outdoor play spaces as well as to a diversity of these estimates of ethnicity in our study. Fifth, our study considered the could be especially important and protective for children’s mental number of outdoor play spaces per 1000 children within 300 m or less health in lower SES areas, indicating strong environmental justice and from each census tract. Testing the impact of the available area of play or health equity benefits. Meanwhile, for children living in high SES areas, quality of the area was not possible due to data limitations. Sixth, there proximity to outdoor play spaces seems to be working as a could be some measurement error due to the linkage of the cross- socio-environmental risk factor vis a vis children’s mental and behav­ sectional play-space proximity for year 2014 with the census tract res­ ioral health outcomes. This difference may reflect spatially bounded idential data of the period 2005–2014. However, due to the nature of class-based differences in the use and meaning of children’s play spaces Barcelona’s little post 2005 urban transformation in terms of the out­ (Pérez del Pulgar et al., 2020) influenced by cultural, social and his­ door play space amenities here analyzed we assume this measurement torical perspectives and ideologies of outdoor recreation as well as error to be potentially very minor. Seventh there might be an under­ material socioeconomic factors (Byrne and Wolch, 2009; Floyd, 2001; reporting of health outcomes, since the used database does not capture Strife and Downey, 2009; Tierney et al., 2001). diagnoses performed in private health centers. Despite Spain has a In this line, previous research suggests that upper- and middle-class universal coverage of healthcare and that children visit their pediatri­ cultures of parenting (Villanueva et al., 2016), time management cians even when they are healthy to follow vaccinations protocols – (Loukaitou-Sideris and Sideris, 2010) and perceptions of safety between others -, our number of diagnoses could be underestimated (Arroyo-Johnson et al., 2016; McCarthy et al., 2017; Pérez del Pulgar given the double coverage of public and private healthcare. However, et al., 2020; Tappe et al., 2013) have a strong impact on the actual use this measurement error would be underestimating the associations we (and likely restorative effect) of these play spaces and could be influ­ observe in our study. encing the negatively associated health benefits we observed for high The study also has several strengths. First, to our knowledge, this is 7 C. Pérez-del-Pulgar et al. Environmental Research 200 (2021) 111326 the first study to report the association between different types of resi­ interests or personal relationships that could have appeared to influence dential outdoor play spaces and diagnosed mental and behavioral dis­ the work reported in this paper. orders in children. Second, this study has a large sample size that enables the exploration of associations and effect modifications without Acknowledgments affecting statistical power. Third, we uniquely include several measures of outdoor play spaces. Fourth, we report objectively assessed health This work was supported by the GREENLULU’s Project under the outcomes and outdoor play space proximity measures. Last, we con­ European Union’s Horizon 2020 research and innovation framework ducted a range of sensitivity analyses of importance for assessing health [grant number 678034]; the NATURVATION Project under the Euro­ equity, including testing the effect of the nationality and area SES index pean Union’s Horizon 2020 research and innovation framework [grant indicators, and our findings were robust across all the analyses. number 730243]; the ECHOCAT project of La Marató de TV3 Founda­ tion [grant number 201621–30] and the the María de Maeztu Program 5. Conclusions