Skip to main content
NCBI home page
GeoHealth logoLink to GeoHealth
. 2025 Oct 15;9(10):e2024GH001099. doi: 10.1029/2024GH001099

Socioeconomic Disparities of Asthma Incidence Attributable to PM2.5 Exposures for Schoolchildren in California

Hyung Joo Lee 1,2,✉, Keita Ebisu 3, Hye‐Youn Park 4
PMCID: PMC12521950  PMID: 41104190

Abstract

This study investigated the socioeconomic disparities of asthma incidence attributable to ambient particulate matter in aerodynamic diameter ≤2.5 μm (PM2.5) exposures among schoolchildren in California, U.S. We found that schoolchildren attending public schools in more vulnerable communities, characterized by higher proportions of people of color, low educational attainment, and poverty, experienced elevated PM2.5 exposures by 2.07–2.96 μg/m3. The disproportionate PM2.5 exposures were likely driven by higher traffic‐related emissions and point‐source facility emissions in these communities. Using school‐specific PM2.5 concentrations, student enrollment numbers, and model‐estimated (not directly observed) baseline age‐specific asthma incidence rates, we calculated that the asthma incidence rate attributable to 2016 PM2.5 exposures was 562 new cases per 100,000 schoolchildren [95% confidence interval (CI) = 311–854]. In absolute terms (i.e., asthma incidence), it was equivalent to 34,537 PM2.5‐related new asthma cases (95% CI = 19,090–52,493) among all schoolchildren. On average, more vulnerable communities experienced 140 excess new asthma cases per 100,000 schoolchildren (i.e., the difference in average asthma cases per 100,000 schoolchildren between more and less vulnerable groups) across all demographic factors considered. Examining health disparities separately by each demographic factor revealed that race/ethnicity was associated with the largest disparities (209 new cases per 100,000 schoolchildren), followed by educational attainment (128) and poverty (85). Our findings indicate the substantial socioeconomic disparities of asthma incidence attributable to PM2.5 among schoolchildren in California. Addressing these health disparities could benefit from sustained and long‐term emission reduction strategies, such as adopting zero‐emission vehicles, which contribute to lower PM2.5 levels.

Keywords: air quality, asthma incidence, PM2.5 , environmental justice, health impact assessment, schoolchildren, socioeconomic status

Plain Language Summary

Asthma in children is a significant public health issue that can affect their health throughout their lives. Understanding and measuring the risk factors for developing asthma are essential to create targeted policy interventions. This study investigated the number of new asthma cases that were linked to exposure to fine particles among schoolchildren in California, U.S. We found that, in 1 year, 34,537 new cases of asthma in schoolchildren could be traced back to fine particle exposure. This study also examined how asthma, caused by fine particle exposure, affected schoolchildren differently based on their community's level of social vulnerability. Particularly in terms of race and ethnicity, our findings showed that communities with higher social vulnerability observed more new asthma cases among schoolchildren (209 additional cases per 100,000 schoolchildren). Because the differences in fine particle levels between more and less vulnerable communities are mainly due to traffic emissions, stricter regulations on vehicle emissions could help lower the number of new asthma cases in schoolchildren and address the unequal health impacts across different communities.

Key Points

  • Schoolchildren in socially vulnerable communities were disproportionately exposed to ambient PM2.5 air pollution

  • The asthma incidence attributable to PM2.5 was 34,537 new cases for schoolchildren

  • The racial disparities of asthma incidence were pronounced by 209 new cases per 100,000 schoolchildren

1. Introduction

Children's asthma is a serious public health concern in the U.S. with approximately 6 million asthmatic children between the ages of 0 and 17 years old (U.S. Centers for Disease Control and Prevention (CDC), 2018). Environmental factors, such as air pollution, have been suggested as risk factors for asthma (Ni et al., 2024; Orellano et al., 2017; Zheng et al., 2015). Health effect studies have reported associations between population PM2.5 exposures and exacerbated asthma symptoms (Mirabelli et al., 2016; Williams et al., 2019). Recent studies have further focused on children's PM2.5 exposures and asthma development and symptoms (McConnell et al., 2010; Strickland et al., 2016; Tétreault et al., 2016; Tsai et al., 2025). Children have developing lungs and immune systems and also high activity levels and air intake, making them more vulnerable to respiratory diseases associated with air pollution exposures than adults (Trasande & Thurston, 2005). According to the U.S. EPA's 2019 Integrated Science Assessment (ISA), there is an “adequate” level of evidence to indicate that children have an increased risk for health impacts from PM2.5 (U.S. EPA, 2019).

PM2.5 exposures have disproportionately affected more vulnerable populations such as people of color, those with low educational attainment, and those living in poverty (Bell & Ebisu, 2012; Colmer et al., 2020; Hajat et al., 2013; Lee, 2019; Nair et al., 2023; Tessum et al., 2021). In California, the PM2.5 disparity is largely caused by on‐road traffic emissions, which are further related to higher road density and traffic volumes in more vulnerable communities (Lee & Park, 2020). The inequities of PM2.5 emissions and thus exposure to PM2.5 air pollution are likely to contribute to the disparities of health outcomes (e.g., asthma) for schoolchildren as well as entire populations. Despite the role of childhood respiratory functions as a life‐long health determinant (McGeachie et al., 2016), the total health disparities of PM2.5 air pollution on asthma for schoolchildren have not been well characterized. Unlike epidemiological studies that report the association between PM2.5 exposures and asthma onsets in children, providing concentration‐response functions (CRFs), health impact assessments quantify the total health impacts among schoolchildren, which help determine policy directions and assign regulatory resources to specific areas. Therefore, health impact assessments focusing on children's asthma attributable to PM2.5 exposures are urgently needed in the context of environmental justice.

This study investigated the long‐term ambient PM2.5 exposures and health disparities among schoolchildren in California, U.S. The source types related to exposure disparities were examined using PM2.5 emission inventories and proxies. We quantified the health impacts of PM2.5 exposures on children's asthma incidence, focusing on the disparity of asthma health outcomes. To our knowledge, this is the first study to investigate PM2.5 exposures and disparities of schoolchildren's asthma incidence through a comprehensive health impact assessment, emphasizing PM2.5‐related disparities of asthma onset. Unlike previous research that typically targets general populations (Hajat et al., 2013; Ma et al., 2023), our study specifically focused on individual schools and schoolchildren attending them, bridging this knowledge gap. This study can aid in revealing if and how the inequities of PM2.5 air pollution exposures contribute to disproportionate health impacts for schoolchildren in California. Furthermore, our school‐specific analysis enables the development of targeted interventions to mitigate the health disparities.

