Introduction

Quality of life after spinal cord injury (SCI) is influenced by multiple factors, including physical wellness, social integration, independent living, and psychological well-being [1, 2]. Mood disorders are prevalent among people with SCI [3,4,5]. The risk of depression among people with SCI is three times the risk reported in the general population [6], and longitudinal studies show that the higher incidence of depression may persist long-term in the years following injury [7, 8]. Depression is also a risk factor for additional negative impacts on quality of life among people with SCI, including secondary medical complications, unemployment, and reduced participation [3, 9].

In the general population, a growing body of research suggests that psychological well-being may be improved by the presence of greenspace in one’s residential community [10,11,12,13]. Greenspace, or “land that is partly or completely covered with grass, trees, shrubs, or other vegetation [14],” includes land cover of both naturally-developed vegetation and human-constructed spaces, such as parks and recreational areas. The presence of greenspace in communities provides local residents with opportunities for recreation, physical activity, and community interaction [11], as well as relaxation and the aesthetic enjoyment of nature [15]. Prior studies from the UK and the Netherlands [12, 16, 17] show that, in the general population, access to naturally-occurring greenspaces may promote better health partially through buffering against the psychological impact of stressful life events. Furthermore, in urban areas, where residents typically have less access to greenspace than people living in suburban and rural communities, the presence of human-constructed greenspaces may promote better mental health [12] and reduced stress [12, 17]. The benefits of greenspace for psychological well-being in disability populations are underexplored, but a recent review [18] found that exposure to greenspaces, including forests and outdoor gardens, can have psychosocial benefits for people with acquired neurological disabilities.

To our knowledge, the psychological effects of residential greenspace have not been studied in populations with mobility disabilities. However, environmental factors may have a special significance to quality of life for people with disabilities. The International Classification of Disability and Functioning (ICF) [19] classifies contextual factors, including social structures and community environments, that may affect the experience of disability, and a growing body of research supports this framework for people living with SCI. Specifically, living in an area with open residential spaces (i.e., fewer built structures) is associated with improved participation and social integration [20] and perceived health [21] among people with SCI, whereas living in an area with greater land use mix is associated with greater functional independence [22]. This investigation builds on this line of research by evaluating the role of residential greenspace in psychological well-being among people with SCI. Because people with SCI are at increased risk for poor psychological well-being compared to people in the general population, understanding how this risk varies across residential communities can inform community integration interventions and public policy aimed at reducing barriers to community life for people with disabilities. We use survey data from a study of SCI-specific quality of life factors linked to administrative data on residential greenspace characteristics. We test the exploratory hypothesis that more residential greenspace is associated with better psychological well-being.

Methods

Materials

This investigation is a secondary analysis of cross-sectional survey data collected for the Spinal Cord Injury-Quality of Life (SCI-QOL) Calibration Study (2009–2010). Data for the SCI-QOL study originally were collected to validate novel participant-reported outcome measures of quality of life specific to people with SCI. A total of 877 people with traumatic SCI who were ages 18 years or older and fluent in English were recruited from six SCI Model Systems (SCIMS) centers in the United States to participate in survey data collection for one or more SCI-QOL domains. Full details of the SCI-QOL development and psychometric calibration have been previously published [23]. Four sites from the original study (New Jersey, Colorado, Michigan, and Illinois) collaborated on the current investigation, which involved re-contacting the original study participants and obtaining consent to geocode their residential address at the time of interview so their SCI-QOL data could be linked to local area spatial and administrative data from the United States Geological Survey (USGS). A total of n = 623 participants from the original SCI-QOL study sample were re-contacted between 2014–2015. Of these participants, six people did not consent to having their address geocoded, 25 could not be successfully re-contacted, and 47 consented but did not have an address that could be successfully matched with the USGS data. This yielded a sample of n = 545 consented participants with geomatched data, which is representative of both the SCI-QOL and national SCIMS participants across demographic (i.e., sex, race, ethnicity, marital status, age) and injury-related (i.e., level of injury and injury completeness, duration of injury) characteristics [24].

