Sage Journals: Discover world-class research

Abstract

Background:

There is a paucity of knowledge regarding the joint and cumulative effects of multiple exposures from the physical, built, and social environments for chronic obstructive pulmonary disease (COPD) outcomes.

Objectives:

We assessed the separate and joint contributions of such factors for initial lung function and annual rate of change in a clinical cohort of patients with COPD.

Design:

Retrospective cohort study using electronic medical records.

Methods:

Data on individuals with an ICD-10 diagnosis of COPD were extracted from the University of Alabama at Birmingham Health System (2010–2020). Analyses were restricted to those who met spirometric criteria for a COPD diagnosis (FEV₁/FVC < 0.7). The predicted FEV₁ values (FEV₁%) were calculated using 2022 GLI race-neutral equations. The earliest residential address during the study period was geocoded and linked to neighborhood characteristics at the Census tract level, including social vulnerability, environmental burden, walkability, residential segregation, rurality, and food access. Analyses adjusted for age, sex, race/ethnicity, marital status, smoking status, obesity, and comorbidities.

Results:

The analytic sample included 5652 patients. Social and environmental exposures differed by race (p < 0.001), with Black patients living in more urban and segregated environments with worse exposures. Adverse exposures were significantly associated with lower baseline FEV₁%: overall environmental justice score, −11.43%; overall social vulnerability score, −12.29%; neighborhood socioeconomic status, −11.96%; residential segregation, −9.44%. Most associations remained significant after adjusting for patient demographic and clinical characteristics. When the multiple aspects of the environment were considered simultaneously, the dominant factor for baseline FEV₁% was the overall social vulnerability (−12.31%, p < 0.001 unadjusted; −9.96%, p < 0.001 adjusted). Residence in a food desert was negatively associated with the annual rate of FEV₁% decline (−0.41% unadjusted, p = 0.004; −0.39% adjusted, p = 0.009).

Conclusion:

Aspects of the physical, built, and social environments affect COPD outcomes. Findings can be used in risk prediction models and for the development of tools that incorporate such exposures at the point of care for individualized disease management.

Keywords

built environment COPD environmental exposures lung function social environment

Introduction

Chronic obstructive pulmonary disease (COPD) is a heterogeneous lung condition characterized by cough, dyspnea, and sputum production due to persistent abnormalities of the airways, including bronchiolitis, chronic bronchitis, and/or emphysema, typically resulting in progressive lung airflow limitation.¹ COPD most often occurs in people aged ⩾65 years with a history of smoking, women, persons with asthma, or those with occupational exposure to dust, gases, vapors, and fumes.^2–4 Socioeconomic disadvantage has also been identified as a risk factor and prognostic indicator, as individuals from less advantaged socioeconomic backgrounds are more likely to have COPD and to experience worse disease outcomes.^5–12 Such disparities are multifactorial, with an extensively documented role of the exposome, or the sum of all environmental exposures throughout the life course, from economic and sociocultural to physical and biochemical.^13–15

There is a wealth of evidence that adverse neighborhood conditions increase the risk of chronic disease, including COPD,^16–22 through multiple pathways, including behavioral (physical activity, diet, sleep) and biological (stress reactivity, endothelial dysfunction, inflammation, oxidative imbalance). For example, features of the neighborhood built environment affect physical activity, diet, and sleep,^23–27 whereas aspects of the sociocultural environment play a role in stress response.^28–31 Prior research shows that neighborhood green space and walkability have protective effects against asthma hospitalizations³² and bronchitis,³³ after accounting for noise and air pollution. Residential segregation predicts asthma burden better than an individual’s own race or ethnicity.³⁴

Importantly, neighborhood characteristics are unequally distributed across socioeconomic and racial/ethnic lines.³⁵ For example, low-income and racial/ethnic minority neighborhoods have limited access to healthy food and green spaces,^36–39 higher rates of crime,^40–43 higher density of tobacco stores,⁴⁴ and greater exposures to air pollution and poorer housing conditions.^44,45 The unequal burden of physical, social, chemical, and biological exposures leads to unequal health outcomes.⁴⁶

Despite such evidence, disparities in COPD incidence and progression are rarely assessed in the context of social and environmental exposures. The majority of prior research on the lung health effects of such exposures has been in respiratory conditions other than COPD, in cross-sectional studies, or with a focus on single exposures. There is a paucity of findings on the joint and cumulative effects of multiple types of exposures from the physical, built, and social environments for COPD outcomes in an exposome framework. To bridge this knowledge gap, we assessed the separate and joint contributions of multiple social and environmental neighborhood factors for initial lung function and annual lung function decline in a clinical cohort of patients diagnosed with COPD.

We used the Environmental Justice Index (EJI) because it includes multiple measures of neighborhood adversity, for a total of 36 variables (Figure 1). These variables are organized into three modules (social vulnerability, environmental burden, and health vulnerability) and 10 domains (racial/ethnic minority status, socioeconomic status, household characteristics, housing type, air pollution, hazardous and toxic sites, built environment, transportation infrastructure, water pollution, and chronic disease burden), which allows us to examine the contributions of these specific aspects of neighborhood context to lung function. In addition, we included measures of neighborhood segregation, walkability, rurality, and food access, which are not part of the EJI but have been associated with COPD outcomes previously.^45,47–50

Figure 1.

The Environmental Justice Index (EJI): modules, domains, and indicators.

Methods

Study design and population

The study’s inclusion criterion was a diagnosis of COPD that met established spirometric criteria (FEV₁/FVC < 0.7). We extracted data on individuals with an ICD-10 diagnosis of COPD from the University of Alabama at Birmingham (UAB) Health System electronic medical records (EMR) for the period from 2010 to 2020. We limited the analyses to those who met spirometric criteria for COPD diagnosis (FEV₁/FVC < 0.7) at any point during the observation period. As this is a health system-based population cohort, no a priori power calculations were performed. The earliest full residential address recorded during the observation period was geocoded and linked to neighborhood characteristics measured at the Census tract level.

Measures

Outcomes

Pre-bronchodilator lung function was measured in outpatient settings, in the same pulmonary function testing laboratory, using spirometry that followed the American Thoracic Society (ATS) standards.^51,52 The analyses used pre-bronchodilator FEV₁ values because they were available for all patients. The % predicted forced expiratory volume in one second (FEV₁%) was calculated using the 2022 GLI race-neutral equations,⁵³ with absolute FEV₁, sex, height, and age extracted from the EMR. Only 7.3% of patients had more than one FEV₁ value from the same encounter. For these individuals, we kept the highest measurement recorded for that encounter.

