If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Leafier Communities, Healthier Hearts: An Australian Cohort Study of 104,725 Adults Tracking Cardiovascular Events and Mortality Across 10 Years of Linked Health Data
Corresponding author at: School of Population Health, Faculty of Medicine and Health, UNSW, NSW 2052, Australia.
Affiliations
School of Population Health, Faculty of Medicine, University of New South Wales, Sydney, NSW, AustraliaSchool of Health and Society, Faculty of Arts, Social Sciences, and Humanities, University of Wollongong, Wollongong, NSW, AustraliaPopulation Wellbeing and Environment Research Lab (PowerLab), Sydney, NSW, AustraliaThe George Institute for Global Health, UNSW Sydney, Sydney, NSW, Australia
School of Health and Society, Faculty of Arts, Social Sciences, and Humanities, University of Wollongong, Wollongong, NSW, AustraliaPopulation Wellbeing and Environment Research Lab (PowerLab), Sydney, NSW, Australia
School of Population Health, Faculty of Medicine, University of New South Wales, Sydney, NSW, AustraliaSchool of Health and Society, Faculty of Arts, Social Sciences, and Humanities, University of Wollongong, Wollongong, NSW, AustraliaPopulation Wellbeing and Environment Research Lab (PowerLab), Sydney, NSW, Australia
Corresponding author at: School of Health and Society, Faculty of Arts, Social Sciences and Humanities, University of Wollongong, Wollongong, NSW 2522, Australia
Affiliations
School of Population Health, Faculty of Medicine, University of New South Wales, Sydney, NSW, AustraliaSchool of Health and Society, Faculty of Arts, Social Sciences, and Humanities, University of Wollongong, Wollongong, NSW, AustraliaPopulation Wellbeing and Environment Research Lab (PowerLab), Sydney, NSW, Australia
Green space reduces cardiovascular disease (CVD) risk, but few studies examine what types of green space matter, which is an important consideration as cities densify and apartments become more common.
Method
Participants were 86,727 in houses and 17,998 in apartments from the 45 and Up Study (Sax Institute) baseline survey with 10 years of linked hospitalisation and death data used to define: (i) all-cause; and (ii) CVD-mortality; (iii) fatal and non-fatal CVD events; and (iv) acute myocardial infarction (AMI). Associations with total green space, tree canopy cover and open grass within 1.6 km buffers were assessed using survival analysis adjusted for potential confounders.
Results
Mean percentage green space indicators were all higher among participants in houses than in apartments. Among residents of houses, a 10% increase in total green space was associated with reduced risk of CVD mortality (HR 0.97, 95%CI 0.95–1.00). A 10% increase in tree canopy cover was associated with reduced risks of all-cause mortality (HR 0.97, 95%CI 0.95–0.99), CVD mortality (HR 0.96, 95%CI 0.93–0.98), and fatal or non-fatal AMI (HR 0.93, 95%CI 0.89–0.96). In contrast, a 10% increase in open grass was associated with an increased risk of fatal or non-fatal AMI (HR 1.15, 95%CI 1.09–1.20) in residents of houses. Among residents of apartments, a 10% increase in total green space was associated with increased risk of all-cause mortality (HR 1.04, 95%CI 1.00–1.08) and CVD mortality (HR 1.03, 95%CI 1.00–1.08).
Conclusions
Urban reforestation may be a population-level intervention to protect cardiovascular health, especially for people living in houses. The intersection of urban greening and cardiovascular health among residents of apartments warrants further investigation.
Restoration in nature: Beyond the conventional narrative.
in: Schutte A.R. Torquati J. Stevens J.R. Nature and psychology: Biological, cognitive, developmental, and social pathways to well-being (Proceedings of the 67th Annual Nebraska Symposium on Motivation). Springer Nature,
Cham, Switzerland2021
Innovations in exposure measurement have permitted research that increases the sophistication of current evidence for urban greening policy and practice. In particular, various studies now report that tree canopy may have the greatest prevention potential for CVDs, rather than any green space per se [
]— 3 trees seen from each home, 30% tree canopy cover in each neighbourhood, each home within 300 metres of the nearest park or green space—which focusses attention on leveling up inequities in access to tree canopy with three inter-related strategies.
