Greenspace Morphology and Preterm Birth: A State-Wide Study in Georgia, United States (2001-2016).

Abstract

Background: Residential greenness is linked to birth outcomes. However, the role of greenspace morphology remains poorly understood. Additionally, evidence is lacking regarding whether these relationships vary by subpopulation.

Objective: We examined the association between preterm birth and residential greenspace morphology, including percentage, shape, connectedness, aggregation, closeness, and fragmentation.

Methods: We analyzed 2,063,444 singleton live births between 2001 to 2016 in Georgia, USA. Thirty-meter resolution landcover data from National Land Cover Databased (2001-2016) were obtained to calculate greenspace morphology metrics for 1,953 census tracts in Georgia. A two-stage logistic regression examined associations between each greenspace morphology metric and preterm birth at individual level. Stratified analysis was conducted by maternal race, ethnicity, education, urbanicity, poverty rate, and greenspace percentage.

Results: Higher greenspace percentage, aggregation, closeness, shape complexity, connectedness, and lower fragmentation were linked to a lower risk of preterm birth. After adjusting for poverty rate, associations with morphology attenuated, except for fragmentation [odds ratio (OR) = 1.014; 95% confidence interval (CI): 1.0001, 1.026] across the entire population. Strongest associations were found among black mothers and in high-poverty areas. Specifically, the odds of preterm birth in the highest quartile of greenspace percentage were 0.962 (95% CI: 0.933, 0.991) times the odds in the lowest quartile. Additionally, a lower risk of preterm birth was associated with higher greenspace aggregation (OR = 0.969; 95% CI: 0.947, 0.992), and a higher risk of preterm birth was associated with higher fragmentation (OR = 1.028; 95% CI: 1.009, 1.047), both in the black mothers group. In tracts with a high poverty rate, a lower risk of preterm birth associate with higher greenspace percentage (OR = 0.953; 95% CI: 0.910, 0.999), aggregation (OR = 0.976; 95% CI: 0.955, 0.997), and lower fragmentation (OR = 0.976; 95% CI: 0.958, 0.994). The association with greenspace morphology was most pronounced in census tracts with a medium level of greenspace percentage.

Discussion: Our study complements other studies by showing the importance and protective effects of greenspace morphology. The observed effects are particularly prominent in census tracts characterized by a moderate level of greenspace percentage, high poverty rates, and among black women. Our findings suggest the need for tailored greenspace planning strategies based on varying levels of greenness in different areas. For locations with low greenness, increasing the greenspace percentage may be prioritized. In areas with a medium level of greenness, strategic enhancement of greenspace morphology is recommended. For areas with high greenness, the focus should be on improving spatial closeness of greenspace. https://doi.org/10.1289/EHP14571.