Urban public parks and mental health in adult women: Mediating and moderating factors

Abstract

Background:

The association of green spaces such as urban public parks and mental health might vary according to personal characteristics and characteristics of the park and be mediated by the use of the park.

Aims:

We investigate the association between urban public park coverage and mental health in adult women, the moderation of this association by personal and park-related characteristics, and the mediation of the association by use of public space.

Methods:

Combining data from a cross-sectional survey of the adult female population of Tijuana (Mexico) in 2014, and a study of public spaces in 2013, we analyzed the association between park coverage in buffers of 400 and 800 m from participants’ homes and score in the Center for Epidemiologic Studies-Depression scale (CES-D). We tested for mediation by use of park and interaction of urban park coverage with personal and park characteristics.

Results:

Urban public park coverage in the 400-m buffer had an inverse association with CES-D score that was moderated by age (significant only for younger participants), with no evidence of mediation. Park coverage in the 800-m buffer also had an inverse association with CES-D score, moderated by age and occupation (significant for younger participants and homemakers), and a mediated association was also observed. There was no interaction between park coverage and park characteristics in their association with CES-D score.

Conclusion:

Our results confirm the potential of public parks to improve mental health and suggest that this effect could be more important at some stages in the life course for women. The upper-middle-income, Latin American country setting adds to the current knowledge that is mostly based on high-income countries.

Keywords

Urban environment public parks mental health women Mexico

Introduction

Public urban green spaces such as parks are important features of the urban environment. Urban parks can improve the quality of life by providing a place for relaxation, social interaction, sports and other health-promoting activities (Douglas, Lennon, & Scott, 2017; Kaczynski & Henderson, 2007; Koohsari et al., 2015; Lee & Maheswaran, 2011; Sugiyama, Leslie, Giles-Corti, & Owen, 2008). The mental health of urban residents could also benefit from contact with green spaces, and three types of mechanisms have been proposed to explain this association (de Vries, Verheij, Groenewegen, & Spreeuwenberg, 2003; Gascon et al., 2015). First, exposure to natural settings is known to improve mood and reduce stress (de Vries, van Dillen, Groenewegen, & Spreeuwenberg, 2013) so that simply having a green space in the vicinity might have a positive effect, regardless of its use. Second, public space facilitates social contact, improving social cohesion in the community and thus mental health and well-being (de la Barrera, Reyes-Paecke, Harris, Bascuñán, & Farías, 2016). Third, green space might increase physical activity by promoting leisure walking, walking through the space when running errands, active playing and sports (Koohsari et al., 2015; Sallis et al., 2016), and physical activity could in turn have a positive effect on mental health. As mental health problems, and especially common mental disorders such as anxiety and depression, are among the main causes of the burden of disease worldwide (Whiteford, Ferrari, Degenhardt, Feigin, & Vos, 2015), and given that the majority of the world’s population is now living in cities (United Nations, 2014), exploring which elements of the urban environment can improve mental health is an important public health issue.

Although the idea of a protective effect of urban green space on mental health is intuitively appealing, the evidence regarding this association is still inconclusive. A review described the evidence for the association of mental health and urban green space as weak and called for more research in this area (Lee & Maheswaran, 2011). According to another review, only three out of six studies of this association in adults found the expected protective relationship, while two reported null findings and one found associations only in certain age and ethnicity groups (Gascon et al., 2015). A third, more recent review reported that most studies of green areas (including parks and other types of green space) had found significant inverse associations with depressive mood, but some had found no association and still others a direct association in some groups (Rautio, Filatova, Lehtiniemi, & Miettunen, 2018).

Divergent results regarding the association of urban green space and mental health can in part be explained by the use of different measurements, statistical analysis, and sets of covariables. Still, another explanation could be that the relationship varies depending on personal characteristics. Age has been reported to modify the association between green space and mental health, so that in one study the protective association was stronger for women 18–24 and over 65 years old, while for women 25–64 years old the association was not significant (Bos, van der Meulen, Wichers, & Jeronimus, 2016). In another research, there was an inverted U–shaped association between green space and mental health in older women (Astell-Burt, Mitchell, & Hartig, 2014). Other variables, such as occupation, marital status or number and ages of children are all related to age, and might also modify the association between parks and mental health. As the life course stages indicated by these variables have particular characteristics for women, a focused exploration of how they modify the effect of green space is guaranteed.

On the other hand, the effect of green space might depend on whether people use it or not (Bos et al., 2016). A cross-sectional study found that those who spent more time in green spaces had better mental health on average, after adjusting by sociodemographic characteristics (van den Berg et al., 2016), and another reported that the proximity of parks was associated with lower odds of depressive symptoms in women, but the effect was significant only among those who visited the park for at least 4 hours each week (Reklaitiene et al., 2014). The effect of green space also seems to be larger for physically active persons as compared to the inactive (Annerstedt et al., 2012; Astell-Burt, Feng, & Kolt, 2013), which also suggests a protective effect on mental health that is mediated by the use of the green space. In this sense, the differential effect of green space by age and other characteristics could be the consequence of time spent in the neighborhood: if younger and older versus middle-age persons spend more time in the local urban environment, they could benefit more from the health-related aspects of public green spaces (de Vries et al., 2003). The mediation of the association of green space and mental health by degree of contact with green space is another subject worth studying.