2. Methods

2.1. Location and Enrollment of Public Schools

We identified the locations of public schools [from kindergarten to 12th grade; K‐12 (ages 5–18)] provided by the California School Campus Database (CSCD) for the year 2016 (California School Campus Database, 2016). The initial CSCD was created for 2016 and subsequently updated for 2018 and 2021. In the CSCD, the point locations and boundaries of each school were mapped using assessor parcel data and aerial imagery. The CSCD data did not include private schools. The number of schoolchildren enrolled in each public school was obtained from the California Department of Education (DOE) (California Department of Education, 2021a). The public school enrollment accounted for 92.6% of the total K‐12 enrollment (the remaining 7.4% from private schools with six or more students) for the years 2015–2016 (California Department of Education, 2021b). Only primary enrollment (students who were enrolled during the entire academic year) was included in this study. Adults enrolled in public schools and schools solely for adult education were not counted. In total, 9,811 public schools and 6,146,236 students in California were included for further analyses.

2.2. Satellite‐Based PM2.5 Concentrations

We used previously published ambient PM2.5 concentrations that were estimated by satellite aerosol optical depth (AOD) from the Multi‐Angle Implementation of Atmospheric Correction (MAIAC) algorithms (1 km resolution), land use parameters, and meteorology in California for the year 2016 (Lee, 2019). Briefly summarizing Lee (2019), a mixed effects model was employed to estimate PM2.5 concentrations. The 10‐fold cross validation (CV) analysis for annual average PM2.5 showed the R 2 of 0.73 and 0.81, mean absolute errors (MAE) of 1.12 and 0.93 μg/m3, and root mean squared errors (RMSE) of 1.40 and 1.17 μg/m3 for site‐based and observation‐based CVs, respectively. The mixed effects model has demonstrated high predictive power in estimating PM2.5 from satellite AOD values (de Hoogh et al., 2018; Lee, 2020; Lee, Liu, et al., 2011; Ma et al., 2016; Xie et al., 2015). Nonetheless, the PM2.5 estimates inevitably included modeling errors, introducing uncertainties in the exposure assessment.

For schoolchildren's PM2.5 exposures, school‐specific PM2.5 levels were determined by average PM2.5 concentrations estimated in the grid cells that contained the point locations of the schools. Because these PM2.5 data were 2016 annual average concentrations, we were not able to consider school‐year PM2.5 averages. We assumed that the exposure misclassification caused by the temporal mismatch was minimized as schoolchildren tended to live close to their schools. We targeted the year 2016 for our analyses because both the high‐quality PM2.5 data and the CSCD were available exclusively for that year. In addition, the atypical PM2.5 levels influenced by the COVID‐19 pandemic (Berman & Ebisu, 2020) made analyzing recent years inappropriate for assessing “business‐as‐usual” PM2.5 concentrations and their associated health impacts.

2.3. Social Vulnerability

To evaluate the extent of social vulnerability, we obtained block‐group level data on social determinants of health, such as demographic factors including people of color, poverty, and low educational attainment, from the American Community Survey (ACS) 2012–2016 5‐year estimates (U.S. Census; block‐group level) (U.S. Census Bureau, 2021). These demographic factors have been widely used to represent multiple aspects of social vulnerability, encompassing demographics (e.g., historical segregation), economic conditions, and social status, which makes the collective use of the factors reasonably comprehensive (Colmer et al., 2020; Jbaily et al., 2022; Lane et al., 2022). Thus, additional factors or indices would likely offer limited insights into our assessment of social vulnerability. Following the definitions used in Lee (2019), the percentage of people of color was defined as [(total population − non‐Hispanic White)/total population] × 100. The percentage of poverty was calculated as [(population with the ratio of income to federal poverty level in the past 12 months < 2)/population with poverty status] × 100. Using the ratio of income to poverty level <2 (i.e., income less than 2 times of federal poverty level) rather than <1 was to reflect comparatively high living expenses in California (Cushing et al., 2018). Finally, the percentage of low educational attainment was calculated as (populations aged 25 or more without a high school diploma/populations aged 25 or more) × 100. These block‐group level data (proxies for “neighborhood” or “community”) were paired with school‐specific average PM2.5 when the schools were located within the block groups. For schoolchildren attending public schools, we supposed that the schools were located near their homes, leading to similar demographic factors for the locations of the homes and the schools. This rationale was based on California's school zoning policies, which manage the distribution of schoolchildren among schools, allocate appropriate educational resources, and ensure that schoolchildren have access to education within a reasonable distance from their homes (California Legislative Information, 1987). Furthermore, He and Giuliano (2018) reported that the median distance from homes to schools was 0.66 miles (1.06 km) for K–6th grades and 1.41 miles (2.27 km) for 7th–12th grades in the Los Angeles region, California (He & Giuliano, 2018). This finding supported our assumption of close proximity between homes and schools although this analysis did not comprehensively represent the entire California. Collectively, we assumed that the demographic factors in each block group were reasonably applicable to all the public schools located in the respective block group.

2.4. PM2.5 Emissions

To investigate the source types of PM2.5 emissions at the schools, we used data on emission inventory and proxies. Considering that traffic emissions comprise a large fraction of total PM2.5 concentrations (Hasheminassab et al., 2014; Lee, Gent, et al., 2011; Squizzato et al., 2018; Zhu et al., 2024), the distance from each school to limited access highways (Environmental Systems Research Institute; ESRI) was calculated, and school‐specific traffic volumes (AADT, California Department of Transportation (CalTrans)) (California Department of Transportation (CalTrans), 2021) were obtained from the nearest traffic counter. The average distance between the schools and the nearest traffic counters was 1.9 (SD = 2.1) km. Although the traffic counters were not densely placed, AADT observed from these counters was able to effectively explain traffic‐related air pollutants in California (Lee et al., 2023). This is likely because (a) traffic volumes observed on arterial roads reflect those on the segments of the same roads closer to schools that are not monitored and (b) the traffic flows on arterial roads are indicative of local traffic dynamics near schools, as they are connected within the road network. Additionally, we used 2016 facility‐level PM2.5 emissions from the California Emissions Inventory Development and Reporting System (CEIDARS) database of the California Air Resources Board (CARB) (California Air Resources Board (CARB), 2019).

2.5. Health Impact Assessment

We assessed the health impacts of long‐term PM2.5 exposures on schoolchildren's asthma incidence (i.e., new cases) (Lee, 2023). It was assumed that all the children attending the same school (i.e., K‐12 schoolchildren) were exposed to the same PM2.5 levels. We used the following health impact function from the U.S. EPA's Environmental Benefits Mapping and Analysis Program (BenMAP) (U.S. EPA, 2021) to calculate the number of new asthma cases attributable to PM2.5 exposures for schoolchildren:

∆y=y0·1–e−β∆PM2.5×Pop

where ∆y is the estimated health impact of PM2.5 on schoolchildren's asthma incidence (i.e., the number of new asthma cases attributable to PM2.5); y 0 is the baseline age‐specific asthma incidence; β is the estimate of the concentration‐response function (CRF) per unit change in PM2.5 concentration; ∆PM2.5 is the change in PM2.5 exposures; and Pop is the number of exposed schoolchildren. The changes in PM2.5 exposures (∆PM2.5) were equivalent to school‐specific PM2.5 concentrations because we assumed no threshold level for health effects and the changes in PM2.5 were considered the difference between school‐specific PM2.5 concentrations and the reference level of 0 μg/m3. This reference level was selected because specific threshold levels had not been established and negative health effects were found even when PM2.5 concentrations fell below the U.S. EPA's PM2.5 standard (U.S. EPA, 2019).

