Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Jan 1.
Published in final edited form as: Soc Sci Med. 2023 Nov 26;340:116448. doi: 10.1016/j.socscimed.2023.116448

Racial/Ethnic Disparities in Exposure to neighborhood violence and Lung Cancer Risk in Chicago

Sage J Kim 1,*, Caroline Kery 2, Jinghua An 3, Georgiy Bobashev 4, Alicia K Matthews 5
PMCID: PMC10836639  NIHMSID: NIHMS1951509  PMID: 38043441

Abstract

Background:

Despite the lower prevalence and frequency of smoking, Black adults are disproportionately affected by lung cancer. Exposure to chronic stress generates heightened immune responses, which creates a cell environment conducive to lung cancer development. Residents in poor and segregated neighborhoods are exposed to increased neighborhood violence, and chronic exposure to violence may have downstream physiological stress responses, which may explain racial disparities in lung cancer in predominantly Black urban communities.

Methods:

We utilized retrospective electronic medical records of patients who underwent a screening or diagnostic test for lung cancer at an academic medical center in Chicago to examine the associations between lung cancer diagnosis and individual characteristics (age, gender, race/ethnicity, and smoking status) and neighborhood-level homicide rate. We then used a synthetic population to estimate the neighborhood-level lung cancer risk to understand spatial clusters of increased homicide rates and lung cancer risk.

Results:

Older age and former/current smoking status were associated with increased odds of lung cancer diagnosis. Hispanic patients were more likely than White patients to be diagnosed with lung cancer, but there was no statistical difference between Black and White patients in lung cancer diagnosis. The odds of being diagnosed with lung cancer were significantly higher for patients living in areas with the third and fourth quartiles of homicide rates compared to the second quartile of homicide rates. Furthermore, significant spatial clusters of increased lung cancer risk and homicide rates were observed on Chicago’s South and West sides.

Conclusions:

Neighborhood violence was associated with an increased risk of lung cancer. Black residents in Chicago are disproportionately exposed to neighborhood violence, which may partially explain the existing racial disparity in lung cancer. Incorporating neighborhood violence exposure into lung cancer risk models may help identify high-risk individuals who could benefit from lung cancer screening.

1. Introduction

Lung cancer is the leading cause of cancer deaths in the United States (U.S.), accounting for nearly 25% of all cancer deaths (American Cancer Society, 2023). More than 238,000 new cases and 127,000 deaths from lung cancer are estimated to occur in 2023. Despite advances in screening and treatment, racial inequities in lung cancer incidence and mortality persist (Sin, 2017). Nationally, the incidence of lung cancer is higher in Blacks (59.0 per 100,000) than Whites (52.1 per 100,000). In Chicago, the third most racially segregated city in the U.S. (Hunt & Whitman, 2015), the lung cancer incidence rate difference between Blacks and Whites (83.5 vs. 64.7 per 100,000) and the mortality rate gap (60.5 vs. 39.7 per 100,000) is even more pronounced than at the national level (Chicago Health Atlas, 2023).

Smoking is the leading preventable risk factor for lung cancer. Although smoking prevalence rates among Black adults are similar to those of White adults (Cornelius et al., 2022), the number of cigarettes smoked per day is much lower for Black (17.1) than for White (21.2) smokers (Choquet et al., 2021). The higher rates of lung cancer incidence and mortality among Blacks despite the lower frequency of smoking, often referred to as the African American Smoking Paradox (Feigelman & Lee, 1995), suggest that factors beyond individual behavioral risks may contribute to racial disparities in lung cancer.

Studies have shown that where people live affects cancer health outcomes (Adie et al., 2020; Krieger et al., 2020). More than half of the variance in one’s health is shown to be determined by one’s zip code (Lawry, 2022). The social determinants of health encompass a complex interplay of social, economic, and environmental factors that significantly influence an individual’s health outcomes and access to healthcare services (Braveman & Gottlieb, 2014). Examining the social determinants of health is critical in reducing health inequities (Penman-Aguilar et al., 2016). Disparities in cancer incidence, prevalence, and survival rates can be attributed to these determinants, reflecting unequal exposure to risk and access to early detection, diagnosis, treatment, and supportive care services. Neighborhood social, economic, and built environmental conditions contribute to cancer disparities in multiple cancer types, including lung, breast, and colorectal cancer (Adie et al., 2020; Danos et al., 2018; Saini et al., 2019). Understanding and addressing these determinants enables the development of more effective interventions and policies, striving for equitable cancer prevention and control that considers the broader context of individuals’ lives (Dankwa-Mullan et al., 2010).

There is a plethora of research findings documenting the relationship between neighborhood disadvantage and poor cancer outcomes (Adie et al., 2020; Danos et al., 2018; Krieger et al., 2020). However, the extant literature is often unclear about the specific paths linking social exposure to biological consequences (Mullan Harris & McDade, 2018). One of the social epigenetic mechanisms of social-biological interplay is downstream stress responses (Gudsnuk & Champagne, 2012). Allostatic load has been the key concept explaining how chronic stress affects physiological systems, leading to detrimental health outcomes (Geronimus et al., 2006). High neighborhood economic deprivation has been documented to be linked to stress response in some studies (Kapuku et al., 2002; Koss & Gunnar, 2017), which others argue that these studies were confounded because neighborhood violence level was not accounted for in their analysis (Busso et al., 2017; Peckins et al., 2020). Chronic exposure to violence is a distinct social stress that has been shown to influence mental and physical health (Goldmann et al., 2011; Peckins et al., 2020). One hypothesized path through which exposure to violence produces physiological stress responses is by dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, which is linked to glucocorticoid (GC) production and pro-inflammatory immune responses (Heath et al., 2013). In addition, a prostaglandin degrading enzyme, 15-hydroxyprostaglandin dehydrogenase (15-PGDH), regulates the ratio of cortisol to cortisone and often decreased 15-PGDH is shown in lung cancer cells compared with adjacent normal lung tissue (Hughes et al., 2008). In lung cancer, 15-PGDH acts as a tumor suppressor (Ding et al., 2005). Simultaneously, the cyclooxygenase 2 (COX-2) enzyme is overexpressed in non-small cell lung cancer (Sandler & Dubinett, 2004). Decreased 15-PGDH expression was correlated with increased COX-2 expression, and increased COX-2 expression frequently occurs in precursor regions of lung cancer (Hosomi et al., 2000).

