Does socioeconomic and environmental burden affect vulnerability to extreme air pollution and heat? A case-crossover study of mortality in California

Climate change poses a significant threat to public health, increasing the frequency and intensity of heat waves and exacerbating air pollution. Extensive research demonstrates the link between high temperatures and increased mortality from cardiovascular and respiratory diseases. Extreme heat can directly cause death, but also worsens pre-existing conditions. Similarly, fine particulate matter (PM2.5) pollution, also impacted by climate change, is associated with various health problems, including respiratory and cardiovascular diseases and mortality. Wildfires, increasing in frequency, are a major source of PM2.5 pollution. While studies have examined the individual effects of heat and air pollution on mortality, fewer have explored their combined impact. Previous research by the authors showed that the co-occurrence of extreme heat and PM2.5 resulted in higher mortality risk than either exposure alone. The relationship between heat and air pollution exposure, socioeconomic status (SES), and adverse health outcomes is complex. Evidence suggests that low-SES communities experience a disproportionately higher health burden due to increased exposure to and vulnerability to both heat and air pollution. This vulnerability is often compounded by other environmental co-exposures, such as proximity to major roads (leading to air and noise pollution) and lack of green spaces (increasing heat vulnerability and air pollution exposure). This complex interplay highlights the importance of understanding how 'socioenvironmental' burden influences vulnerability to environmental exposures. Low-SES communities, often referred to as environmental justice (EJ) communities, face increased environmental exposures and decreased protection due to their economic circumstances. Understanding the impact of historical socioeconomic marginalization on mortality risks from extreme heat and air pollution is crucial given the projected increases in these exposures with climate change. This study aimed to evaluate whether neighborhood socioeconomic or socioenvironmental burden modifies the mortality association with co-exposure to extreme heat and PM2.5, expanding upon previous work by the authors. The study investigated the impact of exposure to extreme heat and PM2.5 alone and in combination, analyzing effect measure modification for all-cause, cardiovascular, and respiratory mortality based on neighborhood socioenvironmental and socioeconomic burden, as well as individual vulnerability.

The literature reveals inconsistent findings regarding the effect modification by socioeconomic factors on the health effects of heat and air pollution. Some studies using individual-level data show significant, albeit inconsistent, results, while studies using area-level SES data often find no significant modifying effects. A systematic review by Laurent et al. (2007) highlighted the importance of the resolution of the socioeconomic variable, suggesting that studies with individual-level data were more likely to show significant effects than those using only area-level data. More recent studies using area-level measures have found significant effect measure modification, showing that areas with higher social deprivation may be more vulnerable to the cardiovascular mortality effects of PM2.5. Another study investigated the association between wildfire-related PM2.5 and out-of-hospital cardiac arrests, finding higher effect estimates in low SES populations. Studies on effect measure modification by SES of heat-related mortality also show mixed results, with some indicating larger effect estimates in lower income neighborhoods and others showing no such effect. The inconsistencies in these findings underscore the need for further research.

This study utilized a time-stratified case-crossover design to analyze mortality data in California from 2014 to 2019, obtained from the California Department of Public Health. After excluding deaths with missing residential information or residing outside of California, the final sample included 1,514,292 deaths. All-cause, cardiovascular, and respiratory mortality were defined using ICD-10 codes. Demographic information was extracted from death certificates. Exposure assessment involved obtaining daily PM2.5 concentrations from the US Environmental Protection Agency's Air Quality System and daily maximum temperatures from a gridded reanalysis dataset. Exposures were assigned to decedents based on their residential census tracts. Extreme exposure was defined using the 90th, 95th, or 97th percentile thresholds for heat and PM2.5. Exposure was parameterized as days with extreme PM2.5 only, extreme heat only, both, or neither. Socioenvironmental burden was assessed using the CalEnviroScreen (CES) score, a measure of socioeconomic and cumulative environmental burden. Socioeconomic burden was assessed with a Social Deprivation Index (SDI). Individual educational attainment was obtained from death certificates. Area-level burden measures were linked to decedents based on their residential census tracts. Conditional logistic regression was used to estimate associations between extreme heat and PM2.5 exposure (alone and combined) and mortality, with product terms used to evaluate effect measure modification by CES and SDI. Analyses were conducted for all-cause, cardiovascular, and respiratory mortality, using various exposure percentile thresholds. Secondary analyses included individual educational attainment as an effect modifier and lag-1 exposure assessment. Sensitivity analyses assessed effect measure modification using alternative burden metrics and continuous exposure data.

The study included 1,514,292 all-cause deaths, with 33% cardiovascular and 9% respiratory. Differences in sociodemographic variables were similar for those with high CES and high SDI scores compared to low scores. The mean age at death was 74.0 years, with lower ages in high CES/SDI tracts. Higher CES/SDI tracts had proportionally more Black and Hispanic populations and fewer White populations. Lower CES/SDI was associated with higher educational attainment. Extreme heat and air pollution, alone and combined, were significantly associated with increased mortality. Combined extreme heat and PM2.5 exposure consistently showed stronger mortality odds ratios than either extreme exposure alone. However, no consistent pattern of increasing or decreasing mortality effect estimates of combined heat and PM2.5 was observed across CES or SDI strata. Similar null modifying effects were seen for extreme heat and PM2.5 exposures separately and for cause-specific cardiovascular and respiratory mortality. Secondary and sensitivity analyses generally yielded null results, showing no significant effect measure modification by individual-level educational attainment, lag-1 exposure, alternative burden metrics (socioeconomic and environmental burden components of CES, HPI), or continuous exposure analysis. Increased daily PM2.5 exposure was observed with increasing CES and SDI, suggesting higher exposure in areas with higher socioenvironmental and socioeconomic burden.

This study found no consistent evidence that neighborhood socioeconomic or socioenvironmental burden modified the mortality risk associated with extreme heat and PM2.5 exposure, either individually or in combination. This contrasts with some previous research showing higher vulnerability in low-SES areas, but aligns with studies that utilized similar area-level metrics. While increased PM2.5 exposure was observed in neighborhoods with high socioenvironmental and socioeconomic burden, this higher exposure didn't translate into a greater mortality risk for this group. The lack of effect modification may be due to several factors, including potential limitations in the exposure assessment (residential census tract instead of personal exposure) and the inability to account for factors like air conditioning use. The study highlights the complexity of the relationship between socioeconomic factors, environmental exposures, and health outcomes and emphasizes the need for further investigation into the specific mechanisms through which socioeconomic factors might influence the impact of extreme heat and PM2.5 on mortality.

This large-scale case-crossover study found no consistent effect modification by socioeconomic or socioenvironmental burden on the mortality risks associated with extreme heat and PM2.5 exposure. While higher exposure to extreme PM2.5 was observed in disadvantaged areas, no evidence of heightened mortality risk was detected for those communities. The study's null findings highlight the complexity of interactions between environmental exposures, socioeconomic factors and mortality. Future research should investigate the factors contributing to observed disparities in PM2.5 exposure. Utilizing more precise individual exposure data and advanced statistical techniques may reveal more complex relationships.

The study's limitations include the use of residential census tract data for exposure assessment, which may introduce exposure measurement error; the lack of data on air conditioning use, a significant factor in heat vulnerability; exclusion of individuals lacking residential information, potentially biasing results towards underrepresentation of the most vulnerable groups; and the use of a possibly crude measure of individual SES. The study's power to detect effect measure modification may also be limited by focusing only on extreme exposure events. The study also notes that its findings may have limited generalizability beyond California's diverse climate and population.