for Units of Excellence of the Spanish Ministry of Science and Innovation [grant number CEX 2019-000940-M]. HC and MTM are funded by Juan Our findings are suggestive of a possible beneficial effect of overall de la Cierva fellowships awarded by the Spanish Ministry of Economy outdoor play space proximity, and its greenness and diversity, on and Competitiveness [grant number FJCI-2018-035322 and FJCI-2017- childhood mental and behavioral health disorders, especially for disor­ 33842]. The sponsors had no role in the design or analysis of this study. ders of psychological development. These benefits, nevertheless, were only found for children living in low SES areas. Appendix A. Supplementary data We recommend future studies to more closely investigate the importance of outdoor play space types and diversity for children’s Supplementary data to this article can be found online at https://doi. disorders of psychological development as well as the pathways through org/10.1016/j.envres.2021.111326. which proximity to outdoor play spaces is associated with higher prev­ alence of children’s disorders of psychological development in high SES References areas. Examining the qualities of the built environment as well as its actual uses a function of gender, physical ability, age or ethnicity and its Alderton, A., et al., 2019. Reducing inequities in early childhood mental health: how might the neighborhood built environment help close the gap? a systematic search impact on children’s mental and behavioral health is a promising and and critical review. Int. J. Environ. Res. Publ. Health 16 (9), 1–23. https://doi.org/ compelling area of study that requires further research. 10.3390/ijerph16091516. Our finding should be of interest for policy makers in planning for American Psychiatric Association, 2013. APA - Diagnostic and Statistical Manual of Mental Disorders DSM-5 Fifth Edition. Available at: https://www.appi.org/Diagnost healthier cities for children through equitable place-based interventions ic_and_Statistical_Manual_of_Mental_Disorders_DSM-5_Fifth_Edition. that aim to remedy urban environments currently criticized for over­ Amoly, E., et al., 2014. ‘Green and blue spaces and behavioral development in Barcelona looking children’s needs for participation and play (Derr and Tarantini, Schoolchildren : the BREATHE project’. Environ. Health Perspect. 122 (12), 1351–1358. https://doi.org/10.1289/ehp.1408215. 2016; Malone, 2013; Moore-Cherry, 2014; Woolcock et al., 2010). Anguelovski, I., 2013. From environmental trauma to safe haven: place attachment and Environmental planners should place particular attention to creating place remaking in three marginalized neighborhoods of Barcelona. Boston, and new play and green space opportunities for children of working-class Havana’, City and Community 12 (3), 211–237. https://doi.org/10.1111/cico.12026. Anguelovski, I., 2014. Neighborhood as Refuge: Community Reconstruction, Place neighborhoods, since those seem to both particularly benefit from Remaking, and Environmental Justice in the City. MIT Press. such interventions while, traditionally, lacking equitable access to those, Anguelovski, I., et al., 2018. Assessing green gentrification in historically disenfranchised and this beyond the case of Barcelona. Placing children at the center of neighborhoods: a longitudinal and spatial analysis of Barcelona. Urban Geogr. 39 public space interventions (including play spaces such as parks or (3), 458–491. https://doi.org/10.1080/02723638.2017.1349987. Anguelovski, I., et al., 2020. Expanding the boundaries of justice in urban greening plazas) is essential for building accessible, green, and healthy cities for scholarship: toward an emancipatory, antisubordination, intersectional, and all. relational approach. Ann. Assoc. Am. Geogr. 