Concerning exposure to neighborhood violence, violent crime results in a severe impact on the health of individuals (Heissel et al., 2018; Sharkey & Sampson, 2015). Violent crimes are offenses that involve force or threat of force, which include forcible rape, robbery, aggravated assault, and homicide (Federal Bureau of Investigation, 2023; National Institute of Justice, 2023), while nonviolent crimes are defined as “property, drug, and public order offenses which do not involve a threat of harm or an actual attack upon a victim,” such as theft and larceny (Durose & Mumola, 2004). Homicide is often used as an indicator of neighborhood violence in studies (County Health Rankings & Roadmaps, 2023) because homicide cases are more accurately reported in crime statistics (Healthy People 2030, 2023; Inter-American Development Bank, 1999) compared to other violent crime types, such as rape or robbery. For example, the Centers for Disease Control and Prevention (CDC) keeps track of homicide deaths as part of the national violent death reporting system (Centers for Disease Control and Prevention, 2023). In Chicago, one of the most segregated cities in the U.S., the crime rate is three times higher (195 per 1,000) in predominantly Black communities than in White communities (65 per 1,000). Residents in high-crime areas, especially those characterized by violent crime, are likelier to experience elevated acute and chronic stress (Heissel et al., 2018; Hureau et al., 2022). Research on exposure to violence has shown significant effects on various health outcomes, including mental and physical health (Browning et al., 2017; Fowler et al., 2009). Furthermore, studies document that stress-related immune responses contribute to many chronic health conditions, including cancer (Bagatini et al., 2018; Gonzalez et al., 2018).

Social stressors, particularly chronic exposure to crime and violence, may contribute to increased lung cancer risk in Black communities (de Groot et al., 2018; Gomez et al., 2015), potentially explaining lung cancer disparities (Lerner et al., 2018). Thus, this study examined the independent effect of neighborhood stress exposure on lung cancer risk among predominantly racial/ethnic minority patient samples in a large urban city. We also estimated the population level lung cancer risk to explore spatial clusters of elevated lung cancer risk within the city. We hypothesized that the neighborhood-level violent crime rate (homicide rate) is associated with the risk of developing lung cancer above and beyond individual-level characteristics such as age, gender, and smoking history and that disproportionate exposure to neighborhood violence among Black residents explains racial disparities in lung cancer.

2. Methods

This retrospective observational study used deidentified electronic medical record (EMR) data from a large academic medical center in Chicago. We first examined the independent contribution of the homicide rates at the patient’s residential location on the likelihood of being diagnosed with lung cancer using EMRs. We then estimated the neighborhood-level risk of lung cancer to explore spatial clusters of increased risk of lung cancer diagnosis by incorporating the RTI Synthetic Household Population (Rineer, 2018) for the city of Chicago, representing Black, White, and Hispanic residents aged over 40 years. The study protocol was reviewed and approved by the University of Illinois Chicago Institutional Review Board (IRB: #2019–0873).

2.1. Identification of patient sample

To create an analytic dataset of high-risk patients, we retrieved EMRs for all patients who lived in Chicago, were over the age of 40 years, and underwent lung cancer screening or diagnostic procedures between January 01, 2013, and December 31, 2018. We retrieved 8,029 unique patient records meeting the criteria. Of those, we excluded 831 cases whose race/ethnicity information was missing, “declined,” “unavailable,” or designated “Other” from the analytic sample. An additional 12 cases that did not have lung cancer diagnosis information were also excluded from the analysis. The final analytic dataset included a total of 7,186 unique patients.

We then obtained the Current Procedural Terminology (CPT) codes, including lung biopsy, imaging, and surgery, and the International Classification of Diseases (ICD)-10 codes, including small, large, and squamous cell carcinoma. ICD-10 codes indicate malignant neoplasm of the lung. Using ICD codes, patients diagnosed with lung cancer during the study period were identified as “being diagnosed with lung cancer.” Patients with ICD codes indicating secondary malignant neoplasm of the lung and a history of malignant carcinoid tumor of the lung were excluded from the lung cancer diagnosis designation. In addition, we appended the patient’s age, race/ethnicity (Black, White, and Hispanic), gender (male and female), smoking history (current, former, and never smoker), and residential address.

2.2. Identification of patient neighborhood characteristics

We geocoded patient residential addresses to append census tract-level neighborhood characteristics. First, to quantify exposure to violence, we used homicide death records from the Cook County Medical Examiner Office (MEO) Case Archive between 2013 and 2018 to compute the homicide rate per 10,000 residents to align with the same time period (January 01, 2013 – December 31, 2018) as the obtained EMR data. The MEO archive contains information about deaths in Cook County, Illinois, within which Chicago is located. Although not all deaths that occur in Cook County are reported to the MEO, human deaths that fall within the categories of criminal violence, suicide, accident, injury or toxic agents from employment, penal institutions, suspicious or unusual circumstances, and unattended by a licensed physician, are investigated by the MEO (Cook County Government, 2022). We also used the Census Bureau American Community Survey (ACS) 5-year estimates for 2013–2017 to identify census tract-level racial/ethnic compositions (% Black, % White, and % Hispanic residents). We also included census tract level poverty rate (% of residents living below the federal poverty line) to examine whether general neighborhood disadvantage instead of violence may have a stronger effect on lung cancer risk. Finally, we obtained community-level adult smoking rates from the Chicago Health Atlas, the most widely used Chicago health data portal (Chicago Health Atlas, 2022).

2.3. Estimating the population level lung cancer risk

To understand how increased homicide rates and lung cancer risk spatially cluster, we estimated the neighborhood-level risk of developing lung cancer. We used Chicago’s RTI Synthetic Population dataset, representing Black, White, and Hispanic residents aged 40 years and older (Rineer, 2018). A synthetic population is a synthesized, geospatially explicit individual-level data that matches the actual marginal distributions of the population, such as age, race/ethnicity, and gender, of each census block group (Wheaton et al., 2009). The RTI synthetic population dataset was created from the U.S. Census Bureau, American Community Surveys data at the census block group level, along with the Public Use Microdata Sample (PUMS) data that are 5% of actual responses of Census long-form questionnaires, thus retaining individual records such as household relationships, individual age, gender, and race/ethnicity (Cajka et al., 2010). Compared to the aggregate data provided by the census, the synthetic population datasets match over 95% of the census-generated data. (RTI International, 2022).

These steps resulted in a synthetic dataset including nearly 1 million people representing the Chicago population and accounting for all individuals residing in Chicago, including patients who visited the hospital, from which we obtained EMRs. Using models from the National Household Survey on Drug Use and Health (NSDUH), a representative national survey (Substance Abuse and Mental Health Services Administration, 2023), we appended the adult smoking rate and the neighborhood homicide rate to persons represented in the synthetic population according to geolocation. This approach allowed us to account for all individuals residing in Chicago, rather than patients who visited the hospital, from which we obtained EMRs, to examine the association between exposure to neighborhood stressors, high levels of homicide rates in this case, and the estimated lung cancer risk at the population level.

The odds of being diagnosed with lung cancer for explanatory variables from the EMR-based regression results were then assigned to persons in the synthetic population according to demographic and neighborhood characteristics. Using this synthetic population dataset, we estimated the population level lung cancer risk at the census tract level to explore spatial clusters and neighborhood-level burden lung cancer risk.