The β and its 95% confidence interval (CI) were adopted from the meta‐analysis performed by Khreis et al. (2017) (Khreis et al., 2017). The meta‐analysis synthesized 41 studies published for the years 1999–2016 that reported the associations between traffic‐related air pollutants (including PM2.5) and childhood asthma development. The meta‐analysis resulted in the odd ratios (OR) per 1 μg/m3 of PM2.5 for ages from birth to 18 years and also for 6 < ages ≤ 18. To closely match the age groups of schoolchildren, we used the pooled estimate of OR = 1.04 (95% CI = 1.02–1.07) for 6 < ages ≤ 18, which resulted in β = 0.03922 (i.e., logOR = β; 95% CI for β = 0.01980–0.06766). As a sensitivity analysis, we also calculated the number of asthma incidence using the hazard ratio (HR) = 1.33 per interquartile range increase (IQR = 6.53 μg/m3; 95% CI = 1.31–1.34) and thus β = 0.04367 (95% CI for β = 0.04135–0.04482), which was derived from Tétreault et al. (2016) for ages 0–12 in Quebec, Canada (Tétreault et al., 2016). It is noted that the ages 0–12 did not entirely overlap with the age groups for schoolchildren (K‐12), which may introduce uncertainties. We primarily used the CRF obtained from Khreis et al. because of the age groups that more closely matched K‐12 schoolchildren included in our study. The 95% CIs reported with the β coefficients represented the uncertainties of CRFs, which were attributed to intrinsic factors related to the epidemiological studies, such as sample size, data quality (both exposure and health data), and the adequacy of controlling confounders.

The baseline asthma incidence rates (per 100,000 people) in California for the year 2016 were obtained from the Institute for Health Metrics and Evaluation (IHME) (Institute for Health Metrics and Evaluation (IHME), 2021). The data depend on reported diagnosis values at the county level. These rates were separated by the following age groups: 1–4, 5–9, 10–14, 15–19, 20–24, and so on. Because schoolchildren (ages 5–18) were the focus of our study and each school included different age groups that were not necessarily the same as the categories of the age groups defined in the IHME, we quantified the incidence rates separately for each age. This approach enabled us to calculate school‐specific asthma incidence rates solely based on the age groups attending each school.

To calculate age‐specific baseline asthma incidence rates, we examined the relationship between age groups defined by IHME and baseline asthma incidence rates, as shown in Figure S1 of the Supporting Information S1. The asthma incidence rate for those under 1 year of age was not provided. From this relationship, we identified the exponential decline of asthma incidence rates with age in children. Thus, we transformed both ages and asthma incidence rates to the natural logarithmic scale. Asthma incidence rates continuously declined from the age group 1–4 (6,264.39 per 100,000 people; maximum) to the age group 30–34 (607.98 per 100,000 people; minimum).

Figure S2A in Supporting Information S1 represents a log‐log linear regression model between baseline asthma incidence rates and age groups from 1–4 to 30–34. We selected these age groups because a monotonic decline of baseline asthma incidence rates was observed only for these age groups including schoolchildren. We assigned an asthma incidence rate representing each age group to the mid‐point of the age group (e.g., asthma incidence rate for ages 5–9 assigned to age 7). The regression model showed R 2 = 0.99 (N = 7), indicating a proper representation of the relationship for the age groups. We preferred log‐transformed age to untransformed age because of higher R 2 with the log‐transformed age (0.99) than with the untransformed age (0.92). As shown in Figure S2B of the Supporting Information S1, when the model included all age groups (1–4 through 95+), the predictive power decreased to R 2 = 0.68 (N = 20). This comparison demonstrates that the log‐log regression model with the age groups from 1–4 to 30–34 was more appropriate to estimate age‐specific asthma incidence rates for schoolchildren. All estimated age‐specific asthma incidence rates for ages 5–18 are presented in Table S1 of the Supporting Information S1. We calculated age‐adjusted school asthma rates by averaging the asthma incidence rates corresponding to age groups specific for each school. Although the age‐specific and school‐specific baseline asthma incidence rates could not be externally validated due to the unavailability of such detailed observation data, their use was reasonably supported by two factors: (a) baseline asthma incidence rates for specific age groups based on reported diagnosis values and (b) a high correlation between age and asthma incidence rates in a log‐log regression model.

2.6. Disparities of PM2.5 Exposures and PM2.5‐Related Asthma Incidence Rates

We employed a generalized additive model (GAM) implemented in the R package “mgcv” to examine non‐linear patterns between schoolchildren's PM2.5 exposures or asthma incidence rates attributable to ambient PM2.5 exposures and demographic factors (Wood, 2006, 2023).

AsthmaorPM2.5exposurei=β0+s(Demographicfactor)i+εi

where Asthma i is asthma incidence rates attributable to PM2.5 exposures at school i; PM2.5 exposure i is average PM2.5 concentrations at school i; s(Demographic factor) i is the smooth function of demographic factors (% people of color, low education, or poverty) in the block groups containing school i; β 0 is the intercept; and ɛ i  ∼ N(0, σ 2) is the error term at school i. The degree of smoothing was selected by generalized cross‐validation (GCV) implemented in the R package “mgcv” (Wood, 2006).

To represent the full scale of disparities by the level of the demographic factors and also prevent outlying values from potentially causing bias, we mainly used a metric of the differences in PM2.5 or asthma incidence rates between 1st and 99th percentiles of each demographic factor. Additionally, we explored those between 5th and 95th percentiles, 10th and 90th percentiles, and 75th and 25th percentiles of each demographic factor. It is noted that we calculated the disparities of PM2.5 exposures and asthma incidence rates separately by each demographic factor. We seek to demonstrate whether and how the PM2.5 exposures or PM2.5‐related asthma incidence rates varied by the level of each demographic factor. We did not aim to prove causal relationships between PM2.5 exposures or PM2.5‐related asthma incidence rates and demographic factors (Lee et al., 2024). Hence, we did not attempt to estimate non‐overlapping variability that was explained by only one demographic factor, while adjusting for the other two demographic factors.

3. Results and Discussion

3.1. Summary Statistics

The average annual PM2.5 concentration across the 9,811 public schools in California for 2016 was 9.2 (SD = 2.0) μg/m3. Across California, the school‐specific average PM2.5 ranged from 1.8 to 13.9 μg/m3. At a regional scale, the highest average PM2.5 at schools was found in the San Joaquin Valley air basin (11.4 μg/m3), followed by the South Coast air basin (10.8 μg/m3) and San Diego County air basin (8.8 μg/m3). The lowest average PM2.5 was observed at the schools in the Lake County air basin (3.4 μg/m3). The South Coast air basin had the largest number of public schools (3,447 schools, 35.1% of the total in California), followed by San Francisco Bay Area (1,682 schools, 17.1%) and San Joaquin Valley (1,386 schools, 14.1%). Air basins are defined by state statutes and regulations to allocate air resources in each region of California (California Air Resources Board (CARB), 2023a). The locations of all public schools analyzed are displayed in Figure 1.