Exposures

Neighborhood-level environmental exposures were assessed with the 2022 Centers for Disease Control and Prevention (CDC) EJI,⁵⁴ calculated with data from the U.S. Census Bureau, the U.S. Environmental Protection Agency, the U.S. Mine Safety and Health Administration, the U.S. Geological Survey, the U.S. Department of Transportation, and the CDC. The EJI ranks all U.S. Census tracts on 36 environmental, social, and health factors grouped into three overarching modules (social vulnerability, environmental burden, and health burden) and 10 different domains (Figure 1). Scores range from 0 to 1, with higher scores indicating a worse ranking of the Census tract compared to other tracts in the United States.

Walkability was assessed with the 2019 National Walkability Index (NWI).⁵⁵ Originally developed for an assessment of neighborhood quality of life,⁵⁶ the NWI provides geographic data ranking walkability as scores within block groups. Scores are based on intersection density, proximity to transit stops, and diversity of land use, where greater density, less distance, and greater diversity indicate a more walkable neighborhood. The score ranges from 0 to 20, with higher values indicating greater walkability.

Residential segregation, derived from the Dissimilarity Index, the Interaction Index, and the Isolation Index, measures segregation within communities by race, ethnicity, socioeconomic status, or sex based on 2010 decennial Census data.⁵⁷ The Dissimilarity Index assesses how groups are distributed across geographic areas within larger areas. A score of zero indicates the groups being compared are evenly distributed, whereas a score of 100 indicates that they are completely segregated, with one group residing entirely in one area and the other entirely in another area. The Interaction Index measures how exposed members of minority groups are to members of the majority group. Scores range from 0 to 1, with higher values indicating greater interaction of minority groups with the majority group. The Isolation Index measures how isolated members of minority groups are from members of the majority group. Scores range from 0 to 1, with higher values indicating greater isolation of the minority groups from the majority group.

Rural/urban status was measured using the 2010 rural-urban commuting (RUCA) codes available from the U.S. Department of Agriculture (USDA) Economic Research Service that tracks urban and rural communities.⁵⁸ RUCA uses 10 classification codes delineating metropolitan, micropolitan, small town, and rural areas, sourced from the U.S. Census tracts and measuring urbanization, population density, and urbanization.⁵⁹ This variable was dichotomized as urban if patients resided in a metropolitan classification code with primary flow within an urbanized area.

Food deserts were measured using the 2019 low-income low-access (LILA) scores developed by the USDA. Census tracts are considered low-income if 20% or more of the households live in poverty or the tract’s median family income is less than or equal to 80% of the state or metropolitan area’s median family income. An area is considered low-access to food if the nearest grocery store is more than 1 mile away from the centroid of the tract in urban areas or 10 miles in rural areas. Tracts were defined as a food desert is they met both criteria.

Covariates

Covariates included age at first observation, sex, race/ethnicity (non-Hispanic White, non-Hispanic Black), marital status (divorced or separated, widowed, married or cohabiting, never married), smoking status (current, former, never), obesity, and number of comorbidities (type 2 diabetes, type 1 diabetes, neoplasms, liver disease, high blood pressure, HIV, heart failure, heart atherosclerosis, chronic kidney disease, and asthma). Individuals were considered to have a comorbidity or obesity if they were diagnosed with it prior to the date of the lung function measure. Individuals were assumed to always have the comorbidity after it was diagnosed.

Statistical analyses

To handle the nested nature of our data both at the geographic level and for multiple observations within an individual, we used multilevel models, which correct for the interdependency created by using observations from multiple individuals in the same Census tract as well as using multiple observations from the same person. We estimated a random intercept at the Census tract level, a random intercept at the individual level, and a random slope (i.e., change in lung function) at the individual level. Time was measured in years since the first lung function measurement during the study period. All spirometry results obtained during the study period were included in the analysis. To account for attrition, we included as an auxiliary covariate the proportion of visits made by an individual based on pulmonary function tests conducted at least every other year, thereby increasing the likelihood that the missingness is more plausibly ignorable.^60,61

Models examined each social or environmental exposure (SEE) (i.e., neighborhood measure) and included both the associations with baseline FEV₁% and annual change in FEV₁%, unadjusted and adjusted for covariates. In addition, we examined the extent to which SEE measures were associated with the outcome after adjusting for the other SEE measures and then adjusting for the covariates. An interaction between time and each neighborhood measure was used to determine if the neighborhood measure impacted the annual change in lung function.

All exposure measures were modeled on the baseline FEV₁%, with the exception of food access, which was modeled on the change in FEV₁% based on model fit statistics. Differences in the association between each SEE and race were examined by including an interaction term; significant interactions are noted in the tables.

To avoid multicollinearity between the exposures, we used the EJI composite module and domain scores rather than the single variables making up the domains. We also estimated the average variance inflation factor (VIF) of the joint adjusted model, which was < 5 (VIF = 2.06), suggesting that multicollinearity was not an issue.⁶²

The SEE measures have multiple editions available for the study period. For example, the Social Vulnerability Index (SVI), which is the social vulnerability component of the EJI, is available for 2010, 2014, 2016, 2018, and 2020, whereas the food desert measure (LILA) is available for 2010, 2015, and 2019. We first assigned to each individual the index scores from the year closest to when they entered the study. For example, those entering between 2010 and 2013 were assigned the 2010 SVI scores, those entering between 2014 and 2015 were assigned the 2014 SVI scores, those entering between 2016 and 2017 were assigned the 2016 SVI scores, those entering between 2018 and 2019 were assigned the 2018 SVI scores, and those entering in 2020 were assigned the 2020 SVI scores. However, using these temporally aligned SVI scores was substantively the same as using the SVI scores in the 2022 EJI. Given our interest in the EJI, we therefore used the 2022 SVI, which is used to calculate the EJI scores. We also used the 2019 LILA scores as they were temporally most closely aligned with the other area-level measures.

The reporting of this study conforms to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.⁶³ We used complete case analysis to analyze the data. Missingness is reported in the STROBE diagram (Figure 2). We also conducted a sensitivity analysis that excluded individuals with an asthma diagnosis. All statistical tests were two-tailed, with the significance level set at 0.05. Statistical analyses were performed using SAS software, version 9.4, © 2004 SAS Institute Inc., Cary, NC, USA.

Figure 2.

STROBE diagram of the patient population.

Results

The complete case sample included 5652 patients contributing 21318 FEV₁% measures. A STROBE diagram of the study population is included in Figure 2. Almost half (48.9%, n = 2765) had only one FEV₁% measurement during the study period and contributed only to the baseline FEV₁% estimates (n = 2368 with COPD only, n = 370 with COPD and asthma). One-fifth (19.1%, n = 1081) had two FEV₁% measurements, with a mean time between measurements of 1.2 years (SD = 1.4). The remaining 32% contributed at least three measurements, and the interval between them decreased at the later encounters. For example, the mean time between the second and third FEV₁% measurement among those with at least three measurements was 0.90 (SD = 0.97) years. Among those with two or more FEV₁% measurements, 1786 had only a COPD diagnosis, and 1128 had a COPD and asthma diagnosis.