However, emerging evidence indicates that some of these greening targets may operate differently according to the type of housing in which people live [
]. For example, an Australian study reported more time spent outdoors, more physical activity overall, and higher proportions of vigorous physical activities with more tree canopy specifically for people living in houses [
]. These benefits were not observed for people living in apartments, for whom more open grass nearby was actually associated with less physical activity. This presents challenges to the implementation of urban densification agendas in many cities around the world where the focus is on apartment-living [
] and may also have downstream impacts on association between green space and CVD risk.
To our knowledge, no previous study has examined whether associations between green space and CVD are modified by housing type. Given the aforementioned evidence, we hypothesised that (i) people living in houses where there is more green space—and tree canopy specifically—will tend to have lower CVD risk, in comparison to their peers living in apartments.
Method
Data
Participants in the Sax Institute’s 45 and Up Study [
] baseline survey were recruited between 2006 and 2009 from the Services Australia (formerly the Department of Human Services and Medicare Australia) enrolment database. The overall response rate to the baseline survey was approximately 18%, with 267,153 people taking part across the Australian state of New South Wales (NSW). All participants gave written informed consent for their responses to be linked to other data for research purposes. Data on all deaths (from the Register of Births, Marriages and Deaths, and the Cause of Death Unit Record File) and a complete census of all admissions to public and private hospitals in NSW (the Admitted Patient Data Collection) from 1 July 2000 up to 31 December 2016 was linked by the Centre for Health Record Linkage (CHeReL) (http://www.cherel.org.au/) using probabilistic methods. The University of New South Wales Human Research Ethics Committee (HREC) approved the 45 and Up Study and ethics approval for the analyses in this study were provided by the University of Wollongong HREC and the NSW Population and Health Services Research Ethics Committee.
This study was restricted to 104,725 participants living in houses and apartments in the cities of Sydney, Newcastle and Wollongong, as associations between green space and CVD in rural countryside settings may be confounded by issues of barred access to farmland and/or prevalence of endocrine disrupting chemicals associated with CVD risk [
]. We stratified the sample to distinguish between city dwellers living in houses (n=86,727) from those in apartments (n=17,998). Participants living in retirement villages, nursing homes, hostels for the aged, mobile homes or other housing types were omitted due to small numbers (total n=5,056). Housing type was identified by how participants answered the following question from the baseline survey: “What best describes your current housing?”
Outcomes
Three types of CVD-events and all-cause mortality constituted the four outcomes of interest. All four were measured up to 31 December 2016 from the date of baseline survey completion. Death records linked from the NSW Register of Births, Marriages and Deaths were used to measure all-cause mortality and CVD mortality, with the latter defined in line with previous research [
] using primary International Classification of Diseases 10th revision-Australian Modification (ICD-10-AM) codes I00 to I99, G45 and G46. The third outcome was fatal and non-fatal CVD events defined by augmenting the CVD mortality data with records of CVD-related hospital admissions using consistent ICD-10-AM codes, focussing on the first CVD event only. The fourth outcome was a subset of the third, comprising hospital admissions and deaths attributable to acute myocardial infarction (AMI). Fatal and non-fatal AMI was measured in line with previous work [
Mortality after admission for acute myocardial infarction in Aboriginal and non-Aboriginal people in New South Wales, Australia: a multilevel data linkage study.
], including events coded as I21 (AMI) as a primary diagnosis, and events classified as both ‘acute care’ and ‘emergency’ while secondarily coded I21 with a primary diagnosis of ischaemic heart disease (I20 or I22 to I25).