Finally, the quality of urban parks is another potential modifier of the association between parks and mental health. Aspects such as good maintenance, pleasant visual features, facilities and vegetation increase the appeal and use of parks, which might in turn reflect on their beneficial effects on mental health (Lee & Maheswaran, 2011). However, only a few studies have considered these variables, and the evidence is inconclusive (Gascon et al., 2015).

In this article, we assess the association between urban public park coverage and mental health, exploring moderation by personal and park characteristics and mediation of the association by the use of park. Our hypotheses for the analysis were that: (1) higher park coverage in the vicinity of participants’ homes would be associated with better mental health; (2) the association would be modified by personal characteristics as well as by the quality of parks; and (3) the association would be mediated by the use of the park.

Participants were adult women living in Tijuana, Mexico, in 2014. Women are disproportionately affected by common mental health issues such as clinical and subclinical depression and anxiety (World Health Organization, 2000), and they can also be more sensitive to the neighborhood environment (Pattyn, Van Praag, Verhaeghe, Levecque, & Bracke, 2011; van Praag, Bracke, Christiaens, Levecque, & Pattyn, 2009). Women are therefore a group for which the effects of urban parks could be especially important. As most research on the subject has been conducted in higher-income countries, by studying the associations of urban parks and mental health among women in a Latin American city, we expect to contribute to a more complete understanding of these associations.

Methods

Site

Tijuana is a city of over 1.5 million inhabitants (Instituto Nacional de Estadística y Geografía (INEGI), 2015) located at the Mexico–United States border. During the past decades, the city experienced uncontrolled growth, resulting in a contrasting pattern of low-income neighborhoods in high-risk areas, housing developments with just the basic services in the outskirts and some better-off areas (Sanchez-Rodriguez, 2011). Although local legislation mandates that urban developments include green spaces, the size and quality of urban parks is limited, and many of them are perceived by neighbors as unsatisfactory (Ojeda-Revah, Bojorquez, & Osuna, 2017). In Tijuana, only 37% of the population have an urban park within 400 m, and most of them have a low percentage of vegetation cover (Huizar & Ojeda-Revah, 2014). The characteristics and distribution of parks in Tijuana allows for assessment of the association between parks and mental health in a context that contrasts with those addressed by the majority of investigations in high-income countries.

Data sources

We combined data from a household survey and from a study of the distribution and quality of public spaces in Tijuana.

The household survey followed a probabilistic sampling design to be representative of adult women living in the city. The primary sampling units were geographical statistical areas (similar to census tracts) as defined by Mexico’s National Institute of Geography and Statistic (INEGI), stratified by the level of marginalization. INEGI assigns to each area a score in an index of marginalization that considers area-level education, health and housing characteristics and classifies the areas into five groups: very high, high, medium, low and very low marginalization. For the survey, the very high and high groups were collapsed into high and the very low and low groups into low to obtain three strata from which geographical statistical areas were sampled with probability proportional to size. In subsequent steps, blocks were selected from each area, all households in the block were visited, and one woman 18–65 years of age was selected from each household to respond an interviewer-applied questionnaire. When more than one woman of the desired age lived in the household, only one of them was randomly selected. Field work was conducted in 2014. The response rate was 94%, and the final unweighted n = 2345. Participants read and signed an informed consent form, and all procedures were revised and approved by the Ethics Committee of El Colegio de la Frontera Norte.

The study of urban parks was conducted in 2013, with the aim of assessing the distribution and quality of all public spaces in Tijuana (Huizar & Ojeda-Revah, 2014). A list of public spaces as defined by Mexico’s Development Ministry (Secretaría de Desarrollo Social (1999)) was obtained from the municipality’s registers, and all parks in the list were visited to conduct ad hoc observation of their characteristics.

Measurements

Urban park coverage

We employed geographic information systems (GIS) to define buffers of 400 and 800 m around the block in which each participant in the survey lived, and calculated park coverage in square meters within each buffer. We chose those buffers as a reasonable walking distance from the participants’ homes (Boone, Buckley, Grove, & Sister, 2009; Reyes & Figueroa, 2010). To make the results easier to interpret, we computed a variable in units of 500 m, so that regression coefficients can be interpreted as the mean effect of having 500 m more of park coverage.

