My thesis mentor was Dr. Timothy Rebbeck at the Harvard T. H. Chan School of Public Health and Dana Farber Cancer Institute. I conducted studies to evaluate the association between natural green neighborhood vegetation (neighborhood greenness) and prostate cancer risk and mortality. In conducting these studies, I also worked with the SpaCE2 (Spatial and Contextual Exposomics and Epidemiology) Lab led by Dr. Francine Laden at 401 Park Drive Our group uses geographic information systems tools, including ArcMap, QGIS, and Google Earth Engine to study neighborhood exposures and a variety of health outcomes.
Using data from the nation-wide American Health Professionals Follow-up Study Cohort, we found no statistically significant association between neighborhood greenness and risk of lethal prostate cancer over 28 years in the total study population. However, when restricting analysis to just men who did not move, we found evidence of modest inverse associations (8% lower risk) with increasing levels of neighborhood greenness. When stratifying by urbanicity, men living in high population density areas experienced 10% lower incidence of lethal prostate cancer over follow-up, but this was not observed among men living in low population density areas. Our study also provided the unique opportunity to compare associations among men for whom greenness was assessed at residdential address compared to work address. We found stronger inverse associations among those for whom neighborhood greenness reflected exposure at work. These findings support an earlier case-control study conducted in a major Canadian city, but raise questions about measurement of exposure and mechanisms. More details about the study are provided in our article published in Environmental Epidemiology.
Racial disparites in mortality among men with prostate cancer are among the largest for any major cancer, and have persisted despite increased access to screening and advances in treatment. We sought to understand whether neighborhood greenness, which has been associated with lower mortality in many large prospective cohort studies, could partially explain these differences. Using data from the population-based Pennsylvania Cancer Registry, we estimated the association between neighborhood greenness and cause-specific mortality (all-cause, prostate, and cardiovascular). We evaluated effect modification by race, and estimated the proportion of disparity that could be eliminated following a hypothetical intervention to fix neighborhood greenness to the level experienced by white men with high levels of neighborhood greenness. Our analysis revealed significant inverse associations between neighborhood greenness and all mortality endpoints, particularly for cardiovascular disease (18% lower rate, 95% CI: 10%, 26% comparing quintile 5 to 1). However, this inverse association was only observed among White but not Black men. When estimating the proportion of disparity that could be eliminated following a hypothetical intervention to increase greenness for all men with prostate cancer, we found non-significant relative reductions in all-cause (-5.3%) and prostate (-23.2%) but not cardiovascular mortality, due to the fact that all gains were experienced by White men but not Black men. These findings raise the importance of crafting policies to increase greenness to specific communities to ensure that benefits are realized. More details can be found in the Environmental Epidemiology publication.