Green Space
We measured the percentage of all green space, tree canopy cover and open grass within road-network buffers of 1.6 km (1 mile) around proxy locations of residence (Mesh Blocks, containing between 30 and 60 dwellings each); 1.6 km represents a reasonable walking distance [
]. We used best available two-metre high-resolution land-use data sourced from Pitney Bowes Ltd (Stamford, CT, USA) on public and private green space, including deciduous and evergreen tree canopy and woodlands. Open grass that was unobstructed by tree canopy (e.g. sports ovals and large grassy reserves) was included, whereas that which was located beneath tree canopy cover was not and as such, the quantity of open grass is likely to be under-estimated. Total green space was the sum percentages of tree canopy cover, open grass, and a third variable referred to as ‘shrub’ (which was not analysed separately due to small values). Each green space variable was analysed separately as continuous variables with 10% increments, aligning with current urban greening strategies used in various cities. For example, 10% of subdivisible land in Perth (Western Australia) has historically been allocated to some form of green space [
Patterning of each outcome and covariate across levels of green space availability were assessed using frequencies and percentages. Cox proportional hazards models were fitted in R (package: Coxph), nesting participants within geographical boundaries using a frailty term. The geographic boundaries were defined by Statistical Area 3 scale, developed by the Australian Bureau of Statistics. Statistical Area 3s are aligned with council areas and each one has a population of between 30,000 to 130,000 residents. These models permitted estimation of hazard ratios (HR) and 95% confidence intervals (95%CI) for each green space variable and outcome while accounting for time-to-event occurrence and adjusting for potential confounding. Confounders were selected to address issues of selection into neighbourhoods containing more green space that are also known correlates of CVD, including sex, age, annual household income, highest educational qualification, work status, relationship status, and region of birth. Interaction terms were fitted to assess potential effect modification of associations between green space and mortality/event outcomes between participants living in houses and apartments. Selected results from interaction analyses were graphed using predicted means and 95% confidence bands to aid interpretation through visualisation. Models were then run on samples stratified by housing type to permit more detailed investigation.
Results
Among participants in houses, all-cause mortality and CVD mortality rates were 12.4% and 7.3%, respectively (Table 1). These rates were lower than those observed among participants in apartments, where all-cause mortality was 17.0% and CVD mortality was 9.9%; 40.8% of people in houses experienced fatal or non-fatal CVD events in comparison to 45.9% in apartments. Fatal and non-fatal AMI was 3.5% and 4.4% in house and apartment dwellers, respectively. Similar demographic profiles for samples in houses and apartments were noted, with main differences being the proportion of household incomes less than AUD$70K per annum (55.9% in houses vs 45.4% in apartments) and how common it was to be living alone (19.7% in houses vs 51.2% in apartments). Mean levels of total green space were higher for those in houses (40.1%) than in apartments (32.2%). Differences in tree canopy cover (22.0% house, 19.6% apartment) and open grass (13.3% house, 9.0% apartment) similarly favoured those in houses.
Cox proportional hazards models were fitted to assess association between each green space indicator and outcome, adjusting for confounding but absent of the housing type variable (Table 2, Models A1-B1-C1). Models A1 indicate that a 10% increase in total green space was not associated with any of the outcomes. Models B1 indicate that a 10% increase in tree canopy cover was associated with a 2% reduced risk of all-cause mortality (HR 0.98, 95%CI 0.96–1.00), a 3% reduced risk of CVD mortality (HR 0.97, 95%CI 0.95–1.00), and a 6% reduced risk of fatal or non-fatal AMI (HR 0.94, 95%CI 0.90–0.97). A 10% increase in open grass was associated with a 13% increased risk of fatal or non-fatal AMI only (HR 1.13, 95%CI 1.08–1.18). Adjusting for housing type did not make substantive impacts on these associations, nor was there much evidence of a difference in outcomes between residents of houses and apartments (Models A2-B2-C2). This was except for the risk of fatal or non-fatal acute myocardial infarction in the presence of adjustment for open grass (C2), wherein apartment dwellers had a 9% increased risk in comparison to peers in houses (HR 1.09, 95%CI 1.00–1.18).
Table 2Cox proportional hazard models for risks of all-cause mortality and CVD-related events/mortality with green space.