2.4. Analysis

First, to summarize differences in the characteristics of the patients by race/ethnicity, we conducted a descriptive analysis. Chi-square tests were used to compare the three race/ethnicity groups (White, Black, and Hispanic) on the lung cancer diagnosis, age category (ages 40 and 50 years, 50 and 60 years), gender (male vs. female), and smoking history (never vs. former smoker and current smoker). ANOVA tests were used to compare continuous variables (age, the homicide rate, and the percentage of poverty) by race/ethnicity. The Census tract-level homicide rate and % poverty variables were grouped into quartiles with equal cases for this analysis.

Second, logistic regression was used to explain lung cancer diagnosis, with individual measures (age, gender, race/ethnicity, and smoking history) and neighborhood-level exposures (homicide rate and % poverty). Due to the concern that the homicide rate might represent a more general neighborhood disadvantage, three logistic regression models were fitted to explore the associations between the homicide rate and neighborhood poverty rate with lung cancer diagnosis. Model I is a model with the homicide rate ordinal variable only. Model II is with the % poverty quartile variable only. Model III includes both homicide rate and % poverty variables together. All three models also accounted for age, gender, race/ethnicity, and smoking history variables.

Third, we also performed a generalized structural equation model to explore paths between a binary outcome (lung cancer diagnosis) and observed factors (race/ethnicity, gender, age, smoking status, neighborhood homicide, and poverty). We fitted a binomial distribution with a logit function to accommodate the binary outcome. While structural equation analysis does not necessarily mean causal relationships among variables, the variables used in this structural equation help support the directions of the proposed paths. For example, gender, the homicide rate, and % poverty may influence one’s smoking behavior, but smoking behavior will not influence one’s gender, or neighborhood homicide or poverty rates. Similarly, one’s age may affect the likelihood of developing lung cancer, but certainly not the other way around.

Fourth, to estimate the population level lung cancer risk, we appended the odds ratio of explanatory variables from the logistic regression to each case in the synthetic population representing Chicago residents. We also estimated the probability of smoking from the NSDUH based on individual demographics for each person in the synthetic population. Because the EMR-based regression results reflect hospital patients with a higher prevalence of lung cancer than the general population, we adjusted the results for the synthetic population. We estimated the probability that a person from the general population would appear in the hospital for each combination of individual characteristics, such as smoking, age, gender, and race/ethnicity. We used the inverse of these probabilities as analytic weights and weighted regression to estimate regression coefficients linking individual characteristics and lung cancer. We rescaled the intercept to reflect population-level prevalence. Because of the random assignments of smoking outcomes based on model-based probabilities, we replicated the estimation process 500 times to estimate 95% simulation uncertainty intervals. The final estimated probability of developing lung cancer was computed by averaging these 500 estimates. We then geographically mapped the estimated lung cancer probability and clusters at the census tract level.

Stata 17 was used for descriptive statistics, logistic regression, and generalized structural equation model (Stata, 2020). Statistical significance was tested at p<.05. ArcGIS Desktop was used for geographic mapping and hotspot analysis (Esri, 2023).

3. Results

3.1. Sample characteristics

Table 1 summarizes sample characteristics. More than 64% of the patients were Black, 22.4% were Hispanic, and 13.2% were White. The mean age of Black patients (61.0 years) was significantly younger than White (62.6 years) and Hispanic patients (63.2 years). Substantially more Black patients were female (60.8%) compared with White (44.5%) and Hispanic (59.7%) patients. Hispanic patients were more likely to be never smokers (49.6%) compared with White (31.2%) and Black patients (31.3%). All Chi-square comparisons of individual characteristics by race/ethnicity were statistically significant at p <.001.

Table 1.

Characteristics of patients by race/ethnicity

Total White Black Hispanic p
N 7,186 951 4,617 1,618
% (100.0) (13.2) (64.2) (22.5) -
Individual characteristics
% Lung cancer 9.8 10.3 11.0 6.1 <.001
Age
 Mean 61.8 62.6 61.0 63.2 <.001
 40–50 years 16.0 12.7 16.8 15.7
 50–60 years 28.7 29.2 29.6 25.8 <.001
 >60 years 55.3 58.1 53.7 58.5
Gender
 Male 43.8 55.5 39.2 50.3
 Female 56.2 44.5 60.8 49.7 <.001
Smoking
 Never smoked 35.3 31.2 31.3 49.6
 Former smoker 29.6 31.2 28.7 31.3 <.001
 Current smoker 35.1 37.6 40.1 19.1
Neighborhood characteristics
Homicide rate/100,000
 Mean 195.4 59.6 266.4 71.4 <.001
 1st quartile 25.6 61.6 11.0 45.9
 2nd quartile 24.6 25.1 19.4 39.2
 3rd quartile 24.5 8.8 32.2 11.6 <.001
 4th quartile 25.4 4.4 37.4 3.2
% Poverty
 Mean 25.5 15.8 29.5 19.8 <.001
 1st quartile 24.7 57.9 15.1 32.6
 2nd quartile 25.4 25.3 19.7 41.8
 3rd quartile 24.8 10.1 30.9 16.1 <.001
 4th quartile 25.1 6.6 34.3 9.5
Open in a new tab

Neighborhood homicide rates ranged from 0 to 1,311 per 100,000 population, with a mean of 197.5 per 100,000 (SD=201.4/100,000). Homicide rates were then grouped into quartiles of an equal number of cases for this analysis. The mean neighborhood homicide rate was significantly higher for Black patients (266.4 per 100,000 population) than for White (59.6/100,000) and Hispanic (71.4/100,000) patients. Less than 15% of White (14.8%) and 13.2% of Hispanic patients lived in neighborhoods of the upper 50th percentile homicide rates; however, close to 69.6% of Black patients resided in neighborhoods in the upper 50th percentile homicide rates (p<.001).

Neighborhood poverty levels ranged from 0.3% to 74.9%, with a mean of 25.5% (SD=13.0%). The poverty rate was also grouped into quartiles. Similar to the homicide rate, there was significant racial/ethnic disparity in the distribution of neighborhood poverty level: 16.7% of White, 25.6% of Hispanic, and 65.2% of Black patients resided in areas in the top 50th percentile poverty rate (p<.001).

Table 2 summarizes the characteristics of patients by lung cancer diagnosis. Of the patients in the analytic sample, 9.8% were diagnosed with lung cancer. The mean age of patients with lung cancer diagnosis (65.6 years) was higher compared to the mean age of those without lung cancer diagnosis (61.4 years) with p<.001. Black patients accounted for 64.2% of the overall sample, but among those with a lung cancer diagnosis, 72.1% were Black. On the other hand, 22.5% of the patients in the analytic sample were Hispanics, while 14.0% of lung cancer patients were Hispanics. Overall, 13.2% of the total sample were White, compared with 13.9% of lung cancer patients being White (p<.001). Those who never smoked were less likely (17.9%), compared with former (43.5%) and current smokers (38.6%), to be diagnosed with lung cancer (p<.001). However, there was no statistically significant difference in lung cancer diagnosis between male and female patients (p<.083). Additionally, a significantly larger proportion of lung cancer patients (57.5%), compared with those who were not diagnosed with lung cancer (49.0%), were living in neighborhoods in the upper 50th percentile for homicide rates (p<.001). On the other hand, there was no statistically significant difference in the proportion of those living in the upper 50th percentile of % poverty (p<.243) between those who had lung cancer (52.8%) and those who did not (49.5%).

