Persistent Neighborhood Poverty and Breast Cancer Outcomes

Key Points Question Is residential persistent poverty associated with breast cancer characteristics, treatment, and mortality? Findings In this cohort study analyzing 312 145 patients in the Surveillance, Epidemiology, and End Results Program, individuals residing in areas with persistent poverty experienced more-aggressive tumor characteristics and underwent higher rates of mastectomy and axillary lymph node dissection compared with individuals residing in areas without persistent poverty. In addition, impoverished individuals had a 10% increased risk of breast cancer–specific mortality and a 13% increased risk of all-cause mortality. Meaning The findings of this study suggest that residing in persistently impoverished neighborhoods is associated with poor tumor characteristics and increased mortality.


Introduction
The social and built environments in which people live have well-established major implications for health behaviors, access to resources, and consequential health outcomes. 13][4][5] Neighborhood racial composition, segregation, and income have also been associated with breast cancer-specific and all-cause mortality. 6rsistent poverty refers to geographic areas with poverty rates greater than or equal to 20% over the last 30 years. 7Persistent poverty, in contrast to transitory or chronic poverty, refers to geographic locations rather than individuals and families with high poverty rates for an extended time. 7,8Areas of persistent poverty are characterized by systemic and structural decay, as property values yielding low investment returns disincentivize property owners from spending money to maintain and/or improve property. 7,8Basic necessities, such as public services (eg, utilities, public transportation, law enforcement), food accessibility, education, health care services, support services, and social programs, are affected.[9] Recent studies have highlighted how people living in geographic areas with persistent poverty experience higher cancer mortality rates, over and above the risk associated with current poverty. 10,11tients with breast cancer living in areas of persistent poverty may have more-advanced disease at diagnosis and differences in treatment; however, additional research is needed. 12The purpose of this study was to evaluate persistent poverty over time at the census tract (CT) level, which more closely resembles neighborhoods, and breast tumor characteristics, treatment, and mortality.

Data Source
The Surveillance, Epidemiology, and End Results (SEER) Program currently represents nearly 50% of the US population with cancer. 13A collection of 18 SEER Program registries was used to identify women aged 18 years or older with stage I to III breast cancer diagnosed from January 1, 2010, to December 31, 2018 (eFigure 1 in Supplement 1). 14This specialized dataset contains information about CT-based measures of rural-urban status and persistent poverty.Patients with missing exposure (ie, CT-persistent poverty), survival, sociodemographic, or clinical data were excluded (203 054 patients of the original 587 107 breast cancer cohort [34.6%]).Additionally, patients with American Joint Committee on Cancer stage IV or SEER summary stage representing distant disease and those who received local tumor surgical treatment were excluded.The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies was followed. 15This study was exempt from institutional review and informed consent was waived according to the Common Rule (45 CFR §46), as the analysis of deidentified, publicly available data is not considered human participant research.

Sociodemographic, Clinical, and Treatment Characteristics
Sociodemographic variables collected included age, race and ethnicity (Black, Hispanic, White, Other), marital status (married/partnered, single, separated/divorced, and widowed), rural-urban status, and Yost Index. 16Due to persistent racial and ethnic disparities in breast cancer outcomes, race and ethnicity were included in the study analysis.Racial and ethnic categories were abstracted from medical records, hindering our ability to identify the method of classification, as the initial collection of this information varies by health care facility and practitioner. 17

Exposure
Persistent poverty was defined as CTs in which 20% or more of the population lived below the poverty level for at least 30 years based on 1990 and 2000 decennial censuses and 2007-2011 and 2015-2019 American Community Survey 5-year estimates. 19Census tracts are subdivisions of a county with an average of 4000 residents that allow for more precise estimates of persistent poverty. 8Use of CTs recently noted an additional 3% of the total US population was living in persistent poverty who otherwise would have been missed with the use of county-level measures. 8tients were categorized as residing vs not residing in a persistently impoverished CT.

Outcomes
The primary outcomes were all-cause mortality and breast cancer-specific mortality.The follow-up period was from date of diagnosis to the event date, defined as the date of death for patients who died or either the date of last follow-up or the end of the study for patients presumed to be alive (December 31, 2020), whichever came first.Patients who did not experience the event were censored at the date of the last follow-up or at the end of the study, whichever occurred earlier.For breast cancer-specific mortality, patients were censored on the non-breast cancer event date if the cause of death was due to causes other than breast cancer.Secondary outcomes included breast surgery type, axillary surgery type, and receipt of breast reconstruction.

