Abstract
Background
Historical redlining, or the Home Owners’ Loan Corporation (HOLC) program’s racially biased mortgage risk monitoring maps in the 1930s, is implicated in shaping modern neighborhoods and health outcomes. This retrospective cohort study evaluates the association between redlining and mortality in young cancer patients.
Methods
Using the California Cancer Registry, we identified patients aged younger than 25 years diagnosed with malignant cancer between 2000 and 2019. HOLC maps were spatially joined with patient address at diagnosis to determine redlining status (A [best], B [still desirable], C [declining], D [hazardous]). Census tract-level US Census and American Community Survey data were appended to determine modern neighborhood characteristics. The Kaplan–Meier method was used to evaluate overall survival and multivariable Cox proportional hazards models to estimate the associations between HOLC grade and mortality, adjusting for clinical and multilevel social drivers of health.
Results
In total, 8108 patients resided in HOLC-graded neighborhoods among 51 084 patients statewide. Overall survival at 5 years was inferior for patients who resided in D-graded neighborhoods at diagnosis vs A-graded neighborhoods (80.3%, 95% confidence interval [CI] = 78.6% to 81.8%, vs 88.5%, 95% CI = 84.3% to 91.6%). Adjusting for clinical characteristics, patients in D-graded neighborhoods experienced greater mortality (hazard ratio [HR] = 1.32, 95% CI = 1.12 to 1.56) compared with those in A- and B-graded neighborhoods. Additional adjustment for insurance attenuated the effect (HR = 1.17, 95% CI = 1.00 to 1.36), and for neighborhood, socioeconomic status marginally attenuated the effect (HR = 0.96, 95% CI = 0.81 to 1.13).
Conclusion
Findings suggest enduring legacy effects of historical redlining on young individuals with cancer potentially mediated social factors, including health insurance.
Introduction
Children, adolescents, and young adults with cancer from marginalized racial and ethnic populations face inferior survival compared with their non-Hispanic White counterparts.1-3 Structural racism is a root cause of racial and ethnic health disparities,4 relevant to populations that are healthy and those with serious health conditions including cancer.5,6 Structural racism encompasses the policies, practices, processes, and norms embedded within the health, justice, education, political, and economic systems that disempower and oppress marginalized racial and ethnic groups.7,8 Historical redlining is one pertinent example of a measurable form of structural racism9 associated with mortality for adults with cancer.10-12
One notable example of redlining is the creation of security maps in the mid-1930s by the Home Owners’ Loan Corporation (HOLC), which graded neighborhoods based on their perceived desirability, primarily driven by racial and ethnic composition.13 Neighborhoods were labeled “best,” “still desirable,” “declining,” and “hazardous,” whereby neighborhoods with higher proportions of Black and immigrant residents were disproportionately labeled unfavorably.10,14 Practical use of these maps appears to be limited, without substantial evidence of widespread use by mortgage lenders or other agencies like the Federal Housing Administration. However, these HOLC maps represent long-standing racist attitudes and beliefs, which fueled discrimination in the housing sector long before and after their creation15,16 and therefore serve as a direct measure of a form of structural racism known as redlining, a term later coined in the 1960s referring to the institutional practice of racially based housing discrimination.16 From a public health perspective, historical redlining is hypothesized to impact individuals today by way of continued unfair housing practices contributing to persistent neighborhood segregation and inequitable access to social and economic opportunity and mobility.10,17,18 Neighborhood segregation limits an individual’s access to high-quality schools, employment, credit, and loans,18 all of which are social drivers of health (SDOH).
Residence at the time of cancer diagnosis in a redlined neighborhood has been associated with disparities across the cancer continuum. Specifically, redlining has been associated with conditions associated with greater cancer risk,19-21 suggesting potential mechanisms for cancer development. Additionally, redlining has been associated with gaps in access to health care including lower lung cancer screening rates22 and advanced-stage cervical, colon, breast, and lung cancer at diagnosis,23 in addition to differential treatment including lower receipt of surgery,24 lymph node evaluation, and adjuvant chemotherapy.25 Finally, redlining has been associated with greater mortality for individuals with cancer.11,25
According to a policy statement released in 2019 by the American Academy of Pediatrics, racism “has a profound impact on the health status of children, adolescents, emerging adults, and their families.”6 Although structural racism plausibly leads to inequitable outcomes for young individuals with cancer, few studies that investigate structural racism and health outcomes exist.26 We fill this gap by conducting a retrospective population-based study of individuals aged younger than 25 years diagnosed with cancer in California to examine the association between historical redlining and mortality.
Methods
Study population
California residents aged 24 years or younger at the time of their first primary malignant cancer diagnosis in 2000-2019 were eligible for this study. All cancer cases are reported to the California Cancer Registry by statewide mandate. Participant demographics including insurance status, cancer diagnosis and stage, date of diagnosis, and vital status data were obtained from the California Cancer Registry. Address at diagnosis is initially collected at time of report and geocoded for cancer surveillance purposes. Vital status is updated through routine follow-up and linkages to state vital statistics, National Death Index, and other data sources. This research was approved for human participants by the University of California institutional review board under the Greater Bay Area Cancer Registry umbrella protocol.
Follow-up
Death ascertainment was complete through December 31, 2020. Patients were censored at date of last contact or the end of the study period, whichever occurred first.
Redlining
To determine the redlining status of each address at diagnosis, the latitude and longitude of each patient’s address was mapped in ArcGIS 3.1 (Esri, Redlands, CA, USA) and spatially joined to HOLC neighborhood maps digitized by Mapping Inequality (Figure 1).27 Each location was assigned to the HOLC grade of the neighborhood that the point fell into (A [best], B [still desirable], C [declining], and D [hazardous]). HOLC security maps were typically created for cities, or in some cases counties, with a population of at least 40 000 residents at the time.28 In total, HOLC-graded areas cover 8 cities in California. We assigned N (not graded) to addresses without HOLC-grade assignment, because of either a location outside of designated maps or within ungraded areas within maps.
