Localized effects of neighborhood park exposure on mental illness mortality in the Pacific Northwest United States

Abstract

Most studies on the mental health effect of greenspace assumed a constant effect of greenspace exposure over the entire study region. We expanded upon literature by hypothesizing spatial disparities in the effect of greenspace exposure on mental illness mortality across census tracts in Oregon (OR) and Washington (WA). Mental illness mortality was calculated using death register data with coordinates of residence over 13 years (2006-2018). Neighborhood park exposure was evaluated as the total area of parks in each census tract. Results indicated that the impact of park areas on mental illness mortality varied across census tracts in OR (Relative Risk=0.985, 95% credible interval=[0.950, 1.020] ‒ 1.030 [0.990, 1.072]) and WA (0.992 [0.952, 1.034] ‒ 1.013 [0.980, 1.048]). These estimates shifted overall fixed effect RRs (OR: 0.976 [0.958, 0.994]; WA: 0.976 [0.956, 0.996]). Protective effects were more likely in areas between Olympia, WA and Portland, OR. The urban-rural contrasts were explained by patients’ preference, differential greenspace accessibility, and limited park supply in national and state forests. This research highlights the disparity in greenspace effect on mental illness mortality and calls for integrating local effect variability into mental health outcomes-greenspace studies.