Is Neighborhood Green Space Associated With a Lower Risk of Type 2 Diabetes in elderly? Evidence From Shanghai, China

Background: The elderly is the main group of patients with type 2 diabetes mellitus (T2DM). The incidence rate of T2DM is closely related to physical activity and environment. There are many literatures that have proved that green space has important influence on the health level of patients with T2DM. However, due to spatial non-stationarity, it is difficult for traditional linear regression models to explain the geographic spatial variation of prevalence and related driving factors. Besides, most of the studies focus on large-scale urban green space, and ignore the green space inside residential area which are also closely related to the daily life of the elderly. Purpose: (1) To investigate association between T2DM and neighborhood green space supply. (2) To assess the current status of community green space supply in Yangpu District for finding problems and proposing corresponding measures to build a healthier community in the Post-COVID-19 World. Method：The research subjects were 318,686 people aged 60 years and older living in 306 neighborhood committees (Administrative Divisions in China), Yangpu District, Shanghai, China. Among them 32,609 were diagnosed with T2DM. The prevalence data comes from the district government statistics, and the green space data extracted by the Shanghai Surveying and Mapping Institute are reliable. Through the API of AMAP, the walking path planning function was called to capture the shortest walk time from the neighborhood to the park, as well as the number of parks within 15 minutes of the neighborhood. Then calculate the comprehensive green space accessibility index through factor analysis. Finally, GWR（Geographically Weighted Regression) was performed in ArcGIS to fit associations between medically diagnosed T2DM and green space supply in order to explore the spatial variation of the prevalence and related driving factors. Findings: The average T2DM in the elderly in the study area is 10%, GWR results showed that there was a certain spatial heterogeneity in the prevalence. The ratio of neighborhood committee green space (- 0.19 < Co < -0.41), the ratio of green space in residential area (- 0.04 < Co < -0.24), the comprehensive green space accessibility index (- 0.03 < Co < -0.11) were had a certain negative impact on T2DM. It also showed that the coverage ratio of neighborhood committee green space had a greater impact on the dependent variables among the three green space supply indicators. The green space in residential area was also correlated with the T2DM, which reflected the characteristics of green space being cherished as a limited resource in the high-density residential environment in Shanghai. In this environment, the accessibility of green space also had a certain importance. The bivariate local Moran’s I statistic results revealed that the “high-low” clusters appeared in some communities in the northern and central regions, which means that although these communities had a higher coverage ratio of green space, the green space accessibility was low. This may cause obstacles to the use of the elderly. Conclusions: This study innovatively uses geographically weighted regression, which embeds the spatial position of the data into the regression equation, to detect the spatial change of T2DM and its driving factors. And we speculate that green space has a mediating effect on the health level of the elderly. Finally, for the central urban area of a city with a high-density built environment like Shanghai, increasing large-scale green space is a huge challenge. What is feasible is that we can improve the supply of urban green space and the mediating effect by increasing the small green spaces in and around residential areas and improving the accessibility of green spaces.