Figure 1.
Home Owners’ Loan Corporation maps of (A) San Francisco, California (1937), and (B) Los Angeles, California (1939). Source: Mapping Inequality, public domain image.
Covariates
Individual-level covariates
Individual-level covariates (age, sex, cancer stage, and insurance) were selected a priori based on previously established associations with mortality for cancer patients. Census tract–level covariates (neighborhood typology, neighborhood socioeconomic status [SES], and racialized economic segregation) based on US Census and American Community Survey data were appended to patient addresses. The most temporally close Decennial Census and American Community Survey data were linked to patient diagnosis year (Table S1). The 1854 (3.5% of 52 938) patients whose address at diagnosis was assigned to census tract based on zip code only, or zip code of P.O. box, or could not be assigned to a census tract, were excluded from further analysis.
Area-level covariates
Modern area-level covariates include racialized economic segregation, neighborhood typology, and neighborhood SES, which represent distinct measures of both racial and economic factors, racial factors, and economic factors, respectively. Area-level characteristics by HOLC grade are available in Table S2. Racialized economic segregation, measured by the Index of Concentration at the Extremes, quantifies the extent to which a neighborhood is concentrated into relative extremes of advantage and disadvantage.29 Index of Concentration at the Extremes has previously been postulated as a modern area-level measure downstream of redlining30 and is measured in 2 dimensions, by race and by income. The index is categorized into quintiles, where the lowest Index of Concentration at the Extremes quintile (1) represents neighborhoods with the greatest concentration of low-income non-Hispanic Black residents, and the highest Index of Concentration at the Extremes quintile (5) represents neighborhoods with the greatest concentration of high-income non-Hispanic White residents.
Neighborhood typology, or the racial or ethnic composition of a neighborhood, was modified from originally published criteria31 to include 5 distinct typologies based on California demographics. Neighborhood typologies were categorized as follows: predominantly Asian and/or Pacific Islander, predominantly Black, predominantly Hispanic, predominantly non-Hispanic White, and multiethnic, whereby “predominantly” indicates at least 50% of the residents in the neighborhood identified with the racial or ethnic identity. Multiethnic neighborhoods were those in which less than 50% of residents identified with any one racial or ethnic identity.
Based on a commonly used measure, neighborhood SES was obtained through the California Neighborhoods Data System,32 a database containing neighborhood-level socioeconomic and built environment data. The neighborhood SES is an index of 7 area-level variables: education attainment, percent persons with a blue-collar job, percent persons employed or unemployed, percent persons above or below 200% of the federal poverty line, median cost of rent, median value of owner-occupied housing unit, and median household income. Neighborhood SES index was categorized into quintiles.
Analysis
Overall survival by HOLC grade was assessed using the Kaplan–Meier method with a log-rank test. Hazard ratios (HRs) for the risk of overall mortality were estimated using Cox proportional hazards regression models for those residing in HOLC-graded neighborhoods. A total of 7 models were fitted as follows. Model 1 is unadjusted. Models 2-4 include individual-level covariates: model 2 (adjusted for age, sex, year of diagnosis), model 3 (additionally adjusted for cancer stage), model 4 (additionally adjusted for insurance). The final 3 models additionally adjusted for a single area-level covariate in each model: model 5 (additionally adjusted for neighborhood SES), model 6 (additionally adjusted for racialized economic segregation), and model 7 (additionally adjusted for neighborhood typology). Stepwise modeling assessed the relative contribution of each covariate. An interaction term testing for the presence of an interaction between race and/or ethnicity and HOLC grade was included in the Cox regression analysis. The models accounted for clustering by HOLC neighborhood and additionally clustering by census tract in models that included census tract–level covariates. The proportional hazards assumption was tested for by assessing Schoenfeld residuals and fulfilled by stratifying by age. Because of the small sample sizes, the reference group consists of HOLC grades A and B combined into 1 category. A sensitivity analysis was performed to test whether inclusion of nongraded neighborhoods yielded differential results. An exploratory analysis used Cox proportional hazards regression models for patients with a truncated survival time of less than 2 years and a conditional analysis of those with a survival time of 2 or more years. A second exploratory analysis stratified by age (0-18 and 19-24 years). An analysis stratified by race and ethnicity (model 4) and a sensitivity analysis including ungraded neighborhoods are available in Tables S3 and S4.
Analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA). A P value less than .05 was considered statistically significant. All statistical tests were 2-sided.
Results
Patient characteristics
Overall, 51 084 individuals aged younger than 25 years were diagnosed with first occurrence of malignant cancer in California between 2000 and 2019. Of these, 16% (n = 8108) resided in areas previously subjected to HOLC grading in 8 California cities (Fresno, Los Angeles, Oakland, Sacramento, San Diego, San Francisco, San Jose, and Stockton) and therefore were included in the analysis (Figure 1). Baseline patient characteristics are reported in Table 1. By age, 4522 (n = 56%) patients were pediatric (aged younger than 18 years) and 3586 (44%) were young adult (aged 18-24 years). Approximately half (n = 4017) of patients lived in high poverty census tracts, and 42% (n = 3380) of patients lived in census tracts of the lowest neighborhood SES quintile. Patient distributions by HOLC grade at the time of diagnosis in order of descending prevalence are as follows: 51% C (n = 4129), 31% D (n = 2505), 14% B (n = 1157), and 4% A (n = 317). Race and ethnicity, insurance, and cancer stage varied by HOLC grade.
Table 1.