Table 1 presents the baseline characteristics of the sample, overall and by race. All social and environmental exposures differed by race at p < 0.001, except for water pollution. Black patients lived in more urban environments and had significantly worse exposures than their White counterparts on all social and environmental measures, except for built environment, housing type, and walkability. Compared to Black counterparts, White patients resided in neighborhoods that were more racially homogenous but had more interaction with and less isolation from other racial groups. In addition, Black patients were younger, less likely to be married, had more chronic conditions, higher prevalence of obesity, lower lung function, and were more likely to be current smokers than White patients.

Table 1.

Baseline characteristics of the sample, overall and by race.

Characteristic	Overall, N = 5652	Black, n = 1254	White, n = 4398	p-value
Characteristic	Mean (SD) or % (N)	Mean (SD) or % (N)	Mean (SD) or % (N)	p-value
Neighborhood social and environmental exposures
Environmental Justice, overall (range 0–1)	0.66 (0.27)	0.86 (0.20)	0.61 (0.27)	<0.001
Social Vulnerability, overall (range 0–1)	0.55 (0.29)	0.73 (0.23)	0.45 (0.26)	<0.001
Theme 1, Socioeconomic status	0.53 (0.29)	0.78 (0.23)	0.46 (0.27)	<0.001
Theme 2, Household characteristics	0.53 (0.28)	0.78 (0.24)	0.52 (0.27)	<0.001
Theme 3, Minority status	0.48 (0.29)	0.59 (0.28)	0.39 (0.23)	<0.001
Theme 4, Housing/Transportation	0.57 (0.28)	0.83 (0.18)	0.59 (0.28)	<0.001
Environmental Burden, overall (range 0–1)	0.62 (0.24)	0.77 (0.20)	0.57 (0.23)	<0.001
Air pollution	0.71 (0.18)	0.84 (0.15)	0.68 (0.17)	<0.001
Hazardous and toxic sites	0.49 (0.31)	0.61 (0.29)	0.45 (0.30)	<0.001
Built environment	0.69 (0.23)	0.57 (0.26)	0.72 (0.21)	<0.001
Transportation pollution	0.45 (0.28)	0.57 (0.26)	0.45 (0.30)	<0.001
Water pollution	0.28 (0.16)	0.28 (0.22)	0.28 (0.14)	0.225
Health Vulnerability (range 0–1)	0.55 (0.33)	0.70 (0.24)	0.51 (0.34)	<0.001
Residential segregation (range 0–1)
Dissimilarity	0.60 (0.18)	0.58 (0.16)	0.61 (0.18)	<0.001
Interaction	0.49 (0.28)	0.16 (0.19)	0.58 (0.22)	<0.001
Isolation	0.46 (0.29)	0.81 (0.21)	0.37 (0.23)	<0.001
Urban, %	62.00 (3,504)	89.00 (1,116)	54.00 (2,375)	<0.001
Walkability (range 0–20)	7.41 (3.90)	10.62 (4.24)	6.49 (3.26)	<0.001
Food desert, %	18.19 (1,028)	46.73 (586)	10.05 (442)	<0.001
Demographic characteristics
Age, years	64.12 (13.10)	61.29 (12.65)	64.93 (13.11)	<0.001
Sex
Male, %	52.34 (2,958)	50.32 (631)	52.91 (2,327)	0.105
Female, %	47.66 (2,694)	49.68 (623)	47.09 (2,071)
Marital status
Divorced or separated, %	14.21 (803)	16.19 (203)	13.64 (600)	<0.001
Married or cohabiting, %	52.39 (2,961)	34.05 (427)	57.62 (2,534)
Single, %	18.88 (1,067)	36.60 (459)	13.82 (608)
Widowed, %	14.53 (821)	13.16 (165)	14.92 (656)
Clinical characteristics
FEV₁%	57.88 (20.87)	52.51 (17.82)	59.42 (21.41)	<0.001
Chronic conditions, n	1.65 (1.60)	2.05 (1.70)	1.53 (1.55)	<0.001
Obesity, %	18.45 (1043)	26.79 (336)	16.08 (707)	<0.001
Smoking status, %
Current	27.12 (1,533)	31.74 (398)	25.81 (1,135)	<0.001
Former	51.72 (2,923)	46.65 (585)	53.16 (2,338)
Never	21.16 (1,196)	21.61 (271)	21.03 (925)

Bold face indicates statistical significance.

Table 2 shows unadjusted and adjusted associations of each neighborhood-level social and environmental exposure with baseline FEV₁%. Each adverse exposure was significantly associated with lower baseline FEV₁% in unadjusted models, except for air pollution, built environment, dissimilarity, and urban neighborhood. The largest effect sizes were for overall environmental justice (−11.43%, p < 0.001), overall social vulnerability (−12.29%, p < 0.001), neighborhood socioeconomic status (−11.96%, p < 0.001), and residential segregation measured by isolation (−9.44%, p < 0.001). Most of the associations remained significant after adjusting for patient-level demographic and clinical characteristics, with attenuation of the effect size. Greater walkability was associated with a lower baseline FEV₁% in unadjusted analyses, but with higher FEV₁% after adjusting for covariates. In addition, adverse built environment was associated with lower baseline FEV₁% only after adjusting for covariates, and urban residence was associated with a higher baseline FEV₁% only after adjusting for covariates. Adjusting for covariates fully accounted for the negative association of neighborhood minority status, overall environmental burden, transportation pollution, and food access with baseline FEV₁% observed in the unadjusted analyses. After adjusting for patient-level characteristics, air pollution was associated with a higher baseline FEV₁%. Similarly, water pollution was associated with higher baseline FEV₁%. Lastly, we examined interactions with race to see if these associations differed between Whites and Black patients. The association between hazardous and toxic sites and water pollution was significantly muted and non-significant for Black patients (data not shown). All other associations were similar between the two groups.

Table 2.

Associations of each neighborhood-level social or environmental exposure with baseline FEV₁%.