Characteristics of park

We employed the scores for park quality obtained in the study of public spaces (Huizar & Ojeda-Revah, 2014). As part of that study, a researcher completed an inventory with nine items (e.g. bathrooms, lighting and playground), and the score was the sum of items present in the park. Thus, a higher score indicates higher park quality in terms of more available features and services, with a possible range of 0 to 9. For the analysis, we assigned to each participant the highest score of those of the different parks in the buffer around the home. The percentage of vegetation cover in each park was also assessed in the previous study, in three categories: 0%, 1%–49% and ⩾50%. From this variable, we computed another with four categories: no park coverage in the buffer, park with 0% vegetation cover, park with 1%–49% vegetation cover and park with ⩾50% vegetation cover. As with the quality variable, when more than one park was present in the buffer, we assigned the highest score of those of the different parks.

Personal characteristics and use of park

We included in the analysis age, marital status (single/married or cohabiting/separated or widowed), occupation (homemaker, retired or not working/working or looking for a job/student), children (no children/at least one child ⩽5 years/all children >5 years), socioeconomic level (score in an index computed from questions about household goods and services) and education attainment (years). As an indicator of park use, we employed the response to the question ‘Do you use a park or public space to practice physical activity (exercise)?’ Response options for this question were ‘never’, ‘occasionally’ or ‘frequently’. For the analysis, we combined the last two into a yes/no indicator of ‘active in a public space’. All of these variables were obtained from the household survey questionnaire.

Mental health

Our dependent variable was mental health problems, as indicated by score in the Center for Epidemiologic Studies-Depression scale (CES-D) (Radloff, 1977). The CES-D measures depressive symptoms and has been amply employed in studies with Mexican population. The household survey’s questionnaire included a Spanish 10-item version of the CES-D (Bojorquez & Salgado, 2009; Tiburcio-Sainz & Natera-Rey, 2007) with possible scores ranging from 0 to 30. A higher score indicates higher frequency and number of symptoms. The CES-D has been previously employed and validated in Mexican population (Bojorquez & Salgado, 2009; Salinas-Rodriguez et al., 2014) and had an internal consistency (Cronbach’s alpha) of 0.80.

Analysis

The analysis was conducted from September 2017 to February 2018. We began by conducting exploratory analysis of each variable and their joint distributions in the sample by means of graphics and simple bivariate tests, to get a sense of the associations and to detect possible outliers. We explored the bivariate associations between the indicators of park coverage and quality, mental health and the associations after controlling for other variables. We explored the possible differential associations by personal characteristics in stratified models, and formally tested for modification of the association between park coverage and CES-D scores by other characteristics with interaction terms in regression models.

Afterwards, we employed a structural equation modeling approach to explore mediation of the main association of interest, testing for association between (1) independent variable an dependent variable (urban park coverage and mental health); (2) independent variable and mediator (urban park coverage and being active in a public space); and (3) mediator and dependent variable (being active in a public space and mental health) (Baron & Kenny, 1986). In the models, we included all covariates of interest and interaction terms that were significant at the p < .10 level. We also tested for moderated mediation, but none of the terms were significant. To accommodate the non-linear distributions of the associations, we employed generalized structural models using negative binomial regression for the paths in which CES-D was the dependent variable and logit regression for the paths in which being active in a public space was the dependent variable. We conducted all analyses with Stata (Stata/SE, version 15.1) and considered the complex sampling design by adjusting standard errors and employing weights by means of Stata’s ‘svy’ commands. The final weights considered the inverse of the selection probability, non-response adjustment, and poststratification based on the distribution of age groups (18–29, 30–65) and occupation (working or studying, homemaker) among women in Tijuana using data from the 2010 census.

Results

As shown in Table 1, the mean coverage of urban public parks in the 400-m buffer around the participants’ homes was 1150 m², and the mean density in the 800-m buffer was 5238 m², with values ranging from 0 to 18,228 for the 400-m buffer and from 0 to 102,005 for the 800-m buffer. The distribution of park density was highly skewed, with 80.9% (95% CI: 70.3, 88.4) of participants having 0 m² in the 400-m buffer and 47.5% (95% CI: 34.6, 60.7) with 0 m² in the 800-m buffer. Out of nine possible points in the score of park quality, the mean score for the 400-m buffer was 1.0 and the mean score for the 800-m buffer 2.7. Most parks had less than 50% of vegetation coverage.

Table 1.