Patient characteristics according to historical redlining exposure (Home Owners’ Loan Corporation grade)a
| Patient characteristics | Home Owners’ Loan Corporation grade |
||||
|---|---|---|---|---|---|
| Total, No. (%) (n = 8108) | Best (A) , No. (%) (n = 317) | Still desirable (B), No. (%) (n = 1157) | Declining (C), No. (%) (n = 4129) | Hazardous (D), No. (%) (n = 2505) | |
| Age at diagnosis, y | |||||
| 0-12 | 3165 (39.0) | 92 (29.0) | 456 (39.4) | 1608 (38.9) | 1009 (40.3) |
| 13-17 | 1357 (16.7) | 55 (17.4) | 189 (16.3) | 690 (16.7) | 423 (16.9) |
| 18-24 | 3586 (44.2) | 170 (53.6) | 512 (44.3) | 1831 (44.3) | 1073 (42.8) |
| Sex | |||||
| Male | 4382 (54.1) | 170 (53.6) | 592 (51.2) | 2247 (54.4) | 1373 (54.8) |
| Female | 3723 (45.9) | 147 (46.4) | 564 (48.8) | 1880 (45.5) | 1132 (45.2) |
| Race and ethnicity | |||||
| American Indian and Alaska Native | 20 (0.3) | 0 (0.0) | 0 (0.0) | 13 (0.3) | <11 D(0.3) |
| Asian | 770 (9.5) | 33 (10.4) | 145 (12.5) | 384 (9.3) | 208 (8.3) |
| Black | 684 (8.4) | 15 (4.7) | 85 (7.4) | 339 (8.2) | 245 (9.8) |
| Hispanic | 4561 (56.3) | 30 (9.5) | 371 (32.1) | 2450 (59.3) | 1710 (68.3) |
| Native Hawaiian and Pacific Islander | 30 (0.4) | 0 (0.0) | <11 (0.3) | 17 (0.4) | <11 (0.4) |
| White | 1938 (23.9) | 219 (69.1) | 526 (45.5) | 889 (21.5) | 304 (12.1) |
| Unknown | 105 (1.3) | 20 (6.3) | 27 (2.3) | 37 (0.9) | 21 (0.8) |
| Insurance | |||||
| No insurance | 286 (3.5) | <11 (1.0) | 32 (2.8) | 136 (3.3) | 115 (4.6) |
| Public insurance | 3703 (45.7) | 29 (9.2) | 328 (28.4) | 2032 (49.2) | 1314 (52.5) |
| Private insurance | 3777 (46.6) | 268 (84.5) | 752 (65.0) | 1787 (43.3) | 970 (38.7) |
| Unknown | 342 (4.2) | 17 (5.4) | 45 (3.9) | 174 (4.2) | 106 (4.2) |
| Year of diagnosis | |||||
| 2000-2009 | 4129 (50.9) | 152 (48.0) | 588 (50.8) | 2111 (51.1) | 1278 (51.0) |
| 2010-2019 | 3979 (49.1) | 165 (52.1) | 569 (49.2) | 2018 (48.9) | 1227 (49.0) |
| Stage | |||||
| Localized | 2964 (36.6) | 134 (42.3) | 480 (41.5) | 1503 (36.4) | 847 (33.8) |
| Regional | 1626 (20.1) | 75 (23.7) | 228 (19.7) | 845 (20.5) | 478 (19.1) |
| Distant sites involved | 3266 (40.3) | 96 (30.3) | 408 (35.3) | 1659 (40.2) | 1103 (44.0) |
| Unknown | 252 (3.1) | 12 (3.8) | 41 (3.5) | 122 (3.0) | 77 (3.1) |
| Cancer type | |||||
| Brain, other nervous system tumor | 914 (11.3) | 41 (12.9) | 141 (12.2) | 472 (11.4) | 260 (10.4) |
| Leukemia, lymphoma | 3073 (37.9) | 117 (36.9) | 408 (35.3) | 1558 (37.7) | 990 (39.5) |
| Solid tumor | 3936 (48.5) | 149 (47.0) | 585 (50.6) | 2013 (48.8) | 1189 (47.5) |
| Other | 185 (2.3) | <11 (3.2) | 23 (2.0) | 86 (2.1) | 66 (2.6) |
In compliance with cancer registry regulations, populations of less than 11 are masked, and exact numeration is unknown. Sex other or unknown participants are not shown because of a count of less than 11 across the sample.
Neighborhood characteristics
Neighborhood typology differed by HOLC grade; C- and D-graded neighborhoods were most frequently multiethnic (60% and 48%, respectively) or predominantly Hispanic (24% and 43%, respectively) neighborhoods, whereas A- and B-graded neighborhoods were most frequently predominantly White (65% and 30%, respectively) and multiethnic (30% and 57%, respectively) neighborhoods (Table S2). Neighborhood SES differed by HOLC grade; C and D neighborhoods were more frequently in the lowest SES quintiles (37% and 48%, respectively) compared with A- and B-graded neighborhoods (0% and 15%, respectively). There was marked racialized economic segregation, whereby 76% of A-graded neighborhoods, 26% of B-graded neighborhoods, 8% of C-graded neighborhoods and 7% of D-graded neighborhoods had the highest concentration of high-income, non-Hispanic White residents, or higher social advantage. Reciprocally, 7% of A-graded neighborhoods, 16% of B-graded neighborhoods, 45% of C-graded neighborhoods, and 56% of D-graded neighborhoods had the highest concentration of low-income, Black residents, or lower social advantage.
Outcomes
Survival data were available for 8108 patients (Figure 2). The estimated 5-year overall survival for patients residing in A-graded neighborhoods was 88.5% (95% confidence interval [CI] = 84.3% to 91.6%); 85.1% (95% CI = 82.8% to 87.1%) for those in B-graded neighborhoods; 81.6% (95% CI = 80.3% to 82.8%) for those in C-graded neighborhoods; and 80.3% (95% CI = 78.6% to 81.8%) for those in D-graded neighborhoods.
