Skip to content

Environmental Health Perspectives

Facebook Page EHP Twitter Feed Open Access icon  

Research April 2018 | Volume 126 | Issue 4

Environ Health Perspect; DOI:10.1289/EHP2429

Green Space Visits among Adolescents: Frequency and Predictors in the PIAMA Birth Cohort Study

Lizan D. Bloemsma,1,2 Ulrike Gehring,2 Jochem O. Klompmaker,1,2 Gerard Hoek,2 Nicole A.H. Janssen,1 Henriëtte A. Smit,3 Judith M. Vonk,4 Bert Brunekreef,2,3 Erik Lebret,1,2 and Alet H. Wijga1
Author Affiliations open

1National Institute for Public Health and the Environment (RIVM), Bilthoven, Netherlands

2Institute for Risk Assessment Sciences (IRAS), Utrecht University, Utrecht, Netherlands

3Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, Netherlands

4Department of Epidemiology, Groningen Research Institute for Asthma and COPD (GRIAC), University of Groningen, University Medical Center Groningen, Groningen, Netherlands

PDF icon PDF Version (351 KB)

  • Background:
    Green space may influence health through several pathways, for example, increased physical activity, enhanced social cohesion, reduced stress, and improved air quality. For green space to increase physical activity and social cohesion, spending time in green spaces is likely to be important.
    Objectives:
    We examined whether adolescents visit green spaces and for what purposes. Furthermore, we assessed the predictors of green space visits.
    Methods:
    In this cross-sectional study, data for 1911 participants of the Dutch PIAMA (Prevention and Incidence of Asthma and Mite Allergy) birth cohort were analyzed. At age 17, adolescents reported how often they visited green spaces for physical activities, social activities, relaxation, and to experience nature and quietness. We assessed the predictors of green space visits altogether and for different purposes by log-binomial regression.
    Results:
    Fifty-three percent of the adolescents visited green spaces at least once a week in summer, mostly for physical and social activities. Adolescents reporting that a green environment was (very) important to them visited green spaces most frequently {adjusted prevalence ratio (PR) [95% confidence interval (CI)] very vs. not important: 6.84 (5.10, 9.17) for physical activities and 4.76 (3.72, 6.09) for social activities}. Boys and adolescents with highly educated fathers visited green spaces more often for physical and social activities. Adolescents who own a dog visited green spaces more often to experience nature and quietness. Green space visits were not associated with the objectively measured quantity of residential green space, i.e., the average normalized difference vegetation index (NDVI) and percentages of urban, agricultural, and natural green space in circular buffers around the adolescents’ homes.
    Conclusions:
    Subjective variables are stronger predictors of green space visits in adolescents than the objectively measured quantity of residential green space. https://doi.org/10.1289/EHP2429
  • Received: 6 July 2017
    Revised: 23 March 2018
    Accepted: 23 March 2018
    Published: 30 April 2018

    Address correspondence to L.D. Bloemsma, National Institute for Public Health and the Environment (RIVM), P.O. Box 1, Bilthoven 3720 BA, Netherlands. Telephone: +31 30 274 2455. Email: lizan.bloemsma@rivm.nl

    Supplemental Material is available online (https://doi.org/10.1289/EHP2429).

    The authors declare that they have no actual or potential competing financial interests.

    Note to readers with disabilities: EHP strives to ensure that all journal content is accessible to all readers. However, some figures and Supplemental Material published in EHP articles may not conform to 508 standards due to the complexity of the information being presented. If you need assistance accessing journal content, please contact ehponline@niehs.nih.gov. Our staff will work with you to assess and meet your accessibility needs within 3 working days.

  • PDF icon Supplemental Material PDF (877 KB)


    Note to readers with disabilities: EHP has provided a 508-conformant table of contents summarizing the Supplemental Material for this article (see below) so readers with disabilities may determine whether they wish to access the full, nonconformant Supplemental Material. If you need assistance accessing journal content, please contact ehponline@niehs.nih.gov. Our staff will work with you to assess and meet your accessibility needs within 3 working days.
    PDF icon Supplemental Table of Contents PDF (76 KB)



Introduction

Exposure to green space may be associated with beneficial health effects, including improved pregnancy outcomes, reduced cardiovascular morbidity and mortality, and improved mental health. However, only few studies examined the effects of green space on the health of adolescents (Banay et al. 2017; Dzhambov et al. 2014; Gascon et al. 2015; James et al. 2015; Nieuwenhuijsen et al. 2017).

Green space may influence health through several pathways. For example, green space may influence health by providing opportunities for physical activity, enhancing social cohesion, reducing stress, decreasing noise levels, and improving air quality (Hartig et al. 2014; Markevych et al. 2017; Nieuwenhuijsen et al. 2017). However, very few studies have examined the contribution of the different pathways between green space and morbidity (Nieuwenhuijsen et al. 2017). Reduced stress, decreased noise levels, and improved air quality can beneficially affect health without individuals consciously engaging with green space (Hartig et al. 2014). In contrast, for green space to increase physical activity and social interaction, actual green space visits are likely to be important. It is therefore important to examine whether people actually spend time in green spaces and for what purposes green spaces are visited to get more insights into the contribution of different pathways.

Generally, objectively measured surrounding greenness and/or access to green space are used in epidemiological studies, assessed by land use maps or remote sensing indices such as the normalized difference vegetation index (NDVI) (Dadvand et al. 2012; Huynh et al. 2013; Markevych et al. 2014). A limitation of assessing exposure merely through the presence of green space is the lack of data on the actual green space visits by the study participants.

To our knowledge, only few studies have examined green space visits by adolescents. A study conducted in the United States found that adolescents used parks more often for physical activity when there was higher perceived park availability, park quality, and park use by friends (Ries et al. 2009). A study in California showed that increasing age was associated with a decreased likelihood of being physically active in parks and that females were less often physically active in parks than males (Babey et al. 2015). However, these studies did not focus on visits to green spaces other than parks that may also affect the health of adolescents. Additionally, the studies mainly focused on physical activity and not on any of the other proposed pathways.