Table 2.

Characteristics of patients by lung cancer diagnosis

Lung Cancer
Total Yes No p
N 7,186 706 6,480
% (100) (9.8) (90.2) -
Individual characteristics
Age
 Mean 61.8 65.6 61.4 <.001
 40–50 16.0 4.4 17.2
 50–60 28.6 24.4 29.1 <.001
 >60 55.4 71.2 53.7
Gender
 Male 43.8 46.3 43.5 .083
 Female 56.2 53.7 56.5
Race/Ethnicity
 White 13.2 13.9 13.2
 Black 64.2 72.1 63.4 <.001
 Latinx 22.5 14.0 23.4
Smoking
 Never smoked 35.3 17.9 37.2
 Former smoker 29.6 43.5 28.1 <.001
 Current smoker 35.1 38.6 34.7
Neighborhood characteristics
Homicide rate/100,000
Mean 195.4 223.9 194.7 <.001
 1st quartile 25.6 22.4 25.9
 2nd quartile 24.6 20.1 25.1
 3rd quartile 24.5 26.5 24.3 <.001
 4th quartile 25.4 31.0 24.7
% Poverty
 Mean 25.5 26.1 25.4 .207
 1st quartile 24.7 23.4 24.8
 2nd quartile 25.4 23.8 25.6
 3rd quartile 24.8 24.8 24.8 .243
 4th quartile 25.1 28.0 24.7
Open in a new tab

3.2. Explaining lung cancer diagnosis

Table 3 summarizes the logistic regression results explaining the likelihood of a lung cancer diagnosis. We first examined the intraclass correlation (ICC) at the census tract to assess a need for multilevel modeling. Two-level hierarchical logistic regression model with the homicide rate at the second level showed that ICC at the census tract level was below 0.5, which conventionally indicates a weak correlation among cases within the same neighborhood. This result indicated lack of substantial neighborhood level clustering of explanatory variables, in this analysis, the homicide rate. Thus, we used logistic regression, treating the census tract-level homicide rate as an individual-level variable in the model. We examined the effects of homicide (Model I), poverty (Model II), and the homicide and poverty rates together (Model III) on the likelihood of being diagnosed with lung cancer. In Model I, patients in the youngest age group (40–50 years) were 63% less likely (OR=0.37, p<.001) compared to patients in the 50–60 age group to be diagnosed with lung cancer and the oldest age group (60 years and over) were 76% more likely (OR=1.76, p<.001) than patients in the 50–60 age group to be diagnosed with lung cancer. Hispanic patients were 38% less likely to be diagnosed with lung cancer than White patients (OR=0.62, p=.003). However, there was no statistically significant difference between Black and White patients (OR=0.89. p=.418) in being diagnosed with lung cancer. The odds of being diagnosed with lung cancer for current/former smokers were 2.33 times greater than for patients who never smoked (p<.001).

Table 3.

Logistic regression explaining lung cancer diagnosis

Model I Model II Model III
Variable OR p OR p OR p
Age
 40 – 50 years 0.37 <.001 0.37 <.001 0.37 <.001
 50 – 60 years - - - - - -
 60 years and older 1.76 <.001 1.74 <.001 1.76 <.001
Gender
 Female - - - - - -
 Male 0.97 .743 0.96 .672 0.97 .722
Race/Ethnicity
 White - - - - - -
 Black 0.89 .418 1.00 .995 0.90 .485
 Hispanic 0.62 .003 0.60 .002 0.62 .003
Smoking history
 Never smoked - - - - - -
 Former/Current smoker 2.33 <.001 2.34 <.001 2.33 <.001
Homicide rate
 1st quartile 1.10 .487 - - 1.10 .517
 2nd quartile - - - - - -
 3rd quartile 1.37 .017 - - 1.39 .014
 4th quartile 1.41 .009 - - 1.48 .007
Poverty rate
 1st quartile - - 0.91 .461 0.94 .656
 2nd quartile - - - - - -
 3rd quartile - - 0.94 .645 0.90 .400
 4th quartile - - 0.98 .868 0.87 .311
−2LL 3,807.026 3,814.995 3,805.890
Open in a new tab

Note: OR = Odds Ratio

There was no significant difference in the likelihood of being diagnosed with lung cancer between the first quartile (<25th) and the second quartile (25th – 50th) of homicide rates (OR=1.10, p=.487). However, compared with patients living in neighborhoods in the second quartile (25th – 50th), the likelihood of being diagnosed with lung cancer was 37% higher (OR=1.37, p=.017) for those in the third quartile (50th – 75th) and 41% higher (OR=1.41, p=.009) for those in the fourth quartile (>75th), after controlling for all other variables.

In Model II, the younger age group (40–50 years) was significantly less likely (OR=0.37, p<.001), and the older age group (>60 years) were significantly more likely (OR=1.74, p<.001) than patients ages between 50 and 60 years to be diagnosed with lung cancer. Hispanics were less likely than Whites to be diagnosed with lung cancer (OR=0.60, p=.002). Former/current smokers were more likely than never smokers to have lung cancer (OR=2.34, p<.001). There was no statistical association between male and female patients in lung cancer diagnosis. In addition, there were no statistically significant associations between % poverty quartiles and lung cancer diagnosis. Model III with the homicide rate and % poverty quartiles showed nearly identical results as Model I with the homicide rate only. The % poverty quartile measure was not associated with lung cancer diagnosis. On the other hand, compared to patients living in the 2nd quartile homicide rates, patients residing in areas with the 3rd quartile homicide rates were 39% more likely (OR=1.39, p=.014) and those living in areas with 4th quartile homicide rates were 48% more likely (OR=1.48, p=.007) to be diagnosed with lung cancer. As shown in Model I, there was no statistical difference in lung cancer diagnosis between patients living in areas with the bottom 25th percentile and the 2nd quartile (25th – 50th percentile) homicide rates.

Figure 1 summarizes the result of a generalized structural equation model, exploring the potential mediation of neighborhood-level homicide and poverty rates between race/ethnicity and lung cancer. Statistically significant paths are shown as solid lines, and non-significant paths are shown as gray dotted lines. Similar to the logistic regression results, patients aged between 40 and 50 years were 63% less likely (OR=0.37, p<.001), and those aged over 60 years were 75% more likely (OR=1.75, p<.001) than patients aged between 50 and 60 years to be diagnosed with lung cancer. The odds of being diagnosed with lung cancer were 2.3 times higher for former and current smokers than never-smokers (p<.001).

Fig. 1.

Fig. 1.