Figure 2.
Overall survival of patients with malignant cancer aged younger than 25 years by HOLC grade of neighborhood at diagnosis, California, 2000-2019. Blue solid line represents residence in HOLC A-graded neighborhoods; red dashed line represents residence in HOLC B-graded neighborhoods; green dashed line represents residence in HOLC C-graded neighborhoods; and brown dashed line represents residence in HOLC D-graded neighborhoods. Abbreviations: A grade = best; B grade = still desirable; C grade = declining; D grade = hazardous; HOLC = Home Owners’ Loan Corporation.
After stratifying by age and adjusting for sex and year of diagnosis, patients residing in D-graded neighborhoods experienced 43% higher mortality compared with those in A- and B-graded neighborhoods (model 2: HR = 1.43, 95% CI = 1.21 to 1.69) (Table 2). With additional adjustment for cancer stage, redlined patients still experienced increased mortality (model 3: HR = 1.32, 95% CI = 1.12 to 1.56), although the relationship was attenuated. With further adjustment for insurance, the relationship between redlining and mortality was further attenuated (model 4: HR = 1.17, 95% CI = 1.00 to 1.36). In models that additionally adjusted for neighborhood SES, racialized economic segregation, or neighborhood typology, the relationship between redlining and mortality remained statistically insignificant as incorporating these variables further attenuated the magnitude of the mortality relationship. An interaction term between race and ethnicity and HOLC grade was statistically significant, although there does not appear to be meaningful patterns of associations for each racial or ethnic group in a stratified analysis (P < .01) (Table S3). In a sensitivity analysis, inclusion of nongraded HOLC areas did not significantly statistically change the results from the analysis excluding nongraded HOLC areas (Table S4).
Table 2.
Mortality of patients with malignant cancer aged younger than 25 years residing in redlined neighborhoods at diagnosis (HOLC graded C and D) compared with nonredlined neighborhoods (HOLC graded A and B combined), California, 2000-2019.
| Modela | HOLC gradeb |
|
|---|---|---|
| C | D | |
| HR (95% CI) | HR (95% CI) | |
| 1 | 1.35 (1.14 to 1.59) | 1.43 (1.21 to 1.69) |
| 2 | 1.36 (1.15 to 1.60) | 1.43 (1.21 to 1.69) |
| 3 | 1.31 (1.12 to 1.54) | 1.32 (1.12 to 1.56) |
| 4 | 1.17 (1.00 to 1.36) | 1.17 (1.00 to 1.36) |
| 5 | 0.99 (0.85 to 1.15) | 0.96 (0.81 to 1.13) |
| 6 | 0.99 (0.85 to 1.16) | 0.96 (0.81 to 1.13) |
| 7 | 1.08 (0.93 to 1.26) | 1.04 (0.89 to 1.22) |
Abbreviations: A grade = best; B grade = still desirable; C grade = declining; CI = confidence interval; D grade = hazardous; HLOC = Home Owners’ Loan Corporation; HR = hazard ratio.
Model 1: unadjusted. Model 2: stratified by age and adjusted for sex and year of diagnosis. Model 3: stratified by age and adjusted for sex, year of diagnosis, and cancer stage. Model 4: stratified by age and adjusted for sex, year of diagnosis, cancer stage, and individual-level insurance. Model 5: stratified by age and adjusted for sex, year of diagnosis, cancer stage, individual-level insurance, and neighborhood socioeconomic status. Model 6: stratified by age and adjusted for sex, year of diagnosis, cancer stage, individual-level insurance, and racialized economic segregation. Model 7: stratified by age and adjusted for sex, year of diagnosis, cancer stage, individual-level insurance, and neighborhood typology.
Bolded results are statistically significant at an alpha level of 0.05. Referent group: combined HOLC grades A and B.
An exploratory survival analysis demonstrated exposure to redlining was associated with increased hazard of death when truncating survival at 2 years (HR = 1.43, 95% CI = 1.17 to 1.75) and in a conditional analysis of patients who survived 2 or more years (HR = 1.47, 95% CI = 1.15 to 1.88) (Table 3). Exposure to redlining was associated with increased mortality for patients aged 0-18 years (HR = 1.31, 95% CI = 1.06 to 1.63) and aged 19-24 years (HR = 1.64, 95% CI = 1.30 to 2.06) (Table 4).
Table 3.
Redlining-associated mortality of patients with malignant cancer, by higher and lower mortality cancer subcohorts, California, 2000-2019a
| Home Owners’ Loan Corporation grade | Truncated follow-up time to 2 years (n = 8069) | Among those who survived ≥2 years (n = 6575) |
|---|---|---|
| HR (95% CI) | HR (95% CI) | |
| A (best) and B (still desirable) (referent) | — | — |
| C (declining) | 1.35 (1.11 to 1.63) | 1.38 (1.09 to 1.74) |
| D (hazardous) | 1.43 (1.17 to 1.75) | 1.47 (1.15 to 1.88) |
Abbreviations: CI = confidence interval; HR = hazard ratio.
Analysis utilizes unadjusted model 1.
Table 4.
Redlining-associated mortality of patients with malignant cancer, by age subcohorts (0-18 and 19-24 years), California, 2000-2019a
| Home Owners’ Loan Corporation grade | Age 0-18 (n = 4859) | Age 19-24 (n = 3249) |
|---|---|---|
| HR (95% CI) | HR (95% CI) | |
| A (best) and B (still desirable) (referent) | — | — |
| C (declining) | 1.21 (0.97 to 1.49) | 1.60 (1.27 to 2.00) |
| D (hazardous) | 1.31 (1.01 to 1.63) | 1.64 (1.30 to 2.06) |
Analysis utilizes unadjusted model 1.