Social-Environmental Exposure	Unadjusted^a				Adjusted^b
Social-Environmental Exposure	Coeff	95% CI		p	Coeff	95% CI		p
Environmental Justice, overall	−11.43	−13.58	−9.28	<0.001	−6.81	−9.07	−4.54	<0.001
Social Vulnerability, overall	−12.29	−14.24	−10.33	<0.001	−7.58	−9.75	−5.41	<0.001
Theme 1: SES	−11.96	−13.91	−10.01	<0.001	−7.10	−9.30	−4.89	<0.001
Theme 2: Household	−6.08	−8.41	−3.76	<0.001	−4.56	−6.68	−2.43	<0.001
Theme 3: Minority	−8.37	−10.54	−6.20	<0.001	−1.87	−4.50	0.75	0.161
Theme 4: Housing	−3.61	−5.92	−1.29	0.002	−3.94	−6.03	−1.86	<0.001
Environmental Burden, overall	−4.38	−7.04	−1.72	0.001	−0.50	−3.11	2.12	0.710
Air pollution	0.79	−2.79	4.37	0.665	6.60	3.09	10.10	<0.001
Hazardous and toxic sites	−4.35	−6.43	−2.27	<0.001	−2.36	−4.31	−0.41	0.018
Built environment	−0.09	−2.97	2.80	0.953	−3.36	−6.01	−0.71	0.013
Transportation pollution	−5.66	−7.97	−3.35	<0.001	−1.72	−3.99	0.56	0.140
Water pollution	4.59	0.51	8.68	0.028	3.80	0.15	7.45	0.042
Health Vulnerability	−7.17	−9.06	−5.29	<0.001	−4.39	−6.20	−2.57	<0.001
Residential segregation
Dissimilarity	−1.30	−4.91	2.31	0.479	−2.42	−5.71	0.87	0.149
Interaction	9.71	7.46	11.96	<0.001	3.45	0.76	6.15	0.012
Isolation	−9.44	−11.57	−7.30	<0.001	−3.59	−6.16	−1.02	0.006
Urban	1.17	−0.17	2.51	0.086	2.77	1.51	4.03	<0.001
Walkability	−0.20	−0.36	−0.03	0.018	0.16	0.00	0.33	0.050
Food desert	−1.50	−2.78	−0.21	0.023	0.14	−1.16	1.44	0.835

Analyses control for the proportion of observations possible by observations seen.

Analyses additionally control for age, sex, race/ethnicity, marital status, obesity, number of chronic conditions, and smoking on baseline FEV₁%; and for sex, race/ethnicity, and obesity on annual change in FEV₁%.

Coeff, coefficient; FEV₁%, forced expiratory volume for 1 second; SES, socioeconomic status.

Bold face indicates statistical significance.

Table 3 shows the unadjusted and adjusted associations of each SEE with the annual change in FEV₁%. In the unadjusted models, adverse overall environmental justice (−0.68, p = 0.042) and transportation pollution (−0.72, p = 0.036) were associated with greater annual decline in FEV₁%, but were no longer significant in the adjusted models. Living in a food desert was associated with a greater annual decline in FEV₁% (−0.80, p = 0.018) in the adjusted models. The interactions between each social-environmental exposure and race were not significant, suggesting that these associations did not vary by race.

Table 3.

Associations of each neighborhood-level social or environmental exposure with annual change in FEV₁%.

Social-environmental exposure	Unadjusted^a				Adjusted^b
Social-environmental exposure	Coeff	95% CI		p	Coeff	95% CI		p
Environmental Justice, overall	−0.68	−1.34	−0.02	0.042	−0.71	−1.43	0.02	0.055
Social Vulnerability, overall	−0.51	−1.14	0.13	0.117	−0.51	−1.22	0.21	0.169
Theme 1: SES	−0.48	−1.11	0.15	0.134	−0.47	−1.20	0.25	0.201
Theme 2: Household	−0.18	−0.86	0.49	0.595	−0.11	−0.79	0.56	0.743
Theme 3: Minority	−0.59	−1.24	0.05	0.073	−0.75	−1.64	0.14	0.100
Theme 4: Housing	0.02	−0.63	0.67	0.954	−0.05	−0.70	0.60	0.880
Environmental Burden, overall	−0.55	−1.34	0.23	0.169	−0.51	−1.35	0.32	0.228
Air pollution	0.07	−1.01	1.14	0.904	0.30	−0.84	1.45	0.603
Hazardous and toxic sites	−0.27	−0.87	0.34	0.386	−0.25	−0.87	0.37	0.429
Built environment	−0.17	−1.00	0.66	0.691	−0.35	−1.20	0.51	0.426
Transportation pollution	−0.72	−1.40	−0.05	0.036	−0.67	−1.40	0.04	0.066
Water pollution	0.10	−1.04	1.24	0.867	0.09	−1.05	1.22	0.882
Health Vulnerability	−0.44	−1.00	0.13	0.130	−0.38	−0.96	0.21	0.208
Residential Segregation
Dissimilarity	1.01	−0.04	2.06	0.061	1.01	−0.04	2.07	0.059
Interaction	0.20	−0.48	0.87	0.566	−0.01	−0.92	0.90	0.985
Isolation	−0.12	−0.76	0.53	0.723	0.15	−0.72	1.02	0.742
Urban	−0.23	−0.62	0.17	0.258	−0.15	−0.56	0.26	0.486
Walkability	−0.02	−0.07	0.03	0.392	−0.01	0.07	0.04	0.693
Food desert	−0.25	−0.59	0.10	0.168	−0.44	−0.80	−0.08	0.018

Analyses control for the proportion of observations possible by observations seen.

Coeff, Coefficient; FEV₁%, forced expiratory volume for 1 second; SES, socioeconomic status.

Bold face indicates statistical significance.

Table 4 displays results from linear mixed models that assess the joint effect of all neighborhood social and environmental exposures, before and after adjusting for patient-level demographic and clinical characteristics. Neighborhood social vulnerability was significantly associated with a lower baseline FEV₁%, both before and after adjusting for covariates (unadjusted −12.31%, p < 0.001; adjusted −9.96%, p < 0.001). Residential segregation measured by isolation was associated with worse baseline FEV₁% in the unadjusted models (−3.31%, p = 0.021), but not after adjusting for patient-level covariates. Better walkability was associated with a higher baseline FEV₁% after accounting for patient-level covariates (0.22, p = 0.021). Residence in a food desert was associated with a greater annual decline in FEV₁% in both adjusted and unadjusted models (unadjusted −0.41%, p = 0.004; adjusted −0.39, p = 0.009). The interaction between food deserts and race was significant, indicating a significant negative association among White patients but no association among Black patients (data not shown).

Table 4.

Joint associations of multiple social or environmental exposures with baseline FEV₁% and annual change in FEV₁%.

Social-environmental exposures	Unadjusted for covariates^a				Adjusted for covariates^b
Social-environmental exposures	Coeff.	95% CI		p	Coeff.	95% CI		p
Baseline FEV₁%
Environmental Burden	−0.95	−3.74	1.84	0.503	−1.20	−3.94	1.54	0.390
Social Vulnerability	−12.31	−16.11	−8.51	<0.001	−9.96	−13.70	−6.22	<0.001
Health Vulnerability	2.45	−0.56	5.45	0.111	2.61	−0.32	5.55	0.081
Residential Segregation: Isolation	−3.31	−6.12	−0.50	0.021	0.72	−2.26	3.71	0.634
Walkability	0.08	−0.11	0.26	0.410	0.22	0.03	0.40	0.021
Urban	0.84	−0.80	2.48	0.316	1.29	−0.32	2.91	0.117
Annual change in FEV₁%
Food desert	−0.41	−0.70	−0.13	0.004	−0.39	−0.69	−0.10	0.009

Includes all social and environmental exposures, not adjusted for covariates.