Models
Variables
All-Cause Mortality
CVD Mortality
Fatal/Non-Fatal CVD Events
Fatal/Non-Fatal AMI
Hazard ratio (95% confidence intervals)
Model A1
Total green space
0.99 (0.97, 1.01)
0.99 (0.97, 1.01)
1.00 (0.98, 1.01)
1.00 (0.97, 1.04)
Model A2
Apartment (ref: house)
1.01 (0.97, 1.06)
0.98 (0.92, 1.03)
1.00 (0.98, 1.03)
1.06 (0.98, 1.16)
Total green space
0.99 (0.97, 1.01)
0.99 (0.96, 1.01)
1.00 (0.98, 1.01)
1.01 (0.97, 1.04)
Model A3
Apartment (ref: house)
0.88 (0.77, 1.02) ˆ
0.76 (0.64, 0.92) x
0.96 (0.88, 1.05)
0.92 (0.70, 1.22)
Total green space (slope for house)
0.98 (0.97, 1.00) ˆ
0.97 (0.95, 1.00) +
0.99 (0.98, 1.01)
1.00 (0.96, 1.04)
Total green space x apartment
1.04 (1.00, 1.08) +
1.07 (1.02, 1.13) x
1.01 (0.99, 1.04)
1.04 (0.97, 1.12)
LR Test of Model 3 vs Model 2
P=0.038
P=0.007
P=0.246
P=0.515
Model B1
Tree canopy cover
0.98 (0.96, 1.00) +
0.97 (0.95, 1.00) +
0.99 (0.98, 1.01)
0.94 (0.90, 0.97) ∗
Model B2
Apartment (ref: house)
1.01 (0.97, 1.06)
0.98 (0.93, 1.03)
1.00 (0.98, 1.03)
1.05 (0.97, 1.14)
Tree canopy cover
0.98 (0.96, 1.00) +
0.97 (0.95, 1.00) +
0.99 (0.98, 1.01)
0.94 (0.90, 0.98) x
Model B3
Apartment (ref: house)
0.93 (0.84, 1.03)
0.84 (0.74, 0.96) x
1.00 (0.94, 1.06)
0.83 (0.69, 1.01) ˆ
Tree canopy cover (slope for house)
0.97 (0.96, 0.99) x
0.96 (0.94, 0.99) x
0.99 (0.98, 1.01)
0.92 (0.89, 0.96) ∗
Tree canopy cover x apartment
1.04 (1.00, 1.09) ˆ
1.08 (1.02, 1.14) +
1.00 (0.97, 1.03)
1.12 (1.03, 1.22) x
LR Test of Model 3 vs Model 2
P=0.052
P=0.012
P=0.969
P=0.009
Model C1
Open grass
1.02 (0.99, 1.04)
1.02 (0.98, 1.06)
1.01 (0.98, 1.03)
1.13 (1.08, 1.18) ∗
Model C2
Apartment (ref: house)
1.02 (0.98, 1.07)
0.98 (0.93, 1.04)
1.01 (0.98, 1.03)
1.09 (1.00, 1.18) ˆ
Open grass
1.02 (0.99, 1.05)
1.02 (0.98, 1.06)
1.01 (0.98, 1.03)
1.13 (1.08, 1.19) ∗
Model C3
Apartment (ref: house)
1.02 (0.95, 1.10)
0.98 (0.89, 1.08)
0.99 (0.95, 1.04)
1.20 (1.04, 1.39) +
Open grass (slope for house)
1.02 (0.99, 1.05)
1.02 (0.98, 1.06)
1.00 (0.98, 1.03)
1.15 (1.09, 1.21) ∗
Open grass x apartment
1.00 (0.95, 1.06)
1.00 (0.93, 1.08)
1.01 (0.98, 1.05)
0.91 (0.81, 1.02) ˆ
LR Test of Model 3 vs Model 2
P=0.888
P=0.744
P=0.344
P=0.074
Green space variables are continuous variables with a unit equal 10% of area.
Apartment (ref: house) is an indicator of dwelling type.
Green space variables (slope for house): change in outcome related to 10% increase in green space variable on house subset.
Green space variables ∗ apartment: denotes interaction term, i.e., change in outcome related to 10% increase in green space variable on apartment subset relative to the house subset.
Models are built on house and flat subset, and fully adjusted for co-variates in Cox proportional hazard regression analysis.
Frailty term was set as intercept associated with statistical area 3 (SA3) boundaries.
Model 1 = Single green space variable, adjusted for confounding.
Model 2 = Single green space variable plus indicator of the type of dwelling (house or apartment), adjusted for confounding.
Model 3 = Interaction of a single green space variable with indicator of the type of dwelling (house or apartment), adjusted for confounding.