Discussion
In this retrospective population-based study, young individuals diagnosed with cancer between 2000 and 2019 in areas previously redlined nearly a century ago experienced greater mortality than those unexposed to historical redlining. Overall, these data align with previous literature demonstrating deleterious health effects for youth residing in redlined areas33-38 and higher mortality for adults with breast and colon cancer residing in redlined areas.11,25 These findings add yet another susceptible population to redlining: children, adolescents, and young adults younger than 25 years with cancer.
This study applied a novel measure of structural racism to the pediatric cancer population: historical redlining. Redlining maps operationalize residential structural racism and provide historical context—a key approach to understanding mechanisms for racial and ethnic inequities.39 Although these maps are not without notable limitations or without controversy around direct culpability given their relatively limited use,16 they offer a unique, measurable capture of widely held, persistent ideologies that informed discriminatory housing practices. These societal ideologies informed the creation of the maps themselves and represent decades of racist attitudes before and after their creation. The role of racism as an SDOH has been highlighted in statements by leading national cancer and pediatric institutions,6,7,40-43 yet few studies interrogating racism exist.44,45 Findings from this investigation add a measure of structural racism for consideration as a novel predictor of adverse health outcomes for young individuals with cancer.
We hypothesized a priori that many of the previously proposed mechanisms for redlining survival disparities in adults with cancer likely apply to young individuals, including individual-level insurance status.23 Insurance is plausibly a downstream mechanism from redlining, as redlined areas are more likely to be low resourced14 and its residents underinsured.46 Indeed, in California there was a clear and striking association between HOLC grades, neighborhood-level characteristics, and patient-level socioeconomic factors. As such, the relationship between historical redlining and mortality was attenuated after adjustment for insurance in this study with half the population eligible for Medicaid and less than 5% uninsured, suggesting underinsurance plays a clinically significant role in the relationship between redlining and mortality. Redlined-exposed populations aged 0-18 years and 19-24 years experienced increased mortality. Although there was a greater effect size for patients aged 19-24 years, which may suggest age-specific barriers, confidence intervals were overlapping.
Young individuals with cancer may face different driving mechanisms from their adult counterparts. For example, exposure to redlining has been associated with late cancer stage in the adult population.23 The diagnosis of pediatric-onset cancers, however, is typically independent from cancer screening. Biologically aggressive disease,21,47 behavior-related diseases,48 and biological weathering47 are hypothesized mechanisms for cancer disparities in adults with cancer, and to what degree these factors apply to younger patients is less well understood. In an exploratory analysis of individuals aged younger than 25 years, in a truncated survival analysis to 2 years and conditional analysis of 2 or more years, exposure to redlining was associated with increased mortality. These findings suggest both early mortality risk factors such as cancer biology or treatment-related mortality and late mortality risk factors such as comorbidities or secondary malignancies may be influential. Future studies are needed to evaluate which factors are predominant and whether they differ by age.
Residential segregation, neighborhood SES, and racialized economic segregation are independently associated with health-related factors and outcomes and plausibly downstream of redlining.49-58 In this study, the effect sizes for redlining on mortality appeared to be marginally attenuated by neighborhood characteristics, suggesting that these particular modern-day neighborhood factors may not be mediators for associations between historical redlining and current-day cancer survival. Although mechanisms by which redlining-associated adverse health outcomes arise are not fully understood, prior research suggests segregation, poverty, and other placed-based factors; social, educational, and economic opportunity; and access to healthcare.14,58 Therefore, proposed interventions to address redlining include direct investment in redlined neighborhoods (ie, sidewalks, bike lanes, greenspace, and public transportation) and local and federal housing, education, health care, economic, and food policies, which expand resources and services for at-risk populations.14 Because characterization of redlining mechanisms directly informs future intervention targets, further evaluation in young cancer populations is needed.
This study includes several limitations to be considered. These findings may not be generalizable outside of California. Patients in ungraded and graded neighborhoods were similar in clinical characteristics, however patients in graded areas were concentrated in urban areas and differed by SDOH. High poverty neighborhoods were overrepresented, as approximately half of the patients resided in neighborhoods that were impoverished, compared with approximately 20% of the national population.59 A sensitivity analysis including ungraded patients yielded similar results to those excluding ungraded patients. Although HOLC classification was assigned using geographic coordinates, misclassification is expected as residence was measured at a single timepoint. The extent to which the HOLC maps were used is debated.16 Because of sample size, conclusions regarding specific racial or ethnic groups are not definitive. With these limitations to consider, this is the largest study evaluating mortality in young cancer patients and yields findings consistent with our previous evaluations in patients aged younger than 40 years in Seattle and Tacoma.38
Our findings suggest that historical redlining remains relevant to children, adolescents, and young adults with cancer today, despite nearly a century’s passing. Fortunately, although the creation of the maps cannot be undone, solutions to mitigate the harmful effects of redlining exist.14 Solutions may take the form of policy change such as those to enforce fair housing, invest in redlined neighborhoods, and improve health-care access.60 Further elucidation of redlining-associated mortality mechanisms will inform which solutions may be most relevant to young individuals with cancer.