Figure 1.

Figure 1

Open in a new tab

Location of public schools (grades K‐12) in California (N = 9,811).

3.2. Social Disparity of Schoolchildren's PM2.5 Exposures

Figure S3 in Supporting Information S1 displays the relationships between schoolchildren's PM2.5 exposures and demographic factors. Because higher population density tends to increase PM2.5 levels and thus affect the relationships between PM2.5 and demographic factors, the GAMs were employed by both including and excluding the variable of population density (Lee, 2019). In the GAMs, school‐specific PM2.5 was significantly associated with all three demographic factors (all p < 0.0001). The disparities of schoolchildren's PM2.5 exposures, while adjusting for population density, were most strongly associated with race/ethnicity (% people of color; 2.96 μg/m3), followed by % low education (2.11 μg/m3) and poverty (2.07 μg/m3).

Previous studies have reported that PM2.5 air pollution disproportionately affects people of color in California and across the U.S. (Lee, 2019; Tessum et al., 2021). Our findings demonstrate that the PM2.5 disparity for the general population is consistently observed for schoolchildren. Our study found larger PM2.5 disparities in relation to race/ethnicity than educational attainment and poverty, implying an independent role of race/ethnicity that was not overlapped with that of educational attainment or poverty as a surrogate of socioeconomic status. The excess PM2.5 impacts in association with race/ethnicity might be attributed to residential segregation by race/ethnicity and the disproportionate distribution of PM2.5 source emissions in residential areas with a comparatively large proportion of people of color (Jones et al., 2014; Woo et al., 2019).

3.3. School Area‐Specific PM2.5 Emissions

Figure 2 shows the distribution of distance to limited access highways as a PM2.5 emission proxy by the level of social vulnerability specific to each school. To improve the comparability across demographic factors that had different average levels, we normalized each of the demographic factors by (% demographic factors/average % demographic factors in a reference group). The average distance from schools to limited access highways was 5.2 (median = 1.7; SD = 12.9) km. Among the 9,811 schools included in this analysis, 493 (5.0%), 1,403 (14.3%), and 3,158 (32.2%) schools were located within 250, 500, and 1,000 m from the highways, respectively. 49 schools (0.5%) were situated even within 100 m from the highways. The schools that were closer to the highways were located in more vulnerable communities, as shown in Figure 2. The highest percentages of people of color, low educational attainment, and poverty were observed at schools located within 250 m from the highways, followed by those within <500, <1,000, and ≥1,000 m. The normalized % demographic factors ranged from 1.19 to 1.23 (<250 m), 1.16 to 1.21 (<500 m), and 1.09 to 1.16 (<1,000 m). The differences in the normalized % demographic factors between each distance category and the reference group (≥1,000 m) were statistically significant (p < 0.05) across all demographic factors. We also found statistically significant differences between <500 and <1,000 m categories for % low education and between <250 and <1,000 m categories and <500 and <1,000 m categories for % poverty (p < 0.05). All the other non‐significant differences were likely due to the progressively broader distance categories. We created these categories that were not mutually exclusive to identify the extent of spatial buffers from the highways, representing high traffic‐impacted areas, in relation to the distributions of more vulnerable communities. This approach, focusing on increasingly comprehensive coverage rather than segmented results, was more directly relevant to informing policy directions related to air pollution and environmental justice.

Figure 2.

Figure 2

Open in a new tab

Relationships between school area‐specific emission proxy (distance to limited access highways) and demographic factors. Normalized demographic factors are calculated by (average % demographic factors/average % demographic factors in the reference group). Means and standard errors are shown.

Figure S4 in Supporting Information S1 presents the distribution of traffic volumes (AADT) by the level of social vulnerability at each school. The average AADT across all schools was 109,955 (median = 86,000; SD = 92,009) vehicles per day. The schools adjacent to higher traffic volumes were located in more vulnerable communities. The schools with the highest traffic volumes (>276,000 vehicles per day or the highest 5% traffic volumes) had the highest percentages of people of color (ratio = 1.44), low education (ratio = 1.65), and poverty (ratio = 1.30), and vice versa. The differences in the normalized % demographic factors between each AADT category and the reference group (≤185,000 vehicles/day) were statistically significant (p < 0.05) across all demographic factors. Additionally, we observed statistically significant differences between >276,000 and >185,000 vehicles/day categories for all three demographic factors and between >276,000 and >247,500 vehicles/day for % low education and poverty (p < 0.05). The remaining four pairs in the figure were not statistically significant.

Figure S5 in Supporting Information S1 represents the relationships between CEIDARS PM2.5 emissions and demographic factors in school areas. Out of the 9,811 schools, 4,938 schools (50.3%) had CEIDARS facilities (PM2.5 emission > 0 tons per year) within a 0.01° (∼1 km) radius. Across these 4,938 schools, the average school‐specific PM2.5 emissions (i.e., the sum of all PM2.5 emissions within the 0.01° radius at each school) were 1.9 (median = 0.05, SD = 14.8) tons per year. Furthermore, 16 schools with a total enrollment of 10,500 students (0.2% of all public school enrollees) were located within 0.01° of high PM2.5 emitters (defined as facilities emitting >100 tons per year). The schools in more vulnerable communities had higher facility‐level PM2.5 emissions nearby. The percentages of people of color, low education, and poverty were higher at the schools located near the upper 5% (>2.48 tons per year), 10% (>0.83 tons per year), and 25% (>0.06 tons per year) PM2.5 emitters than the bottom 75% (≤0.06 tons per year; reference) PM2.5 emitters by 24%–53% (the ratios from 1.24 to 1.53). The differences in the normalized % demographic factors between each CEIDARS emission category and the reference group (≤0.06 tons/year) were statistically significant (p < 0.05) across all demographic factors. However, all the other differences were not statistically significant. Compared to the relationships between either distance from highways or traffic volumes (AADT) and demographic factors, the relationship between CEIDARS PM2.5 emissions and demographic factors was weaker, supported by non‐monotonic decreases in the normalized % demographic factors with decreasing emissions. It was likely due to the sporadically distributed point locations and a wide range of emissions from the facilities, as opposed to the more ubiquitous presence of roads and traffic.