P-values for the significance of regression coefficients are calculated in with Chi-square criterion with P-values ≤ 0.1, 0.05, 0.01, 0.001 coded as 'ˆ', '+', 'x', '∗', respectively.
Likelihood-ratio test (LR Test) for the better fit of the Model 3 with interaction term is calculated in analysis of deviance anova statement.
Augmentation with two-way interaction terms (A3-B3-C3) were found to improve the fit of some of these models according to likelihood ratio tests. Improved models were those testing associations between total green space and (i) all-cause mortality and (ii) CVD mortality, tree canopy cover and (iii) CVD mortality and (iv) fatal or non-fatal AMI (all p<0.05). Components of interaction terms were also statistically significant for (v) tree canopy cover and all-cause mortality, and (vi) open grass and fatal or non-fatal AMI (albeit with likelihood ratio tests of 0.052 and 0.074, respectively). Figure 1 provides an illustration of predicted mean values and 95% confidence bands for the interaction terms obtained from the abovementioned models to aid interpretation. In every graph (Figure 1, A-E) except for open grass and fatal or non-fatal AMI (Figure 1, F), associations were negative between total green space and/or tree canopy cover and risks of all-cause mortality, CVD mortality, fatal or non-fatal AMI. In graphs A-E, the risks of all-cause mortality, CVD mortality and fatal or non-fatal AMI were positively associated with total green space or tree canopy cover. These divergent patterns by housing type were not reflected in graph F, however, in which risks of fatal or non-fatal AMI appeared to increase for all participants, and especially for residents in houses.
Figure 1Cox proportional hazard models for effect modification of housing type on risks of all-cause mortality and CVD-related events/mortality with respect to green space type.
It is notable that in all graphs in Figure 1, 95% confidence bands are wide at higher levels of each green space variable for residents of apartments, so interpretation must be additionally careful and conservative. To provide further insight, we proceeded with models stratified by housing type (Figure 2, A-C ), which provided confirmation of divergent associations. Among the results that reached statistical significance were a 10% increase in total green space being associated with a 3% reduced risk of CVD mortality (HR 0.97, 95%CI 0.95–1.00) among residents of houses, while also being associated with a 6% increased risk of CVD mortality (HR 1.03, 95%CI 1.00–1.08) among residents of apartments. A 10% increase in total green space was also associated with a 4% increased risk of all-cause mortality (HR 1.04, 95%CI 1.00–1.08) in participants living in apartments. Among those in houses, a 10% increase in tree canopy cover was associated with 3% and 4% reduced risks of all-cause mortality (HR 0.97, 95%CI 0.95–0.99) and CVD mortality (HR 0.96, 95%CI 0.93–0.98), respectively. Divergent patterns were also found among participants in the same housing type. Among those in houses, a 10% increase in tree canopy cover was associated with a 7% reduced risk of fatal or non-fatal AMI (HR 0.93, 95%CI 0.89–0.96), whereas a 10% increase in open grass was associated with a 15% increased risk in the same outcome (HR 1.15, 95%CI 1.09–1.20). Stratification of all of these analyses between males and females, or between cohabiters compared with those living alone, did not change the results presented.
Figure 2Cox proportional hazard models for risks of all-cause mortality and CVD-related events/mortality with respect to green space type, stratified by housing type.
People living in houses tended to have lower risks of CVD mortality with more green space overall. They also had lower risks of all-cause mortality, CVD mortality and fatal/non-fatal AMI events with more tree canopy, but also a higher risk of fatal/non-fatal AMI events with more open grass. Results for people living in apartments were mostly statistically non-significant, with higher risks of all-cause and CVD mortality with more green space overall. These results confirm that the potential cardiovascular benefits of green space tend to be more strongly associated with tree canopy cover [
]. Additionally, they also indicate that these benefits may be disproportionately experienced by people living in houses. This contingency of green space and housing types may explain some of the varied and generally small effect sizes reported in previous studies that have been unable to differentiate between trees and open grass, and houses and apartments as key contexts for contingency [
]. The differences in association reported in our paper underline the need for future studies and meta-analyses to reach for more realistically complex understandings of the intersection between green space and cardiovascular health as cities evolve and, in many cases, attempt to ‘green’ as well as densify.