Supplementary Material
Acknowledgments
We acknowledge support by the California Department of Public Health pursuant to California Health and Safety Code Section 103885; Centers for Disease Control and Prevention’s (CDC) National Program of Cancer Registries, under cooperative agreement 1NU58DP007156; the National Cancer Institute’s Surveillance, Epidemiology and End Results Program under contract HHSN261201800032I awarded to the University of California, San Francisco, contract HHSN261201800015I awarded to the University of Southern California, and contract HHSN261201800009I awarded to the Public Health Institute, Cancer Registry of Greater California. These supported collection of cancer registry data. The ideas and opinions expressed herein are those of the author(s) and do not necessarily reflect the opinions of the State of California, Department of Public Health, the National Cancer Institute, and the Centers for Disease Control and Prevention or their Contractors and Subcontractors. The aforementioned UCSF contract, ASCO Conquer Cancer, Seattle Children’s Hospital, Robert A. Winn Excellence in Clinical Trials Award Program, and American Cancer Society supported PI effort as listed in funding section and did not participate in study design, data collection, analysis, interpretation, writing, or submission of the manuscript.
These data were presented in poster abstract form at ASCO Annual Conference 2024.
Contributor Information
Kristine A Karvonen, Public Health Sciences Division, Fred Hutchinson Cancer Center, Seattle, WA, United States; Division of Hematology/Oncology, Department of Pediatrics, University of Washington School of Medicine, Seattle, WA, United States.
Annie Vu, Department of Epidemiology and Biostatistics, University of California San Francisco, San Francisco, CA, United States; Greater Bay Area Cancer Registry, University of California San Francisco, San Francisco, CA, United States.
Katherine Lin, Department of Epidemiology and Biostatistics, University of California San Francisco, San Francisco, CA, United States; Greater Bay Area Cancer Registry, University of California San Francisco, San Francisco, CA, United States.
Joseph Gibbons, Department of Sociology, San Diego State University, San Diego, CA, United States.
Jason A Mendoza, Public Health Sciences Division, Fred Hutchinson Cancer Center, Seattle, WA, United States.
Eric J Chow, Public Health Sciences Division, Fred Hutchinson Cancer Center, Seattle, WA, United States; Division of Hematology/Oncology, Department of Pediatrics, University of Washington School of Medicine, Seattle, WA, United States.
Lena E Winestone, Helen Diller Family Comprehensive Cancer Center, University of California San Francisco, San Francisco, CA, United States; Division of Allergy, Immunology, and Blood & Marrow Transplant, Department of Pediatrics, University of California San Francisco Benioff Children’s Hospitals, San Francisco, CA, United States.
Scarlett L Gomez, Department of Epidemiology and Biostatistics, University of California San Francisco, San Francisco, CA, United States; Greater Bay Area Cancer Registry, University of California San Francisco, San Francisco, CA, United States; Helen Diller Family Comprehensive Cancer Center, University of California San Francisco, San Francisco, CA, United States.
Author contributions
Kristine A. Karvonen (Conceptualization, Formal analysis, Methodology, Project administration, Visualization, Writing—original draft), Annie Vu (Data curation, Investigation, Methodology, Software, Visualization, Writing—review & editing), Katherine Lin (Formal analysis, Investigation, Methodology, Software, Supervision, Writing—review & editing), Joseph Gibbons (Formal analysis, Methodology, Supervision, Writing—review & editing), Jason A. Mendoza (Conceptualization, Formal analysis, Methodology, Supervision, Writing—review & editing), Eric J. Chow (Formal analysis, Methodology, Supervision, Writing—review & editing), Lena E. Winestone (Data curation, Formal analysis, Methodology, Supervision, Writing—review & editing), and Scarlett L. Gomez (Conceptualization, Data curation, Formal analysis, Methodology, Resources, Software, Supervision, Writing—review & editing).
Supplementary material
Supplementary material is available at JNCI: Journal of the National Cancer Institute online.
Funding
California Department of Public Health pursuant to California Health and Safety Code Section 103885; Centers for Disease Control and Prevention’s (CDC) National Program of Cancer Registries, under cooperative agreement 1NU58DP007156; the National Cancer Institute’s Surveillance, Epidemiology and End Results Program under contract HHSN261201800032I awarded to the University of California, San Francisco, contract HHSN261201800015I awarded to the University of Southern California, and contract HHSN261201800009I awarded to the Public Health Institute, Cancer Registry of Greater California; ASCO Conquer Cancer Young Investigator Award 2023 (PI: Karvonen, AWD00000577); Seattle Children’s Hospital Center for Diversity and Health Equity Mentored Scholars Program (PI: Karvonen, HE-FY24-MS-07); Robert A. Winn Excellence in Clinical Trials Award Program (PI: Karvonen, 50012779); and American Cancer Society Clinician Scientist Development Grant (PI: Winestone).
Conflicts of interest
The authors have no conflicts of interest to report.
S.L.G., who is a JNCI associate editor and co-author on this paper, was not involved in the editorial review or decision to publish the manuscript.
Data availability
Deidentified data will be made available on reasonable request to the corresponding author.