3.4. Asthma Onsets Attributable to PM2.5 for Schoolchildren

The asthma incidence rate attributable to PM2.5 exposures was 562 new cases per 100,000 schoolchildren (95% CI = 311–854) in California. In an absolute term (i.e., asthma incidence), there were 34,537 PM2.5‐related new cases among all schoolchildren (95% CI = 19,090–52,493). Figure 3 demonstrates the schoolchildren's asthma onsets attributable to PM2.5 in each air basin of California. It is noted that Figures 3a and 3b represent the number of new asthma cases attributable to PM2.5 exposures per 100,000 schoolchildren (i.e., incidence rate) and the total number of new asthma cases attributable to PM2.5 exposures among all schoolchildren (i.e., incidence) in each air basin, respectively. The asthma incidence rates were the highest in the San Joaquin Valley [667 (95% CI = 375–992)], followed by the South Coast [621 (95% CI = 346–933)] and San Diego County [529 (95% CI = 290–813)] in the unit of new cases per 100,000 schoolchildren. Compared to Figure 3b, the smaller differences between air basins in Figure 3a were attributed to the number of schoolchildren that was factored into the calculation of the incidence rates. We further presented PM2.5‐related asthma incidence to evaluate the overall health impact of asthma onsets in each air basin. The incidence substantially varied by air basin. The highest asthma incidence attributable to PM2.5 was found in the air basin of South Coast (15,923 new cases; 95% CI = 8,876–23,917), followed by the San Joaquin Valley (5,494 new cases; 95% CI = 3,086–8,169) and San Francisco Bay (4,677 new cases; 95% CI = 2,537–7,288). These results suggested that asthma onsets attributable to schoolchildren's PM2.5 exposures were substantial, requiring the prioritized management of PM2.5 air quality and asthmatic children in the air basins with greater risks.

Figure 3.

Figure 3

Open in a new tab

Asthma onsets attributable to ambient PM2.5 exposures by air basin: (a) new asthma cases per 100,000 schoolchildren (incidence rates) and (b) total new asthma cases (incidence). Air basins including at least 100 public schools are displayed. Means and 95% CI are shown. The air basins are ordered by the number of new asthma cases per 100,000 schoolchildren in (a) and the number of new asthma cases among all schoolchildren in (b). Note: The number of schools located in each air basin is following: San Joaquin Valley (1,386), South Coast (3,447), San Diego County (740), Salton Sea (146), South Central Coast (408), Mountain Counties (199), San Francisco Bay (1,681), Sacramento Valley (905), Mojave Desert (296), North Coast (202), and North Central Coast (236).

Figure S6 in Supporting Information S1 illustrates results from main and sensitivity analyses that employed CRFs obtained from Khreis et al. and Tétreault et al. (Khreis et al., 2017; Tétreault et al., 2016). In the sensitivity analysis, the number of asthma incidence attributable to PM2.5 was estimated to be 37,683 (95% CI = 36,060–38,475). On average, this health impact was 9.1% higher than 34,537 new cases shown in our main analysis. Among the air basins, the highest number of asthma incidence attributable to PM2.5 exposures was identified in the South Coast air basin [17,340 cases (95% CI = 16,609–17,696)], followed by the San Joaquin Valley [5,973 cases (95% CI = 5,726–6,093)] and San Francisco Bay [5,124 cases (95% CI = 4,893–5,237)]. Per 100,000 students, the highest asthma incidence rate was found in the San Joaquin Valley [726 cases (95% CI = 696–740)], while the lowest rate was observed in the North Central Coast [423 cases (95% CI = 403–432)]. These findings demonstrate that the use of the additional CRF resulted in the increases in the asthma incidence by less than 10%, accompanied by narrower CIs.

3.5. Inequities of Asthma Incidence for Schoolchildren

Figure 4 shows the asthma incidence rates attributable to PM2.5 exposures in association with demographic factors. In the GAMs, the asthma cases were significantly associated with each of the demographic factors (p < 0.0001), showing higher asthma cases for schoolchildren attending public schools located in more vulnerable communities. Despite the generally positive correlations, we identified non‐linear patterns between the asthma incidence and % low education and poverty. While the relationship between the asthma incidence and % people of color was almost linear for the entire range of % people of color (0%–100%), those between the asthma incidence and % low education or poverty demonstrate a declining pattern for >65% of low education and a fairly constant pattern for 0%–40% of poverty. The near‐linear pattern for % people of color suggests a stronger association than for % low education or poverty showing non‐linear patterns. For % low education, the declining pattern was largely attributed to the small number of observations over 65% of low education, supported by the substantially expanded 95% CI. The relatively constant pattern within the 0%–40% of poverty might be due to similar PM2.5 levels within the range, as shown in Figure S3 of the Supporting Information S1. The steep increase observed for >40% of poverty indicates threshold levels that begin to demonstrate health effects, disproportionately impacting schoolchildren in impoverished communities. In addition, this could be caused by effect modifiers, such as nutritional or behavioral factors, which interact with poverty.

Figure 4.

Figure 4

Open in a new tab

Asthma incidence rates attributable to ambient PM2.5 exposures per 100,000 schoolchildren in association with demographic factors: (a) % people of color, (b) % low education, and (c) % poverty. Means and 95% confidence intervals are displayed by solid lines and shaded areas, respectively. Values on the y‐axis represent the deviation of the number of new asthma cases relative to the mean of new asthma cases (per 100,000 schoolchildren) across the levels of each demographic factor.

Figure 5 illustrates the disparities of the asthma incidence rates attributable to PM2.5 per 100,000 schoolchildren. For all three demographic factors considered, the asthma incidence rates were higher for schoolchildren attending public schools located in more vulnerable communities by 140 new cases per 100,000 schoolchildren on average (range = 85–209). The disparities of schoolchildren's asthma incidence rates were most strongly associated with % people of color, followed by % low education and poverty. On average, the disparities of asthma incidence rates associated with % people of color (209 new cases per 100,000 schoolchildren) were 2.0 (range = 1.6–2.5) times higher than those associated with % low education (128) and poverty (85). The near‐linear pattern for % people of color, as shown in Figure 4, contributed to the highest disparities of asthma onsets among schoolchildren, while the observed thresholds (approximately 40%) for % poverty resulted in the lowest disparities. The substantially greater health disparities related to % people of color, compared to those related to % low education and poverty, necessitate targeted policy interventions to address asthma onsets among schoolchildren in the communities of color. As expected, the absolute magnitude of the disparities decreased using additional metrics of 95th–5th, 90th–10th, and 75th–25th percentiles of each demographic factor by order.

Figure 5.

Figure 5

Open in a new tab

Disparity of schoolchildren's asthma incidence rates attributable to ambient PM2.5 exposures per 100,000 schoolchildren by the level of each demographic factor. The disparity is represented by the difference in asthma incidence rates between percentiles. Higher percentiles of the demographic factors indicate schoolchildren attending public schools located in more vulnerable communities. Means and standard errors are displayed.

Our findings suggest that schoolchildren living in communities with a higher proportion of people of color have a higher asthma incidence attributable to long‐term PM2.5 exposure. Since traffic‐related emissions are the major source of PM2.5 exposure disparities in California (Koolik et al., 2024; Lee & Park, 2020), regulatory efforts to mitigate traffic emissions, such as electrifying gasoline and diesel vehicles and implementing increasingly strict fuel economy standards, would be effective to reduce disparities in asthma incidence for schoolchildren. In 2022, CARB set targets to achieve 100% zero‐emission cars and light trucks for all new sales by 2035 (California Air Resources Board (CARB), 2023b). These long‐term emission strategies are expected to improve the equity of PM2.5 health impacts as a co‐benefit of lowering PM2.5 levels. Because communities located closer to facilities (including farms) emitting PM2.5 tend to be more vulnerable, as shown in Figure S5 of the Supporting Information S1, regulatory actions to control emissions from these facilities would further enhance the PM2.5 health equities, while reducing PM2.5 exposure levels and health outcomes for schoolchildren in these fence line communities.

