Repository logo

Browsing by Author "Rojas-Rueda, David, advisor"

Now showing 1 - 2 of 2
  • Thumbnail Image
    ItemOpen Access
    Colorado EnviroScreen as a predictor of mortality: an ecological analysis of 2019 county-level data
    (Colorado State University. Libraries, 2024) Pusker, Stephanie, author; Rojas-Rueda, David, advisor; Clark, Maggie, advisor; Gutilla, Molly, committee member
    Background In today's rapidly evolving landscape of environmental awareness in public health, Colorado stands at the forefront of innovation with its Environmental Justice (EJ) mapping and health screening tool, Colorado EnviroScreen (Colorado EnviroScreen, 2022). This tool, developed by Colorado Department of Public Health and Environment (CDPHE), empowers governmental agencies, research institutions, and the broader public to quantify and understand the interplay between environmental factors and community health by calculating an "EnviroScreen Score" (Colorado EnviroScreen, 2022). The higher the EnviroScreen score, the more likely the area will be affected by environmental health injustices at the census block group, census tract, and/or county levels (Colorado EnviroScreen, 2022). The purpose of this study is to bridge a gap in the current research landscape by exploring the association between aggregate county-level data derived from an EJ mapping tool and all-cause mortality rates. Specifically, we aim to investigate the relationship between the CO EnviroScreen score and the component scores – Demographics, Sensitive Populations, Climate Vulnerability, Environmental Effects, and Environmental Exposures – and all-cause mortality rates at the county level in Colorado in 2019. By conducting this ecological analysis, we seek to provide valuable insights into the potential impact of environmental justice factors on public health outcomes, thereby contributing to a more comprehensive understanding of the interaction between environmental conditions and mortality rates within communities. Methods An ecological study was conducted at the county-level spatial scale using a generalized linear model to assess the association between three EnviroScreen component score percentiles (Demographics, Environmental Exposures, and Climate Vulnerability) and age-standardized all-cause mortality at the county level. These three score percentiles were selected due to correlation with other scores, as well as the indicators included in some of the component scores being more comprehensive than others. County-level covariates included in the model were insufficient sleep, alcohol overindulgence, physical inactivity, and smoking. In addition to the full model, secondary models were created, including Demographics, Environmental Exposures, and Climate Vulnerability as independent predictors. Furthermore, the total EnviroScreen score percentile, which includes all component scores, was used in the analysis. Results In the fully adjusted model, a 10% increase in the EnviroScreen Environmental Exposures component score was associated with a 3% increase in all-cause mortality rate at the county level in Colorado in 2019 (95% CI: 1.00, 1.05). In the crude model, a 10% increase in EnviroScreen score was associated with a 5% increase in all-cause mortality rate at the county level in Colorado in 2019 (95% CI: 1.03, 1.07). Neither Demographics nor Climate Vulnerability component scores percentile were associated with an increase or decrease in all-cause mortality rates. Discussion This study suggests that there is a potential association between a higher EnviroScreen component score and an increase in age-standardized, all-cause mortality at the county level in Colorado. This is the first study to estimate the association between aggregate environmental and health-related scores from CO EnviroScreen with mortality. This study supports the notion of cumulative impacts as a tool to monitor possible health disparities and environmental injustice.
  • Thumbnail Image
    ItemEmbargo
    Quantitative health impact assessments as a tool for exploring public health dimensions of environmental exposures
    (Colorado State University. Libraries, 2024) Dean, Daniel, author; Rojas-Rueda, David, advisor; Anderson, G. Brooke, advisor; Peel, Jennifer, committee member; Hurrell, James, committee member
    Public health is influenced by a population's built and natural environment in both negative (e.g., natural disasters or ongoing stress from heat) and positive (for instance, heat-moderating effects of vegetation) ways, as well as interactively with behavioral and social dynamics. One framing of policy priorities and urban resilience is a "triad" consisting of exposure reduction (limiting the extent to which community members are exposed to environmental hazards—including "ambient" ones like stressful temperatures), vulnerability reduction (mitigating the impacts of sustained hazards), and hazard reduction (actively reducing the frequency or intensity of hazards) (Hoegh-Guldberg et al. 2018). Because any such measures carry tradeoffs in financial and other resources, it is important that policymakers and other stakeholders weigh comparative benefits of potential environmental hazards or interventions with consistent, quantifiable metrics. In this body of work, we applied quantitative health impact assessments, an epidemiology framework that provides a valuable tool here, allowing researchers to project health outcome changes for a population of interest given predicted changes in a relevant exposure and using epidemiological evidence, including exposure-response functions (exposure-response functions), which link exposure and health outcomes. In this body of work, we use HIAs to explore three different resilience-relevant systems spanning a range of intervention types, environmental systems, and spatiotemporal scales: Project 1: Health Impacts of Future Tropical Cyclones in the Eastern United States: While tropical cyclones are among the most damaging natural disasters faced by the United States, the temporal and spatial rarity of these events impedes traditional frequency-based estimates for public health and related risk projections, leading to potential oversights in risk characterization. In addition, mortality associated with tropical cyclones may not be readily apparent between delayed onset and indirect causes (e.g. stress, disrupted medical care, infections), meaning that immediate mortality counts often underestimate full attributable mortality. In this project, we performed a pilot quantitative health impact assessment designed to address aspects of these limitations. First, we tested extending the historical tropical storm dataset using a pool of 10,000 simulated, or "synthetic" tropical cyclone seasons from the widely used and open-source STORM algorithm, trained from and intended to represent the "gold standard" of historical International Best Tracks Archive for Climate Stewardship (IBTrACS) data. To the extent that STORM represents real-world conditions, this vastly expanded 'sample population provided information