Statistical Analysis
Data analysis was performed from August 2023 to March 2024.Sociodemographic, clinical, and treatment characteristics were summarized using means (SDs) for continuous variables and frequencies and percentages for categorical variables.Differences in sociodemographic, clinical, and treatment characteristics were compared between patients living in CTs with and without persistent poverty, using t tests for continuous variables and χ 2 tests for categorical variables.With 2-sided, unpaired testing, P < .05 was considered statistically significant.
Differences in all-cause and breast cancer-specific mortality were assessed using Kaplan-Meier analysis.Crude and adjusted Cox proportional hazards regression models were fitted to assess the association between persistent poverty and all-cause mortality and between breast cancer-specific mortality and persistent poverty.Proportional hazards assumptions were confirmed visually using the negative-log curve of the survival distribution function and by including a time-dependent covariate in the regression model.Final models were adjusted for age, race and ethnicity, marital status, Rural Urban Commuting Area code, pathologic grade, SEER stage, molecular subtype, receipt of chemotherapy and radiotherapy, and breast and axillary surgery type.
All-cause mortality and breast cancer-specific mortality rates were estimated at 3, 6, and 9 years of follow-up by dividing the number of observed deaths by the total person-years of follow-up for patients living in CTs with and without persistent poverty.Mortality rate differences were calculated by subtracting the mortality rate for patients living in CTs without persistent poverty from mortality rates for patients living in CTs with persistent poverty.Crude rate ratios (RRs) were calculated by dividing the mortality rate for patients living in CTs with persistent poverty by mortality rates for patients living in CTs without persistent poverty.Analysis was conducted with SAS, version 9.4 software (SAS Institute).

Surgical Treatment
Regardless of persistent poverty status, use of lumpectomy increased from 2010 to 2018, whereas use of mastectomy decreased (eFigure 2 in Supplement 1).However, patients living in persistently impoverished areas were consistently more likely to undergo a mastectomy rather than lumpectomy.
Similarly, rates of SLNB increased, whereas performance of ALND decreased over time, but patients living in areas with persistent poverty were more likely to undergo ALND vs SLNB throughout the study period (eFigure 3 in Supplement 1).Although overall use of breast reconstruction increased over time, patients residing in persistently impoverished areas remained consistently less likely to receive reconstruction (eFigure 4 in Supplement 1).

Discussion
Patients in this cohort with stage I to III breast cancer living in persistently impoverished neighborhoods were more likely to identify with historically marginalized and minoritized populations (ie, Black race or Hispanic ethnicity), live in more rural regions, and present with moreaggressive tumor characteristics, including higher grade, molecular subtypes with worse prognostication, and greater disease burden.Although surgical management remained similar between groups, patients residing in persistently impoverished areas were consistently more likely to undergo ALND.Furthermore, patients living in persistently impoverished areas undergoing mastectomy were less likely to undergo breast reconstruction.A larger proportion of patients residing in areas with persistent poverty had a higher likelihood of receiving chemotherapy.Both all-cause and breast cancer-specific mortality risk remained higher among those living in CTs with persistent poverty despite receiving more aggressive treatments.
The role of persistent poverty in cancer outcomes has only recently gained attention in the literature, and our findings are consistent with existing research. 20Moss et al 10 noted that patients living in counties with persistent poverty were disproportionately non-Hispanic Black and Hispanic, with a 1.6 per 100 000 person-year increase in cancer mortality rate compared with currently impoverished (but not persistently impoverished) counties, following adjustment of sociodemographic factors.Similarly, Papageorge et al 12 noted that a higher proportion of patients living in persistent poverty were categorized as non-Hispanic Black and less likely to have HR-positive disease.Patients living in persistently impoverished counties were also more likely to undergo a  radical mastectomy with higher risk of cancer-specific mortality after adjusting for demographic characteristics, molecular subtype, and stage of disease.Papageorge et al 12 only included patients with breast cancers diagnosed between 2012 and 2016 and evaluated persistent poverty at a county level, which inherently leads to a larger degree of imprecision with both overinclusion and underinclusion of targeted areas. 21Their study also only evaluated breast surgery without consideration of axillary or reconstructive surgery or receipt of chemotherapy or radiotherapy, all of which have major implications on survival. 22The present study noted that patients living in persistently impoverished CTs had increased risks of all-cause and breast cancer-specific mortality despite controlling for all clinical and treatment factors, suggesting the likely presence of an additional distinct and likely nuanced mechanistic relationship associating persistent poverty with mortality beyond treatment receipt.
We also noted widening mortality gaps for both all-cause and breast cancer-specific mortality based on persistent poverty status, despite similar patterns and stable differences in surgical   23 Prior studies have established an association between relative poverty (rather than absolute poverty) and health outcomes, where counties with higher levels of relative income inequality experience greater cancer mortality. 24Greater income inequality has been hypothesized to encourage "resentment…hopelessness, and alienation," leading to a "sense of injustice, discontent, and distrust" that erodes social cohesion. 25Social buffers (eg, social networks) disintegrate, which leads to withdrawal and disorganization of community structure and departure of businesses with loss of job opportunities, all of which fuel a cycle of declining social capital and further disinvestment. 25storical institutional policies and private practices, such as the Jim Crow laws, which enforced residential racial segregation and differential resource allocation in the South, where most CTs with persistent poverty are located, further concentrated areas with poverty such that residents must cope with the consequences of both their own lack of income and lack of neighborhood income (eg, lack of labor force role models, high education dropout rates, exposure to unsupervised peer group activity). 8,25Moreover, the structural forces that contribute to the creation of persistent poverty limit any plausibility of upward economic mobility, particularly among low-income Black families, which only perpetuates intergenerational poverty. 21,26,27In rural areas, where there is already a shortage of health care personnel and supplies, lower insurance coverage, and longer distance to health care facilities, the collapsed infrastructure in persistently impoverished areas further restricts access to healthy food, recreational facilities, and healthy behaviors. 28e mechanism in which living in persistent poverty is associated with increased mortality risk warrants further evaluation.One plausible mechanism involves the chronic activation of the hypothalamic-pituitary-adrenal axis and the sympathetic-adrenal-medullary pathway, often measured through allostatic load. 29Prolonged exposure to adverse socioenvironmental stressors, such as neighborhood deprivation and racialized segregation, has been associated with high allostatic load burden. 30,31Studies have also evaluated associations between allostatic load with childhood and adult socioeconomic status and the impact of upward economic mobility. 32,335][36] However, studies evaluating the role of persistent poverty on allostatic load are needed.