In the present study, we aim to examine whether adolescents visit green spaces and for what purposes. We also aim to identify the predictors of green space visits.

Methods

Study Design and Population

This study used data from the Dutch Prevention and Incidence of Asthma and Mite Allergy (PIAMA) birth cohort. Detailed descriptions of the PIAMA study have been published previously (Brunekreef et al. 2002; Wijga et al. 2014). In brief, pregnant women were recruited from the general population in three different parts of Netherlands during their second trimester of pregnancy. Their children were born in 1996–1997 (n=3,963 at baseline) and have been followed from birth up to the age of 17 y. Questionnaires were sent to the participating parents during pregnancy, 3 mo after the child was born, when the child was 1 y old, and yearly thereafter until the child was 8. When the children were 11, 14, and 17 y old, both parents and children were asked to complete a questionnaire. The study protocol has been approved by the medical ethical committees of the participating institutes, and written informed consent was obtained from the parents of all participants.

The present cross-sectional study used data from the questionnaires completed by parents and adolescents at the age of 17 y. At this age, 3,109 families (78.5% of 3,963) were still participating in the study and received questionnaires: one questionnaire for the adolescents and one questionnaire for the parents. In total, 2,096 adolescents (67.4% of 3,109; 52.9% of 3,963) and 1,875 parents completed their questionnaire. Most of the information used in the present study was obtained from the questionnaire for the adolescents. However, information on some of the potential predictors of green space visits, such as dog ownership and parental level of education, was obtained from the parental questionnaire. Adolescents with complete information on green space visits and potential predictors thereof were included in this study. This resulted in a study population of 1,911 adolescents.

Definition of Green Space Visits as the Outcome Variable

The frequency of green space visits was assessed with the following question: “How often do you intentionally go to a green environment (not your own garden) for the following activities: physical activities (e.g., walking, cycling, doing sports), social activities (e.g., meeting friends, having a picnic), relaxation (e.g., reading, resting, watching people, sunbathing), and to experience nature and quietness?” Answering options were “never,” “less than once a month,” “1–3 times a month,” “once a week,” and “more than once a week.” We assumed that green spaces may be beneficial to health when they are visited regularly. Therefore, five binary outcome variables were created: Visiting green space at least once a week for a) physical activities (yes/no); b) social activities (yes/no); c) relaxation (yes/no); d) experiencing nature and quietness (yes/no); and e) any of the types of activity mentioned before (yes/no). Participants reported how often they visited green spaces in summer and in winter. Since only a small percentage of adolescents visited green spaces in winter (Figure 1), we only assessed predictors of green space visits in summer.

Bar graph plotting percentage of adolescents visiting green spaces at least once a week (y-axis) across types of activities, namely, physical activities, social activities, relaxation, experiencing nature, and any of the types (x-axis) during winter, summer, summer in urban areas, and summer in nonurban areas.

Figure 1. Percentage of adolescents visiting green spaces at least once a week according to type of activity, winter or summer season, and level of urbanization for visits during the summer. Urban area: 1,000 addresses per km2; nonurban area: <1,000 addresses per km2.

Definition of Potential Predictors

Sociodemographic characteristics.

We included the child’s sex and several indicators of socioeconomic status (SES): maternal and paternal level of education (obtained from the 1-y questionnaire), the participant’s level of education (obtained from the 17-y questionnaire), and area-level SES [based on status scores of four-digit postal code areas of 2014 from Statistics Netherlands (Knol 2012)]. Status scores include the average income, the percentage of residents with a low income, the percentage of low-educated residents, and the percentage unemployed subjects in a postal code area. A higher status score indicates a higher neighborhood SES (Knol 2012). Maternal and paternal education were divided into three categories: primary school only or lower secondary or lower vocational education (low), intermediate vocational education or intermediate or higher secondary education (intermediate), and higher vocational education or university (high). The educational level of the adolescents was divided into two categories: lower secondary or lower vocational education, intermediate secondary, or intermediate vocational education (low/intermediate), higher secondary education or higher vocational education or university (high).

Urbanization.

Data on address density as an indicator of the degree of urbanization per four-digit postal code for 2011 were obtained from Statistics Netherlands (Centraal Bureau voor de Statistiek 2016). Statistics Netherlands divides address density into five categories: 2,500 addresses per km2; 1,500 to 2,500 addresses per km2; 1,000 to 1,500 addresses per km2; 500 to 1,000 addresses per km2; and <500 addresses per km2.

Dog ownership.

Parents of the adolescents answered the following questions: “Do you have a dog at home?” and if yes, “How many hours per week does your child walk the dog outside?” Responses to the questions were used to create a variable with the following categories: does not own a dog, owns a dog and adolescent walks the dog 1 h or less per week, and owns a dog and adolescent walks the dog more than 1 h per week. Since information on dog ownership was missing for 271 adolescents (14.2%), we created a fourth category, “no information available about dog ownership,” to avoid that these adolescents would be excluded from the analyses.

Perceived importance of a green environment.

For each type of activity (i.e., physical activities, social activities, relaxation, and to experience nature and quietness), adolescents reported whether a green environment was “not important,” “somewhat important,” “important,” or “very important” to them. Our questionnaire did not contain a question about the importance of a green environment for green space visits in general, i.e., for any type of activity. We therefore combined the responses to the four separate activities to create a variable with three categories for the outcome “visiting green space at least once a week for any type of activity”: not important, somewhat important, and important.

Perceived neighborhood greenness.

Adolescents were asked to classify their neighborhood as very green, green, moderately green, little green, or not green. Since only a small group of adolescents reported that their neighborhood was not green, we combined the latter two categories into one: little to no green.

Quantity of residential green space.