Participants

Participants were eligible for the current investigation of greenspace and psychological well-being if they were (1) community-living at the time of interview, (2) at least one year post-injury, and (3) residing in an urban or suburban area. Participants who were non-community living (n = 129) or less than one year post-injury (n = 65) were excluded from the analysis due to their more limited tenure in a residential environment. Rural area residents (n = 63) were also excluded to aid comparability with prior approaches in the general population. The final analytic sample for this investigation is n = 313 people with chronic SCI, representing 20 U.S. states, 61 counties, and 351 Census tracts. Bivariate tests indicated that there were fewer non-Hispanic white participants in the final analytic sample compared to excluded cases (62.4% vs. 75%; χ2 = 14.2, p = 0.001).

Measures

Dependent variables

This analysis includes two domains from the SCI-QOL study measuring psychological well-being: positive affect and depressive symptoms. Positive affect is based on 28 items assessing sense of well-being, life satisfaction, and sense of purpose and meaning. Depressive symptoms are measured by 28 items assessing sadness or despair and/or loss of interest in things, and feelings of hopelessness, helplessness, or worthlessness. The inclusion of both positive affect and depressive symptoms in the analysis is motivated by the development of the field of positive psychology, which emphasizes not only the absence of pathology but the endorsement of positive emotions as important components of psychological well-being [25]. Participants rated items for all subdomains on a 5-point Likert-type scale indicating level of agreement (1 = never; 5 = always). Participants were asked to think about how they had felt “lately” for the positive affect domain and “in the past 7 days” for the depressive symptoms domain. Final domain scores are reported as T-score distributions with a mean of 50 and standard deviation of 10 and are normally distributed. Higher scores correspond to higher levels of positive affect and depressive symptoms, respectively.

Key independent variables: greenspace

Following prior approaches [11], greenspace consisting of natural land cover and developed (i.e., “useable”) open space were treated as distinct measures in the analysis. Natural greenspace includes the total proportion of land covered by USGS land cover categories supporting vegetation, including forest (deciduous, evergreen, or mixed), scrub/shrubland, herbaceous/grasslands, cultivated land (pasture, crops), and wetlands (herbaceous, woody). Open water, or “bluespace,” was not included in this measure. Developed open space measures the total proportion of land covered by a human-built mixture of constructed material and vegetation, often constructed for recreational or aesthetic purposes (i.e., parks, golf courses), in which impervious surfaces (i.e., concrete, asphalt, buildings) account for less than 20% of total land cover. Participant addresses were coded to latitude and longitude using the U.S. Census Bureau’s publicly available Geocoder tool (https://geocoding.geo.census.gov/). Consistent with prior research on residential environments and health [26,27,28], including individuals with SCI [22], half- and five-mile buffers around participants’ addresses were constructed using ArcMap to define neighborhood and community environments, respectively. USGS shapefile data on land cover in 2010—the most closely aligned year to the time of interview for the original participant-reported outcomes data— were used to calculate proportions of land cover types within the buffers using Geospatial Modelling Environment. Census tract-level socioeconomic indicators were obtained from the 2009–2013 American Community Survey (ACS) five-year pooled estimates. Distributions of greenspace land cover at both the community and neighborhood level were divided into tertiles of “low,” “moderate,” or “high” intensity of land cover.

Covariates

We control for demographic, injury-related, and community socioeconomic covariates due to variation in the odds of depression after SCI across these characteristics (see Hartoonian and colleagues for an overview) [4], and also to account for demographic and socioeconomic variation across communities with more or less greenspace [29]. Participants self-reported their injury (paraplegia/tetraplegia) and impairment type (complete/incomplete). Time since injury was categorized as 1–2 years versus three or more years. Wheelchair use was assessed with a binary measure based on participants’ self-reported use of a power or manual wheelchair (vs. ambulation.) Participants’ sex, race/ethnicity, marital status, and age were assessed for the SCI-QOL study using standard survey items. The socioeconomic status (SES) of the participants’ residential areas was measured by an index of six Census tract indicators (median household income; median home value; proportion of residents receiving interest, dividend, or rental income; high school degrees; college degrees; and employment in high-status occupations) and used as a proxy for individual SES. Higher index scores represent higher SES [30].

Statistical analysis

All data analysis was conducted with Stata/SE version 16.1. Descriptive and bivariate statistics (i.e., Chi-square tests, t-tests) were used to assess the distribution of key variables and compare differences across demographic and injury-related characteristics. Ordinary least squares (OLS) regression was used to estimate the associations between community and neighborhood greenspace and psychological well-being, controlling for injury-related, demographic, and community SES characteristics. All regression models were estimated using unstandardized coefficients with robust standard errors. The models are presented in terms of unstandardized coefficients and 95% confidence intervals.