Characteristics of the sample.^a

Variable	Mean, % or range	95% CI
Characteristics of urban public parks
Park coverage: m² in 400-m buffer (mean)	1150	566, 1734
Park coverage: m² in 400-m buffer (range)	0–18,228
Park coverage: m² in 800-m buffer (mean)	5238	2978, 7498
Park coverage: m² in 800-m buffer (range)	0–102,005
Park quality score in 400-m buffer (mean)	1.0	0.6, 1.4
Park quality score in 800-m buffer (mean)	2.7	1.9, 3.5
Vegetation coverage in 400-m buffer (%)
No park in buffer	80.9	70.3, 88.4
Park with 0% vegetation coverage	0.2	0.0, 1.8
Park with 1%–49% vegetation coverage	13.9	7.6, 24.2
Park with ⩾50% vegetation coverage	4.9	2.0, 11.6
Vegetation coverage in 800-m buffer (%)
No park in buffer	47.5	34.6, 60.7
Park with 0% vegetation coverage	2.6	0.6, 10.3
Park with 1%–49% vegetation coverage	35.0	22.7, 49.6
Park with ⩾50% vegetation coverage	14.9	8.3, 25.3
Characteristics of participants
Age (years) (mean)	37.0	36.5, 37.4
Marital status (%)
Single	22.6	21.1, 24.2
Married or cohabiting	66.7	64.7, 68.6
Separated or widowed	10.7	9.5, 12.1
Occupation (%)
Homemaker, retired or not working	50.0	49.3, 50.8
Working or looking for a job	43.0	41.5, 44.4
Student	7.0	6.1, 8.1
Children (%)
No children	21.3	19.0, 23.9
At least one child ⩽5 years old	31.2	28.9, 33.5
All children >5 years old	47.5	45.8, 49.2
Socioeconomic level	0.1	−0.1, 0.3
Education attainment (years)	9.5	9.1, 9.8
Active in a public space (%)	21.4	18.6, 24.5
Score in CES-D (mean)	5.5	5.1, 5.9

CES-D: Center for Epidemiologic Studies-Depression scale (Radloff, 1977).

Unweighted n = 2345 (varies for some variables because of missing values). All figures in table are weighted and consider sample design.

Participant’s mean age was 37 years and most were married/cohabiting. Half were homemakers or not working for other reasons, and slightly under half were working or looking for a job. About one-fifth (21.4%, 95% CI: 18.5, 24.4) of participants reported using a public space for physical activity. The mean CES-D score was 5.5 (Table 1).

In Table 2, we show the associations between mental health and each park-related characteristic, including park coverage. More park coverage in the 400-m buffer was associated with lower CES-D score, even after adjusting by potential confounders (β = –.01, 95% CI: –.01, –.00). Having a park with 0% vegetation coverage in the 400-m buffer, as opposed to having no park in that buffer, was also associated with lower CES-D score. There was no gradient in the estimated associations between the category of vegetation coverage and CES-D score. All park-related variables were strongly correlated, with Spearman’s correlation coefficients all ⩾.90 (not shown in Table). In the regression models, there were no significant interactions between park coverage and park characteristics in their association with CES-D score. We present the results of analysis with only park coverage as the main independent variable, but similar results were found with other park-related variables.

Table 2.

Associations between park-related variables and mental health.^a

Independent variables	Regression coefficient (95% CI), unadjusted	Regression coefficient (95% CI), adjusted
Park coverage in 400-m buffer (500 m²)	−.01 (–.01, –.00)	−.01 (–.01, –.00)
Park coverage in 800-m buffer (500 m²)	−.00 (–.01, –.00)	−.00 (–.01, .00)
Park quality score in 400-m buffer	−.01 (–.04, .01)	−.02 (–.04, .01)
Park quality score in 800-m buffer	−.00 (–.03, .02)	.00 (–.02, .02)
Vegetation coverage in 400-m buffer
No park in buffer	Ref.	Ref.
Park with 0% vegetation coverage	−.15 (–.23, –.08)	−.27 (–.35, –.19)
Park with 1%-49% vegetation coverage	−.17 (–.35, .02)	−.13 (–.26, –.00)
Park with ⩾50% vegetation coverage	.02 (–.11, .16)	−.03 (–.15, .09)
Vegetation coverage in 800-m buffer
No park in buffer	Ref.	Ref.
Park with 0% vegetation coverage	.06 (–.03, .16)	.08 (–.09, .25)
Park with 1%-49% vegetation coverage	−.00 (–.17, .17)	−.02 (–.16, .12)
Park with ⩾ 50% vegetation coverage	−.00 (–.13, .12)	.04 (–.12, .20)

Unweighted n = 2298 (varies for some models because of missing values). Adjusted regression includes socioeconomic level, education attainment, age (years), marital status, occupation and children.

Models in which terms for the interaction between park coverage and each of the personal characteristics were introduced separately showed that the association was significant for younger, single participants and participants without children (Figure 1). When various interaction terms were simultaneously added to the model, some of the interactions in Figure 1 lost significance. The models reported in Table 3 include the significant interaction between park coverage and age for the 400-m buffer and the significant three-way interaction between park coverage, age and occupation for the 800-m buffer. There was a negative-sign direct (i. e. not mediated) association between park coverage and CES-D score in both models, but the association was only significant for younger participants for the 400-m buffer and for younger participants and homemakers for the 800-m buffer. As park coverage was not associated with being active in a public space for the 400-m buffer (β = .01, 95% CI: –.01, .02), there was no mediation of the association between park coverage and CES-D score for this buffer. In contrast, park density in the 800-m buffer was associated with being active in a public space (β = .01, 95% CI: .01, .02) and being active in a public space was associated with lower CES-D score (mediated association).