Different indicators were used to objectively assess the quantity of green space within certain distances of the adolescent’s home. To assess surrounding greenness levels, we used the NDVI, derived from Landsat 5 Thematic Mapper (United States Geological Service, https://earthexplorer.usgs.gov/) data at 30 m×30 m resolution. NDVI values range from 1 to 1, with higher values indicating more greenness (Weier and Herring 2000). Negative values correspond to water and were set to zero so that the effects of water surfaces do not negate the presence of green space (Markevych et al. 2014). Several options to handle negative NDVI values are available (Markevych et al. 2017). We could have artificially reduced the average NDVI by recoding negative values to zero, as compared to removing negative NDVI values. However, the average percentage of water in a buffer of 3,000 m was only 6.29 in our study population, so we assume that this has not substantially affected our estimates of greenness exposure. We generated a map of Netherlands by combining cloud-free images of the summer of 2010. For each adolescent, surrounding greenness was assessed by calculating the average NDVI in buffers of 100 m, 300 m, 500 m, 1,000 m, and 3,000 m around his/her home. The buffers of 100 m, 300 m, and 500 m represent the quantity of green space near the adolescent’s home, whereas the buffers of 1,000 m and 3,000 m represent the quantity of green space in a larger area around the adolescent’s home. We hypothesize that green spaces close to home may have a different effect on the frequency of green space visits than green spaces in a larger area around the home.

We hypothesized that different types of green space may have different effects on green space visits. Top10NL (https://www.kadaster.nl/brt) is a detailed land-use map of Netherlands (Kadaster 2014). In contrast to the NDVI, street greenery and private green property (such as gardens) are not included in the Top10NL. We used Top10NL of 2015 to assess the percentages of urban green, natural green, and agricultural green in buffers of 100 m, 300 m, 500 m, 1,000 m, and 3,000 m around the residential addresses. To distinguish between different types of green space, we defined all green spaces within a “population cluster” as “urban green space.” A population cluster is defined as a locality with at least 25 predominantly residential buildings (Vliegen et al. 2006). The remaining green spaces were classified as “agricultural green space” (arable land, fruit or tree nurseries, orchards, or grassland) and “natural green space” (forests or heather). A total of 57% of the adolescents had no natural green space in the 300-m buffer around their homes (Table 1). Therefore, a binary variable was created: natural green space in a buffer of 300 m: yes/no. Both the average NDVI and the percentages of urban, agricultural, and natural green space in different buffer sizes were highly correlated (Table S1). We therefore decided to include the buffers of 300 m, 1,000 m, and 3,000 m only in the present analyses, and did not perform analyses with the indicators of the quantity of green space in buffers of 100 m and 500 m.

Table 1. Characteristics of the study population and potential predictors of green space visits (n=1,911).
Characteristic n (%), mean±SD, or median (25th–75th percentiles)
Sex
 Girl 974 (51.0)
 Boy 937 (49.0)
Age in years 17.8±(0.3)
Maternal level of education
 Low 339 (17.7)
 Intermediate 780 (40.8)
 High 792 (41.4)
Paternal level of education
 Low 400 (20.9)
 Intermediate 638 (33.4)
 High 873 (45.7)
Educational level of adolescent
 Low/intermediate 1,015 (53.1)
 High 896 (46.9)
Neighborhood SESa 0.5 (0.1  to 1.3)
Urbanization
2500 addresses per km2 153 (8.0)
 1,500–2,500 addresses per km2 638 (33.4)
 1,000–1,500 addresses per km2 378 (19.8)
 500–1,000 addresses per km2 443 (23.2)
<500 addresses per km2 299 (15.7)
Owning a dog
 Does not own a dog 1,198 (62.7)
 Owns a dog, and walks the dog 1 h per week 267 (14.0)
 Owns a dog, and walks the dog >1 h per week 175 (9.2)
 No information available about dog ownership 271 (14.2)
Importance of a green environment
 Not important 589 (30.8)
 Somewhat important 757 (39.6)
 Important 565 (29.6)
Perceived neighborhood greenness
 Very green 297 (15.5)
 Green 871 (45.6)
 Moderately green 643 (33.7)
 Little to no green 100 (5.2)
Distance from home to the nearest park
300 m 645 (33.8)
300–1,000 m 845 (44.2)
>1,000 m 421 (22.0)
Average NDVI in 300-m bufferb 0.55 (0.48–0.61)
Average NDVI in 1,000-m bufferb 0.58 (0.51–0.65)
Average NDVI in 3,000-m bufferb 0.62 (0.56–0.68)
Percentage urban green in 300-m buffer 9.7 (4.4–15.6)
 Buffers that have no urban green 112 (5.9)
Percentage urban green in 1,000-m buffer 9.2 (5.1–13.9)
 Buffers that have no urban green 54 (2.8)
Percentage urban green in 3,000-m buffer 6.0 (2.9–9.7)
 Buffers that have no urban green 7 (0.4)
Percentage agricultural green in 300-m buffer 1.1 (0.0–16.3)
 Buffers that have no agricultural green 829 (43.4)
Percentage agricultural green in 1,000-m buffer 18.8 (4.9–39.8)
 Buffers that have no agricultural green 173 (9.1)
Percentage agricultural green in 3,000-m buffer 39.9 (23.7–55.4)
 Buffers that have no agricultural green 0.0 (0.0)
Percentage natural green in 300-m buffer 0.0 (0.0–1.3)
 Buffers that have no natural green 1091 (57.1)
Percentage natural green in 1,000-m buffer 1.8 (0.3–5.7)
 Buffers that have no natural green 244 (12.8)
Percentage natural green in 3,000-m buffer 4.1 (2.1–10.0)
 Buffers that have no natural green 0.0 (0.0)

Note: NDVI = normalized difference vegetation index; SD, standard deviation; SES, socioeconomic status.

aA higher score indicates a higher SES.

bNDVI values range from 0 to 1, with higher values indicating more greenness.

Bestand Bodemgebruik is another land-use map of Netherlands (Centraal Bureau voor de Statistiek 2008). It is less detailed than Top10NL (it contains fewer land-use categories), but in contrast to Top10NL, it contains a separate category for parks defined as public green spaces that are used for relaxation. With a detailed map covering all roads and paths of Netherlands, we estimated the distance along roads (i.e., network distance) in meters from the adolescents’ homes to the nearest park. A categorical variable was created: has a park within 300 m of the residential address, has a park within 300 m to 1,000 m of the residential address, and has no park within 1000 m of the residential address.