Results

Table 1 presents the sample characteristics of the study participants. Consistent with the overall SCI population in the United States, the majority of participants were male (76.4%) and non-Hispanic white (62.4%). Participants were classified by age in 15-year intervals and include relatively equal proportions of individuals 18–35 years (32.9%), 36-50 years (33.5%), and greater than 50 years old (33.5%). The majority of participants were classified as having neurologically incomplete tetraplegia (31.5%), followed by complete paraplegia (25.4%), incomplete paraplegia (21.5%), and complete tetraplegia (21.5%). The majority of the sample represents people who have been injured at least three years (65.2%) and who reported using a wheelchair as their primary means of mobility (79.3%). At the community (5 mi.) level, the amount of natural greenspace ranged from 0.04% to 94.5%, and the total amount of open developed space ranged from 0.8% to 53.1%. At the neighborhood (0.5 mi.) level, natural greenspace covered between 0.00 and 95.6% of land, whereas developed open space covered between 0.00 and 74.9% of land.

Table 1 Sample characteristics [mean (SD) or frequency (%)] (N = 313).

Table 2 presents the OLS regression models assessing the effects of greenspace in communities and neighborhoods on positive affect, adjusted for all covariates. Contrary to our hypothesis, living in communities with a higher percentage of natural greenspace was associated with lower levels of positive affect. In the community (5 mi.) model, living within the moderate tertile of natural greenspace was associated with a 2.25 decrease in positive affect T-scores compared to living in the low natural greenspace tertile (b = −2.25; 95% CI = −4.44 – −0.06.) There was no significant difference in positive affect scores between those living in the highest community greenspace tertile compared to the lowest tertile. Differences in positive affect were not influenced by natural greenspace in the smaller, neighborhood area. The models testing the effects of developed open space in the larger community indicated no differences in positive affect. However, the presence of developed open space in the neighborhood (0.5 mi. buffer) demonstrated a negative association with positive affect. Specifically, living within the moderate developed open space tertile was associated with a 2.59 decrease in positive affect T-scores (b = −2.59; 95% CI = −5.06 – −1.13), whereas there was no significant difference between the high and low tertiles. Taken together, the models show that people living in areas with moderate levels of natural greenspace in the larger surrounding community, as well as those with moderate levels of developed open spaces in their immediate neighborhoods, were less likely to report positive psychological well-being.

Table 2 OLS regression: associations of positive affect with community and neighborhood greenspace (N = 313).

Table 3 presents the effects of natural greenspace and developed open space on depressive symptoms adjusted for all covariates. In the community, natural greenspace was associated with greater levels of depressive symptoms, in that living in the moderate tertile of natural greenspace was associated with a 2.84 increase in depressive symptom T-scores compared to living in a low greenspace area (b = 2.84, 95% CI = 0.10–5.58). There was no significant difference in depressive symptom scores between those living in the highest community greenspace tertile compared to the lowest tertile. Natural greenspace was not significantly associated with depressive symptoms at the neighborhood level. Developed open space, by contrast, was significantly associated with depressive symptoms at the neighborhood level, but not the community level. Living in a neighborhood within the moderate developed open space tertile was associated with a 3.63 increase in depressive symptoms T-scores compared to living within the low tertile (b = 3.63; 95% CI = 0.59–6.66.) There were no significant differences in depressive symptom scores between the high and low greenspace tertiles. In summary, the results show that natural and developed greenspace are associated with depressive symptoms in a similar pattern to that of positive affect, in that moderate levels of natural greenspace at the community level and moderate levels of developed greenspace at the neighborhood level were both associated with poorer psychological well-being.

Table 3 OLS regression: associations of depressive symptoms with community and neighborhood greenspace (N = 313).

Discussion

This investigation adds to the growing literature on community environment predictors of quality of life after SCI by describing the association between residential greenspace and psychological well-being. By combining self-report and administrative data in a large, geographically diverse sample of people with SCI, we find evidence that natural greenspace at the community level and developed open space at the neighborhood level are associated with psychological well-being for people with SCI, albeit not in the expected direction.