Open in a new tab

Structural equation model explaining lung cancer diagnosis with odds ratios

The odds of being diagnosed with lung cancer were 1.45 times higher for those living in areas with homicide rates above the mean (>195.4/100,000) compared to those living in areas with below the mean homicide rates (p<.001). Additionally, Hispanic residents were 38% less likely than White residents to be diagnosed with lung cancer, while there was no statistical difference in the odds between Black and White individuals. However, Black individuals were 10.5 times more likely than Whites to reside in areas with above the mean homicide rates (p<.001). At the same time, there was no statistical difference between Hispanic and White individuals in the odds of residing in high homicide areas. Regarding the odds of living in high-poverty areas, the odds were 9.3 times higher for Blacks (p<.001) and 1.7 times higher for Hispanics (p<.001) compared to Whites. Furthermore, those living in high-poverty areas (above the mean % poverty, 25.5%) were 5.6 times more likely to also be in high homicide areas (p<.001); in other words, high-poverty areas were more likely to have increased homicide rates.

While a history of smoking had the strongest association with a lung cancer diagnosis, male individuals were 2.4 times more likely than female individuals to be former/current smokers (p<.001); those living in high-poverty areas were 1.2 times more likely than those in low poverty areas to be former/current smokers (p<.001); and those living in high homicide areas were 1.3 times more likely than those in low homicide areas to be former/current smokers (p<.001).

3.3. Estimating population-level lung cancer risk

Using the synthetic population of Black, White, and Hispanic residents in Chicago who are 40 years and older, we appended corresponding odds ratios from the logistic regression to persons according to residential census tract, age, race/ethnicity, gender, and smoking status. Figure 2 shows the spatial distribution of homicide rates per 100,000 population (2.A), the estimated probability of developing lung cancer (2.B), and the hotspots (2.C) indicating high lung cancer and high homicide rate (red) vs. low lung cancer and low homicide rate (blue). Mostly, there is an increased probability of developing lung cancer clusters on the west and south sides of the city. The majority of these community areas are also affected by increased homicide rates. Figure 2.C displays hotspots (red) of areas where both the lung cancer and homicide rates are high, located on the city’s west and south sides. Figure 2.D shows the racial/ethnic distribution of the Chicago population. Predominantly Black communities are mostly located on the west and south sides of the city (green), and Hispanic communities are located on the north and southwest sides (red dots). Predominantly White communities are primarily located on the city’s north side (blue dots). As Figure 2 illustrates, an elevated risk of lung cancer, for the most part, is found in predominantly Black communities on the west and south sides of the city.

Fig. 2.

Fig. 2.

Open in a new tab

Spatial distributions of lung cancer risk, homicide rate, and racial/ethnic composition of the city of Chicago

Figure 3 shows the relationship between lung cancer risk, homicide rate, and % of Black residents. As shown, community areas in the top 25th percentile of lung cancer risk had over 83.3% Black residents, with 30% living below the 100% federal poverty line. The average homicide rate for these high lung cancer risk areas was 345.2 per 100,000 population, which is 8.4 times higher than the homicide rate for the bottom 25th of lung cancer risk areas. On the other hand, the bottom 25th percentile lung cancer areas had 5.2% Black residents with just 16.2% poverty level, and the homicide rate was 41.3 per 100,000 population. The poverty and homicide rates were higher for increased lung cancer risk areas. However, the homicide rates (red line, right Y axis) increased significantly more for the top two quartiles, while the poverty rates (blue line, left Y axis) increased gradually.

Fig. 3.

Fig. 3.

Open in a new tab

Comparison between neighborhood characteristics and lung cancer probability

4. Discussion

This study sought to understand the association between exposure to neighborhood violence and lung cancer risk and whether current racial/ethnic inequities in neighborhood conditions may explain lung cancer disparity. The overarching premise of this study was that chronic exposure to social stress propagates immune responses that may contribute to the risk of developing lung cancer. We found that individuals who lived in neighborhoods that were affected by high rates of violence, in this case, increased homicide rates, had a greater risk of being diagnosed with lung cancer, even after controlling for other individual-level risk factors such as smoking history, older age, race/ethnicity, and gender as well as neighborhood-level poverty.

We also found that significantly more Black residents were exposed to neighborhood violence than White or Hispanic residents, which may explain racial disparities in lung cancer in Black communities (Zhang et al., 2020). We showed that neighborhoods in the top 25th percentile of lung cancer risk had over 90% Black residents, with 30% of residents living below the 100% federal poverty line. Racial residential segregation in the city of Chicago has resulted in residents being exposed to significantly different social, economic, and environmental conditions. Our findings showed that living in areas with homicide rates above the mean was associated with a 36% increase in the estimated risk of lung cancer compared with those living in areas with below-average homicide rates, a burden shared unequally by Black residents in the city. In fact, among Chicago residents who were over the age of 60 and had a history of smoking, 13.6% of Black residents lived in communities with a high prevalence of violence, an estimated 19,000 Black residents, compared to only 0.7% of Hispanic residents and 0.4% of White residents, that is less than 600 residents each. The implication of such significant differences in social exposure for the Black community must be acknowledged concerning racial disparities in lung cancer.

White and Hispanic residents are significantly less likely to live in areas above the mean homicide rate. However, we found that regardless of race/ethnicity, living in high homicide areas was independently associated with an increased risk of being diagnosed with lung cancer. This finding indicates that it is not race/ethnicity, per se, but differential social exposure to stress increases the risk of lung cancer among Black Americans. We showed that neighborhood violence rates rose substantially higher for the top 50th percentile lung cancer areas, while the poverty rate, while still higher in the higher lung cancer areas, was not significantly higher for areas of the top 50th percentile lung cancer risk. No research is available comparing the effects of violence and poverty on stress response. However, chronic exposure to the ambient threat of physical harm due to neighborhood violence presents a different, more immediate stress and physiological response than a general economic disadvantage. While communities with higher poverty levels often tend to suffer from high levels of crime and violence, further research is required to examine the differential effects of violence and poverty on the stress response.

Our findings showed substantial differences between Black and Hispanic communities in Chicago regarding neighborhood violence, poverty, and lung cancer risk. From our EMR of patients who received any screening or diagnostic tests for lung cancer, Hispanic patients were significantly less likely to be current/former smokers (50.4%) compared to Black patients (68.8%). This finding is consistent with current literature, where the overall tobacco smoking rate is higher for Blacks (14.9%) than the rate for Hispanics (9.9%) in the U.S. (Wang et al., 2018). The lung cancer rate among patients included in our EMR was significantly higher for Black patients (11%) than for Hispanic patients (6.1%). The overall U.S. population’s lung cancer rate is 35.7 per 100,000 for Black Americans and 14.9 per 100,000 for Hispanic Americans (American Lung Association, 2021). Furthermore, the mean % poverty for Black residents in Chicago is 27.7%, while the mean % poverty for Hispanic residents is 16.4%. Similarly, the overall homicide rate was 79.5/100,000 for Black residents, compared to 9.7 per 100,000 for Hispanic residents in Chicago (Chicago Health Atlas, 2023). Our findings suggest that although Black and Hispanic Americans are both experiencing social and economic disadvantages in America, complex conditions affecting Black and Hispanic Americans need to be further examined (Williams et al., 2010).

