Association between greenspace and time spent in nature with subjective wellbeing: a cross-sectional data linkage study

Abstract

Background Evidence that greenspaces are related to mental health and wellbeing mostly relies on residential exposure and few studies have considered actual use. We aimed to link the National Survey for Wales (NSW) to environmental metrics to determine whether there is an association between increased residential exposure to greenspace and subjective wellbeing, and whether this is mediated by visits to outdoor spaces. MethodsIn this cross-sectional data linkage study, we linked NSW data (2016-17 and 2018-19; repeat cross-sectional) to the Secure Anonymised Information Linkage (SAIL) Databank. Survey data included the Warwick and Edinburgh Mental Wellbeing Scale (WEMWBS), and self-reported time spent on leisure visits to open spaces in Wales (including bluespaces; derived from visit frequency in the past 4 weeks and duration of the main activity on the most recent visit). Linkage at individual level augmented this data with home neighbourhood greenspace data (Enhanced Vegetation Index [EVI] derived from satellite imagery). Using multivariate linear regression models, we estimated associations between home or visit exposures and WEMWBS, adjusting for various covariates including area level (via Welsh Index of Multiple Deprivation) and individual deprivation indicators. EVI, weekly time outdoors, and WEMWBS were standardised and both linear and quadratic terms were included in analyses. The e-cohort uses anonymised data and was approved by the SAIL Information Governance Review Panel. Findings Among NSW respondents providing outcome measures (n=5971), EVI was significantly related to both WEMWBS (U-shaped relationship, EVI beta -0•02, p=0•13; EVI² beta 0•02, p= 0•0098), and weekly time outdoors (EVI beta 0•04, p=0•020). Time outdoors was also significantly related to WEMWBS (time outdoors beta 0•17, p<0•0001; time outdoors 2 beta -0•04, p=0•018). Despite the conditions for potential mediation being met-i.e., EVI predicting both time outdoors and WEMWBS and time outdoors predicting WEMWBS-there was no evidence that time outdoors mediated the relationship between EVI and WEMWBS. EVI coefficients were not attenuated after including time outdoors (EVI beta -0•03, p=0•076; EVI² beta 0•02, p=0•018).Interpretation Living near and time spent visiting greenspaces and bluespaces were both related to better subjective wellbeing. The absence of mediation suggests that better wellbeing associated with EVI occurs through a different mechanism than visiting.