We included the following objectively measured indicators of the quantity of residential green space in the analyses: a) the average NDVI in buffers of 300 m, 1,000 m, and 3,000 m around the adolescent’s home; b) the percentage of urban, agricultural, and natural green space in buffers of 300 m, 1,000 m, and 3,000 m around the adolescent’s home; and c) the distance from the home address to the nearest park. These indicators of the quantity of residential green space were determined in ArcGIS (version 10.2.2; Esri).

Statistical Analyses

First, we investigated the shape of the unadjusted relationships between continuous predictors and visiting green space at least once a week (yes/no) by generalized additive models with integrated smoothness estimation and a log link [GAM function of R (version 2.8.0; R Development Core Team]. Since most of the associations were found to be linear or almost linear, the continuous predictors were not transformed. We used log-binomial regression models to calculate prevalence ratios (PRs) and 95% confidence intervals (95% CIs) (Spiegelman and Hertzmark 2005). We performed analyses of the unadjusted associations between each of the potential predictors and the five outcomes (visiting green space at least once a week for any type of activity, for physical activities, for social activities, for relaxation, and for experiencing nature and quietness). Predictors that were associated with at least one outcome with a p-value of 0.10 were selected for multivariable modeling with backward variable selection. Simultaneous inclusion of the objectively measured indicators of the quantity of green space in buffers of 300 m, 1,000 m, and 3,000 m in the same model resulted in multicollinearity problems (variance inflation factor >4.5). We therefore decided to include only buffers of 300 m and 3,000 m in multivariable modeling. The Akaike information criterion was used to determine for each of the five outcomes the model that best fit the data, i.e., five outcome-specific models were made (Burnham and Anderson 2002). In the end, we included all predictors that were selected into at least one of the five outcome-specific models in our final model to facilitate the comparison of models between outcomes.

Additionally, stratified analyses by level of urbanization were performed. The level of urbanization was divided into two categories: urban (1,000 addresses per km2) and nonurban (<1,000 addresses per km2).

Several sensitivity analyses were performed. We repeated the analysis of the predictors of green space visits in summer with the frequency of green space visits in five categories (never, less than once a month, one to three times a month, once a week, more than once a week) by polytomous logistic regression with “never” as the reference category. We also performed sensitivity analyses without the predictor “perceived importance of a green environment” to examine whether other predictors were associated with green space visits in summer when this predictor was omitted. Moreover, we assessed whether the predictors of green space visits in winter differed from the predictors of green space visits in summer. The percentage of adolescents that visited a green space at least once a week in winter for relaxation or to experience nature and quietness was too low to perform the analyses. Therefore, the sensitivity analyses were limited to the frequency of green space visits in winter for physical activities, social activities, and any type of activity. Since log-binomial regression models failed to converge, we assessed the associations between the frequency of green space visits in winter and potential predictors with Poisson regression (Spiegelman and Hertzmark 2005).

The statistical analyses, except the generalized additive models with integrated smoothness estimation, were performed with SAS (version 9.4; SAS Institute Inc.).

Results

Population Characteristics and Frequencies of Green Space Visits

Characteristics of the study population are presented in Table 1. In total, 41% of mothers and 46% of fathers were highly educated, 61% of the adolescents classified their neighborhood as (very) green, and 78% of the adolescents had a park within 1,000 m of their home. The median average NDVI in buffers of 300 m, 1,000 m, and 3,000 m increased with increasing levels of “perceived neighborhood greenness,” whereas no such trend was observed for the median percentage of urban green space (Figure S1).

Adolescents reported visiting green spaces mostly for physical activities and social activities and less often for relaxation and to experience nature and quietness (Figure 1). A total of 53% of adolescents visited a green space at least once a week for any type of activity in summer, whereas this percentage was 26 in winter. Participants living in urban areas and participants living in nonurban areas hardly differed in how often they visited green spaces in summer (Figure 1).

Predictors of Green Space Visits

Table S2 displays the unadjusted associations between visiting green spaces at least once a week for different activities and potential predictors of green space visits. The perceived importance of a green environment is the only predictor that was strongly and consistently associated with all five outcomes. Different predictors were associated with the frequency of green space visits for different purposes. For example, adolescents who owned a dog were more likely to visit green spaces at least once a week for physical activities, relaxation, and to experience nature and quietness, while boys visited green spaces more often for physical activities than girls.

Table 2. Predictors associated with visiting green spaces at least once a week (yes/no) from multivariable regression analysis.
Predictor Physical activities Social activities Relaxation Experiencing nature and quietness Any type of activity
PR (95% CI) PR (95% CI) PR (95% CI) PR (95% CI) PR (95% CI)
Sex
 Girl 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
 Boy 1.12 (1.01, 1.24) 1.15 (1.02, 1.28) 0.87 (0.74, 1.02) 0.86 (0.70, 1.07) 1.06 (0.99, 1.15)
Paternal level of education
 Low 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
 Intermediate 1.09 (0.91, 1.29) 1.09 (0.92, 1.29) 1.00 (0.82, 1.23) 1.08 (0.80, 1.45) 1.03 (0.92, 1.16)
 High 1.25 (1.06, 1.48) 1.22 (1.03, 1.44) 0.92 (0.75, 1.14) 1.01 (0.75, 1.36) 1.13 (1.01, 1.26)
Educational level of adolescent
 Low/intermediate 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
 High 1.02 (0.91, 1.14) 0.85 (0.75, 0.96) 0.84 (0.71, 0.99) 0.82 (0.65, 1.03) 0.95 (0.88, 1.04)
Owning a dog
 Does not own a dog 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
 Owns a dog, and walks the dog 1 hour per week 1.01 (0.86, 1.19) 1.05 (0.90, 1.23) 1.15 (0.92, 1.43) 1.47 (1.10, 1.97) 1.01 (0.90, 1.14)
 Owns a dog, and walks the dog >1 hour per week 1.07 (0.93, 1.24) 1.03 (0.85, 1.25) 1.16 (0.91, 1.47) 1.67 (1.23, 2.25) 1.05 (0.92, 1.20)
 No information available about  dog ownership 0.95 (0.81, 1.10) 1.17 (1.01, 1.34) 1.15 (0.93, 1.42) 1.29 (0.96, 1.73) 1.13 (1.02, 1.25)
Importance of a green environment
 Not important 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
 Somewhat important 2.53 (1.85, 3.45) 2.04 (1.59, 2.61) 1.85 (1.35, 2.52) 3.54 (1.93, 6.50) 1.82 (1.60, 2.08)
 Important 4.55 (3.41, 6.09) 3.81 (3.03, 4.80) 3.48 (2.60, 4.65) 9.62 (5.48, 16.88) 2.16 (1.89, 2.46)
 Very important 6.84 (5.10, 9.17) 4.76 (3.72, 6.09) 4.87 (3.53, 6.71) 18.76 (10.69, 32.92) N/A

