Green qualities in the neighbourhood and mental health – results from a longitudinal cohort study in Southern Sweden

In spite of general health improvements and an increasing average lifespan in most European countries, the prevalence of mental disorders is rising cross-nationally [1]. Mental and behavioural disorders are estimated to account for 12% of the global burden of disease [2]. Gender differences reported throughout the world remain partly etiologically unclear, but the 2:1 ratio (women:men) for major depression seems to be rather consistent cross-culturally [3].

According to a survey performed in Southern Sweden in 2005 the prevalence of mental ill health, based on self-assessment methods, was 15% among men, and 21% among women (2004) [4]. This is mirrored by a growing number of people who are on sick leave due to mental disorders. The most common diagnoses are stress related states (e.g. burnout, depression, anxiety), and the prevalence as well as the increase are higher among women [4, 5].

The modern concept of health is a complex interaction of environmental, organizational, and personal factors within the contexts and places that people live their lives. Thus broader environmental issues must be considered in matters of population health, something that has for example been reflected in studies of migrant populations where significant health effects by changing life environment have been demonstrated [6, 7].

The salutogenic approach focuses on health factors (e.g. physical activity and healthy diet) as ways of maintaining good health [8]; access to nature is one such health factor that has received greater attention of late [9–11].

Some research on associations between nature and health has been based on evolutionary hypotheses, claiming that we have a genetic, inherent need for nature which by instinct makes us calm and less stressed in such settings [12–14].

Another hypothesis maintains that we are prone to mental fatigue, due to an overload of directed attention in our concentration demanding daily lives. In a natural setting, where spontaneous attention (fascination) is activated, this mental fatigue can be alleviated, allowing our brains to rest and recover [15, 16].

Both these hypotheses link nature’s positive health outcome to reduced stress or mental fatigue, and there is good precedence for believing in the validity of these hypotheses. Since at least 50 years empirical studies have shown improved psychological wellbeing, reduced stress, and beneficial effects on pulse and blood pressure by contact with green spaces [17–20]. Data suggest that access and interaction with nature settings (e.g. parks, community gardens, urban greenways, forests, playing fields, and river corridors) has independent health effects and increase vitality and perceived general health [21, 22]. Recently it was concluded that mental health is the realm of health that is probably most affected by green space in people’s neighbourhoods [23].

Apart from acting as a buffer between stressful life events and health, other mechanisms between nature and health have been considered. One such mechanism that has been suggested is physical activity. Access to nearby green areas is expected to increase physical activity and consequently also for example reduce levels of obesity [24–26].

Reduced socioeconomic health inequalities by access to green environments is another assumed mechanism for the link between nature and health [10], as well as increased social cohesion [27].

There is some consistency in the findings of access to nature as a positive factor for population health, but studies linking nature with health behaviour or health outcome has traditionally been mainly cross-sectional, and partly incoherent [28]. With few exceptions [29], there is also a lack of deeper exploration into what kind of nature is particularly beneficial to health. As an answer to this deficiency, attempts to assessing and categorizing certain nature characteristics have been considered. By storing such landscape assessments in Geographic Information Systems (GIS) a higher level of transparency is achieved.

GIS has emerged as an important computer-based tool in understanding spatial and temporal variation in human exposure to environmental and social determinants of health [30]. With GIS data specific qualities of nature can be mapped and integrated in epidemiological research in order to investigate the impact of place on health (hence referring to ‘place’ in the classic epidemiological triad of causation – person, place, time) [26, 31, 32].

In previous studies of connections between nature and health the effect of nature is often rather discrete. This is an expected phenomenon given the multifactorial background to many disorders. Hence spatial data may be considered as complementary in identifying health or risk factors. Complementary factors can be investigated by studying potential synergistic effects between theoretically plausible interactions, for example between “place” factors and “person” factors (referring to the triad of causation). Socioeconomic traits and physical activity are examples of person factors influencing health; in particular these factors are also suggested to correlate with the connections between green areas and population health [33–35].

Given that mental health may be affected by access to green spaces, and given that a few other person factors are suggested to interact and correlate with the connections between nature and health, a few questions arise. In this study the main research question is whether or not an inverse relationship exists between green qualities in the neighbourhood and development of mental disorder.

A secondary question is if there exist any synergistic mental health effects of nature (“place” factor) and a few life style related factors (“person” factors): financial stress, living conditions, and physical activity. Theoretically, those in financial stress or those living in apartments would benefit the most from nature, and the physically active population would have an even more beneficial effect on mental effect (than what would be expected as an outcome from the physical activity itself) if access to nature. We also aimed at scrutinizing any difference between each specific green quality, in terms of health effect and the synergistic effect.