Several potential factors may be contributing to the largely null findings for people in apartments, as well as the counterintuitive findings for overall green space and both mortality indicators coupled with the increased AMI risk with more open grass. One speculation is that green space provision in higher density areas composed of apartments generally contain fewer qualities that contribute to cardiovascular health in comparison with those in areas dominated by housing. Perceptions of size, access, amenity, biodiversity, facilities, safety and various other quality domains are important for attracting people to spend time outdoors, exercise, socialise and feel a sense of restoration [
Perceived qualities, visitation and felt benefits of preferred nature spaces during the COVID-19 Pandemic in Australia: a nationally-representative cross-sectional study of 2940 adults.
]. This is not to say that all green spaces around apartments lack quality, but an incongruence between the qualities that are available relative to those which members of the community need is hypothesised to undermine the potential mental, physical and social health benefits [
]. For example, given the typical absence of private gardens and backyards, people in apartments are likely, on average, to be more dependent and therefore influenced by the qualities of nearby green spaces than their peers in houses. Further research to link indicators of green space qualities is needed to define whether there are systematic differences in exposure between people living in houses from those in apartments, and thereby to test whether this explains the divergent results by housing type.
Perhaps a key quality of green space within the context of housing type is green space provision per capita, since two areas may have similar levels of green space but that which is in a higher density community has greater risk of overcrowding. A study set in Vienna, Austria, reported greater levels of overcrowding in urban green spaces located in high density areas [
]. Small and overcrowded green space may constrain potential for regular exercise. This points to the actual density of the area as having a possible modifying role, as well as other contextual factors experienced by people living in apartments having a conditioning role. Higher levels of air pollution and noise are possibilities [
] and generally smaller internal spaces within apartments may amplify the psychosocial impacts.
Selection may play a role in the findings, over and above adjustment for factors influencing choice in the housing market (income, education, work status, etc.). Apartment living is the norm in many urbanised societies, but uncommon in Australian cities which remain dominated by low-density suburbs [
]. Just 10% of people in Australia were residing in apartments according to the 2016 census, with the most common age group aged 25–34 years, and only 7% of 45–54 year-olds occupying apartments [
]. Ownership of a relatively large block of land containing an expansive open plan home is traditionally perceived by some to support middle class identity and values [
]; a perception that is not afforded to apartment living. Some researchers have suggested a culture of apartment living tends towards people experiencing transitory periods rather than long-term residential ambitions [
]. There is a documented tendency to stigmatise people in apartments as being dangerous, deviant, bad for children and families, and for apartment-living to be considered the domain of singles, divorcees, and empty-nesters [
]. Meanwhile, although it is common for residents of houses to have access to some form of private outdoor space, this is a rarity for people living in apartments, who then miss out on opportunities for sustaining cardiovascular health through gardening [
]. Thus, there may be many characteristics of apartment dwellers that condition the ability to access, experience and benefit from nearby green space that we do not yet fully understand and are currently unable to measure. This indicates a strong need for more qualitative research to examine the diverse range of lived experiences of people in apartments and how they interact with green spaces.
On this point, housing type was self-reported in the 45 and Up Study and while useful as a broad indicator, it did not permit further information that could be important, such as changes in prior living circumstances (e.g. related to relationship status) and how many levels a house has, or how high up the apartment is. As mentioned earlier, there was no data on whether either housing type had access to private gardens and courtyards. In a similar vein, we selected a road network distance (a 1.6 km buffer) to measure our green space indicators within the context of sprawling, low-density Australian cities. It is important to note that some work has shown associations with some green space indicators measured for 1.6 km buffers are not observed when testing smaller buffers (e.g. 400 m or 800 m radii) [
]. Furthermore, emerging evidence indicates that larger buffers are more likely to reflect the real lengths to which many people are willing to travel to visit green spaces [
Amegbor PM, Dalgaard R, Nainggolan D, Jensen A, Sabel C, Panduro TE, et al. Spatial modelling of psychosocial benefits of favourite places in Denmark: a tale of two cities. Available at SSRN 3995572 2021.