Note: Results are derived from log-binomial regression analysis. The prevalence ratios are adjusted for all variables presented in this table. Predictors that are included in at least one of the five outcome-specific models that best fit the data are presented. CI, confidence interval; PR, prevalence ratio.

Table 2 shows the results from the multivariable log-binomial regression analyses. The perceived importance of a green environment remained the strongest predictor of green space visits in multivariable analyses. Adolescents who reported that a green environment was (very) important to them visited green spaces more often than adolescents for whom a green environment was not important [PR (95% CI): 6.84 (5.10, 9.17) for physical activities and 4.76 (3.72, 6.09) for social activities]. Boys visited green spaces more often for physical and social activities than girls [PR (95% CI) 1.12 (1.01, 1.24); PR (95% CI) 1.15 (1.02, 1.28), respectively] and adolescents who owned a dog were 1.5–1.7 times more likely to visit green spaces at least once a week to experience nature and quietness. Adolescents with a high level of education visited green spaces less often for social activities [PR 0.85 (95% CI: 0.75, 0.96)] and relaxation [PR 0.84 (95% CI: 0.71, 0.99)] than adolescents with a low to intermediate level of education. Participants with highly educated fathers were more likely to visit green spaces at least once a week for physical and social activities and any type of activity compared to adolescents with fathers who were less educated [PR 1.25 (95% CI: 1.06, 1.48); PR 1.22 (95% CI: 1.03, 1.44); PR 1.13 (95% CI: 1.01, 1.26), respectively].

Maternal level of education, area-level SES, and the objectively measured quantity of green space (i.e., the average NDVI and percentages of urban, agricultural, and natural green space in buffers of 300 m and 3,000 m around the participants’ homes and the distance to the nearest park) were not significantly associated with green space visits in multivariable analyses (data not shown). We could not add all objectively measured indicators of the quantity of green space at the same time to the models in Table 2 because of multicollinearity problems. The NDVI is the most frequently used indicator to assess exposure to greenness in epidemiological studies. We have therefore decided to only add the average NDVI in buffers of 300 m and 3,000 m at the same time to the models that are displayed in Table 2. The association between the average NDVI and green space visits remained nonsignificant (Table S3).

When we stratified by level of urbanization, we found that the associations between adolescents living in urban areas and adolescents living in nonurban areas were generally very similar (Figure 2). However, adolescents with highly educated fathers only visited green spaces more often for physical and social activities and any type of activity when they lived in an urban area. We found no associations between paternal level of education and green space visits for social activities and any type of activity in adolescents who lived in nonurban areas.

Plot marking prevalence ratios with 95 percent confidence intervals for physical activities, social activities, relaxation, experiencing nature, and any types in urban and nonurban areas. The study population is stratified as follows: sex (boy versus girl), education father (intermediate versus high; high versus low), education adolescent (high versus low; intermediate), dog ownership (walks dog more than or equal to 1 hour per week versus no dog, walks dog more than 1 hour per week versus no dog, and no info versus no dog), importance of green environment (somewhat versus not important, important versus not important, and very important versus not important).

Figure 2. Predictors associated with visiting green space (yes/no) stratified by level of urbanization from multivariable log-binomial regression analysis. The prevalence ratios are adjusted for all variables presented in this figure. Note: The level of urbanization is divided into two categories: urban (1,000 addresses per km2) and nonurban (<1,000 addresses per km2).

The perceived importance of a green environment was also the strongest predictor of green space visits when the frequency of green space visits was divided into five categories (Tables S4–S8). The results of these sensitivity analyses were generally similar to those of the main analyses except that adolescents with a higher level of education visited green spaces more often for any type of activity than adolescents with a low/intermediate level of education [odds ratio (OR) 1–3 times a month vs. never: 2.57 (95% CI: 1.66, 3.98); OR once a week vs. never: 2.21 (95% CI: 1.42, 3.44); OR more than once a week vs. never: 1.60 (95% CI: 1.03, 2.50)] (Table S8). Tables S9 and S10 show the associations between the frequency of green space visits in winter and potential predictors of green space visits. The strongest predictor of the frequency of green space visits in winter was the perceived importance of a green environment, which is consistent with our results for green space visits in summer. We found some differences between the predictors of green space visits in summer and green space visits in winter. Perceived neighborhood greenness and the percentage of agricultural green space in a buffer of 3,000 m around the adolescent’s home were significantly associated with green space visits in winter, whereas paternal level of education was not (Table S10). Adolescents who classified their neighborhood as “green” or “very green” visited green spaces more often for any type of activity in winter, whereas a higher percentage of agricultural green space in a buffer of 3,000 m was associated with a lower likelihood of visiting green space at least once a week for physical activities or any type of activity (Table S10). Sensitivity analyses without the predictor “perceived importance of a green environment” yielded results similar to those of the main analyses. However, perceived neighborhood greenness was significantly associated with green space visits for physical activity only when “perceived importance of a green environment” was omitted (Table S11). Adolescents who classified their neighborhood as “very green” visited green spaces more often for physical activities than adolescents who classified their neighborhood as “little to no green” [PR: 1.40 (95% CI: 1.00, 1.94)]. We also observed this positive trend of perceived neighborhood greenness for the other outcome variables, but these associations were nonsignificant.