110 (6), 1743–1769. https://doi.org/ 10.1080/24694452.2020.1740579. Arroyo-Johnson, C., et al., 2016. Still separate, still unequal: social determinants of Author contribution playground safety and proximity disparities in st. Louis. J. Urban Health 93 (4), 627–638. https://doi.org/10.1007/s11524-016-0063-8. Carmen Pérez-del-Pulgar: Conceptualization, Methodology, Formal Balseviciene, B., et al., 2014. ‘Impact of residential greenness on preschool children’s emotional and behavioral problems’. Int. J. Environ. Res. Publ. Health 11 (7), analysis, Investigation, Writing – original draft, Project administration, 6757–6770. https://doi.org/10.3390/ijerph110706757. Isabelle Anguelovski: Conceptualization, Resources, Writing – review & Barros, A.J.D., Hirakata, V.N., 2003. Alternatives for logistic regression in cross-sectional editing, Funding acquisition, Helen V. S. Cole: Conceptualization, studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med. Res. Methodol. 3, 1–13. https://doi.org/10.1186/1471-2288-3-21. Methodology, Writing – review & editing, Jeroen de Bont: Methodology, Biernacka, M., Kronenberg, J., 2018. Classification of institutional barriers affecting the Resources, Data curation, Writing – review & editing, James Connolly: availability, accessibility and attractiveness of urban green spaces. Urban For. Urban Conceptualization, Methodology, Writing – review & editing, Francesc Green. 36, 22–33. https://doi.org/10.1016/j.ufug.2018.09.007. Bijnens, E.M., et al., 2020. ‘Residential green space and child intelligence and behavior Baró; : Methodology, Writing – review & editing, Yesika Díaz: Resources, across urban, suburban, and rural areas in Belgium: a longitudinal birth cohort study Data curation, Writing – review & editing, Mario Fontán-Vela: Writing – of twins’, PLOS Medicine. Edited by I. Markevych 17 (8), e1003213. https://doi.org/ review & editing, Talita Duarte-Salles: Methodology, Resources, Writing 10.1371/journal.pmed.1003213. Bolíbar, B., et al., 2012. Base de datos SIDIAP: La historia clínica informatizada de – review & editing, Funding acquisition, Supervision, Margarita Atención Primaria como fuente de información para la investigación epidemiológica. Triguero-Mas: Conceptualization, Methodology, Formal analysis, Med. Clínica 138 (14), 617–621. https://doi.org/10.1016/j.medcli.2012.01.020. Writing – review & editing, Supervision Britannica, T. E. of E, 2020. ‘Child development’, encyclopedia britannica. Available at: https://www.britannica.com/science/child-development-process. Busquets, J., 2006. Barcelona: the Urban Evolution of a Compact City. Nicolodi & Approval and funding sources Harvard University Graduate School of Design. Byrne, J., Wolch, J., 2009. Nature, race, and parks: past research and future directions for Approval: This study was approved by the ethics committee of the geographic research. Prog. Hum. Geogr. 33 (6), 743–765. https://doi.org/10.1177/ 0309132509103156. Jordi Gol i Gurina Institute for Research in Primary Care (IDIAPJGol, Cederborg, A.C., 2020. ‘Young children’s play: a matter of advanced strategies among 20/163). peers’. Early Child. Dev. Care 190 (5), 778–790. https://doi.org/10.1080/ 03004430.2018.1491561. Christian, H., et al., 2015. The influence of the neighborhood physical environment on Declaration of competing interest early child health and development: a review and call for research. Health Place 33, 25–36. https://doi.org/10.1016/j.healthplace.2015.01.005. The authors declare that they have no known competing financial 8 C. Pérez-del-Pulgar et al. Environmental Research 200 (2021) 111326 Compton-Lilly, C., et al., 2017. Intersectional identity negotiation. J. Literacy Res. 49 (1), Mnich, C., et al., 2019. ‘Psychosocial and physiological health outcomes of green exercise 115–140. https://doi.org/10.1177/1086296X16683421. in children and adolescents—a systematic review’. Int. J. Environ. Res. Publ. Health Derr, V., Chawla, L., van Vliet, W., 2017. Children as natural change agents. In: 16 (21). https://doi.org/10.3390/ijerph16214266. Bishop, K., Corkery, L. (Eds.), Designing Cities with Children and Young People- Módenes, J.