Ecological context modifies estimated effects of residential greenspace on mental illness mortality: a case study of Washington State, USA

Residential greenspace has been associated with lower mortality in large population-based studies, yet most analyses assume that identical levels of greenness have comparable health effects across ecologically diverse landscapes. We investigated whether the association between residential greenspace and mental illness mortality varies across ecological subregions when greenspace is defined using within-region exposure thresholds. Death certificate records from Washington State in 2018 were used to construct an analytical cohort of 31,245 decedents, including 233 deaths attributed to mental illness (ICD-10 codes F20–F69 and F99). Residential greenspace exposure was measured as the five-year mean Normalized Difference Vegetation Index (NDVI) within a 15-minute pedestrian walkshed around each residence. Binary treatment indicators were defined using the 50th, 60th, 70th, 80th, and 90th percentiles of the NDVI distribution within each ecological subregion. Average treatment effects were estimated using covariate-balancing propensity score matching followed by logistic regression. Statewide treatment effect estimates varied across NDVI thresholds without a consistent protective association. In contrast, subregion-specific analyses revealed substantial heterogeneity in both the direction and magnitude of estimated effects across ecological settings. Region-specific NDVI thresholds produced more stable and interpretable estimates than statewide absolute NDVI cutoffs, indicating that associations between residential greenspace and mental illness mortality depend on local ecological context. The association between residential greenspace and mental illness mortality differed across ecological subregions, suggesting that the health relevance of a given NDVI value is context dependent. These findings highlight ecological stratification as an important consideration in environmental health research and support the use of context-specific exposure definitions when evaluating the health effects of residential greenspace. Larger longitudinal studies with improved exposure assessment are needed to clarify these relationships.