Mortality rates associated with lung cancer are significantly reduced by lung cancer early detection screening (US Preventive Services Task Force, March 09, 2021). However, the uptake of lung cancer screening among Black smokers has been limited (Sosa et al., 2021). Future intervention efforts are needed to strengthen lung cancer screening and early detection, particularly for those who live in areas with elevated levels of violence, to mitigate the uneven distribution of lung cancer risk. In clinical settings, many clinics and hospitals have recently adapted the social determinants of health (SDOH) screening in their EMR systems, which could provide a unique opportunity to design system-level interventions that incorporate neighborhood context into SDOH assessments and recommend lung cancer screening for those who may not meet the criteria if they are exposed to high neighborhood violence. This clinic-level intervention can be performed by augmenting risk calculations for lung cancer by automatically populating patient addresses from EMR and alerting care providers. However, these communities have long been underserved due to lack of access to healthcare and social services. Thus, SDOH screening within clinics must be combined with neighborhood-level interventions to address suboptimal access to care, care utilization, and, ultimately, broader neighborhood investments. Consequently, policy interventions must address the lack of community safety, economic opportunity, and social capital in highly segregated communities, affecting residents’ health and well-being. Reducing the racial gap in exposure to violence in different communities would require policies at the macro level that aim to achieve social and economic equity and at the micro level that improve access to care and more robust gun control.

Further research is needed to establish lung cancer screening eligibility guidelines that equitably identify high-risk smokers, regardless of race/ethnicity (Li et al., 2019). Currently, the U.S. Preventive Services Task Force (USPSTF) lung cancer screening guidelines suggest that adults aged 50 to 80 years with a 20-pack-year smoking history (current smokers or those who have quit within the past 15 years) are eligible for low-dose computed tomography (LDCT) every year. However, many Black smokers are not eligible for LDCT because their smoking pattern differs from White smokers. Black smokers are likelier to be low-frequency smokers (Li et al., 2019). The national lung cancer trial (NLST) of heavy smokers found that 1.1% of participants were diagnosed with lung cancer among those who received LDCT (The National Lung Screening Trial Research Team, 2011). While this national trial included more than 26,000 participants in the LDCT group, just 4.5% were Black. Conversely, LDCT results from a minority-serving medical center in Chicago, the same institution where the current research was conducted, showed that the lung cancer rate, 2.6%, was more than double the NLST rate (Pasquinelli et al., 2018). In that sample, nearly 70% of the Chicago participants were Black. Our findings provide insights into what may contribute to the high rates of lung cancer in the Chicago study with predominantly Black participants by exploring a neighborhood stressor and exposure to violence, which may explain the increased risk of developing lung cancer in Black communities. National-level lung cancer screening guidelines can be refined by including neighborhood context, particularly exposure to violent crime. The revision of current screening guidelines should include neighborhood violence as an additional risk factor for lung cancer to capture low-frequency smokers who may still have an increased risk of lung cancer.

4.1. Limitations

This study has some limitations that must be considered when interpreting the findings. First, our initial estimates were based on EMRs, which are biased regarding lung cancer risk and diagnosis since the EMR included those seeking screening and/or diagnostic tests. We used analytic weights and adjusted the intercept coefficient to the population-level lung cancer rate to rescale our estimates to match lung cancer prevalence in Chicago. Nonetheless, we could not explore whether the relationships among predictors might change in the general population. This approach assumes that the risk factors in the general population and EMR sample are the same. It is possible, however, that the relationship between smoking and the risk of lung cancer may be nonlinear. Although this may or may not be the case, we did not have ways to validate this potential bias due to data availability.

Another limitation related to EMR is the relatively limited information about smoking history. In our analysis, smoking status included current, former, and never smokers and did not include a measure of smoking frequency and intensity, such as a pack-year measure. Thus, we could not tease apart racial differences in the impact of pack-year on lung cancer risk. Further research with more detailed smoking patterns by race/ethnicity will be beneficial.

We did not include individual-level access to care and poverty status in the analyses, which limited our ability to explore how differences in access to healthcare between White and Black patients may affect racial disparity in early detection of lung cancer. It is also possible that there might be an underestimation of lung cancer cases among Black patients at the time of our data acquisition due to potentially limited access to care and subsequent late-stage diagnosis. A future analysis incorporating lung cancer screening access at neighborhood and individual levels may help expand our understanding of individual and contextual health disparity.

This study examined the association between the neighborhood violence level and the estimated risk of lung cancer. While we compare neighborhood violence with poverty to tease out more generalized neighborhood disadvantages, other unobserved contextual factors might have influenced the relationship between violence and lung cancer risk. For example, environmental toxins, air pollutants, or occupational exposures could be other explanatory factors in future research. Environmental exposure such as radon and particulate matter (PM 2.5) and social and economic risk factors must be examined for comprehensive profiles of neighborhood contextual social determinants of lung cancer.

Finally, this study was conducted in a single geographical location, Chicago, the third largest urban city with extreme racial segregation. Thus, our findings may not fully represent other cities with different social and spatial organizations. This limitation does not reduce the significance of the study findings, however. Broader social and economic changes at the national or global level constantly interact with the local context, producing local-specific conditions. Interpreting and generalizing the current study’s findings and the findings of other studies need to be accompanied by understanding local social, economic, and political conditions. Additionally, owing to its intensified social and spatial conditions, Chicago may offer a unique opportunity to detect patterns that may not be evident elsewhere.

5. Conclusions

The neighborhood context, specifically exposure to neighborhood social stressors, including high violence and crime rates, increases the risk of developing lung cancer. Historical patterns of racial residential segregation differentially affect Black Americans residing in disadvantaged neighborhoods and thus experience disproportionate exposure to social stress and subsequent health consequences. Incorporating social exposure into the current lung cancer screening guidelines could improve the early detection of lung cancer and reduce racial disparities in lung cancer outcomes.

Current NLST trials have not explored neighborhood exposure as a risk indicator for lung cancer. This limitation excludes individuals who may not meet the smoking criteria but have increased social stress. Neighborhood-level interventions are needed to address disproportionate exposure to neighborhood social, economic, and environmental risk factors. However, it is far more challenging to design and implement neighborhood interventions. It requires rethinking resource allocation, fair community investment, and disruption of crime and violence in the most disadvantaged neighborhoods in our cities. Social policies promoting equitable neighborhood development are ultimately necessary to achieve health equity.

Highlights.

  • Neighborhood violence was associated with an increased risk of lung cancer.

  • Neighborhood violence and lung cancer risk were spatially clustered.

  • Increased neighborhood violence may contribute to racial disparity in lung cancer.

Acknowledgements

Sage Kim’s research was supported in part by the National Institute on Minority Health and Health Disparities (NIMHD: R01MD014839 and U54MD012523-05S1).

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Contributor Information

Sage J. Kim, University of Illinois at Chicago, School of Public Health, Division of Health Policy and Administration. Chicago, IL.