We find that the people living in areas with moderate levels of greenspace report lower levels of positive affect and higher levels of depressive symptoms compared to those living in areas with low levels of greenspace. This is inconsistent with previous reports that greenspace can lower stress levels and improve mental well-being [10,11,12,13], and raises the question of whether people with disabilities may experience more limited benefits of nature. One factor that may explain this pattern is participants’ access to and engagement with greenspaces. People with disabilities may use greenspace less than their able-bodied counterparts due to concerns about safety and limited accessibility [31]. Among people with neurological disabilities, lack of access to community greenspace may promote greater negative emotions [18]. Inaccessible public spaces remain a major barrier to full community participation for people with SCI in the United States, where an estimated 60–80% of people with SCI use a wheelchair or scooter as their primary means of mobility [24]. Having a moderate amount of inaccessible greenspaces nearby may have a negative impact on psychological well-being.

Future research might explore the role of other built environment characteristics in these patterns. Low greenspace areas may have greater land use mix and more destinations, providing more opportunities to access public spaces and improving community participation for people with SCI. Furthermore, the availability of accessible transportation is a persistent problem in the SCI community, and areas of moderate greenspace in urban areas may represent areas with greater sprawl, less street connectivity, and greater dependence on private transportation to access, which could negatively affect psychological well-being among those without reliable and accessible transportation.

The roles of social connectedness and community integration in these patterns are additional factors that may also be explored in future research to understand how and why greenspace influences psychological well-being among people with SCI. Reports in the general population have found that the positive association of greenspace with mental health was explained by greater social cohesion in neighborhoods with more greenspace [12]. People with SCI are at increased risk for social isolation and decreased community integration after injury [32], and this is a contributing factor to poorer psychological well-being [3]. The extent to which the negative effects of greenspace on psychological well-being can be explained by poor social connectedness and community integration is a factor to be explored in future research.

Finally, we find different effects on psychological well-being by type and distance of greenspace measured. Specifically, moderate natural greenspace is associated with negative psychological outcomes at the community level (i.e., within a 5 mi. buffer of participants’ home addresses), whereas developed open space has an effect at the neighborhood level (i.e., within a 0.5 mi. buffer of participants’ home addresses.) This divergence is consistent with previous reports measuring natural and developed greenspace as separate constructs [11]. The negative effects of developed open or “useable” greenspace at the neighborhood level, but not at the larger community level, further emphasize the importance of understanding how people with SCI can or cannot use greenspace in order to understand these associations with psychological well-being.

Study limitations

This investigation has several limitations. Neighborhood effects on health research benefits from longitudinal data that can disentangle the effects of neighborhood features from the characteristics of individuals who live there. The SCI-QOL study design was cross-sectional, which precluded an assessment of neighborhood selection efforts and changes in the association of residential greenspace on psychological well-being over time. The sample is based on four regional sites and is not intended to be representative of people living with SCI in different areas of the United States. A larger sample size would have also allowed for comparisons of urban and rural areas, which was not possible in this analysis due to a relatively low proportion of people from rural areas. The USGS land cover categories used to create the greenspace measures are relatively coarse measures of geography that cannot account for variation in features that may make greenspace more or less accessible, such as paved paths or public access. Controlling for additional environmental factors that could affect how people experience greenspace, such as weather-related patterns (e.g., precipitation, temperature), may have provided further insight into how the effects of greenspace on mood vary across climates and seasons, but these measures were not available. We excluded bluespace, consistent with prior studies on “green” residential characteristics, though people may find open water a pleasurable aspect of the natural environment. Finally, as a secondary analysis of existing data, the study was limited to available measures in the source material. The SCI-QOL study’s aim was to develop new participant-centered quality of life outcomes measures for the SCI population, and while a limited number of participant demographic measures were collected, a full assessment of factors that influence quality of life was beyond the scope of the study. As such, the role of other key factors associated with psychological well-being among people with SCI, including pain [33], secondary complications [3, 9], self-efficacy [3, 34], and social support [3], are not included in the analysis.

Conclusion

The psychological benefits of living near greenspace may not extend to people living with SCI. The potential negative effects of greenspace on psychological well-being should be investigated further to understand the community environment factors that shape quality of life for people with SCI. Understanding how and why greenspace is associated with psychological well-being can inform public policy and urban planning strategies aimed at ensuring natural spaces are accessible and beneficial to all community members.