This is the code and data for an R Shiny dashboard, which is associated with the manuscript 'Urban Heat Island Impacts on Heat-Related Cardiovascular Morbidity: A Time Series Analysis of Older Adults in US Metropolitan Areas'. The dashboard is an interactive supplemental material and allows for interaction with the manuscript's primary results - the heat-related cardiovascular risk and burden, in UHI-affected areas compared to areas unaffected by UHIs, in 120 contiguous US metropolitan statistical areas (MSAs), 2000-2017. The results can be viewed for the entire study population, different subpopulations, and in low and high urban heat island intensity (UHII) areas. It also allows for interaction with the MSA-level heat-related risk and burden, overall and in low and high UHII areas, and exploration of the ZIP code-level temperature, UHII, and Medicare hospitalization data used in the analyses. The dashboard displays a variety of interactive tables and figures and the results displayed in the dashboard can be downloaded. Details on the datasets and methods used as well as a discussion of all results and sensitivity analyses can be found in the associated manuscript. At present, the dashboard can be viewed here: https://shiny.stat.ncsu.edu/Heat-CVD-UHI-Dashboard/.
Many United States (US) cities are experiencing urban heat islands (UHIs) and climate change-driven temperature increases. Extreme heat increases cardiovascular disease (CVD) risk, yet little is known about how this association varies with UHI intensity (UHII) within and between cities. We aimed to identify the urban populations most at-risk of and burdened by heat-related CVD morbidity in UHI-affected areas compared to unaffected areas. ZIP code-level daily counts of CVD hospitalizations among Medicare enrollees, aged 65-114, were obtained for 120 US metropolitan statistical areas (MSAs) between 2000-2017. Mean ambient temperature exposure was estimated by interpolating daily weather station observations. ZIP codes were classified as low and high UHII using the first and fourth quartiles of an existing surface UHII metric, weighted to each have 25% of all CVD hospitalizations. MSA-specific associations between ambient temperature and CVD hospitalization were estimated using quasi-Poisson regression with distributed lag non-linear models and pooled via multivariate meta-analyses. Across the US, extreme heat (MSA-specific 99th percentile, on average 28.6 °C) increased the risk of CVD hospitalization by 1.5% (95% CI: 0.4%, 2.6%), with considerable variation among MSAs. Extreme heat-related CVD hospitalization risk in high UHII areas (2.4% [95% CI: 0.4%, 4.3%]) exceeded that in low UHII areas (1.0% [95% CI: -0.8%, 2.8%]), with upwards of a 10% difference in some MSAs. During the 18-year study period, there were an estimated 37,028 (95% CI: 35,741, 37,988) heat-attributable CVD admissions. High UHII areas accounted for 35% of the total heat-related CVD burden, while low UHII areas accounted for 4%. High UHII disproportionately impacted already heat-vulnerable populations; females, individuals aged 75-114, and those with chronic conditions living in high UHII areas experienced the largest heat-related CVD impacts. Overall, extreme heat increased cardiovascular morbidity risk and burden in older urban populations, with UHIs exacerbating these impacts among those with existing vulnerabilities.