School of Public Health

Catherine Knott

Longitudinal Association between Greenspace and Type 2 Diabetes: Evidence from MESA

This study aimed to evaluate the longitudinal association between the abundance of nearby greenspace in urban and suburban environments and incident type 2 diabetes mellitus (T2DM). Data was used from the Multi-Ethnic Study of Atherosclerosis (MESA). Normal Difference Vegetation Index (NDVI) was used as a proxy for greenspace. Exposure categories used NDVI median values from two time-periods (annual and high-vegetation season) and three buffer radii (500m, 1km, and 2.5km). Incident T2DM cases were identified from fasting glucose levels and medication review. Using Cox proportional hazards regression, we estimated hazard ratios (HRs) of incident T2DM associated with greenspace, adjusting for covariates including demographics, family history of T2DM, socioeconomic status, and neighborhood factors. The skewed distribution of NDVI values prompted the exclusion of participants from Forsyth County, NC. There were 659 incident T2DM cases (11.2%). For 1km exposure categories excluding NC participants, HRs for incidence T2DM were 0.95 (0.74 – 1.21), 0.78 (0.55 – 1.07), and 0.82 (0.55 – 1.24) for high-vegetation season median and 0.93 (0.72 – 1.20), 0.79 (0.58 – 1.07), and 0.72 (0.50 – 1.06) for annual median for increasing quartile of greenspace exposure compared to the lowest quartile. Overall, the results of this study were null, but showed a protective effect in point estimates with increasing greenspace exposure. The null results could be because there is no association between greenspace and T2DM or could be due to study limitations including insufficient statistical power, NDVI measurement error, or the validity of NDVI as a proxy for greenspace. More research is needed to give a better estimate of the effect of greenspace on T2DM.