-A., López-Colás, J., 2007. Second homes and compact cities in Spain: two Beyond Playgrounds and Skate Parks. Routledge, New York, p. 12. elements of the same system? Tijdschr. Econ. Soc. Geogr. 98 (3), 325–335. https:// Derr, V., Tarantini, E., 2016. ‘“Because we are all people”: outcomes and reflections from doi.org/10.1111/j.1467-9663.2007.00400.x. young people’s participation in the planning and design of child-friendly public Moore-Cherry, N., 2014. Creating child-friendly cities: reinstating kids in the city. Plann. spaces’. Local Environ. 21 (12), 1534–1556. https://doi.org/10.1080/ Theor. Pract. 15 (1), 144–146. https://doi.org/10.1080/14649357.2013.873230. 13549839.2016.1145643. Perez-del-Pulgar, C., Anguelovski, I., Connolly, J., 2020. ‘Toward a green and playful Domínguez-Berjón, M.F., et al., 2008. ‘Construcción de un índice de privación a partir de city : understanding the social and political production of children ’ s relational datos censales en grandes ciudades españolas: (Proyecto MEDEA)’. Available at: Gac. wellbeing in Barcelona’. Cities 96 (August), 102438. https://doi.org/10.1016/j. Sanit. 22 (3) http://scielo.isciii.es/scielo.php?script=sci_arttext&pid=S0213-9111 cities.2019.102438. 2008000300002. Putra, I.G.N.E., et al., 2021. Association between caregiver perceived green space quality Dunton, G.F., et al., 2014. Neighborhood park use by children: use of accelerometry and and the development of prosocial behaviour from childhood to adolescence: latent global positioning systems. Am. J. Prev. Med. 46 (2), 136–142. https://doi.org/ class trajectory and multilevel longitudinal analyses of Australian children over 10 10.1016/j.amepre.2013.10.009. years. J. Environ. Psychol. 74, 101579. https://doi.org/10.1016/j. Dyment, J., O’Connell, T.S., 2013. The impact of playground design on play choices and jenvp.2021.101579. behaviors of pre-school children. Child Geogr. 11 (3), 263–280. https://doi.org/ Rigolon, A., Flohr, T.L., 2014. Access to parks for youth as an environmental justice issue: 10.1080/14733285.2013.812272. access inequalities and possible solutions. Buildings 4 (2), 69–94. https://doi.org/ Engemann, K., et al., 2019. Residential green space in childhood is associated with lower 10.3390/buildings4020069. risk of psychiatric disorders from adolescence into adulthood. Proc. Natl. Acad. Sci. Schoeppe, S., et al., 2016. ‘Too far from home? Adult attitudes on children’s independent U.S.A. 116 (11), 5188–5193. https://doi.org/10.1073/pnas.1807504116. mobility range’. Child Geogr. 14 (4), 482–489. https://doi.org/10.1080/ Flies, E.J., et al., 2019. ‘Urban-associated Diseases: Candidate Diseases, Environmental 14733285.2015.1116685. Risk Factors, and a Path Forward’, Environment International. Elsevier Ltd, p. 105187. Sobko, T., Jia, Z., Brown, G., 2018. Measuring connectedness to nature in preschool https://doi.org/10.1016/j.envint.2019.105187. children in an urban setting and its relation to psychological functioning. PloS One Flouri, E., Midouhas, E., Joshi, H., 2014. ‘The role of urban neighbourhood green space 13 (11), 1–17. https://doi.org/10.1371/journal.pone.0207057. in children’s emotional and behavioural resilience’. J. Environ. Psychol. 40, Spellerberg, I.F., Fedor, P.J., 2003. ‘A tribute to Claude-Shannon (1916-2001) and a plea 179–186. https://doi.org/10.1016/j.jenvp.2014.06.007. for more rigorous use of species richness, species diversity and the “Shannon- Floyd, M.F., 2001. ‘Managing National Parks in a Multicultural Society: Searching for Wiener” Index’. Global Ecol. Biogeogr. 