Caroline Kery, RTI International, Research Triangle Park, NC.

Jinghua An, Rutgers Cancer Institute of New Jersey, New Brunswick, NJ.

Georgiy Bobashev, RTI International, Research Triangle Park, NC.

Alicia K. Matthews, Columbia University, School of Nursing. New York, NY.

References

  1. Adie Y, Kats D, Tlimat A, Perzynski A, Dalton J, Gunzler D, et al. (2020). Neighborhood Disadvantage and Lung Cancer Incidence in Ever-Smokers at a Safety Net Health-Care System: A Retrospective Study. CHEST, 157, 1021–1029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. American Cancer Society. (2023). Key Statistics for Lung Cancer. American Cancer Society. [Google Scholar]
  3. American Lung Association. (2021). Lung Cancer Deaths and Rates by Race/Ethnicity, Sex and Year. Chicago, IL: American Lung Association. [Google Scholar]
  4. Bagatini M, Cardoso A, Reschke C, & Carvalho F (2018). Immune System and Chronic Diseases 2018. J Immunol Res, 2018, 8653572. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Braveman P, & Gottlieb L (2014). The social determinants of health: it’s time to consider the causes of the causes. Public health reports, 129, 19–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Browning C, Calder C, Ford J, Bethany B, Smith A, & Haynie D (2017). Understanding Racial Differences in Exposure to Violent Areas: Integrating Survey, Smartphone, and Administrative Data Resources. Ann Am Acad Pol Soc Sci, 669, 41–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Busso D, McLaughlin K, & Sheridan M (2017). Dimensions of Adversity, Physiological Reactivity, and Externalizing Psychopathology in Adolescence: Deprivation and Threat. Psychosomatic Medicine, 79, 162–171. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cajka J, Cooley P, & Wheaton W (2010). Attribute Assignment to a Synthetic Population in Support of Agent-Based Disease Modeling. Methods Rep RTI Press, 19, 1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Centers for Disease Control and Prevention. (2023). National Violent Death Reporting System (NVDRS). Atlanta, GA: Centers for Disease Control and Prevention. [Google Scholar]
  10. Chicago Health Atlas. (2022). Adult smoking. Chicago, IL: Chicago Health Atlas. [Google Scholar]
  11. Chicago Health Atlas. (2023). Indicators. Chicago, IL: Chicago Department of Public Health. [Google Scholar]
  12. Choquet H, Yin J, & Jorgenson E (2021). Cigarette smoking behaviors and the importance of ethnicity and genetic ancestry. Translational Psychiatry, 11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Cook County Government. (2022). About the Medical Examiner’s Office. Chicago IL: Cook County Government. [Google Scholar]
  14. Cornelius M, Loretan C, Wang T, Jamal A, & Homa D (2022). Tobacco Product Use Among Adults — United States, 2020. Morbidity and Mortality Weekly Report, 71, 97–405 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. County Health Rankings & Roadmaps. (2023). Homicides. Madison, WI: University of Wisconsin Population Health Institute. [Google Scholar]
  16. Dankwa-Mullan I, Rhee K, Williams K, Sanchez I, Sy F, Stinson N Jr, et al. (2010). The science of eliminating health disparities: summary and analysis of the NIH summit recommendations. American Journal of Public Health, 100, S12–S18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Danos D, Ferguson T, Simonsen N, Leonardi C, Yu Q, Wu X-C, et al. (2018). Neighborhood disadvantage and racial disparities in colorectal cancer incidence: a population-based study in Louisiana. Annals of Epidemiology, 28, 316–321.e312. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. de Groot P, Wu C, Carter B, & Munden R (2018). The epidemiology of lung cancer. Transl Lung Cancer Res, 7, 220–233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Ding Y, TOng M, Liu S, Moscow J, & Tai H-H (2005). NAD+-linked 15-hydroxyprostaglandin dehydrogenase (15-PGDH) behaves as a tumor suppressor in lung cancer. Carcinogenesis, 26, 65–72. [DOI] [PubMed] [Google Scholar]
  20. Durose M, & Mumola C (2004). Profile of Nonviolent Offenders Exiting State Prisons. Washington, DC: U.S. Department of Justice. [Google Scholar]
  21. Esri. (2023). ArcGIS. Redlands, CA: Environmental Systems Research Institute. [Google Scholar]
  22. Federal Bureau of Investigation. (2023). Violent Crime. Washington, DC: US Department of Justice. [Google Scholar]
  23. Feigelman W, & Lee J (1995). Probing the paradoxical pattern of cigarette smoking among African-Americans: low teenage consumption and high adult use. J Drug Educ, 25, 307–320. [DOI] [PubMed] [Google Scholar]
  24. Fowler P, Tompsett C, Braciszewski J, Jacques-Tiura A, & Baltes B (2009). Community violence: A meta-analysis on the effect of exposure and mental health outcomes of children and adolescents. Development and Psychopathology, 21, 227–259. [DOI] [PubMed] [Google Scholar]
  25. Geronimus A, Hicken M, Keene D, & Bound J (2006). “Weathering” and Age Patterns of Allostatic Load Scores Among Blacks and Whites in the United States. American Journal of Public Health, 96, 826–833. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Goldmann E, Aiello A, Uddin M, Delva J, Koenen K, Gant L, et al. (2011). Pervasive exposure to violence and posttraumatic stress disorder in a predominantly African American Urban Community: The Detroit neighborhood health study. Journal of traumatic stress, 24, 747–751. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Gomez S, Shariff-Marco S, DeRouen M, Keegan T, Yen I, Mujahid M, et al. (2015). The impact of neighborhood social and built environment factors across the cancer continuum: Current research, methodological considerations, and future directions. Cancer, 121, 2314–2330. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Gonzalez H, Hagerling C, & Werb Z (2018). Roles of the immune system in cancer: from tumor initiation to metastatic progression. Genes Dev, 32, 1267–1284. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Gudsnuk K, & Champagne F (2012). Epigenetic Influence of Stress and the Social Environment. ILAR Journal, 53, 279–288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Healthy People 2030. (2023). Crime and Violence. Office of Disease Prevention and Health Promotion. [Google Scholar]
  31. Heath NM, Chesney SA, Gerhart JI, Goldsmith RE, Luborsky JL, Stevens NR, et al. (2013). Interpersonal violence, PTSD, and inflammation: potential psychogenic pathways to higher C-reactive protein levels. Cytokine, 63, 172–178. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Heissel J, Sharkey P, Torrats-Espinosa G, Grant K, & Adam E (2018). Violence and Vigilance: The Acute Effects of Community Violent Crime on Sleep and Cortisol. Child Dev, 89, e323–e331. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Hosomi Y, Yokose T, KNakajima R, Nagai K, Nishiwaki Y, & Ochiai A (2000). Increased cyclooxygenase 2 (COX-2) expression occurs frequently in precursor lesions of human adenocarcinoma of the lung. Lung Cancer, 30, 73–81. [DOI] [PubMed] [Google Scholar]