3.6. Study Limitations

The limitations of our study are worth discussing. First, the PM2.5 exposures at 1 km resolution might not capture fine‐scale PM2.5 spikes around the school locations. Though 1 km resolution was the finest regarding satellite‐based AOD data, this approach could potentially underestimate schoolchildren's PM2.5 exposures. Second, the baseline asthma incidence rates were not specific to each block group but based on the entire California. Such fine‐scale asthma data were not available. Because the baseline asthma incidence rates are likely to differ by block group or region, our health impact assessment could become more comprehensive with the information. Third, the time‐activity patterns of schoolchildren were not accounted for. The information about schoolchildren's locations after school hours was not available. Hence, we assumed that schoolchildren stayed close to schools and school‐specific PM2.5 levels were applicable for their long‐term PM2.5 exposures. As discussed above, this assumption was supported by California's school zoning policies and the short median distance from homes to schools (0.66 and 1.41 miles for K–6th and 7th–12th grades, respectively) (California Legislative Information, 1987; He & Giuliano, 2018). Nonetheless, without individual survey data on the travel distance and its associated after‐school locations, the assumption cannot be definitely confirmed. Fourth, our study did not account for CRFs specific to schoolchildren within each demographic subgroup because such detailed CRFs were not available. Previous studies have reported that PM2.5‐related CRFs for mortality are different by the subgroups of demographic factors, such as race/ethnicity (Di et al., 2017; Ma et al., 2023). When available, incorporating subgroup‐specific CRFs for asthma onsets would refine our health estimates and disparities. Fifth, short‐term PM2.5 exposure is also a crucial aspect of asthma, which was not addressed in our study. Our study did not examine short‐term PM2.5 exposure because it is more closely associated with the exacerbation of asthma symptoms and respiratory‐related hospital admissions and emergency department visits, rather than the onset of asthma itself (U.S. EPA, 2019).

4. Conclusions

This study underscores significant socioeconomic disparities associated with asthma incidence attributable to ambient PM2.5 exposures among schoolchildren in California. Specifically, schoolchildren attending public schools in more vulnerable communities had disproportionately higher health impacts in relation to PM2.5 exposures. We focused on schoolchildren's asthma incidence as it is a critical determinant of lifelong health conditions. Our calculation of baseline asthma incidence rates for each age (rather than age group) helped refine the PM2.5 impacts on asthma incidence because the baseline asthma incidence rates tend to vary within each age group. Moreover, the use of high‐resolution PM2.5 concentrations with 1 km resolution enabled the spatially resolved determination of schoolchildren's PM2.5 exposures. Since traffic‐related emissions are a major source type of PM2.5 exposure disparities (Lee & Park, 2020), the implementation of targeted mitigation strategies such as the conversion of combustion engines to electric ones and the adoption of emission reduction technologies for conventional combustion engines can continue to reduce PM2.5 health disparities. Finally, a comprehensive approach that accounts for multiple risk factors related to socioeconomic status would be essential for achieving public health benefits effectively.

Conflict of Interest

The authors declare no conflicts of interest relevant to this study.

Supporting information

Supporting Information S1

GH2-9-e2024GH001099-s001.docx (183.8KB, docx)

Acknowledgments

The authors thank Dr. May Bhetraratana at the California Air Resources Board (CARB) for her feedback on the PM2.5‐related health calculations. The statements and conclusions in this manuscript are those of the authors and do not represent the official views of California's Office of Environmental Health Hazard Assessment (OEHHA) or the California Air Resources Board (CARB).

Lee, H. J. , Ebisu, K. , & Park, H.‐Y. (2025). Socioeconomic disparities of asthma incidence attributable to PM2.5 exposures for schoolchildren in California. GeoHealth, 9, e2024GH001099. 10.1029/2024GH001099

Data Availability Statement

The data used for calculating asthma incidence attributable to ambient PM2.5 exposure among schoolchildren in California are available in Lee (2023). The version 9.4 of the SAS used for calculating schoolchildren's asthma incidence requires a paid license (DOI not available). The version 4.3.1 of the R software and its package “mgcv” to calculate the disparities of PM2.5 exposure and asthma incidence is available in Wood (2023).