Discussion

Main Findings

This study found that in Netherlands, 53% of adolescents aged 17 y visited a green space at least once a week in summer. Adolescents reported visiting green space mostly for physical activities and social activities and less often for relaxation and to experience nature and quietness. The strongest predictor of green space visits in adolescents was the perceived importance of a green environment. Boys and adolescents with highly educated fathers visited green spaces more often for physical and social activities. Adolescents who own a dog visited green spaces more often to experience nature and quietness. The frequency of green space visits was not significantly associated with the objectively measured quantity of residential green space.

Comparison with Other Studies

Previous studies in children and adolescents have focused on the frequency and predictors of park visits. To our knowledge, this is the first study that has examined the frequency and predictors of green space visits in general (not only park visits) in adolescents. Veitch et al. examined park visits of 8- to 16-y-old children living in disadvantaged areas of Victoria, Australia (Veitch et al. 2014). In this study, 75% of the children reported visiting parks. Among these children, 37% visited their “usual park” at least once a week and 69% at least several times per month (Veitch et al. 2014). Another study found in California that 78% of adolescents aged 12–17 y reported that they visited a park in the past 30 d (Babey et al. 2015). Flowers et al. examined 2,079 working-age adults in the United Kingdom and found that 68% of participants visited the green space closest to their home at least a few times a month (Flowers et al. 2016).

The present study showed that adolescents who own a dog visited green spaces more often to experience nature and quietness. Our results are in line with previous research that has shown that dog walkers are frequent users of green space (Lachowycz and Jones 2013).

We found that the perceived importance of a green environment was the strongest predictor of green space visits in adolescents. The frequency of green space visits was not associated with the objectively measured quantity of residential green space.

Our results are in line with a study by Flowers et al. that has shown that subjective predictors, such as nature relatedness (individual levels of connectedness with the natural world), are associated with the use of local green space in adults (Flowers et al. 2016). That study also found that the objectively measured percentage of local green spaces was not associated with the use of local green space (Flowers et al. 2016). Our findings are also consistent with a study that showed that the number of parks within a 1-mi radius around the adolescents’ homes was not associated with adolescents’ park use for physical activity (Ries et al. 2009).

In contrast, a study that examined 135 low- to middle-income children aged 8–14 y in southern California found that children used neighborhood parks more often when parks were closer to the children’s homes (Dunton et al. 2014). The discrepancy between the current study and the study in southern California may be due to differences in study populations: Our study population was, on average, 17.8 y old and mainly consisted of middle- and highly educated families. Furthermore, park use was measured differently in the two studies. In the Californian study, park use of participants was measured by Global Positioning System (GPS) over a 7-d period, while the present study used questionnaires to assess the frequency of green space visits. Our results are also not consistent with two studies that examined green space visits in adults. Giles-Corti et al. showed that the likelihood of using public open spaces increased with increasing levels of access to public open spaces in 1,803 adults in Australia (Giles-Corti et al. 2005). Another study in the United States found that residents living closer to parks had a higher number of weekly park visits (Sturm and Cohen 2014).

Interpretation and Implications of Findings

We found that adolescents with fathers who are highly educated were more likely to visit green spaces at least once a week for physical and social activities and any type of activity. In contrast, highly educated adolescents were less likely to visit green spaces for social activities and relaxation than adolescents with a low/intermediate level of education. Maternal level of education was not associated with green space visits in adolescents. The educational level of the mother, father, and adolescent may be indicators of different constructs. Paternal level of education tends to be associated with family income and standard of living, whereas the educational level of the adolescent is more likely to be an indicator of the attitudes, preferences, and behaviors of his/her peer group. Maternal level of education may be an indicator of family lifestyle and health-related behaviors. This may explain the discrepancy between the associations of paternal level of education, maternal level of education, and the educational level of the adolescent.

The present study showed that the perceived importance of a green environment was the strongest predictor of green space visits in adolescents, suggesting that it is the adolescents’ attitude towards a green environment that impacts green space visits. The frequency of green space visits was not associated with the objectively measured quantity of residential green space. For adolescents, other environmental attributes may influence the frequency of green space visits, like the quality of green space. No information on the quality of green space was available in the present study. Other explanations for the lack of an association between green space visits and the distance to the nearest park are the relatively short distances from the homes to parks and the frequent use of bicycles in our study population. Nearly 80% of the adolescents had a park within 1,000 m of their homes. In Netherlands, teenagers bike on average 2,000 km per year (Centraal Bureau voor de Statistiek 2015). In other words: parks were generally available and accessible (by bike) for our study population. It is therefore possible that the adolescents’ attitudes towards green space (i.e., did the adolescents want to visit a green space?) influenced the frequency of green space visits more than the actual distance to residential green space.

The NDVI also includes street greenery and private green property (such as gardens), which are not included in our definition of green space visits. This may explain the absence of an association between the average NDVI in several buffers and the frequency of green space visits in our study. However, we also found no relation between the percentages of urban, agricultural, and natural green spaces with the frequency of green space visits. These percentages of green space are based on Top10NL, which does not include street greenery and private green space.

In the present study, adolescents reported how often they intentionally visit green spaces for specific purposes. We did not find an association between the objectively measured quantity of residential green space and visiting green spaces for physical activities. However, when the quantity of residential green space is higher, adolescents may use active modes of travel instead of passive modes of travel. The quantity of residential green space could therefore influence physical activity levels in adolescents.