  34. Hughes D, Otani T, Yang P, Newman R, Yantiss R, Altorki N, et al. (2008). NAD+-dependent 15-hydroxyprostaglandin dehydrogenase regulates levels of bioactive lipids in non-small cell lung cancer. Cancer Prev Res, 1, 241–249. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Hunt B, & Whitman S (2015). Black:White Health Disparities in the United States and Chicago: 1990–2010. Journal of Racial and Ethnic Health Disparities, 2, 93–100. [DOI] [PubMed] [Google Scholar]
  36. Hureau D, Wilson T, Jackl H, Arthur J, Patterson C, & Papachristos A (2022). Exposure to gun violence among the population of Chicago community violence interventionists. Sci Adv, 8, eabq7027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Inter-American Development Bank. (1999). How is violence measured? Technical Note. New York, NY: Inter-American Development Bank. [Google Scholar]
  38. Kapuku G, Treiber F, & Davis H (2002). Annals of Behavioral Medicine. Annals of Behavioral Medicine, 24, 320–325. [DOI] [PubMed] [Google Scholar]
  39. Koss K, & Gunnar M (2017). Annual Research Review: Early adversity, the hypothalamic–pituitary–adrenocortical axis, and child psychopathology. Journal of Child Psychology and Psychiatry, 59, 327–346. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Krieger N, Wright E, Chen J, Waterman P, Huntley E, & Arcaya M (2020). Cancer Stage at Diagnosis, Historical Redlining, and Current Neighborhood Characteristics: Breast, Cervical, Lung, and Colorectal Cancers, Massachusetts, 2001–2015. American Journal of Epidemiology, 189, 1065–1075. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Lawry T (2022). Genetic Code Vs. Zip Code: The Social Determinants Of Health Jersey City, NJ: Forbes Technology Council. [Google Scholar]
  42. Lerner L, Winn R, & Hulbert A (2018). Lung cancer early detection and health disparities: The intersection of epigenetics and ethnicity. Journal of Thoracic Disease, 10, 2498–2507. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Li CC, Matthews AK, Rywant MM, Hallgren E, & Shah RC (2019). Racial disparities in eligibility for low-dose computed tomography lung cancer screening among older adults with a history of smoking. Cancer Causes & Control, 30, 235–240. [DOI] [PubMed] [Google Scholar]
  44. Mullan Harris K, & McDade T (2018). The Biosocial Approach to Human Development, Behavior, and Health Across the Life Course. Russell Sage Foundation Journal of the Social Sciences, 4, 2–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. National Institute of Justice. (2023). Violent Crime. Washington, DC: National Institute of Justice. [Google Scholar]
  46. Pasquinelli M, Kovitz K, Koshy M, Menchaca M, Liu L, Winn R, et al. (2018). Outcomes From a Minority-Based Lung Cancer Screening Program vs the National Lung Screening Trial. JAMA Oncol, 4, 1291–1293. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Peckins M, Roberts A, Hein T, Hyde L, Mitchell C, Brooks-Gunn J, et al. (2020). Pervasive exposure to violence and posttraumatic stress disorder in a predominantly African American Urban Community: The Detroit neighborhood health study. Psychoneuroendocrinology, 111, 104426.31639588 [Google Scholar]
  48. Penman-Aguilar A, Talih M, Huang D, Moonesinghe R, Bouye K, & Beckles G (2016). Measurement of health disparities, health inequities, and social determinants of health to support the advancement of health equity. Journal of Public Health Management and Practice, 22, S33–S42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Rineer JI, Bobashev GV, Jones KR, Hadley EC, Kery C, et al. (2018). RTI Synthetic Population. In R. International. (Ed.). [Google Scholar]
  50. RTI International. (2022). RTI U.S. Synthetic Household Population™,. Chapel Hill, NC: RTI International. [Google Scholar]
  51. Saini G, Ogden A, McCullough L, Torres M, Rida P, & Aneja R (2019). Disadvantaged neighborhoods and racial disparity in breast cancer outcomes: the biological link. Cancer Causes & Control, 30, pages 677–686. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Sandler A, & Dubinett S (2004). COX-2 inhibition and lung cancer. Semin Oncol, 31, 45–52. [DOI] [PubMed] [Google Scholar]
  53. Sharkey P, & Sampson R (2015). Violence, Cognition, and Neighborhood Inequality in America. In Schutt R, Seidman L, & Keshavan M (Eds.), Social Neuroscience: Brain, Mind, and Society. [Google Scholar]; Sharkey Patrick S., and Sampson Robert J.. 2015. “Violence, Cognition, and Neighborhood Inequality in America.” Social Neuroscience: Brain, Mind, and Society, edited by Schutt Russell K., Seidman Larry J., and Keshavan Matcheri S.. Cambridge, MA: Harvard University Press.: Harvard University Press. [Google Scholar]
  54. Sin MK (2017). Lung cancer disparities and African-Americans. Public Health Nursing, 34, 359–362. [DOI] [PubMed] [Google Scholar]
  55. Sosa E, D’Souza G, Akhtar A, Sur M, Love K, Duffels J, et al. (2021). Racial and socioeconomic disparities in lung cancer screening in the United States: a systematic review. CA: A cancer journal for clinicians, 71, 299–314. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Stata. (2020). Your data tell a story. College Station, TX: StataCorp. [Google Scholar]
  57. Substance Abuse and Mental Health Services Administration. (2023). National Survey on Drug Use and Health (NSDUH) Washington, DC: Substance Abuse and Mental Health Services Administration. [Google Scholar]
  58. The National Lung Screening Trial Research Team (2011). Reduced Lung-Cancer Mortality with Low-Dose Computed Tomographic Screening. N Engl J Med, 365, 395–409. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. US Preventive Services Task Force. (March 09, 2021). Final Recommendation Statement: Lung Cancer Screening. Rockville, MD: US Preventive Services Task Force (USPSTF). [Google Scholar]
  60. Wang T, Asman K, Gentzke A, Cullen K, Holder-Hayes E, Reyes-Guzman C, et al. (2018). Tobacco Product Use Among Adults — United States, 2017. MMWR Morb Mortal Wkly Rep, 67, 1225–1232. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Wheaton W, Cajka J, Chasteen B, Wagener D, Cooley P, Ganapathi L, et al. (2009). Synthesized Population Databases: A US Geospatial Database for Agent-Based Models. Research Triangle Park, NC: Research Triangle Institute. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Williams D, Mohammed S, Leavell J, & Collins C (2010). Race, Socioeconomic Status and Health: Complexities, Ongoing Challenges and Research Opportunities. Ann NY Acad Sci, 1186, 69–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Zhang L, Pan J, Chen W, Jiang J, & Huang J (2020). hronic stress-induced immune dysregulation in cancer: implications for initiation, progression, metastasis, and treatment. American Journal of Cancer Research, 10, 1294–1307. [PMC free article] [PubMed] [Google Scholar]

RESOURCES