Figure 1.

Predicted effects of urban public park coverage, by personal characteristics: (a) 400-m buffer and (b) 800-m buffer.

Table 3.

Associations between urban public park coverage, park use and mental health.^a

Dependent variable: CES-D scoreIndependent variables	400-m bufferRegression coefficient (95% CI)	800-m bufferRegression coefficient (95% CI)
Park coverage (500 m²)	−.03 (–.05,–.01)	−.01 (–.01,–.00)
Active in a public space	−.17 (–.30,–.05)	−.17 (–.29,–.06)
Age (years)	−.00 (–.01, .00)	−.00 (–.01, .00)
Marital status
Single	Ref.	Ref.
Married/cohabiting	−.13 (–.25,–.02)	−.15 (–.26,–.03)
Separated/widow	.11 (–.05, .26)	.09 (–.06, .24)
Occupation
Homemaker/retired/not working	Ref.	Ref.
Working	−.04 (–.14, .05)	−.04 (–.34, .26)
Student	−.04 (–.24, .16)	.86 (–.18, 1.89)
Children
No children	Ref.	Ref.
Some children ⩽5 years old	.04 (–.10, .19)	.04 (–.09, .18)
All children >5 years old	.03 (–.10, .16)	.03 (–.10, .16)
Socioeconomic level	−.05 (–.07,–.02)	−.04 (–.07,–.02)
Education attainment (years)	−.02 (–.04,–.01)	−.02 (–.04,–.01)
Coverage*Age	.00 (.00, .00)	.00 (–.00, .00)
Coverage*occupation
Working		.00 (–.00, .02)
Student		−.08 (–.14,–.01)
Age*occupation
Working		−.00 (–.01, .01)
Student		−.05 (–.09, .00)
Coverageageoccupation
Working		.00 (–.00, .00)
Student		.00 (.00, .01)
Park coverage (500 m²)	.01 (–.01, .02)	.01 (.01, .02)
Age (years)	−.01 (–.02, .01)	−.01 (–.02, .01)
Marital status
Single	Ref.	Ref.
Married/cohabiting	−.18 (–.57, .21)	−.20 (–.59, .20)
Separated/widow	−.69 (–1.26,–.13)	−.76 (–1.32,–.20)
Occupation
Homemaker/retired/not working	Ref.	Ref.
Working	.19 (–.11, .48)	.21 (–.10, .51)
Student	−.23 (–.78, .31)	−.28 (–.82, .25)
Children
No children	Ref.	Ref.
Some children ⩽5 years old	−.03 (–.34, .28)	−.01 (–.31, .29)
All children >5 years old	.15 (–.24, .55)	.19 (–.20, .58)
Socioeconomic level	.20 (.12, .27)	.20 (.13, .27)
Education attainment (years)	.08 (.04, .13)	.07 (.03,.12)

CES-D: Center for Epidemiologic Studies-Depression scale (Radloff, 1977).

Negative binomial regression for CES-D score, unweighted n = 2055; logit regression for active in public space, unweighted n = 2087.

As for the associations of other covariates with the mediator and dependent variables, being separated or widowed was associated with lower odds of being active in a public space but not with CES-D scores. Higher socioeconomic level and education attainment were associated with higher odds of being active in a public space and with lower CES-D scores.

Discussion

In our study, park coverage had an inverse association with CES-D score that was significant only for some participants. Other authors have also found a stronger negative-sign association among younger women (Bos et al., 2016) and people not in full-time employment (Ruijsbroek et al., 2017). The other variables that modified the association in our exploratory analyses were marital status, occupation and children. Taken together with age, they point to life course stages, and trying to disentangle their respective effects is difficult. In regression models that simultaneously control for all of them, the variables face the problem of multicollinearity, and while some combinations are common (e.g. young, single and without children vs. middle age, married and with children >5 years old) others include only a few cases, reducing the statistical power to explore their main effects and interactions. The authors’ decisions as to which variables and interactions should be kept in the model probably influence results, and the sample size for each different combination of variables can make the results sample dependent. We choose to emphasize the effect of age because the main effects and interaction terms for age were significant in almost all models but also considering that age indexes other aspects of the life course. However, age and other variables are part of clusters of life course indicators, and so the idea of ‘independent effects’ is still problematic from a theoretical point of view.