Our finding that the quantity of residential green space was not associated with green space visits may indicate that self-selection bias, i.e., individuals choose to reside in neighborhoods that align with their preferences for green space visits, does not play a critical role in studies examining the health effects of green space in adolescents. Our findings may also have implications for the interpretation of studies examining the health effects of green space. Those studies mostly use objective measures to assess surrounding greenness, such as the NDVI, as a proxy for greenness exposure. In those studies, there is no information on the actual green space visits by the study participants. Yet some proposed pathways through which green space may affect health require actual green space visits (Hartig et al. 2014). Our results suggest that the quantity of residential green space as measured by the NDVI or land-use maps may not be a suitable proxy for visiting green space in adolescents because the quantity of residential green space was not associated with the frequency of green space visits. It is therefore likely that pathways that do not require actual green space visits are involved in the associations between objectively measured green space and health in adolescents that have been reported in the literature.

There is a possibility of reverse causation in our study: It is unclear whether the perceived importance of a green environment actually causes a higher frequency of green space visits or whether a higher frequency of green space visits influences adolescents’ attitudes towards a green environment. However, our finding that the perceived importance of a green environment was strongly associated with the frequency of green space visits may be relevant for public health policy. It indicates that not only the availability of residential green space, but also attitudes towards green space might be relevant targets for public health strategies. We were not able to examine the predictors of the perceived importance of a green environment, i.e., environmental attitudes. Future epidemiological studies are needed to explore the predictors of proenvironmental attitudes so that public health strategies to promote such attitudes could be implemented.

Strengths and Limitations

To our knowledge, this is the first study that has examined the perceived importance of a green environment as a predictor of green space visits in adolescents. Furthermore, we included several objective measures (the average NDVI; the percentage of urban, agricultural, and natural green space; and the distance to the nearest park) in several buffers to assess the quantity of residential green space in addition to perceived neighborhood greenness in the analyses.

However, this study has some limitations. The frequency of green space visits was self-reported and not objectively measured by, for example, GPS devices. We used the following question to assess the frequency of green space visits: “How often do you intentionally go to a green environment (not your own garden)?” No definition of a green environment was given to the study participants. It is possible that the adolescents interpreted this term in different ways. For example, some, but not all adolescents may have considered sports fields as green spaces. Since there is no universally accepted definition of green space, it is not possible to assess whether this has resulted in an over- or underestimation of the frequency of green space visits. The lack of a definition of a green environment may have resulted in differences in the reported frequencies of green space visits between adolescents that are no actual differences but caused by a different interpretation of the term “green environment.”

Furthermore, information about the quality of green space was unavailable in the present study. Both perceived and objective quality of green space may be associated with the frequency of green space visits (Flowers et al. 2016; Lee and Maheswaran 2011; Ries et al. 2009).

Of the baseline PIAMA study population, 53% completed the questionnaire at the age of 17 y. There was selective loss to follow-up of children with lower paternal and maternal education (Wijga et al. 2014). This loss to follow-up may have influenced our observed frequencies of green space visits, since a higher level of paternal education was associated with more frequent green space visits in our study. However, we assume that the associations between potential predictors of green space visits and the frequency of green space visits would not be different in the general population of Dutch adolescents.

Our study population mainly consisted of adolescents who live in a house with a garden, which is similar to the general Dutch population (Ministerie van Binnenlandse Zaken en Koninkrijksrelaties 2013). People who live in a house without a garden may visit green spaces more often. The results of our study may therefore not be generalizable to adolescents in other countries, where the percentage of homes with a garden is lower.

The present study examined the frequency and predictors of green space visits as reported among adolescents aged 17 y. Future studies are needed to assess these associations in other age groups as well.

Conclusion

This study found that more than half of the adolescents visited a green space at least once a week in summer, mostly for physical and social activities. The strongest predictor of green space visits among adolescents was the perceived importance of a green environment. The objectively measured quantity of residential green space was not associated with green space visits. Our results suggest that subjective variables are stronger predictors of green space visits than the objectively measured quantity of residential green space.

Acknowledgments

The PIAMA study has been supported by project grants from Netherlands Organization for Health Research and Development; Netherlands Organization for Scientific Research; the Lung Foundation Netherlands (formerly Netherlands Asthma Fund); Netherlands Ministry of Spatial Planning, Housing, and the Environment; and Netherlands Ministry of Health, Welfare, and Sport. The research leading to these results was supported by the Strategic Program of the National Institute for Public Health and the Environment.

References

Babey SH, Tan D, Wolstein J, Diamant AL. 2015. Neighborhood, family and individual characteristics related to adolescent park-based physical activity. Prev Med 76:31–36, PMID: 25869220, 10.1016/j.ypmed.2015.04.001.

Banay RF, Bezold CP, James P, Hart JE, Laden F. 2017. Residential greenness: current perspectives on its impact on maternal health and pregnancy outcomes. Int J Womens Health 9:133–144, PMID: 28280395, 10.2147/IJWH.S125358.

Brunekreef B, Smit J, de Jongste J, Neijens H, Gerritsen J, Postma D, et al. 2002. The prevention and incidence of asthma and mite allergy (PIAMA) birth cohort study: design and first results. Pediatr Allergy Immunol 13(suppl 15):55–60, 10.1034/j.1399-3038.13.s.15.1.x.

Burnham KP, Anderson DR. 2002. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach, 2nd Edition. New York, NY:Springer-Verlag.

Centraal Bureau voor de Statistiek. 2008. Bestand bodemgebruik productbeschrijving [in Dutch]. https://www.cbs.nl/nl-nl/onze-diensten/methoden/classificaties/overig/bestand-bodemgebruik-bbg/classificaties/classificatie-bodemgebruik [accessed 19 October 2015].

Centraal Bureau voor de Statistiek. 2015. Factsheet nederland fietsland [in Dutch]. https://www.cbs.nl/nl-nl/achtergrond/2015/27/factsheet-nederland-fietsland [accessed 18 May 2017].

Centraal Bureau voor de Statistiek. 2016. Begrippen [in Dutch]. https://www.cbs.nl/nl-nl/onze-diensten/methoden/begrippen?tab=s#id=stedelijkheid–van-een-gebied [accessed 18 February 2016].