Includes all social and environmental exposures, adjusted for covariates: age, sex, race/ethnicity, marital status, obesity, number of chronic conditions, and smoking on baseline FEV₁%; and for sex, race/ethnicity, and obesity on annual change in FEV₁%.

Coeff, Coefficient; FEV₁%, forced expiratory volume for 1 second.

Bold fact indicates statistical significance.

We also conducted a sensitivity analysis by excluding those who had an asthma diagnosis (26.5%, n = 1498). Supplemental Table A1 presents the baseline FEV₁%, and Supplemental Table A2 presents the change in FEV₁%. The results for baseline FEV₁% in this subset (n = 4154) mirror the results for baseline FEV₁% in the full sample (N = 5652). The results for FEV₁% change are very similar between the two samples, except that transportation pollution is associated with FEV₁% decline, whereas the association between food desert and FEV₁% decline is not significant in this subset. However, the association between these two measures and FEV₁% is not significantly different for those with an asthma diagnosis compared to those without an asthma diagnosis (p = 0.194 for transportation pollution, p = 0.922 for food desert). In addition, individuals with COPD without an asthma diagnosis had fewer observations than individuals with both COPD and an asthma diagnosis. Among those with COPD without asthma, 57% had only one observation compared to only 25% of those with both COPD and asthma.

Discussion

Using longitudinal data from a single health system clinical cohort over the course of 10 years, we examined the separate and joint contributions of multiple social and environmental exposures on baseline lung function and annual lung function decline in patients diagnosed with COPD. Adverse neighborhood conditions had a significant negative association with baseline lung function, while limited food access affected the annual rate of lung function decline. When exposures were considered jointly, the neighborhood’s social vulnerability was the largest contributor to lower baseline lung function, whereas living in a food desert was associated with greater lung function decline.

Black patients with COPD resided predominantly in urban settings and worse living environments characterized by adverse social and environmental exposures, with the exception of water pollution. These results add to multiple prior reports of racial disparities in harmful environmental exposures,⁶⁴ from industrial contamination and lack of green space to limited food access and substandard housing and living conditions.^65–67 Our data also show that, compared to White patients, Black patients lived in more segregated neighborhoods with high isolation from and little interaction with other racial groups. Several studies have highlighted the negative health effects of residential segregation. For example, among people with or at risk of COPD, urban Black residents in segregated neighborhoods had worse lung function and clinical disease severity and a higher rate of exacerbations than their counterparts in non-segregated neighborhoods.⁴⁷

In our cohort, residence in a food desert was associated with greater annual decline in lung function. This finding is not surprising. Nutritional deficiencies are a known risk factor for lung disease through chronic inflammation and compromised immune response pathways. For example, a systematic review reported that a Western-style diet is associated with an increased risk of COPD and an accelerated decline of pulmonary function.⁶⁸ Consumption of health foods (fruit, fish, dairy products) has been associated with less decline of FEV₁% in patients with COPD,⁶⁹ whereas vitamin D deficiency has been associated with a greater decline.⁷⁰

The natural, built, and social features of neighborhoods do not exist in isolation. Understanding how they interact with each other and with biological, genetic, and behavioral factors over the life course to affect disease onset, progression, and outcomes is of increasingly recognized importance.^71,72 Our study shows that, when multiple aspects of the environment – natural, built, and social – are considered simultaneously, the dominant factor for lung function in people with COPD is the socioeconomic status of their neighborhood. This conclusion is in agreement with a recent ecological analysis that identified social vulnerability as a primary driver of higher COPD prevalence rates in disadvantaged neighborhoods.⁷³

Our findings are congruent with a wealth of sociological literature on socioeconomic status as a “fundamental cause” of disease that puts people “at risk of risk.”⁷⁴ Indeed, poverty generates a host of proximal risk factors, including increased exposure to hazards in the natural, built, and social environments.^46,75–78 Since social factors put individuals at risk for multiple proximal factors, it is difficult to disentangle the mechanisms that link social factors to poor lung health. For example, low education limits one’s occupational choices and may increase occupational hazards, affects health literacy, limits adherence to prescribed treatments, and has implications for other health-related behaviors, such as diet, physical activity, and tobacco use. Low income, on the other hand, affects one’s housing options, living conditions, diet quality, schooling and transportation options, and health insurance coverage. On a neighborhood level, residence in a socioeconomically deprived area increases exposure to unhealthy food, tobacco, and liquor stores while limiting access to healthy food options, exposing residents to industrial and transportation-related pollution, crime, and neighborhood disorder, while limiting access to green space, recreational facilities, and essential services. In our study, each adverse neighborhood exposure was associated with a lower baseline FEV₁%, and most of the associations remained significant after controlling for patient-level characteristics. The negative effects were as large as 12.29% (overall social vulnerability), 11.96% (socioeconomic status), 11.43% (overall environmental justice), and 9.44% (residential segregation measured by isolation). Some of the known biological mechanisms linking these exposures to lung disease involve physiological stress response,^28–31,79 physiologic dysregulation and chronic inflammation,^80,81 and other complex pathogenic mechanisms including induction of oxidative stress and activation of inflammatory pathways.

A notable exception to the adverse exposures was air pollution, which showed a positive association with baseline FEV₁% after adjusting for patient-level covariates, in contrast to multiple prior research documenting that air pollution contributes to the incidence, prevalence, and mortality of lung diseases.^82,83 The counterintuitive results of our study are likely due to a number of factors, including the lack of precision of the air pollution data used in the analyses. The 2022 EJI calculates the air pollution score from levels of ozone, PM2.5, diesel particulate matter, and air toxics (hazardous air pollutants, such as benzene, dioxin, formaldehyde, and ethylene oxide). For ozone and PM2.5, data are derived from the EPA Air Quality System, which includes the mean annual percent of days where the maximum 8-hour average ozone concentration and the daily 24 h average PM2.5 concentration exceed the National Ambient Air Quality Standard, both averaged across 3 years (2018–2020). For diesel particulate matter and air toxics, data are obtained from the EPA AirToxScreen for 2019. All data come from EPA air quality monitors, which are both sparse and unequally distributed.⁸⁴ A recent national study highlighted the profound racial and ethnic disparities in the spatial distribution of the monitors, which results in sampling bias and potentially incorrect conclusions about the air quality.⁸⁵ Rural areas, comprising 37% of our sample, are outside the bounds of air quality monitoring networks altogether, leaving rural populations without available air quality data.⁸⁶ Therefore, integration of multiple data sources, including satellite data, is critical to fill gaps in air quality monitoring and generate more reliable air quality data.

