Effects of greenspace morphology on mortality at the neighbourhood level: a cross-sectional ecological study

Abstract

The association between urban greenspace and mortality risk is well known, but less is known about how the spatial arrangement of greenspace affects population health. We aimed to investigate the relation between urban greenspace distribution and mortality risk. We did a cross-sectional study in Philadelphia, PA, USA, using high-resolution landcover data for 2008 from the Pennsylvania Spatial Data Access database. We calculated landscape metrics to measure the greenness, fragmentation, connectedness, aggregation, and shape of greenspace, including and omitting green areas 83·6 m2 or smaller, using Geographical Information System and spatial pattern analysis programs. We analysed all-cause and cause-specific mortality (related to heart disease, chronic lower respiratory diseases, and neoplasms) recorded in 2006 for 369 census tracts (small geographical areas with a population of 2500-8000 people). We did negative binomial regression and principal component analyses to assess associations between landscape spatial metrics and mortality, controlling for geographical, demographic, and socioeconomic factors. A 1% increase in the percentage of greenspace was predicted to reduce all-cause mortality by 0·419% (95% CI 0·050-0·777), with no effect on cause-specific mortality. All-cause mortality was negatively associated with the area of greenspace. A 1 m2 increase in the mean area of greenspace led to a 0·011% (95% CI 0·004-0·018) fall in all-cause mortality and a 0·019% (0·007-0·032) decrease in cardiac mortality; considering only green areas larger than 83·6 m2 would contribute to a 0·002% (95% CI 0·001-0·003) decrease in all-cause mortality and a 0·003% (0·001-0·006) reduction in cardiac deaths. Census tracts with more connected, aggregated, coherent, and complex shape greenspaces had a lower risk of all-cause and cause-specific mortality. The negative association between articulated landscape parcels and all-cause mortality varied with age and education, such that the relation was stronger for census tracts with a higher percentage of older and less well-educated adults. A significant modest association exists between the spatial distribution of greenspace in cities and mortality risk. The overall amount of greenspace alone is probably failing to capture significant variance in local health outcomes and, thus, environment-based health planning should consider the shape, form, and function of greenspace. None.