12 (3), 177–179. https://doi.org/10.1046/ Common Ground’, the George Wright Forum. George Wright Society, pp. 41–51. j.1466-822X.2003.00015.x. https://doi.org/10.2307/43597755. Stanley, E., 2011. The place of outdoor play in a school community: a case study of recess Frost, J., 1992. Play and Playscapes. values. Child. Youth Environ. 21 (1), 185–211. https://doi.org/10.7721/ Gibson, J.J., 1979. The Ecological Approach to Visual Perception. Psychology Press & chilyoutenvi.21.1.0185. Routledge Classic Editions. Strife, S., Downey, L., 2009. Childhood development and access to nature’. Organ. Hart, R., 1979. ‘Children’s experience of place. Geogr. Rev. 229. https://doi.org/ Environ. 22 (1), 99–122. https://doi.org/10.1177/1086026609333340. 10.2307/214444. Tappe, K.A., et al., 2013. ‘Children’s physical activity and parents’ perception of the Holt, N.L., et al., 2015. ‘“Eyes on where children play”: a retrospective study of active neighborhood environment: neighborhood impact on kids study’. Int. J. Behav. Nutr. free play’. Child Geogr. 13 (1), 73–88. https://doi.org/10.1080/ Phys. Activ. 10, 1. https://doi.org/10.1186/1479-5868-10-39. 14733285.2013.828449. Taylor, A.F., Kuo, F.E.M., 2011. ‘Could exposure to everyday green spaces help treat Hughey, S.M., et al., 2016. Using an environmental justice approach to examine the adhd? Evidence from children’s play settings’. Appl. Psychol.: Health and Well-Being relationships between park availability and quality indicators, neighborhood 3 (3), 281–303. https://doi.org/10.1111/j.1758-0854.2011.01052.x. disadvantage, and racial/ethnic composition. Landsc. Urban Plann. 148, 159–169. Taylor, A.F., Kuo, F.E., Sullivan, W.C., 2002. Views of nature and self-discipline: https://doi.org/10.1016/j.landurbplan.2015.12.016. evidence from inner city children. J. Environ. Psychol. 22 (1–2), 49–63. https://doi. Huynh, Q., et al., 2013. Exposure to public natural space as a protective factor for org/10.1006/jevp.2001.0241. emotional well-being among young people in Canada. BMC Publ. Health 13 (1). Thygesen, M., et al., 2020. The association between residential green space in childhood https://doi.org/10.1186/1471-2458-13-407. and development of attention deficit hyperactivity disorder: a population-based Kamel, A.A., Ford, P.B., Kaczynski, A.T., 2014. Disparities in park availability, features, cohort study. Environ. Health Perspect. 128 (12), 127011. https://doi.org/10.1289/ and characteristics by social determinants of health within a U.S.-Mexico border EHP6729. urban area. Prev. Med. 69 (S), S111–S113. https://doi.org/10.1016/j. Tierney, P.T., Dahl, R., Chavez, D., 2001. Cultural diversity in use of undeveloped natural ypmed.2014.10.001. areas by Los Angeles county residents. Tourism Manag. 22 (3), 271–277. https://doi. Lilius, J., 2019. Reclaiming cities as spaces of middle class parenthood. Reclaiming Cities org/10.1016/S0261-5177(00)00058-3. as Spaces of Middle Class Parenthood. https://doi.org/10.1007/978-981-10-9010-3. Tillmann, S., et al., 2018. Mental health benefits of interactions with nature in children Loukaitou-Sideris, A., Sideris, A., 2010. ‘What brings children to the park? Analysis and and teenagers: a systematic review. J. Epidemiol. Community 1–9. https://doi.org/ measurement of the variables affecting children’s use of parks’. J. Am. Plann. Assoc. 10.1136/jech-2018-210436. 76 (1), 89–107. https://doi.org/10.1080/01944360903418338. Timperio, A., et al., 2008. Features of public open spaces and physical activity among Luken, E., Carr, V., Brown, R.D., 2011. ‘Playscapes: Designs for Play, Exploration and children: findings from the CLAN study. Prev. Med. 47 (5), 514–518. https://doi. Science Inquiry’, Children, Youth And Environments. University of Cincinnati, org/10.1016/j.ypmed.2008.07.015. pp. 325–337. https://doi.org/10.7721/chilyoutenvi.21.2.0325. Ulrich, R.S., et al., 1991. Stress recovery during exposure to natural and urban Lynch, K., 1977. Growing up in Cities. MIT Press, Cambridge. environments. J. Environ. Psychol. 