The study has several limitations. First, our data are derived from a single health system, and no a priori power calculation is given, given the population-based design. Therefore, our findings may not be generalizable to other geographic regions and populations and may have limited power to assess associations within specific subgroups or interactions. Second, because we utilize clinically obtained spirometry, we cannot fully standardize how the measurement was obtained. However, all tests were done in the same pulmonary function testing lab, which should minimize variability. Data also suggest that while the diagnosis of airflow obstruction can differ between techniques, longitudinal outcomes are similar.^87,88 Further, we have no way of discerning the clinical indication for the spirometry, and it is possible that results were affected by differences in the participant’s clinical condition that prompted the clinician to order the test. Thus, it is also possible that some disparity exists between groups with regard to the indication that could affect our results. We also acknowledge differential longitudinal observation by asthma status, with implications for precision and potential follow-up selection bias. Finally, we used the 2022 EJI, which is calculated from data obtained in the 5-year period from 2018 to 2022, whereas spirometric data were collected between 2010 and 2020. Measuring the outcome prior to the exposure violates the epidemiologic principle of temporality⁸⁹ and further limits our ability to establish causation.

The main contribution of this work is the assessment of multiple types of exposures from the natural, built, and social environment to determine their separate and joint contributions to lung function and annual lung function decline in a real-world clinical sample of patients with a COPD diagnosis. Despite evidence for the effect of social and environmental exposures on lung health and healthcare use,^90,91 exposure data are not captured in the medical record and are seldom considered during disease management. Our findings can be used in risk prediction models and for the development of tools to assess exposures at the point of care for individualized disease management and post-hospital care plans.^92–94

Conclusion

Adverse neighborhood conditions have a negative association with baseline lung function, while limited food access is associated with greater annual lung function decline in the full cohort. When exposures are considered jointly, the neighborhood’s social vulnerability is the largest contributor to lower baseline lung function, whereas living in a food desert is associated with greater lung function decline.

Supplemental Material

sj-docx-1-tar-10.1177_17534666261436228 – Supplemental material for Contribution of social and environmental exposures to lung function decline in chronic obstructive pulmonary disease: longitudinal analysis of electronic medical records data

Supplemental material, sj-docx-1-tar-10.1177_17534666261436228 for Contribution of social and environmental exposures to lung function decline in chronic obstructive pulmonary disease: longitudinal analysis of electronic medical records data by Elizabeth H. Baker, Lucia D. Juarez, Ariann Nassel, Trisha M. Parekh, Crystal T. Stephens, Mark T. Dransfield and Gabriela R. Oates in Therapeutic Advances in Respiratory Disease

Supplemental Material

sj-docx-2-tar-10.1177_17534666261436228 – Supplemental material for Contribution of social and environmental exposures to lung function decline in chronic obstructive pulmonary disease: longitudinal analysis of electronic medical records data

Supplemental material, sj-docx-2-tar-10.1177_17534666261436228 for Contribution of social and environmental exposures to lung function decline in chronic obstructive pulmonary disease: longitudinal analysis of electronic medical records data by Elizabeth H. Baker, Lucia D. Juarez, Ariann Nassel, Trisha M. Parekh, Crystal T. Stephens, Mark T. Dransfield and Gabriela R. Oates in Therapeutic Advances in Respiratory Disease

Footnotes

Acknowledgements

None.

Declarations

ORCID iDs

Lucia D. Juarez

Crystal T. Stephens

Gabriela R. Oates

Supplemental material

Supplemental material for this article is available online.

References

Celli

Fabbri

Criner

, et al. Definition and nomenclature of chronic obstructive pulmonary disease: time for its revision. Am J Respir Crit Care Med 2022; 206(11): 1317–1325.

Liu

Carlson

Watson

, et al. Trends in the prevalence of chronic obstructive pulmonary disease among adults aged ⩾18 years–United States, 2011–2021. MMWR Morb Mortal Wkly Rep 2023; 72(46): 1250–1256.

Centers for Disease Control and Prevention. BRFSS web enabled analysis tool (WEAT). https://nccd.cdc.gov/weat/#/ (accessed 18 September 2024).

Grahn

Gustavsson

Andersson

, et al. Occupational exposure to particles and increased risk of developing chronic obstructive pulmonary disease (COPD): a population-based cohort study in Stockholm, Sweden. Environ Res 2021; 200: 111739.

Eisner

Blanc

Omachi

, et al. Socioeconomic status, race and COPD health outcomes. J Epidemiol Community Health 2011; 65(1): 26–34.

Disano

Goulet

Muhajarine

, et al. Social-economic status and rates of hospital admission for chronic disease in urban Canada. Can Nurse 2010; 106(1): 24–29.

Lewis

Clegg

Johnson

NJ.

Lung disease mortality in the United States: the National Longitudinal Mortality Study. Int J Tuberc Lung Dis 2009; 13(8): 1008–1014.

Gershon

Dolmage

Stephenson

, et al. Chronic obstructive pulmonary disease and socioeconomic status: a systematic review. COPD 2012; 9(3): 216–226.

Oates

Jackson

Partridge

, et al. Sociodemographic patterns of chronic disease: how the mid-south region compares to the rest of the country. Am J Prev Med 2017; 52(1S1): S31–S39.

10.

Pleasants

Riley

Mannino

DM.

Defining and targeting health disparities in chronic obstructive pulmonary disease. Int J Chron Obstruct Pulmon Dis 2016; 11: 2475–2496.

11.

Lowe

Make

Crapo

, et al. Association of low income with pulmonary disease progression in smokers with and without chronic obstructive pulmonary disease. ERJ Open Res 2018; 4(4): 00069–2018

12.

Schraufnagel

Blasi

Kraft

, et al. An official American Thoracic Society and European Respiratory Society policy statement: disparities in respiratory health. Eur Respir J 2013; 42(4): 906–915.

13.

Cosselman

Navas-Acien

Kaufman

JD.

Environmental factors in cardiovascular disease. Nat Rev Cardiol 2015; 12(11): 627–642.

14.

Kolak

Bhatt

Park

, et al. Quantification of neighborhood-level social determinants of health in the continental United States. JAMA Network Open 2020; 3(1): e1919928 (1–17).

15.

Powell-Wiley

Baumer

Baah

, et al. Social determinants of cardiovascular disease. Circ Res 2022; 130(5): 782–799.

16.

Holden

Lee

Hawcutt

, et al. The impact of poor housing and indoor air quality on respiratory health in children. Breathe (Sheff) 2023; 19(2): 230058.

17.

Wimalasena

Chang-Richards

, et al. Housing risk factors associated with respiratory disease: a systematic review Int J Environ Res Public Health. 2021; 18(6): 2815.

18.

Kim

Rossen

Stierman

, et al. Federal housing assistance and chronic disease among US adults, 2005–2018. Prev Chronic Dis. 2023; 20: E111.

