The impact of green space on nonaccidental and cause-specific mortality in the Adventist Health Study-2 population.

Abstract

There is growing interest in evaluating the long-term health effects of neighborhood environments, particularly green space. However, only a limited body of research further incorporates multiple ambient air pollutants. This study looked at the relationship between green space, as measured by the Normalized Difference Vegetation Index, and mortality adjusted by key confounders in the Adventist Health Study-2, a longitudinal cohort study from 2002 to 2015, across the contiguous United States (N = 67,400). We used Cox proportional hazard regression models to assess the risk of nonaccidental, cardiovascular disease (CVD), ischemic heart disease (IHD), and respiratory disease mortality from green space around subjects' home address under multiple covariate and pollutant adjustments. We found a 0.1 unit increase in the Normalized Difference Vegetation Index was associated with nonaccidental (hazard ratio [HR]: 0.96 [95% confidence interval (CI): 0.93, 0.99]), CVD (HR: 0.94 [95% CI: 0.90, 0.98]), and IHD (HR: 0.87 [95% CI: 0.81, 0.94]) mortality, with the greatest precision in fully adjusted three-pollutant models using the 1000-m buffer. Effect estimates were strengthened in urban areas, when incorporating seasons, and for females. However, all associations between green space and respiratory mortality were null. This study supports evidence that increased neighborhood green space is inversely associated with nonaccidental, CVD, and IHD mortality, where the inclusion of multiple environmental covariates had a greater impact on effect estimate magnitude and precision than adjustment by individual lifestyle and health factors.