References

  1. Bell, M. L. , & Ebisu, K. (2012). Environmental inequality in exposures to airborne particulate matter components in the United States. Environmental Health Perspectives, 120(12), 1699–1704. 10.1289/ehp.1205201 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Berman, J. D. , & Ebisu, K. (2020). Changes in U.S. air pollution during the COVID‐19 pandemic. Science of the Total Environment, 739, 139864. 10.1016/j.scitotenv.2020.139864 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. California Air Resources Board (CARB) . (2019). CEIDARS database structure. Retrieved from https://ww3.arb.ca.gov/ei/drei/maintain/dbstruct.htm
  4. California Department of Transportation (CalTrans) . (2021). Traffic census program (AADT). Retrieved from https://dot.ca.gov/programs/traffic‐operations/census
  5. California Air Resources Board (CARB) . (2023a). Emissions by air basin. Retrieved from https://ww2.arb.ca.gov/applications/emissions‐air‐basin
  6. California Air Resources Board (CARB) . (2023b). Zero‐emission vehicle program. Retrieved from https://ww2.arb.ca.gov/our‐work/programs/zero‐emission‐vehicle‐program/about
  7. California Department of Education . (2021a). Census day enrollment by school. Retrieved from https://www.cde.ca.gov/ds/ad/filesenr.asp
  8. California Department of Education . (2021b). Private schools—CalEdFacts. Retrieved from https://www.cde.ca.gov/sp/ps/cefprivinstr.asp
  9. California Legislative Information . (1987). Education code‐ compulsory education law. Retrieved from https://leginfo.legislature.ca.gov/faces/codes_displaySection.xhtml?lawCode=EDC&sectionNum=48200
  10. California School Campus Database . (2016). Location of public schools in California. Retrieved from www.californiaschoolcampusdatabase.org/
  11. Colmer, J. , Hardman, I. , Shimshack, J. , & Voorheis, J. (2020). Disparities in PM2.5 air pollution in the United States. Science, 369(6503), 575–578. 10.1126/science.aaz9353 [DOI] [PubMed] [Google Scholar]
  12. Cushing, L. , Blaustein‐Rejto, D. , Wander, M. , Pastor, M. , Sadd, J. , Zhu, A. , & Morello‐Frosch, R. (2018). Carbon trading, co‐pollutants, and environmental equity: Evidence from California’s cap‐and‐trade program (2011–2015). PLoS Medicine, 15(7), e1002604. 10.1371/journal.pmed.1002604 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. de Hoogh, K. , Héritier, H. , Stafoggia, M. , Künzli, N. , & Kloog, I. (2018). Modelling daily PM2.5 concentrations at high spatio‐temporal resolution across Switzerland. Environmental Pollution, 233, 1147–1154. 10.1016/j.envpol.2017.10.025 [DOI] [PubMed] [Google Scholar]
  14. Di, Q. , Wang, Y. , Zanobetti, A. , Wang, Y. , Koutrakis, P. , Choirat, C. , et al. (2017). Air pollution and mortality in the medicare population. New England Journal of Medicine, 376(26), 2513–2522. 10.1056/nejmoa1702747 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hajat, A. , Diez‐Roux, A. V. , Adar, S. D. , Auchincloss, A. H. , Lovasi, G. S. , O’Neill, M. S. , et al. (2013). Air pollution and individual and neighborhood socioeconomic status: Evidence from the multi‐ethnic study of atherosclerosis (MESA). Environmental Health Perspectives, 121(11–12), 1325–1333. 10.1289/ehp.1206337 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hasheminassab, S. , Daher, N. , Ostro, B. D. , & Sioutas, C. (2014). Long‐term source apportionment of ambient fine particulate matter (PM2.5) in the Los Angeles Basin: A focus on emissions reduction from vehicular sources. Environmental Pollution, 193, 54–64. 10.1016/j.envpol.2014.06.012 [DOI] [PubMed] [Google Scholar]
  17. He, S. Y. , & Giuliano, G. (2018). School choice: Understanding the trade‐off between travel distance and school quality. Transportation, 45(5), 1475–1498. 10.1007/s11116-017-9773-3 [DOI] [Google Scholar]
  18. Institute for Health Metrics and Evaluation (IHME) . (2021). GBD results tool. Retrieved from https://vizhub.healthdata.org/gbd‐results/
  19. Jbaily, A. , Zhou, X. , Liu, J. , Lee, T.‐H. , Kamareddine, L. , Verguet, S. , & Dominici, F. (2022). Air pollution exposure disparities across US population and income groups. Nature, 601(7892), 228–233. 10.1038/s41586-021-04190-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Jones, M. R. , Diez‐Roux, A. V. , Hajat, A. , Kershaw, K. N. , O’Neill, M. S. , Guallar, E. , et al. (2014). Race/ethnicity, residential segregation, and exposure to ambient air pollution: The multi‐ethnic study of atherosclerosis (MESA). American Journal of Public Health, 104(11), 2130–2137. 10.2105/ajph.2014.302135 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Khreis, H. , Kelly, C. , Tate, J. , Parslow, R. , Lucas, K. , & Nieuwenhuijsen, M. (2017). Exposure to traffic‐related air pollution and risk of development of childhood asthma: A systematic review and meta‐analysis. Environment International, 100, 1–31. 10.1016/j.envint.2016.11.012 [DOI] [PubMed] [Google Scholar]
  22. Koolik, L. H. , Alvarado, Á. , Budahn, A. , Plummer, L. , Marshall, J. D. , & Apte, J. S. (2024). PM2.5 exposure disparities persist despite strict vehicle emissions controls in California. Science Advances, 10(37), eadn8544. 10.1126/sciadv.adn8544 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lane, H. M. , Morello‐Frosch, R. , Marshall, J. D. , & Apte, J. S. (2022). Historical redlining is associated with present‐day air pollution disparities in U.S. cities. Environmental Science and Technology Letters, 9(4), 345–350. 10.1021/acs.estlett.1c01012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lee, H. J. (2019). Benefits of high resolution PM2.5 prediction using satellite MAIAC AOD and land use regression for exposure assessment: California examples. Environmental Science & Technology, 53(21), 12774–12783. 10.1021/acs.est.9b03799 [DOI] [PubMed] [Google Scholar]
  25. Lee, H. J. (2020). Advancing exposure assessment of PM2.5 using satellite remote sensing: A review. Asian Journal of Atmospheric Environment, 14(4), 319–334. 10.5572/ajae.2020.14.4.319 [DOI] [Google Scholar]
  26. Lee, H. J. (2023). California school PM2.5 and demographic factors [Dataset]. 10.6084/m9.figshare.24637434.v1 [DOI]
  27. Lee, H. J. , Gent, J. F. , Leaderer, B. P. , & Koutrakis, P. (2011). Spatial and temporal variability of fine particle composition and source types in five cities of Connecticut and Massachusetts. Science of the Total Environment, 409(11), 2133–2142. 10.1016/j.scitotenv.2011.02.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lee, H. J. , Kuwayama, T. , & FitzGibbon, M. (2024). Simultaneous decreases in NO2 levels and disparities in California during the COVID‐19 pandemic. Atmospheric Environment, 318, 120214. 10.1016/j.atmosenv.2023.120214 [DOI] [Google Scholar]
  29. Lee, H. J. , Liu, Y. , & Chatfield, R. B. (2023). Neighborhood‐scale ambient NO2 concentrations using TROPOMI NO2 data: Applications for spatially comprehensive exposure assessment. Science of the Total Environment, 857, 159342. 10.1016/j.scitotenv.2022.159342 [DOI] [PubMed] [Google Scholar]
  30. Lee, H. J. , Liu, Y. , Coull, B. A. , Schwartz, J. , & Koutrakis, P. (2011). A novel calibration approach of MODIS AOD data to predict PM2.5 concentrations. Atmospheric Chemistry and Physics, 11(15), 7991–8002. 10.5194/acp-11-7991-2011 [DOI] [Google Scholar]
  31. Lee, H. J. , & Park, H.‐Y. (2020). Prioritizing the control of emission sources to mitigate PM2.5 disparity in California. Atmospheric Environment, 224, 117316. 10.1016/j.atmosenv.2020.117316 [DOI] [Google Scholar]