19.

Krieger

Higgins

DL.

Housing and health: time again for public health action. Am J Public Health 2002; 92(5): 758–768.

20.

Fullilove

3rd . What’s housing got to do with it? Am J Public Health 2000; 90(2): 183–184.

21.

Digenis-Bury

Brooks

Chen

, et al. Use of a population-based survey to describe the health of Boston public housing residents. Am J Public Health 2008; 98(1): 85–91.

22.

Fertig

Reingold

DA.

Public housing, health and health behaviors: is there a connection?

J Policy Anal Manage Autumn 2007; 26(4): 831–859.

23.

Sallis

Floyd

Rodríguez

, et al. Role of built environments in physical activity, obesity, and cardiovascular disease. Circulation 2012; 125(5): 729–737.

24.

Smith

Hosking

Woodward

, et al. Systematic literature review of built environment effects on physical activity and active transport – an update and new findings on health equity. Int Behav Nutr Phys Act 2017; 14(1): 158.

25.

Gomez

Shariff-Marco

DeRouen

, et al. The impact of neighborhood social and built environment factors across the cancer continuum: current research, methodological considerations, and future directions. Cancer 2015; 121(14): 2314–2330.

26.

Kondo

Fluehr

McKeon

, et al. Urban green space and its impact on human health. Int J Environ Res Public Health 2018; 15(3): 445.

27.

South

Kondo

Cheney

, et al. Neighborhood blight, stress, and health: a walking trial of urban greening and ambulatory heart rate. Am J Public Health 2015; 105(5): 909–913.

28.

Gracia

Neighborhood disorder. In: Michalos

(ed). Encyclopedia of quality of life and well-being research. Dordrecht: Springer Netherlands, 2014: 4325–4328.

29.

Butel

Braun

KL.

The role of collective efficacy in reducing health disparities: a systematic review. Fam Community Health 2019; 42(1): 8.

30.

Cohen

Finch

Bower

, et al. Collective efficacy and obesity: the potential influence of social factors on health. Soc Sci Med 2006; 62(3): 769–778.

31.

Theall

Chaparro

Denstel

, et al. Childhood obesity and the associated roles of neighborhood and biologic stress. Prev Med Rep 2019; 14: 100849.

32.

Alcock

White

Cherrie

, et al. Land cover and air pollution are associated with asthma hospitalisations: a cross-sectional study. Environ Int 2017; 109: 29–41.

33.

Tischer

Gascon

Fernandez-Somoano

, et al. Urban green and grey space in relation to respiratory health in children. Eur Respir J 2017; 49(6): 1502112.

34.

Alexander

Currie

Is it who you are or where you live? Residential segregation and racial gaps in childhood asthma. J Health Econ 2017; 55: 186–200.

35.

Anderson KM, Institute of Medicine (U.S.). Roundtable on the promotion of health equity and the elimination of health disparities. How far have we come in reducing health disparities? Progress since 2000: workshop summary. Washington, DC: National Academies Press, 2012: xii, 106 pages.

36.

Franco

Diez Roux

Glass

, et al. Neighborhood characteristics and availability of healthy foods in Baltimore. Am J Prev Med 2008; 35(6): 561–567.

37.

Moore

Diez Roux

AV.

Associations of neighborhood characteristics with the location and type of food stores. Am J Public Health 2006; 96(2): 325–31.

38.

Miller

Middendorf

Wood

Food availability in the heartland: exploring the effects of neighborhood racial and income composition. Rural Sociol 2015; 80(3): 340–361.

39.

Jennings

Baptiste

Osborne Jelks

, et al. Urban green space and the pursuit of health equity in parts of the United States. Int J Environ Res Public Health 2017; 14(11): 1432

40.

Sampson

Morenoff

Raudenbush

Social anatomy of racial and ethnic disparities in violence. Am J Public Health 2005; 95(2): 224–232.

41.

Salas-Wright

Nelson

Vaughn

, et al. Trends in fighting and violence among adolescents in the United States, 2002–2014. Am J Public Health 2017; 107(6): 977–982.

42.

Bogar

Beyer

KM.

Green space, violence, and crime: a systematic review. Trauma Violence Abuse 2016; 17(2): 160–171.

43.

Han

Cohen

Derose

, et al. Violent crime and park use in low-income urban neighborhoods. Am J Prev Med 2018; 54(3): 352–358.

44.

Levy

Quiros-Alcala

Fabian

, et al. Established and emerging environmental contributors to disparities in asthma and chronic obstructive pulmonary disease. Curr Epidemiol Rep 2018; 5(2): 114–124.

45.

Balmes

JR.

Place matters: residential racial segregation and chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2021; 204(5): 496–498.

46.

Humphrey

Lindstrom

Barton

, et al. Social and environmental neighborhood typologies and lung function in a low-income, urban population. Int J Environ Res Public Health 2019; 16(7): 1133

47.

Woo

Brigham

Allbright

, et al. Racial segregation and respiratory outcomes among urban black residents with and at risk of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2021; 204(5): 536–545.

48.

Shen

Yang

, et al. Association of walkability and fine particulate matter with chronic obstructive pulmonary disease: a cohort study in China. Sci Total Environ 2023; 858(Pt. 1): 159780.

49.

Burkes

Gassett

Ceppe

, et al. Current, former investigators of the SPIROMICS reading centers. Rural residence and chronic obstructive pulmonary disease exacerbations. Analysis of the SPIROMICS Cohort. Ann Am Thorac Soc 2018; 15(7): 808–816.

50.

Moughames

Woo

Galiatsatos

, et al. Disparities in access to food and chronic obstructive pulmonary disease (COPD)-related outcomes: a cross-sectional analysis. BMC Pulm Med 2021; 21(1): 139.

51.

Miller

Hankinson

Brusasco

, et al. Standardisation of spirometry. Eur Respir J 2005; 26(2): 319–338.

52.

Graham

Steenbruggen

Miller

, et al. Standardization of spirometry 2019 update. An official American thoracic society and European respiratory society technical statement. Am J Respir Crit Care Med 2019; 200(8): e70–e88.

53.

Bowerman

Bhakta

Brazzale

, et al. A race-neutral approach to the interpretation of lung function measurements. Am J Respir Crit Care Med 2023; 207(6): 768–774.

54.

Agency for Toxic Substances and Disease Registry. Environmental Justice Index (EJI). U.S. Department of Health and Human Services, https://www.atsdr.cdc.gov/place-health/php/eji/index.html (2022).

55.

U.S. Environmental Protection Agency. National walkability index user guide and methodology. https://www.epa.gov/smartgrowth/national-walkability-index-user-guide-and-methodology (accessed 8 January 2025).

56.