Dadvand P, Sunyer J, Basagaña X, Ballester F, Lertxundi A, Fernandez-Somoano A, et al. 2012. Surrounding greenness and pregnancy outcomes in four Spanish birth cohorts. Environ Health Perspect 120(10):1481–1487, PMID: 22899599, 10.1289/ehp.1205244.

Dunton GF, Almanza E, Jerrett M, Wolch J, Pentz MA. 2014. Neighborhood park use by children: use of accelerometry and global positioning systems. Am J Prev Med 46(2):136–142, PMID: 24439346, 10.1016/j.amepre.2013.10.009.

Dzhambov AM, Dimitrova DD, Dimitrakova ED. 2014. Association between residential greenness and birth weight: systematic review and meta-analysis. Urban For Urban Green 13(4):621–629, 10.1016/j.ufug.2014.09.004.

Flowers EP, Freeman P, Gladwell VF. 2016. A cross-sectional study examining predictors of visit frequency to local green space and the impact this has on physical activity levels. BMC Public Health 16:420, PMID: 27207300, 10.1186/s12889-016-3050-9.

Gascon M, Triguero-Mas M, Martinez D, Dadvand P, Forns J, Plasència A, et al. 2015. Mental health benefits of long-term exposure to residential green and blue spaces: a systematic review. Int J Environ Res Public Health 12(4):4354–4379, PMID: 25913182, 10.3390/ijerph120404354.

Giles-Corti B, Broomhall MH, Knuiman M, Collins C, Douglas K, Ng K, et al. 2005. Increasing walking: how important is distance to, attractiveness, and size of public open space?. Am J Prev Med 28(2 suppl 2):169–176, PMID: 15694525, 10.1016/j.amepre.2004.10.018.

Hartig T, Mitchell R, de Vries S, Frumkin H. 2014. Nature and health. Annu Rev Public Health 35:207–228, PMID: 24387090, 10.1146/annurev-publhealth-032013-182443.

Huynh Q, Craig W, Janssen I, Pickett W. 2013. Exposure to public natural space as a protective factor for emotional well-being among young people in Canada. BMC Public Health 13:407, PMID: 23627738, 10.1186/1471-2458-13-407.

James P, Banay RF, Hart JE, Laden F. 2015. A review of the health benefits of greenness. Curr Epidemiol Rep 2(2):131–142, PMID: 26185745, 10.1007/s40471-015-0043-7.

Kadaster. 2014. Basisregistratie topografie: catalogus en productspecificaties [in Dutch]. https://www.kadaster.nl/brt [accessed 19 October 2015].

Knol F. 2012. “Statusontwikkeling Van Wijken in Nederland 1998–2010.” Den Haag:Sociaal en Cultureel Planbureau.

Lachowycz K, Jones AP. 2013. Towards a better understanding of the relationship between greenspace and health: development of a theoretical framework. Landsc Urban Plan 118:62–69, 10.1016/j.landurbplan.2012.10.012.

Lee AC, Maheswaran R. 2011. The health benefits of urban green spaces: a review of the evidence. J Public Health (Oxf) 33(2):212–222, PMID: 20833671, 10.1093/pubmed/fdq068.

Markevych I, Fuertes E, Tiesler CM, Birk M, Bauer CP, Koletzko S, et al. 2014. Surrounding greenness and birth weight: results from the GINIplus and LISAplus birth cohorts in Munich. Health Place 26:39–46, PMID: 24361636, 10.1016/j.healthplace.2013.12.001.

Markevych I, Schoierer J, Hartig T, Chudnovsky A, Hystad P, Dzhambov AM, et al. 2017. Exploring pathways linking greenspace to health: theoretical and methodological guidance. Environ Res 158:301–317, PMID: 28672128, 10.1016/j.envres.2017.06.028.

Ministerie van Binnenlandse Zaken en Koninkrijksrelaties. 2013. Cijfers over wonen en bouwen 2013 [in Dutch]. Den Haag:Ministerie van Binnenlandse Zaken en Koninkrijksrelaties.

Nieuwenhuijsen MJ, Khreis H, Triguero-Mas M, Gascon M, Dadvand P. 2017. Fifty shades of green: pathway to healthy urban living. Epidemiology 28(1):63–71, PMID: 27525811, 10.1097/EDE.0000000000000549.

Ries AV, Voorhees CC, Roche KM, Gittelsohn J, Yan AF, Astone NM. 2009. A quantitative examination of park characteristics related to park use and physical activity among urban youth. J Adolesc Health 45(3 Suppl):S64–S70, PMID: 19699439, 10.1016/j.jadohealth.2009.04.020.

Spiegelman D, Hertzmark E. 2005. Easy SAS calculations for risk or prevalence ratios and differences. Am J Epidemiol 162(3):199–200, PMID: 15987728, 10.1093/aje/kwi188.

Sturm R, Cohen D. 2014. Proximity to urban parks and mental health. J Ment Health Policy Econ 17:19–24, PMID: 24864118.

Veitch J, Carver A, Hume C, Crawford D, Timperio A, Ball K, et al. 2014. Are independent mobility and territorial range associated with park visitation among youth?. Int J Behav Nutr Phys Act 11:73, PMID: 24909862, 10.1186/1479-5868-11-73.

Vliegen M, van Leeuwen N, Kerkvliet F. 2006. Bevolkingskernen in nederland opnieuw afgebakend [in Dutch]. Voorburg/Heerlen:Centraal Bureau voor de Statistiek.

Weier J, Herring D. 2000. Measuring vegetation (NDVI & EVI). http://earthobservatory.nasa.gov/Features/MeasuringVegetation/ [accessed 19 October 2015].

Wijga AH, Kerkhof M, Gehring U, de Jongste JC, Postma DS, Aalberse RC, et al. 2014. Cohort profile: the prevention and incidence of asthma and mite allergy (PIAMA) birth cohort. Int J Epidemiol 43(2):527–535, PMID: 23315435, 10.1093/ije/dys231 [accessed 19 October 2015].


WP-Backgrounds Lite by InoPlugs Web Design and Juwelier Schönmann 1010 Wien