References
- 1. Bhatia S. Disparities in cancer outcomes: lessons learned from children with cancer. Pediatr Blood Cancer. 2011;56:994-1002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Delavar A, Barnes JM, Wang X, et al. Associations between race/ethnicity and US childhood and adolescent cancer survival by treatment amenability. JAMA Pediatr. 2020;174:428-436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Moke DJ, et al. Emerging cancer survival trends, disparities, and priorities in adolescents and young adults: a California Cancer Registry-based study. JNCI Cancer Spectr. 2019;3:pkz031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Bailey ZD, Krieger N, Agénor M, et al. Structural racism and health inequities in the USA: evidence and interventions. Lancet. 2017;389:1453-1463. [DOI] [PubMed] [Google Scholar]
- 5. Jayasekera J, El Kefi S, Fernandez JR, et al. Opportunities, challenges, and future directions for simulation modeling the effects of structural racism on cancer mortality in the United States: a scoping review. J Natl Cancer Inst Monogr. 2023;2023:231-245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Trent M, Dooley DG, Dougé J, et al. ; COMMITTEE ON ADOLESCENCE. The impact of racism on child and adolescent health. Pediatrics. 2019;144:1-3. [DOI] [PubMed] [Google Scholar]
- 7. Best AL, Roberson ML, Plascak JJ, et al. Structural racism and cancer: calls to action for cancer researchers to address racial/ethnic cancer inequity in the United States. Cancer Epidemiol Biomarkers Prev. 2022;31:1243-1246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Braveman PA, Arkin E, Proctor D, et al. Systemic and structural racism: definitions, examples, health damages, and approaches to dismantling. Health Aff (Millwood). 2022;41:171-178. [DOI] [PubMed] [Google Scholar]
- 9. Rothstein R. The Color of Law: A Forgotten History of How Our Government Segregated America. 1st ed. Liveright Publishing Corporation; 2017. [Google Scholar]
- 10. Swope CB, Hernandez D, Cushing LJ. The relationship of historical redlining with present-day neighborhood environmental and health outcomes: a scoping review and conceptual model. J Urban Health. 2022;99:959-983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Plascak JJ, Beyer K, Xu X, et al. Association between residence in historically redlined districts indicative of structural racism and racial and ethnic disparities in breast cancer outcomes. JAMA Netw Open. 2022;5:e2220908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Lynch EE, Malcoe LH, Laurent SE, et al. The legacy of structural racism: associations between historic redlining, current mortgage lending, and health. SSM Popul Health. 2021;14:100793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Aaronson D, Faber J, Hartley D, et al. The long-run effects of the 1930s HOLC “redlining” maps on place-based measures of economic opportunity and socioeconomic success. Reg Sci Urban Econ. 2021;86:103622. [Google Scholar]
- 14. Egede LE, Walker RJ, Campbell JA, et al. Modern day consequences of historic redlining: finding a path forward. J Gen Intern Med. 2023;38:1534-1537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Fishback PV, LaVoice J, Shertzer A, et al. The HOLC maps: how race and poverty influenced real estate professionals’ evaluation of lending risk in the 1930s. J Econ Hist. 2023;83:1019-1056. [Google Scholar]
- 16. Markley S. Federal ‘redlining’ maps: a critical reappraisal. Urban Stud. 2024;61:195-213. [Google Scholar]
- 17. Bemanian A, Cassidy LD, Fraser R, et al. Racial disparities of liver cancer mortality in Wisconsin. Cancer Causes Control. 2019;30:1277-1282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Faber JW. We built this: consequences of new deal era intervention in America’s racial geography. Am Sociol Rev. 2020;85:739-775. [Google Scholar]
- 19. Richardson J, Mitchell BC, Meier HCS, Lynch E, Edlebi J. Redlining and Neighborhood Health. National Community Reinvestment Coalition; 2024.
- 20. Namin S, Xu W, Zhou Y, et al. The legacy of the Home Owners’ Loan Corporation and the political ecology of urban trees and air pollution in the United States. Soc Sci Med. 2020;246:112758. [DOI] [PubMed] [Google Scholar]
- 21. Wright E, Waterman PD, Testa C, et al. Breast cancer incidence, hormone receptor status, historical redlining, and current neighborhood characteristics in Massachusetts, 2005-2015. JNCI Cancer Spectr. 2022;6:1-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Poulson MR, Kenzik KM, Singh S, et al. Redlining, structural racism, and lung cancer screening disparities. J Thorac Cardiovasc Surg. 2022;163:1920-1930 e2. [DOI] [PubMed] [Google Scholar]
- 23. Krieger N, Wright E, Chen JT, et al. Cancer stage at diagnosis, historical redlining, and current neighborhood characteristics: breast, cervical, lung, and colorectal cancers, Massachusetts, 2001-2015. Am J Epidemiol. 2020;189:1065-1075. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Bikomeye JC, Zhou Y, McGinley EL, et al. Historical redlining and breast cancer treatment and survival among older women in the United States. J Natl Cancer Inst. 2023;115:652-661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Hussaini SMQ, Fan Q, Barrow LCJ, et al. Association of historical housing discrimination and colon cancer treatment and outcomes in the United States. J Clin Oncol Oncol Pract. 2024;20:678-687. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Umaretiya PJ, Vinci RJ, Bona K. A structural racism framework to guide health equity interventions in pediatric oncology. Pediatrics. 2022;149:1-5. [DOI] [PubMed] [Google Scholar]
- 27. Nelson RK, Winling L, Madron J, et al. Mapping Inequality: Redlining in New Deal America. D.S. Lab, Editor. 2023.