11 (3), 201–230. https://doi.org/10.1016/ Maas, J., et al., 2009. Morbidity is related to a green living environment. J. Epidemiol. S0272-4944(05)80184-7. Community 63 (12), 967–973. https://doi.org/10.1136/jech.2008.079038. Vaden-Kiernan, M., et al., 2010. Neighborhoods as a developmental context: a multilevel Madzia, J., et al., 2019. Residential greenspace association with childhood behavioral analysis of neighborhood effects on head start families and children. Am. J. outcomes. J. Pediatr. 207, 233–240. https://doi.org/10.1016/j.jpeds.2018.10.061. Community Psychol. 45 (1), 49–67. https://doi.org/10.1007/s10464-009-9279-z. Malone, K., 2013. ‘“The future lies in our hands”: children as researchers and Vanaken, G.J., Danckaerts, M., 2018. ‘Impact of green space exposure on children’s and environmental change agents in designing a child-friendly neighbourhood’. Local adolescents’ mental health: a systematic review’. Int. J. Environ. Res. Publ. Health Environ. 18 (3), 372–395. https://doi.org/10.1080/13549839.2012.719020. 15 (12). https://doi.org/10.3390/ijerph15122668. Markevych, I., et al., 2014. Access to Urban Green Spaces and Behavioural Problems in Vaughan, K.B., et al., 2013. Exploring the distribution of park availability, features, and Children: Results from the GINIplus and LISAplus Studies, vol. 71. Environment quality across Kansas City, Missouri by income and race/ethnicity: an environmental International, pp. 29–35. https://doi.org/10.1016/j.envint.2014.06.002. justice investigation. Ann. Behav. Med. 45 (Suppl. 1), 28–38. https://doi.org/ Markevych, I., et al., 2018. Outdoor air pollution, greenspace, and incidence of ADHD: a 10.1007/s12160-012-9425-y. semi-individual study. Sci. Total Environ. 642, 1362–1368. https://doi.org/ Villanueva, K., et al., 2016. ‘Can the neighborhood built environment make a difference 10.1016/j.scitotenv.2018.06.167. in children’s development? Building the research agenda to create evidence for Mårtensson, F., et al., 2009. Outdoor environmental assessment of attention promoting place-based children’s policy’. Academic Pediatrics 16 (1), 10–19. https://doi.org/ settings for preschool children. Health Place 15 (4), 1149–1157. https://doi.org/ 10.1016/j.acap.2015.09.006. 10.1016/j.healthplace.2009.07.002. Wells, N.M., 2000. ‘At home with nature: effects of “greenness” on children’s cognitive Mayer, F.S., et al., 2009. Why is nature beneficial? The role of connectedness to nature’, functioning’. Environ. Behav. 32 (6), 775–795. https://doi.org/10.1177/ Environment and Behavior 41 (5), 607–643. https://doi.org/10.1177/ 00139160021972793. 0013916508319745. WHO, 2001. WHO | international Classification of functioning, Disability and health (ICF), McCarthy, S.M., Hughey, S.M., Kaczynski, A.T., 2017. Examining sociodemographic WHO. World health organization. Available at: http://www.who.int/classifications differences in playground availability and quality and associations with childhood /icf/en/. obesity. Child. Obes. 13 (4), 324–331. https://doi.org/10.1089/chi.2016.0239. Woolcock, G., Gleeson, B., Randolph, B., 2010. Urban research and child-friendly cities: a McCormick, R., 2017. Does access to green space impact the mental well-being of new australian outline. Child Geogr. 8 (2), 177–192. https://doi.org/10.1080/ children: a systematic review. J. Pediatr. Nurs. 37, 3–7. https://doi.org/10.1016/j. 14733281003691426. pedn.2017.08.027. Wu, C. Da, et al., 2014. ‘Linking student performance in Massachusetts elementary schools with the “greenness” of school surroundings using remote sensing’. PloS One 9 (10), 1–9. https://doi.org/10.1371/journal.pone.0108548. 9