], so focussing on that which is nearby can majorly underestimate what environments people are actually exposed to and benefit from. While it is unlikely that any particular buffer size is uniquely best aligned with a particular mediating pathway, use of a larger buffer such as 1.6 km will help to avoid arbitrarily limiting (and inducing measurement bias of) the cumulative range of opportunities for contact with green space that a person may enjoy; the so-called “local trap” [
Other limitations warrant acknowledgement. The 45 and Up Study response rate was approximately 18%, which is low. However, the responding sample had a demographic profile that was broadly representative of the population aged >45 years in Australia [
]. The green space data was high resolution, but also only available in 2016. Loss of green space can be expected in some areas over time due to various factors such as road-widening, suburban sprawl, and urban infill. Thus, linking green space data measured in 2016 with baseline residential data spanning 2006–2009 may mean some participants are classified with lower levels of green space than they actually had available at the time. We had no information on which floor people in apartments were living on, or whether they had access to communal areas with greenery. Our study benefits from having best available mortality and hospitalisation data linked to the 45 and Up Study and high resolution green space variables, with which tree canopy and open grass can be distinguished for separate testing. Other longitudinal cohort studies have previously reported people with more green space per se have reduced odds of CVD morbidity and/or mortality [
], so our ability to identify whether particular types of green space underpin these benefits advances the field and could help inform budget-constrained healthy place making.
Conclusion
This is the first longitudinal cohort study we are aware of that has examined differences in associations between green space and cardiovascular health separately for people living in houses and apartments. Unexpectedly, evidence of potential benefits was almost entirely among occupants of houses with none among apartment dwellers. Most notably, a 10% increase in tree canopy cover within 1.6 km was associated with a 3% reduced risk of all-cause mortality, 4% reduced risk of CVD mortality, and 7% reduced risk of fatal or non-fatal AMI among participants living in houses. The equivalent associations for participants living in apartments were not statistically significant. In addition, our use of high resolution spatial data to facilitate assessment of associations between health indicators and quantities of different green space types (expressed as percentages) attends to common policy-relevant limitations of previous work [
]. Thus, our study provides foundational results and indicates many avenues for further research. Given the rising number of people living in apartments around the world, this work could be consequential for developing urban greening strategies that benefit everyone and help to narrow urban health inequities.
Funding Sources
This study was supported by the Green Cities Fund - Hort Innovation Limited, with coinvestment from the University of Wollongong Faculty of Social Sciences, the University of Wollongong Global Challenges initiative, and the Australian Government project GC15005. Astell-Burt was supported by a National Health and Medical Research Council Boosting Dementia Research Leader Fellowship 1140317. Feng was supported by a National Health and Medical Research Council Career Development Fellowship 1148792. The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.
Declaration of Interest
None.
Acknowledgements
This research was completed using data collected through the 45 and Up Study (www.saxinstitute.org.au). The 45 and Up Study is managed by the Sax Institute in collaboration with major partner Cancer Council NSW; and partners: the National Heart Foundation of Australia (NSW Division); NSW Ministry of Health; NSW Government Family & Community Services – Ageing, Carers and the Disability Council NSW; and the Australian Red Cross Blood Service. We thank the many thousands of people participating in the 45 and Up Study.
References
Roth G.A.
Johnson C.
Abajobir A.
Abd-Allah F.
Abera S.F.
Abyu G.
et al.
Global, regional, and national burden of cardiovascular diseases for 10 causes, 1990 to 2015.
Restoration in nature: Beyond the conventional narrative.
in: Schutte A.R. Torquati J. Stevens J.R. Nature and psychology: Biological, cognitive, developmental, and social pathways to well-being (Proceedings of the 67th Annual Nebraska Symposium on Motivation). Springer Nature,
Cham, Switzerland2021
Mortality after admission for acute myocardial infarction in Aboriginal and non-Aboriginal people in New South Wales, Australia: a multilevel data linkage study.
Perceived qualities, visitation and felt benefits of preferred nature spaces during the COVID-19 Pandemic in Australia: a nationally-representative cross-sectional study of 2940 adults.
Amegbor PM, Dalgaard R, Nainggolan D, Jensen A, Sabel C, Panduro TE, et al. Spatial modelling of psychosocial benefits of favourite places in Denmark: a tale of two cities. Available at SSRN 3995572 2021.
01148-9&id=gr1.jpg){kind=link}
01148-9&id=gr2.jpg){kind=link}