We can advance some ideas on why park coverage was associated with mental health only among participants with a certain profile. As other authors suggest, urban parks might have more effect on the mental health of people who spend more time at home or in the neighborhood (Bos et al., 2016; de Vries et al., 2003), as might be the case for younger homemakers. Conversely, working women might be in the area only a few hours each day, and therefore be less exposed to local urban parks. Middle-age women might also have less contact with urban parks, given that in this life period individuals are busier with work and family responsibilities, and their children might already be too old to be taken to the park for playing. As the age range in our sample was 18–65 years. we were not able to explore if the association was significant for older adults. In any case, it would be useful to conduct a deeper exploration of the effect of urban parks on mental health at different life stages, clarifying which elements of the life context of women are more relevant for their experience of urban parks.

In our study, we found that park coverage in the 800-m buffer was associated with being active in a public space, which seems to indicate that at least the larger parks (which also were the better equipped ones, having more features such as walking trails, sports courts, etc.) were used for physical activity. As physical activity is a known correlate of mental health (Penedo & Dahn, 2005), this supports the idea of park use as a mediator of the association between park coverage and mental health. A similar finding was reported by Wood, Hooper, Foster, and Bull (2017), who found that larger parks had the stronger association with mental health in an Australian city. However, the direct effects of park coverage observed for both buffers also support the conclusion that simply having a green space close by could result in better mental health.

When comparing our results with those of other authors, it is important to notice that the distribution of urban park coverage in our study was different from the ones in high-income countries. Studies in those countries usually report much larger areas of green space, and sometimes take only the bigger ones into account in analysis (Bos et al., 2016; Nutsford, Pearson, & Kingham, 2013; Reklaitiene et al., 2014). As an example, in the study by Triguero-Mas et al. (2015), the mean coverage of greenness in a 300-m buffer was 8793 m, and 60% of participants had access to a green space, as compared with slightly over 1000 m and 20%, respectively, in our sample for a 400-m buffer. Another study (Wood et al., 2017) reported a mean coverage of 25.91 ha in a 1600-m distance from participants’ homes, while in ours, the mean coverage was of less than 1 ha for the 800-m buffer. The association between urban parks and mental health in places with low coverage of green areas might be different, as compared to the association in cities with more coverage, as larger parks tend to have more services and be suitable for varied activities (Lee & Maheswaran, 2011; Ojeda-Revah et al., 2017), and larger green areas might have more effect on mental health (Nutsford et al., 2013). A similar conclusion was reached by Bos et al. (2016), who found no association between green spaces and mental health for a 1-km buffer but did find one for a 3-km buffer. Those authors suggested that it was the size of spaces in the ampler buffer that explained the positive association.

Strengths and weaknesses

Having only cross-sectional data limits our conclusions, as the observed associations might be due to reverse causality. Also, when analyzing mediation by use of park, we had the important limitation that our indicator of use did not refer specifically to the local parks and referred only to exercise-related use. It is also important to consider that urban park coverage measures only one aspect of the urban green environment. Other elements, such as streetscape greenery, have been associated with better mental health (Araya et al., 2007; de Vries et al., 2013; Rautio et al., 2018; Triguero-Mas et al., 2015), and it is possible that the general visual aspect of the surrounding urban landscape has an important effect over mental health that we were not able to measure in this study (Araya et al., 2007). Likewise, our indicator of urban park quality was limited, and this could explain why we found no modification of the effect by this variable, while other authors have reported that the quality of public open space is more strongly associated with mental health than its quantity (Francis, Wood, Knuiman, & Giles-Corti, 2012).

On the other hand, a strength of our study was that we controlled for socioeconomic level and education when testing the associations of interest. Before adding these variables to the models, the statistical association between park coverage and mental health was highly significant, but after controlling for them the association decreased. Socioeconomic level and education are related to place of living (and therefore to urban park coverage) as well as to mental health, and so these variables are important confounders that should be controlled for in this type of analysis.

Conclusion

To conclude, our study supports the notion that urban parks are beneficial for mental health at least in some subpopulations, even in an urban environment where public spaces are scarce and relatively small. However, to provide more useful information for urban and public health planning, there is a need for more research about the specific life course, park-related and contextual conditions that make this association significant.

Footnotes

Acknowledgements

The authors thank Rafael Vela, deceased 12 December 2017, for assistance with GIS analysis.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Consejo Nacional de Ciencia y Tecnología/Secretaría de Educación Pública (grant no. 153536).

ORCID iD

Ietza Bojorquez

References

Annerstedt

Ostergren

P. O.

Bjork

Grahn

Skarback

Wahrborg

(2012). Green qualities in the neighbourhood and mental health – Results from a longitudinal cohort study in Southern Sweden. BMC Public Health, 12, 337. doi:10.1186/1471-2458-12-337

Araya

Montgomery

Rojas

Fritsch

Solis

Signorelli

Lewis

(2007). Common mental disorders and the built environment in Santiago, Chile. British Journal of Psychiatry, 190, 394–401. doi:10.1192/bjp.bp.106.024596

Astell-Burt

Feng

Kolt

G. S.

(2013). Mental health benefits of neighbourhood green space are stronger among physically active adults in middle-to-older age: Evidence from 260,061 Australians. Preventive Medicine, 57, 601–606. doi:10.1016/j.ypmed.2013.08.017

Astell-Burt

Mitchell

Hartig

(2014). The association between green space and mental health varies across the lifecourse: A longitudinal study. Journal of Epidemiology and Community Health, 68, 578–583. doi:10.1136/jech-2013-203767

Baron

R. M.

Kenny

D. A.

(1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182.