on potential tropical cyclone exposure risk than would be possible from historical data alone. For the second challenge of accounting for delayed and indirect attributable mortality, we combined the synthetic data with a recently-developed exposure-response function: an integrated Bayesian causal-predictive model trained on Medicare Claims data (simplified to Americans aged 65 and older), featuring an integrated model approach to combine a whole-population causal inference model for central trends with county-specific predictive models for give county-specific estimates. This model also tracked up to 21 days of "lag time" in health outcomes after a TC exposure to capture delayed mortality. This combination of methodologies promises a comprehensive, county level picture of tropical cyclone-associated all-cause mortality risk among older adults. This approach provided insights including regions of the country at the greatest risk for tropical cyclone-related exposures among older adults. However, as our study represented a new application of the STORM algorithm (in particular, our emphasis focusing on post-landfall behavior of tropical cyclones), we also assessed the level of agreement between STORM and the historical dataset, finding some discrepancies including lower overall frequency, and considerably 'smoother' spatial distribution in exposures; some discrepancies were in line with previously noted limitations. This project used recent innovations in atmospheric science and epidemiology modeling to explore the utility of a quantitative health impact assessment framework for present-day risk and could inform policy and planning decisions in terms of tropical cyclone preparedness and response measures. Project 2: Health impacts of Urban Tree Canopy policy scenarios in Denver and Phoenix: We explored potential health impacts (in terms of all-cause mortality, stroke, and dementia) of standing policy goals in Denver, Colorado and Phoenix, Arizona, for increasing the urban tree canopy coverage in these relatively arid cities. We projected health benefits (in terms of reduced attributable all-cause mortality, stroke, and dementia incidence) at a census block group level using several existing exposure-response functions based on the widely used Normalized Difference Vegetation Index (NDVI). Because the cities expressed policy goals in terms of percentage urban tree canopy, we generated predictive models to "translate" between this metric used in policy goals and the NDVI metric. We modeled the public health impacts of proposed real-world policies for near-future policy interventions in the form of increasing urban tree canopy, using current populations, and modeling an "overnight" change in exposure, with policy scenario benefits modeled for populations in year 2020, rather than demographic projections for the 2030 (Phoenix) and 2050 (Denver) dates in the policy goal timelines. We also considered socioeconomic dimensions by using the census-based Social Vulnerability Index to trace the equity of current UTC and NDVI exposures, as well as of potential benefits. We determined that each city could, by reaching its standing policy goals, could avert hundreds of all-cause mortality cases, with even a partial attainment scenario (halfway between current and desired UTC levels) having appreciable benefits, with roughly half the captured mortality prevention; with respect to equity of UTC access, more-vulnerable communities in the cities saw lower access to current canopy cover, and consequently greater potential per-capita benefits under successful intervention scenarios. Project 3: Health Impacts of Future Temperature Extremes Under a Solar Climate Intervention Scenario. In this project, we explored potential all-cause mortality implications of a proposed climate intervention effort intended to counteract anthropogenic warming, modeling the years 2050-2060 under alternate climate scenarios. Specifically, we projected temperature-associated mortality under a stratospheric aerosol injection (SAI) intervention scenario, as well as a corresponding scenario of "middle-of-the-road" climate change. We used a study population of 65-and-older Americans in eight major US cities (Seattle, Chicago, New York, Philadelphia, Los Angeles, Phoenix, Houston, and Miami) spanning a range of local climates. We built our analysis on widely used models and the shared socioeconomic pathway platform, allowing our two scenarios to be compatible, differing only in the SAI intervention itself. We focused on two age groups (65-75, and 75+) to reflect elevated heat- and cold-associated mortality risks among this population, finding broadly similar trends between age groups. We explored city-specific exposure-response functions for the temperature-mortality association, using a widely used modeling, comparing the anticipated number of cold- and heat-related deaths under each scenario, and highlighted tradeoffs for either policy scenario, finding considerable heterogeneity in trends between these cities. To make our analysis more specific to the mid-21st century, we incorporated existing estimates for population growth and mortality rate changes based on the same climate modeling scenarios as the SAU exposure scenarios. We observed dramatic variability in minimum mortality temperatures and temperature-attributable mortality between cities and found that SAI was not associated with decisive reductions in all-cause mortality among either age group. While SAI did effectively reduce heat-attributable mortality, lower cold-attributable mortality under the warmer, non-SAI scenario counterbalanced this effect, yielding a weak net impact in central tendencies. This observation could help inform planning and resilience efforts as far as types of temperature-related stress under each scenario, as well as provide insights for larger cost-benefit analyses for the overall proposition of SAI. Together, these projects demonstrated how quantitative health impact assessments can help form a methodological foundation for exploring epidemiology and resilience-relevant systems. The variety of projects covered demonstrated the utility of this methodology in a variety of spatial scales, ranging from census block groups (comparable to neighborhoods) in Project 2 to county-level characterizations of tropical cyclone-associated risk for much of the eastern United States in Project 1. We also explored a range of time periods, ranging from Project 1's focus on characterizing tropical cyclone risk representative of the past several decades (as represented by the STORM resampling algorithm), through our attempts to explicitly model mid-21st century populations and temperature-related mortality trends using both climate and demographic projections. The modularity of the quantitative health impact assessment framework enabled our projects to leverage of existing research and datasets for low-cost, comparatively rapid assessments, as well as to lay infrastructure for future research and introduce several specific innovations in their respective designs.