- 28. Michney TM. How the city survey’s redlining maps were made: a closer look at HOLC’s mortgagee rehabilitation division. J Plan Hist. 2022;21:316-344. [Google Scholar]
- 29. Krieger N, Kim R, Feldman J, et al. Using the Index of Concentration at the Extremes at multiple geographical levels to monitor health inequities in an era of growing spatial social polarization: Massachusetts, USA (2010–14). Int J Epidemiol. 2018;47:788-819. [DOI] [PubMed] [Google Scholar]
- 30. Krieger N, Van Wye G, Huynh M, et al. Structural racism, historical redlining, and risk of preterm birth in New York City, 2013-2017. Am J Public Health. 2020;110:1046-1053. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Gibbons J, Yang TC. Self-rated health and residential segregation: how does race/ethnicity matter? J Urban Health. 2014;91:648-660. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Gomez SL, Glaser SL, McClure LA, et al. The California Neighborhoods Data System: a new resource for examining the impact of neighborhood characteristics on cancer incidence and outcomes in populations. Cancer Causes Control. 2011;22:631-647. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Karp RJ. Redlining and lead poisoning: causes and consequences. J Health Care Poor Underserved. 2023;34:431-446. [DOI] [PubMed] [Google Scholar]
- 34. Nardone AL, Casey JA, Rudolph KE, et al. Associations between historical redlining and birth outcomes from 2006 through 2015 in California. PLoS One. 2020;15:e0237241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Fanta M, Ladzekpo D, Unaka N. Racism and pediatric health outcomes. Curr Probl Pediatr Adolesc Health Care. 2021;51:101087. [DOI] [PubMed] [Google Scholar]
- 36. Ellis DA, Cutchin MP, Carcone AI, et al. Racial residential segregation and the health of Black youth with type 1 diabetes. Pediatrics. 2023;151:1-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Blatt LR, Sadler RC, Jones EJ, et al. Historical structural racism in the built environment and contemporary children’s opportunities. Pediatrics. 2024;153: [DOI] [PubMed] [Google Scholar]
- 38. Karvonen K, Doody DR, Barry D, et al. Historical redlining and survival among children, adolescents, and young adults with cancer diagnosed between 2000-2019 in Seattle and Tacoma, Washington. Cancer. 2024. [DOI] [PubMed]
- 39. Adkins-Jackson PB, Chantarat T, Bailey ZD, et al. Measuring structural racism: a guide for epidemiologists and other health researchers. Am J Epidemiol. 2022;191:539-547. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Hematology/Oncology T. ASPHO Statements and Press Releases. ASPHO Condemns Racism, Supports Equity in Care for All 2020. https://aspho.org/about/aspho-statements.
- 41. Patel MI, Lopez AM, Blackstock W, et al. Cancer disparities and health equity: a policy statement from the American Society of Clinical Oncology. J Clin Oncol. 2020;38:3439-3448. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. National Cancer Institute. NCI statement on ending structural racism in biomedical research. 2021. https://www.cancer.gov/news-events/press-releases/2021/nci-statement-unite-end-structural-racism#:∼:text=NCI%20statement%20on%20ending%20structural%20racism%20in%20biomedical%20research,-Posted%3A%20March%202&text=NCI%20is%20committed%20to%20supporting, most%20innovative%20ideas%20against%20cancer.&text=As%20one%20of%20the%2027, entire%20NIH%20in%20supporting%20UNITE
- 43. American Society of Hematology. ASH/ASH RC/CMSS Statement on Racism in Healthcare. 2020. [11 December 2024]. https://www.hematology.org/newsroom/press-releases/2020/ash-cmss-statement-on-racism-in-healthcare.
- 44. Landrine H, Corral I, Lee JGL, et al. Residential segregation and racial cancer disparities: a systematic review. J Racial Ethn Health Disparities. 2017;4:1195-1205. [DOI] [PubMed] [Google Scholar]
- 45. Abraham IE, Rauscher GH, Patel AA, et al. Structural racism is a mediator of disparities in acute myeloid leukemia outcomes. Blood. 2022;139:2212-2226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Semprini J, Ali AK, Benavidez GA. Medicaid expansion lowered uninsurance rates among nonelderly adults in the most heavily redlined areas. Health Affairs. 2023;42:1439-1447. [DOI] [PubMed] [Google Scholar]
- 47. Miller-Kleinhenz JM, Moubadder L, Beyer KM, et al. Redlining-associated methylation in breast tumors: the impact of contemporary structural racism on the tumor epigenome. Front Oncol. 2023;13:1154554. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Valencia CI, Gachupin FC, Molina Y, Batai K. Interrogating patterns of cancer disparities by expanding the social determinants of health framework to include biological pathways of social experiences. Int J Environ Res Public Health. 2022;19:1-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Robinette JW, Charles ST, Gruenewald TL. Neighborhood socioeconomic status and health: a longitudinal analysis. J Commun Health. 2017;42:865-871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50. Diez Roux AV, Mair C. Neighborhoods and health. Ann N Y Acad Sci. 2010;1186:125-145. [DOI] [PubMed] [Google Scholar]
- 51. Warren Andersen S, Blot WJ, Shu X-O, et al. Associations between neighborhood environment, health behaviors, and mortality. Am J Prev Med. 2018;54:87-95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52. Khan SS, McGowan C, Seegmiller LE, et al. Associations between neighborhood-level racial residential segregation, socioeconomic factors, and life expectancy in the US. JAMA Health Forum. 2023;4:e231805. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53. Gibbons J. Evaluating the association between Home Owners’ Loan Corporation redlining and concentrated Black poverty. J Urban Affairs. 2023;47:799-812. [Google Scholar]
- 54. Wang G, Schwartz GL, Kershaw KN, et al. The association of residential racial segregation with health among U.S. children: a nationwide longitudinal study. SSM Popul Health. 2022;19:101250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55. Larrabee Sonderlund A, Charifson M, Schoenthaler A, et al. Racialized economic segregation and health outcomes: a systematic review of studies that use the Index of Concentration at the Extremes for race, income, and their interaction. PLoS One. 2022;17:e0262962. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56. Karvonen KL, McKenzie-Sampson S, Baer RJ, et al. Structural racism is associated with adverse postnatal outcomes among Black preterm infants. Pediatr Res. 2023;94:371-377. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57. Dyer L, Chambers BD, Crear-Perry J, et al. The Index of Concentration at the Extremes (ICE) and pregnancy-associated mortality in Louisiana, 2016–2017. Matern Child Health J. 2022;26:814-822. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58. Swope CB, Hernández D. Housing as a determinant of health equity: a conceptual model. Soc Sci Med. 2019;243:112571. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59. U.S.C. Bureau. American Community Survey 5-Year Estimates. 2015-2019. Accessed November 11, 2024. https://data.census.gov/
- 60. Arcaya M, Ellen I, Steil J. Neighborhoods and health: interventions at the neighborhood level could help advance health equity. Health Aff. 2024;43:156-163. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Deidentified data will be made available on reasonable request to the corresponding author.