Mahajan

Larkins-Pettigrew

Racial demographics and COVID-19 confirmed cases and deaths: a correlational analysis of 2886 US counties. J Public Health 2020; 42: 445–447.

57.

Allen

Burgess

Davidson

, et al. More reliable inference for the dissimilarity index of segregation. Econom J 2015; 18(1): 40–66.

58.

U.S. Department of Agriculture. Rural-urban commuting area codes. Updated 2020. https://www.Ers.usda.gov/data-products/rural-urban-commuting-area-codes/ (accessed 14 May 2023)

59.

U.S. Department of Agriculture, Economic Research Service. Rural-Urban Continuum Codes. Updated January 2024. https://www.ers.usda.gov/data-products/rural-urban-continuum-codes/ (accessed 8 Sept 2024)

60.

Plewis

Auxiliary variables and adjustments for missingness in longitudinal studies. Cathie March Centre Census Survey Res 2011; 138810.

61.

Wang

Clustered longitudinal data analysis. Louisville: University of Louisville, 2008.

62.

DeMaris

Regression with social data: modeling continuous and limited response variables. Hoboken, NJ: John Wiley & Sons, 2004.

63.

von Elm

Altman

Egger

, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 2008; 61(4): 344–349.

64.

Jbaily

Zhou

Liu

, et al. Air pollution exposure disparities across US population and income groups. Nature 2022; 601(7892): 228–233.

65.

Braubach

Fairburn

Social inequities in environmental risks associated with housing and residential location–a review of evidence. Eur J Public Health 2010; 20(1): 36–42.

66.

Solomon

Morello-Frosch

Zeise

, et al. Cumulative environmental impacts: science and policy to protect communities. Annu Rev Public Health 2016; 37: 83–96.

67.

Ash

Fetter

TR.

Who lives on the wrong side of the environmental tracks? Evidence from the EPA’s risk-screening environmental indicators model. Soc Sci Q. 2004; 85(2): 441–462.

68.

van Iersel

LEJ

Beijers

Gosker

, et al. Nutrition as a modifiable factor in the onset and progression of pulmonary function impairment in COPD: a systematic review. Nutr Rev 2022; 80(6): 1434–1444.

69.

Hanson

Sayles

Rutten

, et al. The association between dietary intake and phenotypical characteristics of COPD in the ECLIPSE cohort. Chronic Obstr Pulm Dis 2014; 1(1): 115–124.

70.

Burkes

Ceppe

Doerschuk

, et al. SPIROMICS Investigators. Associations among 25-hydroxyvitamin D levels, lung function, and exacerbation outcomes in COPD: an analysis of the SPIROMICS cohort. Chest 2020; 157(4): 856–865.

71.

Juarez

Hood

Song

, et al. Use of an exposome approach to understand the effects of exposures from the natural, built, and social environments on cardio-vascular disease onset, progression, and outcomes. Front Public Health 2020; 8: 379.

72.

Vineis

Robinson

Chadeau-Hyam

, et al. What is new in the exposome? Environ Int 2020; 143: 105887.

73.

Mein

Liu

Marinacci

, et al. Neighborhood exposome and prevalence of asthma and COPD in the United States. Ann Am Thorac Soc 2025; 22:797–801

74.

Phelan

Link

Tehranifar

Social conditions as fundamental causes of health inequalities: theory, evidence, and policy implications. J Health Soc Behav 2010; 51 Suppl: S28–S40.

75.

LaVeist

Pollack

Thorpe

Jr , et al. Place, not race: disparities dissipate in southwest Baltimore when blacks and whites live under similar conditions. Health Aff (Millwood) 2011; 30(10): 1880–1887.

76.

Massey

Gross

Shibuya

Migration, segregation, and the geographic concentration of poverty. Am Soc Rev 1994; 59: 425–445.

77.

Lakerveld

Mackenbach

The upstream determinants of adult obesity. Obes Facts 2017; 10(3): 216–222.

78.

Pool

Carnethon

Goff

, et al. Longitudinal associations of neighborhood-level racial residential segregation with obesity among blacks. Epidemiol 2018; 29(2): 207–214.

79.

Pearlin

Bierman

Current issues and future directions in research into the stress process. In: Aneshensel

Phelan

Bierman

(eds). Handbook of the sociology of mental health. Dordrecht, Netherlands: Springer, 2013: 325–340.

80.

Davis

Latimer

Rice

Biomarkers of stress and inflammation in children. Biol Res Nurs 2023; 25: 10998004231168805.

81.

Noushad

Ahmed

Ansari

, et al. Physiological biomarkers of chronic stress: a systematic review. Int J Health Sci (Qassim) 2021; 15(5): 46–59.

82.

Canaday

Georas

Croft

DP.

Examining the impact of air pollution, climate change, and social determinants of health on asthma and environmental justice. Curr Opin Pulm Med 2024; 30(3): 276–280.

83.

Franchini

Mannucci

Harari

, et al. The health and economic burden of air pollution. Am J Med 2015; 128(9): 931–932.

84.

Technical information about fused air quality surface using downscaling tool: metadata description. U.S. Environmental Protection Agency (July 2016). https://www.epa.gov/hesc/rsig-related-downloadable-data-files.

85.

Kelly

Cova

Debbink

, et al. Racial and ethnic disparities in regulatory air quality monitor locations in the US. JAMA Netw Open 2024; 7(12): e2449005.

86.

Roque

Andrews

Santos-Lozada

AR.

Identifying air quality monitoring deserts in the United States. Proc Natl Acad Sci U S A 2025; 122(17): e2425310122.

87.

Mannino

Diaz-Guzman

Buist

Pre- and post-bronchodilator lung function as predictors of mortality in the Lung Health Study. Respir Res 2011; 12(1): 136.

88.

Malinovschi

Zhou

Andersson

, et al. Consequences of using post- or prebronchodilator reference values in interpreting spirometry. Am J Respir Crit Care Med 2023; 208(4): 461–471.

89.

Hill

AB.

The environment and disease: association or causation? 1965. J R Soc Med 2015; 108(1): 32–37.

90.

Kind

Jencks

Brock

, et al. Neighborhood socioeconomic disadvantage and 30-day rehospitalization: a retrospective cohort study. Ann Intern Med 2014; 161(11): 765–774.

91.

Kind

AJH

Nerenz

Area deprivation index predicts readmission risk at an urban teaching hospital. Am J Med Qual 2018; 33(5): 493–501.

92.

Kansagara

Englander

Salanitro

, et al. Risk prediction models for hospital readmission: a systematic review. JAMA 2011; 306(15): 1688–1698.

93.

Joynt

Jha

AK.

A path forward on Medicare readmissions. N Engl J Med 2013; 368(13): 1175–1177.

94.

Fleegler

Lieu

Wise

, et al. Families’ health-related social problems and missed referral opportunities. Pediatrics 2007; 119(6): e1332–41.