In this cohort study we did not find a simple inverse relationship between green qualities in the neighbourhood and development of mental disorder.

However, we found a synergistic mental health effect through interaction between certain green qualities and physical activity. The risk of having poor mental health at follow-up decreased 80% if having access to Serene and being physically active and 70% if access to Space and physically active, compared to not having access to either of these qualities and being physically inactive. These effects were statistically significant for women, but not for men. However, the tendencies were the same for men.

Regarding the other three green qualities as well as amount of qualities no statistically certain synergistic effects were found.

The strength of this study is the longitudinal perspective in a relatively large cohort. Data were achieved from a health survey that was broad and contained a validated instrument (GHQ12) for measuring mental health. The extensiveness of the questionnaire made detailed confounding control possible.

Another advantage was the objective measures of nature by the storage of predefined green qualities in GIS. This enabled transparent environmental neighbourhood assessments that could be correlated to estimates of the participants’ mental health. Due to technical restrictions we were only able to assess five of the original eight qualities, but among these five, the three qualities that have been considered the most important in aspects of stress relief (Wild, Serene, and Space) were included [17, 46].

Our study is geographically restricted to Southern Sweden. The particular area studied is in general quite limited in aspects of natural resources and green qualities; the average amount of green qualities in the population was small (μ = 0.72 at baseline and 0.71 at follow-up). As in any epidemiological study we must also be aware of the risk for neglected confounders. Other limitations include that the exposure variable financial stress was constructed from one single question about problems with paying bills. However, this construct has been used in former studies, and has also been found to correlate with poor self-rated health [26, 60]. We were also dependent on existing data from the survey on physical activity, and the measure is not validated.

The selective attrition on the basis of mental health at baseline poses a risk for selection bias. However, this should not affect the relative risk estimate, since we assume that loss to follow-up applied equally to the exposed (to green qualities and physical activity) and non-exposed group [61]. There seemed to be no selection bias concerning level of physical activity.

The statistical significance levels found for the additive effect of Serene and Spacious nature to physical activity among women, were 0.050 and 0.045 respectively, thus the risk for false positive results must be considered. However, the tendency of a measurable effect for certain qualities seemed to be consistent.

Regarding the indicated gender difference concerning the benefits from surrounding nature the mechanism is obscure. It must be acknowledged that the effect trends were the same for women and men; only the significance was weaker for men. Theoretically possible explanations for plausible gender discrepancies might be varied use of nature between the sexes, or variance in response to stress, mental issues, and necessary restorative experiences. Gender differences in perceiving and experiencing natural landscapes may also exist. From brain-imaging studies gender-related differences in the neural correlates of aesthetic preference have been found [62].

A common assumption in research on nature and health has been “the more the merrier”, in the sense that you would presume more nature to result in an increased health outcome. We did not find any support for this assumption in our study, but it suggested that the specific quality of greenery might be more important than quantity, and that this quality can actually be specified. This is in line with former studies, which have shown that the quantity as well as the quality of neighbourhood greenspace seem relevant with regard to health [29].

Given the multifactorial background to mental disorders and other non-communicable diseases the concept of synergies and interaction effects is interesting. This potentially also allows for detection of influential factors with otherwise too small effect sizes. Physical activity is increasingly being recognized as not only beneficial to physical, but also to mental health [63–65]. It is noteworthy that moderator variables have previously been demonstrated to influence the nature and magnitude of the relationship between exercise and different mental health outcomes [66]. In animal studies so called “enriched environments”, where exercise is regarded as one of the components, have been found to promote neurogenesis and enhance memory functions [67].

We did not find a direct connection between green qualities and mental health in this study, but it does not necessarily mean that this connection does not exist. As previously mentioned the studied area is rather low in landscape diversity, hence the studied differences in access become rather subtle, something that might require larger power of the study in order to detect significant effects. There is ongoing research on the qualities to strengthen their validity and consequently applicability. More stringent precision and validations of the methods would hypothetically enable the revealing of significant connections between certain green qualities and certain health outcome. In this perspective it is interesting that we did find tendencies of health effects from green qualities, given the non-validated landscape data and the sub-optimal geographic area in this study. This stresses the relevance of further studies since a validated landscape assessment tool would be an utterly important method for landscape planners and population health workers.

A considerable body of research has shown links between health and nature, but until date the implications for policy and decision making, whether it concerns population health issues or landscape planning, have been scarce. In this study we have found that in interaction with physical activity the qualities Serene and Space have some risk-reducing effect on mental health disorders for women, an impact that seems to overshadow the mere amount of nature. This in turn might be considered in the practical design and management of everyday environments. Notwithstanding some limitations and restrictions we believe that this study may bring us closer to positive and efficient implications and use of green spaces in relation to population health.