The short-term effects of extreme weather such as hurricanes on cardiovascular disease (CVD) are well-docu- mented. However, the long-term CVD effects are unknown, likely leading to adverse outcomes, particularly for older, low socioeconomic status (SES), and functionally impaired residents, whose limited resources often mean staying in place. After Hurricane Katrina, the US Department of Health and Human Services launched the emPOWER Program to identify community-based, high-risk, functionally impaired Medicare beneficiaries who rely on essential medical services and durable medical equipment in order to improve pre-disaster evacu- ation planning and post-disaster management. One potentially significant limitation for both the general and high-risk emPOWER population is that their estimated disaster-related needs are based on short-term out- comes since we lack knowledge about longer-term consequences. The objective of this study is to receive mentored training to quantify the long-term adverse CVD outcomes associated with hurricanes. I hypothesize that the short-term worsening of CVD outcomes endures beyond the first year, with higher rates occurring in vulnerable groups that are yet to be rigorously defined. I will evaluate Hurricane Sandy’s impact on New York City (NYC) in 2012 because of the substantial physical damage wrought, its large population and neighbor- hood-level SES disparities, and the fact that no further hurricanes affected NYC. Building on disaster-related clinical expertise and research skills as a physician-scientist trained in health services research, I will learn and apply new spatial, epidemiological, and machine learning analytical skills to accomplish the following aims: 1) estimate the hurricane’s impact on the ≥ 65 years old population over the short- (1 year) and long-term (5 years) on census tract-level CVD prevalence using spatiotemporal modeling; 2) compare the hurricane’s short- and long-term impact on CVD event rates for current high-risk Medicare beneficiaries (emPOWER) compared to all other beneficiaries ≥ 65 years old using survival analyses and Cox models; and 3) develop and validate a CVD risk prediction tool that expands the definition of high-risk Medicare beneficiaries impacted by hurricanes beyond emPOWER using machine learning and risk prediction methods. In a future R01 application, I will ex- ternally validate the risk prediction model using 100% Medicare claims from the 2017 hurricane season (the costliest in US history). Together, this work will allow public health authorities to better tailor pre-disaster evac- uation and post-disaster CVD management that address disaster-related morbidity for high-risk groups into the longer term. I will be mentored and advised by a team of nationally recognized experts: Dr. Martin Shapiro (ca- reer mentor), Dr. Monika Safford (CVD and social epidemiology), Dr. David Abramson (disaster science and policy), Dr. Charles DiMaggio (spatial analysis), and Dr. Samprit...