Bojorquez

Salgado

(2009). Características psicométricas de la Escala Center for Epidemiological Studies-depression (CES-D), versiones de 20 y 10 reactivos, en mujeres de una zona rural mexicana. Salud Mental, 32, 299–307.

Boone

C. G.

Buckley

G. L.

Grove

J. M.

Sister

(2009). Parks and people: An environmental justice inquiry in Baltimore, Maryland. Annals of the Association of American Geographers, 99, 767–787.

Bos

E. H.

van der Meulen

Wichers

Jeronimus

B. F.

(2016). A primrose path? Moderating effects of age and gender in the association between green space and mental health. International Journal of Environmental Reserach and Public Health, 13, pii: E492. doi:10.3390/ijerph13050492

de la Barrera

Reyes-Paecke

Harris

Bascuñán

Farías

J. M.

(2016). People’s perception influences on the use of green spaces insocio-economically differentiated neighborhoods. Urban Forestry & Urban Greening, 20, 254–264. doi:10.1016/j.ufug.2016.09.007

10.

de Vries

van Dillen

S. M.

Groenewegen

P. P.

Spreeuwenberg

. (2013). Streetscape greenery and health: Stress, social cohesion and physical activity as mediators. Social Science and Medicine, 94, 26–33. doi:10.1016/j.socscimed.2013.06.030

11.

de Vries

Verheij

R. A.

Groenewegen

P. P.

Spreeuwenberg

. (2003). Natural environments-healthy environments? An exploratory analysis of the relationship between greenspace and health. Environment and Planning A, 35, 1717–1731.

12.

Douglas

Lennon

Scott

(2017). Green space benefits for health and well-being: A life-course approach for urban planning, design and management. Cities, 66, 53–62.

13.

Francis

Wood

L. J.

Knuiman

Giles-Corti

(2012). Quality or quantity? Exploring the relationship between public open space attributes and mental health in Perth, Western Australia. Social Science & Medicine, 74, 1570–1577. doi:10.1016/j.socscimed.2012.01.032

14.

Gascon

Triguero-Mas

Martinez

Dadvand

Forns

Plasencia

Nieuwenhuijsen

M. J.

(2015). Mental health benefits of long-term exposure to residential green and blue spaces: A systematic review. International Journal of Environmental Reserach and Public Health, 12, 4354–4379. doi:10.3390/ijerph120404354

15.

Huizar

Ojeda-Revah

(2014). Los Parques de Tijuana: una perspectiva de justicia ambiental. In Ojeda-Revah

Espejel

(Eds.), Cuando las áreas verdes se transforman en paisaje. La visión de Baja California [When green areas become landscape. The Baja California vision] (pp. 87–120). Tijuana, Mexico: Colegio de la Frontera Norte.

16.

Instituto Nacional de Estadística y Geografía. (2015). Encuesta intercensal. Principales resultados. Mexico City, Mexico: INEGI.

17.

Kaczynski

A. T.

Henderson

K. A.

(2007). Environmental correlates of physical activity: A review of evidence about parks and recreation. Leisure Sciences: An Interdisciplinary Journal, 29, 315–354.

18.

Koohsari

M. J.

Mavoa

Villanueva

Sugiyama

Badland

Kaczynski

A. T.

Giles-Corti

(2015). Public open space, physical activity, urban design and public health: Concepts, methods and research agenda. Health & Place, 33, 75–82. doi:10.1016/j.healthplace.2015.02.009

19.

Lee

A. C.

Maheswaran

(2011). The health benefits of urban green spaces: A review of the evidence. Journal of Public Health, 33, 212–222. doi:10.1093/pubmed/fdq068

20.

Nutsford

Pearson

A. L.

Kingham

(2013). An ecological study investigating the association between access to urban green space and mental health. Public Health, 127, 1005–1011. doi:10.1016/j.puhe.2013.08.016

21.

Ojeda-Revah

Bojorquez

Osuna

J. C.

(2017). How the legal framework for urban parks design affects user satisfaction in a Latin American city. Cities, 69, 12–19. doi:10.1016/j.cities.2017.05.016

22.

Pattyn

Van Praag

Verhaeghe

Levecque

Bracke

(2011). The association between residential area characteristics and mental health outcomes among men and women in Belgium. Archives of Public Health, 69, 3. doi:10.1186/0778-7367-69-3

23.

