Life expectancy (LE) is a crucial indicator of public health, and disparities in LE are often linked to sociodemographic factors and exposure to environmental hazards. In developed countries like the U.S., exposure to toxic chemicals from sources such as Superfund sites is a potential contributor to mortality and reduced LE. While previous research has explored the link between hazardous waste sites and specific health outcomes, comprehensive nationwide analyses on the impact of Superfund sites on LE at a fine-grained geographic scale are limited. This study addresses this gap by examining the association between the presence of Superfund sites and LE at the census tract level, considering both independent effects and interactions with sociodemographic factors. The study also explores how site characteristics like flooding potential and cleanup status modulate this association. The overarching goal is to provide a national-level assessment of the impact of Superfund sites on general health, using LE as a proxy, and to identify populations most vulnerable to this risk. The specific research questions focus on quantifying the independent and combined effects of Superfund site presence and sociodemographic factors on LE, examining how these effects vary across different sociodemographic groups, and exploring the influence of Superfund site characteristics, particularly flooding and cleanup status, on LE.

Existing literature reveals inconsistent findings regarding the health effects of proximity to hazardous waste sites. Some studies demonstrate a correlation between residential proximity to Superfund sites and increased mortality rates or specific health issues like Non-Hodgkin's lymphoma, particularly among males. Other studies report no significant association between proximity to these sites and maternal-fetal death or show varied results depending on factors such as remediation status. Studies examining benzene emissions from industrial sources in the European Union showed a significant positive correlation between exposure and mortality rates. Despite these findings, a comprehensive national-level analysis of the impact of Superfund sites on LE, particularly at a resolution finer than the county level, has been lacking. This research aims to fill this gap by investigating the relationship at the census tract level, encompassing a broader range of sites beyond those on the National Priorities List (NPL).

This study used a nationwide geocoded statistical modeling approach to assess the relationship between Superfund sites and life expectancy (LE). Data were obtained from multiple sources including: 1) the National Historical Geographic Information System (NHGIS) for census data (demographics, income, education, etc.); 2) the National Center for Health Statistics (NCHS) database for LE data (2010-2015); and 3) the EPA Superfund Enterprise Management System database for information on Superfund sites (including their location, NPL status, and cleanup status). Flood hazard data were obtained from the Federal Emergency Management Agency (FEMA). A total of 11,989 uniquely identified Superfund sites in the contiguous U.S. were included. A 322-meter buffer was created around each site, and spatial join analysis was used to link Superfund site information to census tracts. The study employed a matching technique, comparing life expectancy in census tracts with Superfund sites to the median LE of their nearest neighboring tracts without Superfund sites. The Mann-Whitney U test was used to evaluate differences in LE and other sociodemographic variables between these groups. To quantify the impact of Superfund sites, ordinary least squares (OLS) regression models were developed, incorporating both the presence of Superfund sites (binary variable) and multiple sociodemographic variables to account for confounding effects. The performance of the OLS model was compared to a Random Forests machine learning algorithm. Quantile regression was used to analyze the effect of Superfund site presence and sociodemographic variables on different LE percentiles. Finally, effect modification analyses explored how the impact of Superfund site presence varied across subgroups defined by sociodemographic factors and across Superfund sites with different characteristics (NPL status, flooding potential, cleanup status).

The analysis revealed a statistically significant difference in LE between census tracts with at least one Superfund site and their neighboring tracts without sites. The presence of a Superfund site was associated with a decrease of -0.186 ± 0.027 years in LE, independent of sociodemographic factors. This negative effect, however, was significantly amplified in tracts with higher levels of sociodemographic disadvantage. For instance, in tracts with income below the national median, the LE reduction could reach -0.58 years, and in tracts in the lowest 10% income percentile, it could be as high as -1.223 years. Higher income, on the other hand, showed a buffering effect, potentially offsetting the negative impact of Superfund sites. Further, characteristics of Superfund sites, including flooding potential and lack of cleanup, amplified their adverse effects on LE. Tracts with Superfund sites prone to flooding showed a substantially larger decrease in LE (-0.36 years) compared to tracts with sites at minimal flood risk (-0.034 years). The NPL status of the sites was also influential; sites not on the NPL showed a more pronounced negative impact on LE than NPL sites, possibly reflecting the effects of EPA-led remediation and redevelopment efforts. While the Random Forests model showed slightly better performance in predicting LE compared to the OLS model, both models yielded similar overall conclusions. The quantile regression analysis indicated that the effects of both sociodemographic variables and Superfund sites were more pronounced in census tracts with lower LE.

This study provides compelling evidence supporting the association between the presence of Superfund sites and reduced LE in the U.S., even after accounting for sociodemographic factors. The findings underscore the disproportionate impact of Superfund sites on vulnerable populations with greater sociodemographic disadvantage. The amplified negative effects associated with site characteristics like flooding potential and lack of cleanup highlight the need for comprehensive remediation strategies and proactive mitigation efforts to protect public health. The study's findings reinforce the importance of considering environmental justice issues when addressing the legacy of pollution from hazardous waste sites. The results suggest the need for targeted interventions in communities disproportionately burdened by environmental hazards and sociodemographic disadvantage.

This study demonstrates a significant negative association between the presence of Superfund sites and life expectancy in the U.S., particularly within socioeconomically disadvantaged communities and sites with specific characteristics. The findings highlight the need for further research investigating the long-term health impacts of exposure to Superfund sites, considering population mobility and site history. Further investigation into residual confounding effects and more detailed site-specific analyses are recommended to refine risk assessments and inform effective remediation strategies.

This study's cross-sectional design limits the ability to establish definitive causality. The assumption of uniform exposure to Superfund sites, regardless of specific contaminants or exposure pathways, may oversimplify the complex relationship between site characteristics and health outcomes. Furthermore, the study relies on publicly available data, which may contain measurement error or incomplete information. Future research should incorporate a longitudinal design and more detailed data on exposure pathways to better elucidate the causal relationships involved.