Limitations
Interpretation of our findings is not without limitations.We are limited by the information SEER provides, such as the inability to distinguish between individuals who did not receive chemotherapy and/or radiotherapy from those for whom this information is unknown, which may inadvertently lead to misclassification bias.Comorbidity data are also not available in SEER, limiting our ability to control for the role of comorbidities in all-cause mortality.Persistent poverty was depicted through CTs, which better aligns with the association between neighborhoods and individual outcomes. 21wever, CTs lack administrative boundaries and are too small and numerous for targeting economic development-related interventions. 21Additionally, complete geographic addresses are not always available for geocoding residential CTs, which may limit the quality of the persistent poverty measure.Individual addresses at the time of diagnosis similarly do not account for population mobility and residential history.Focusing on persistent poverty may also underestimate and not fully capture the multidimensional nature of the barriers that families living in persistently disadvantaged areas may experience. 37Nevertheless, the findings from our study add to the literature advocating for the necessity to advance equity for patients residing in persistently impoverished areas.

Figure 1 .
Figure 1.Overall Survival by Residence in Census Tracts With and Without Persistent Poverty (N = 312 145) 100 American Indian or Alaska Native, Asian or Pacific Islander, and Other were collapsed into an Other category given small sample sizes.Racial categories in SEER are a social construct as genetic ancestry is not available in JAMA Network Open.2024;7(8):e2427755.doi:10.1001/jamanetworkopen.2024.27755(Reprinted) August 29, 2024 2/11 Downloaded from jamanetwork.comby guest on 09/26/2024

Table 1 .
Sociodemographic, Clinical, and Treatment Characteristics by Residence in Persistently Impoverished CTs (continued) a P values calculated using t test for continuous variables and χ 2 tests for categorical variables.bThe Other category includes patients identified as American Indian or Alaska Native, Asian or Pacific Islander, and Other, collapsed due to small sample sizes.

Table 2 .
Association Between Residence in a Persistently Impoverished Census Tract and All-Cause Mortality a Adjusted for age; race and ethnicity; marital status; 11ss et al11similarly reported widening health disparities for mortality outcomes, particularly for Black patients living in persistently impoverished rural counties.While the cause of this expanding gap requires further investigation, the ever-growing wealth gap between the wealthiest and poorest families in the US has more than doubled between 1989 and 2016.
Study Schema Showing Patient Inclusion/Exclusion Criteria eTable 1.Crude Incidence, Rate Difference and Risk Ratio of Mortality At 3, 6, and 9 Years Between Patients Living in Persistently Impoverished Census Tracts vs Patients Living in Non-Persistently Impoverished Census Tracts eTable 2. Crude Incidence, Rate Difference and Risk Ratio of Breast Cancer-Specific Mortality at 3, 6, and 9 Years Between Patients Living in Persistently Impoverished Census Tracts vs Patients Living in Non-Persistently Impoverished Census Tracts eFigure 2. Breast Surgery by Residence in Persistently Impoverished Census Tracts Over Time eFigure 3. Axillary Surgery by Residence in Persistently Impoverished Census Tracts Over Time eFigure 4. Reconstruction Surgery by Residence in Persistently Impoverished Census Tracts Over Time