Penedo

F. J.

Dahn

J. R.

(2005). Exercise and well-being: A review of mental and physical health benefits associated with physical activity. Current Opinion in Psychiatry, 18, 189–193.

24.

Radloff

(1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401.

25.

Rautio

Filatova

Lehtiniemi

Miettunen

(2018). Living environment and its relationship to depressive mood: A systematic review. International Journal of Social Psychiatry, 64, 92–103. doi:10.1177/0020764017744582

26.

Reklaitiene

Grazuleviciene

Dedele

Virviciute

Vensloviene

Tamosiunas

Nieuwenhuijsen

M. J.

(2014). The relationship of green space, depressive symptoms and perceived general health in urban population. Scandinavian Journal of Public Health, 42, 669–676. doi:10.1177/1403494814544494

27.

Reyes

Figueroa

I. M.

(2010). Distribución, superficie y accesibilidad de las áreas verdes en Santiago de Chile. EURE, 36, 89–110.

28.

Ruijsbroek

Droomers

Kruize

van Kempen

Gidlow

C. J.

Hurst

Groenewegen

P. P.

(2017). Does the health impact of exposure to neighbourhood green space differ between population groups? An explorative study in four European cities. International Journal of Environmental Reserach and Public Health, 14, pii: E618. doi:10.3390/ijerph14060618

29.

Salinas-Rodriguez

Manrique-Espinoza

Acosta-Castillo

G. I.

Franco-Nunez

Rosas-Carrasco

Gutierrez-Robledo

L. M.

Sosa-Ortiz

A. L.

(2014). [Validation of a cutoff point for the short version of the Depression Scale of the Center for Epidemiologic Studies in older Mexican adults]. Salud Publica De Mexico, 56, 279–285.

30.

Sallis

J. F.

Cerin

Conway

T. L.

Adams

M. A.

Frank

L. D.

Pratt

Owen

(2016). Physical activity in relation to urban environments in 14 cities worldwide: A cross-sectional study. Lancet, 387, 2207–2217. doi:10.1016/S0140-6736(15)01284-2

31.

Sanchez-Rodriguez

(2011). Urban and social vulnerability to climate variability in Tijuana, Mexico. In Kasperson

R. E.

Berberian

(Eds.), Integrating science and policy: Vulnerability and resilience in global environmental change (pp. 187–214). New York, NY: Routledge.

32.

Secretaría de Desarrollo Social. (1999). Sistema normativo de equipamiento urbano, tomo V, Recreación y Deporte. Retrieved from http://www.inapam.gob.mx/es/SEDESOL/Documentos

33.

Sugiyama

Leslie

Giles-Corti

Owen

(2008). Associations of neighbourhood greenness with physical and mental health: Do walking, social coherence and local social interaction explain the relationships? Journal of Epidemiology and Community Health, 62, e9.

34.

Tiburcio-Sainz

Natera-Rey

(2007). Adaptación al contexto ñahñú del Cuestionario de Enfrentamientos (CQ), la Escala de Síntomas (SRT) y la Escala de Depresión del Centro de Estudios Epidemiológicos (CES-D). Salud Mental, 30, 48–58.

35.

Triguero-Mas

Dadvand

Cirach

Martínez

Medina

Mompart

Nieuwenhuijsen

M. J.

(2015). Natural outdoor environments and mental and physical health: Relationships and mechanisms. Environment International, 77, 35–41.

36.

United Nations. (2014). World’s population increasingly urban with more than half living in urban areas. Retrieved from http://www.un.org/en/development/desa/news/population/world-urbanization-prospects-2014.html

37.

van den Berg

van Poppel

van Kamp

Andrusaityte

Balseviciene

Cirach

Maas

. (2016). Visiting green space is associated with mental health and vitality: A cross-sectional study in Four European Cities. Health & Place, 38, 8–15. doi:10.1016/j.healthplace.2016.01.003

38.

van Praag

Bracke

Christiaens

Levecque

Pattyn

. (2009). Mental health in a gendered context: Gendered community effect on depression and problem drinking. Health & Place, 15, 990–998. doi:10.1016/j.healthplace.2009.04.003

39.

Whiteford

H. A.

Ferrari

A. J.

Degenhardt

Feigin

Vos

(2015). The global burden of mental, neurological and substance use disorders: An analysis from the Global Burden of Disease Study 2010. PLoS ONE, 10(2), e0116820. doi:10.1371/journal.pone.0116820

40.

Wood

Hooper

Foster

Bull

(2017). Public green spaces and positive mental health – Investigating the relationship between access, quantity and types of parks and mental wellbeing. Health & Place, 48, 63–71. doi:10.1016/j.healthplace.2017.09.002

41.

World Health Organization. (2000). Women’s mental health: An evidence based review. Geneva, Switzerland: Author.