Study shows inverse association between density of London street trees and rates of antidepressant prescription
Greater density of trees in London neighborhoods had an inversion association with the use of antidepressant medication among residents in a recent cross-sectional analysis.
Taylor MS, Wheeler BW, White MP, Economou T, Osborne NJ. Research note: Urban street tree density and antidepressant prescription rates—A cross-sectional study in London, UK. Landsc Urban Plann. 2015 Apr;136:174-179.
Cross-sectional analysis comparing datasets of pharmaceutical antidepressant prescription (aD-Rx) rates with density of urban street trees (UST) in London, England, United Kingdom (UK). Both variables were analyzed based on geographic borough, each of which contains roughly 250,000 residents and covers roughly 20 square miles (50 km2).
Residents of London (population ~8 million) who utilized the British National Health Service (NHS) during the fiscal year 2009-2010: It is estimated that approximately 90% of the British population uses NHS for their healthcare,1 providing a highly representative and robust sample of Londoners. All 33 London boroughs were included in this study.
Exposure to UST was determined by calculating mean density of street trees per km of roadway for each borough, via data available from a standard available greenspace dataset from the Greater London Authority. Detailed record-keeping by UK municipal services department ensured that every street tree in the city of London was included.
This study looked at rates of NHS aD-Rx, averaged for each borough. Due to the universal tracking of modern electronic medical records (EMR), the near ubiquity of NHS enrollment, and the requirement of prescription for obtaining pharmaceutical antidepressant medication, rate of aD-Rx is considered an adequate and complete proxy of rates of depression across London.
Control for potentially confounding factors included borough mean 1) socioeconomic status (SES), 2) employment status, 3) tobacco use, and 4) age. Bayesian data analysis methods were utilized to adjust for other unmeasured confounding factors and account for uncertainty in the effect models.
After controlling for confounding variables, rates of aD-Rx were inversely associated with UST density using standard linear regression modeling. Specifically, the addition of 1 UST was associated with 1.18 fewer aD-Rx per 1,000 people per borough (‒1.18, 95% confidence interval [CI]: ‒2.45-0.00). This is a slight change from the unadjusted model (‒1.38, 95% CI: 2.72-0.03), demonstrating that while confounders such as income and employment status influenced aD-Rx rates, these SES factors do not completely account for all of the between-borough differences.
The relationship between environmental factors (natural features in particular) and mental health status is well established in the research literature2,3 and has been previously discussed in this journal.4,5 This brief research study adds to that body of work by using an established EMR dataset as the outcome measure of interest. The use of prescription rates for antidepressant medication is a unique way to assess health outcomes and adds support to previous self-reported measures of mental health that show an association with UST density.6 While the use of aD-Rx as a proxy for mental health has its limitations (see below), the robustness of the data minimizes many of these concerns.
Currently, up to 12% of Americans report having at least 1 episode of major depression in the past year, costing 23 billion USD annually in worker productivity due to absenteeism alone.
Regardless of the method, this study is valuable in that it investigates a potential solution to the modern epidemic of clinical depression. Currently, up to 12% of Americans report having at least 1 episode of major depression in the past year, costing 23 billion USD annually in worker productivity due to absenteeism alone.7 In terms of disease burden, depression is currently ranked as the second most disabling health condition in the world according to the World Health Organization, with projections that it will reach number 1 by the year 2050.8
This rise in depression prevalence has numerous contributing causes, including the accelerating pace of urban development and the subsequent “nature deficit disorder” that has been previously discussed.4 These “urban stress” effects9 have been demonstrated by experimental functional magnetic resonance imaging studies,10,11 as well as large epidemiological urban-vs-rural comparisons. A 2010 meta-analysis of 20 studies (N=143,894) from economically developed countries revealed a statistically significant increase in mood disorders (odds ratio [OR]:1.28; 95% CI:1.13-1.44; P<0.001) and anxiety disorders (OR:1.13; 95%CI:1.00-1.28; P=0.06) but not in substance-use disorders for residents of urban vs rural settings.12 Like the current aD-Rx study, these results were consistent even after controlling for income and multiple other SES factors.
All of these data support the holistic “settings approach” to health and wellness that is becoming a major aspect of modern public health and sustainable urban development practice.13 It also falls easily under the purview of natural medicine, in that exposure to more natural elements like street trees in the urban environment reduces the impact of disease,14 while cultivating positive health states that contribute to overall well-being.15
As with any research study, this one has its limitations. As mentioned, this was a cross-sectional study utilizing existing datasets. It can only establish spatial correlation and cannot demonstrate causality or direct effect of UST on depression prevalence or aD-Rx. Furthermore, the study utilizes aggregated mean data for each borough and therefore any attempts to ascribe individual relationships or benefits fall prey to the “ecological inference fallacy” that posits what is true for the group as a whole is necessarily true for each person in the group. However, as noted above, other experimental studies6,16 have demonstrated individual-level effects.
In addition, this study examined only the density of street trees. It did not account for other aspects of urban green space (UGS), such as yard trees, home or allotment gardens, or proximity to public parks or urban forests that have been shown to influence measures of mental health.17-19 However, it is likely that these other components provide additive mental health benefits, and therefore their absence should be seen as an incomplete assessment rather than lack of potential confounding of UGS effects.
Lastly, rate of aD-Rx is only a proxy measure of population depression prevalence. It is well established that not every person with clinical depression is medically diagnosed, nor does everyone seek or utilize pharmaceutical interventions. Some unknown percentage of the study population likely utilizes complementary and integrative health (formerly called complementary and alternative medicine)20 treatments or self-medicates for symptom management; this subpopulation would be therefore outside the NHS data. A more rigorous future study could consider direct assessment of individual depression criteria or prevalence of clinical depression diagnosis, as other studies have done.19 However, the vast size of the utilized dataset and Bayesian analysis methods account for some of these study design limitations and suggest the analyses are valid enough to warrant further investigation into the relationship between UST and depression.
This study adds to the evidence base for support of UGS as a health-promotion tool and suggests that increased exposure to nature may be useful to residents of modern urban environments for improving mental health.