  32. Ma, Y. , Zang, E. , Opara, I. , Lu, Y. , Krumholz, H. M. , & Chen, K. (2023). Racial/ethnic disparities in PM2.5‐attributable cardiovascular mortality burden in the United States. Nature Human Behaviour, 7(12), 2074–2083. 10.1038/s41562-023-01694-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Ma, Z. , Hu, X. , Sayer, A. M. , Levy, R. , Zhang, Q. , Xue, Y. , et al. (2016). Satellite‐based spatiotemporal trends in PM2.5 concentrations: China, 2004‐2013. Environmental Health Perspectives, 124(2), 184–192. 10.1289/ehp.1409481 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. McConnell, R. , Islam, T. , Shankardass, K. , Jerrett, M. , Lurmann, F. , Gilliland, F. , et al. (2010). Childhood incident asthma and traffic‐related air pollution at home and school. Environmental Health Perspectives, 118(7), 1021–1026. 10.1289/ehp.0901232 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. McGeachie, M. J. , Yates, K. P. , Zhou, X. , Guo, F. , Sternberg, A. L. , Van Natta, M. L. , et al. (2016). Patterns of growth and decline in lung function in persistent childhood asthma. New England Journal of Medicine, 374(19), 1842–1852. 10.1056/nejmoa1513737 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Mirabelli, M. C. , Vaidyanathan, A. , Flanders, W. D. , Qin, X. T. , & Garbe, P. (2016). Outdoor PM2.5, ambient air temperature, and asthma symptoms in the past 14 days among adults with active asthma. Environmental Health Perspectives, 124(12), 1882–1890. 10.1289/ehp92 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Nair, A. A. , Lin, S. , Luo, G. , Ryan, I. , Qi, Q. , Deng, X. , & Yu, F. (2023). Environmental exposure disparities in ultrafine particles and PM2.5 by urbanicity and socio‐demographics in New York state, 2013–2020. Environmental Research, 239, 117246. 10.1016/j.envres.2023.117246 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Ni, R. , Su, H. , Burnett, R. T. , Guo, Y. , & Cheng, Y. (2024). Long‐term exposure to PM2.5 has significant adverse effects on childhood and adult asthma: A global meta‐analysis and health impact assessment. One Earth, 7(11), 1953–1969. 10.1016/j.oneear.2024.09.022 [DOI] [Google Scholar]
  39. Orellano, P. , Quaranta, N. , Reynoso, J. , Balbi, B. , & Vasquez, J. (2017). Effect of outdoor air pollution on asthma exacerbations in children and adults: Systematic review and multilevel meta‐analysis. PLoS One, 12(3), e0174050. 10.1371/journal.pone.0174050 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Squizzato, S. , Masiol, M. , Rich, D. Q. , & Hopke, P. K. (2018). A long‐term source apportionment of PM2.5 in New York State during 2005–2016. Atmospheric Environment, 192, 35–47. 10.1016/j.atmosenv.2018.08.044 [DOI] [Google Scholar]
  41. Strickland, M. J. , Hao, H. , Hu, X. , Chang, H. H. , Darrow, L. A. , & Liu, Y. (2016). Pediatric emergency visits and short‐term changes in PM2.5 concentrations in the U.S. State of Georgia. Environmental Health Perspectives, 124(5), 690–696. 10.1289/ehp.1509856 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Tessum, C. W. , Paolella, D. A. , Chambliss, S. E. , Apte, J. S. , Hill, J. D. , & Marshall, J. D. (2021). PM2.5 polluters disproportionately and systemically affect people of color in the United States. Science Advances, 7(18), eabf4491. 10.1126/sciadv.abf4491 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Tétreault, L.‐F. , Doucet, M. , Gamache, P. , Fournier, M. , Brand, A. , Kosatsky, T. , & Smargiassi, A. (2016). Childhood exposure to ambient air pollutants and the onset of asthma: An administrative cohort study in Quebec. Environmental Health Perspectives, 124(8), 1276–1282. 10.1289/ehp.1509838 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Trasande, L. , & Thurston, G. D. (2005). The role of air pollution in asthma and other pediatric morbidities. Journal of Allergy and Clinical Immunology, 115(4), 689–699. 10.1016/j.jaci.2005.01.056 [DOI] [PubMed] [Google Scholar]
  45. Tsai, Y.‐G. , Chio, C.‐P. , Yang, K. D. , Lin, C.‐H. , Yeh, Y.‐P. , Chang, Y.‐J. , et al. (2025). Long‐term PM2.5 exposure is associated with asthma prevalence and exhaled nitric oxide levels in children. Pediatric Research, 97(1), 370–377. 10.1038/s41390-023-02977-5 [DOI] [PubMed] [Google Scholar]
  46. U.S. Census Bureau . (2021). American Community Survey (ACS). Retrieved from https://www.census.gov/acs/www/data/data‐tables‐and‐tools/data‐profiles/2016/
  47. U.S. Centers for Disease Control and Prevention (CDC) . (2018). Asthma in children. Retrieved from https://www.cdc.gov/vitalsigns/childhood‐asthma/index.html [Google Scholar]
  48. U.S. EPA . (2019). Integrated Science Assessment (ISA) for particulate matter. Retrieved from https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=347534#tab‐3
  49. U.S. EPA . (2021). BenMAP‐CE manual and appendices. Retrieved from https://www.epa.gov/benmap/benmap‐ce‐manual‐and‐appendices
  50. Williams, A. M. , Phaneuf, D. J. , Barrett, M. A. , & Su, J. G. (2019). Short‐term impact of PM2.5 on contemporaneous asthma medication use: Behavior and the value of pollution reductions. Proceedings of the National Academy of Sciences of the United States of America, 116(12), 5246–5253. 10.1073/pnas.1805647115 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Woo, B. , Kravitz‐Wirtz, N. , Sass, V. , Crowder, K. , Teixeira, S. , & Takeuchi, D. T. (2019). Residential segregation and racial/ethnic disparities in ambient air pollution. Race and Social Problems, 11(1), 60–67. 10.1007/s12552-018-9254-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Wood, S. (2023). R package “mgcv” [Software]. https://cran.r‐project.org/web/packages/mgcv/index.html
  53. Wood, S. N. (2006). Generalized additive models: An introduction with R. Chapman and Hall/CRC. [Google Scholar]
  54. Xie, Y. , Wang, Y. , Zhang, K. , Dong, W. , Lv, B. , & Bai, Y. (2015). Daily estimation of ground‐level PM2.5 concentrations over Beijing using 3 km resolution MODIS AOD. Environmental Science & Technology, 49(20), 12280–12288. 10.1021/acs.est.5b01413 [DOI] [PubMed] [Google Scholar]
  55. Zheng, X.‐Y. , Ding, H. , Jiang, L.‐N. , Chen, S.‐W. , Zheng, J.‐P. , Qiu, M. , et al. (2015). Association between air pollutants and asthma emergency room visits and hospital admissions in time series studies: A systematic review and meta‐analysis. PLoS One, 10(9), e0138146. 10.1371/journal.pone.0138146 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Zhu, Q. , Liu, Y. , & Hasheminassab, S. (2024). Long‐term source apportionment of PM2.5 across the contiguous United States (2000‐2019) using a multilinear engine model. Journal of Hazardous Materials, 472, 134550. 10.1016/j.jhazmat.2024.134550 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Citations

  1. Lee, H. J. (2023). California school PM2.5 and demographic factors [Dataset]. 10.6084/m9.figshare.24637434.v1 [DOI]
  2. Wood, S. (2023). R package “mgcv” [Software]. https://cran.r‐project.org/web/packages/mgcv/index.html

Supplementary Materials

Supporting Information S1

GH2-9-e2024GH001099-s001.docx (183.8KB, docx)

Data Availability Statement

The data used for calculating asthma incidence attributable to ambient PM2.5 exposure among schoolchildren in California are available in Lee (2023). The version 9.4 of the SAS used for calculating schoolchildren's asthma incidence requires a paid license (DOI not available). The version 4.3.1 of the R software and its package “mgcv” to calculate the disparities of PM2.5 exposure and asthma incidence is available in Wood (2023).

Articles from GeoHealth are provided here courtesy of Wiley

RESOURCES