In a cohort of breast cancer patients from New Castle County, Delaware, a geographic area with among the highest rates of TNBC in the US, we tested whether different versions of the ICE metric could efficiently identify census tracts with greater odds of TNBC relative to other invasive subtypes of breast cancer. Consistent with prior epidemiological findings, women with TNBC were younger, twice as likely to be Black, more likely to have Medicaid or no insurance, and twice as likely to present with a late-stage cancer. As hypothesized, ICE-Race, -Income, and ICE-Race/Income metrics were associated with the odds of TNBC on a bivariate basis. However, contrary to our hypotheses, only the ICE-Race metric was significantly associated with higher odds of TNBC in multilevel models that adjusted for patient-level age and race. To our knowledge, this is the first multilevel study that evaluated the use of ICE metrics in the context of TNBC.
By including both patient- and census tract-level measures of race, our results help to clarify the extent to which the higher rates of TNBC observed in predominantly Black neighborhoods can be attributed to compositional or contextual effects. That is, if the relationship between ICE-Race and census tract odds of TNBC became nonsignificant when adjusting for patient-level race, we might infer that the apparent neighborhood effects were likely an artifact of the neighborhood composition. By contrast, our results suggest that the relationship between TNBC and area-level measures of race are a function of both composition and context. Stated differently, compared to predominantly White neighborhoods, predominantly Black neighborhoods may differ in some important ways that have relevance for TNBC risk. This interpretation is bolstered by the multilevel results stratified by patient-level race, which showed that White patients who lived in predominantly Black census tracts were also at greater odds of TNBC than White patients living in predominantly White census tracts.
These results are in partial disagreement with the findings from two prior studies, which found higher odds of TNBC for Black women living in lower-SES, predominantly White neighbourhoods.[15,64] The authors of these reports reasoned that Black women who reside in predominantly Black neighborhoods may have social support systems that mitigate the effects of living in a low-SES environment. However, the potentially protective effects of social support derived from racial/ethnic enclaves on cancer outcomes has not been well-studied in Black populations compared to Hispanic or Asian populations. Not only did we find that census tracts with a higher proportion of Black residents were associated with higher odds of TNBC, we also did not observe any significant patient-level race by ICE-Income cross-level interactions. Our results are more consistent with a segregation-based disparate exposure hypothesis. That is, through a series of historical laws and policies (e.g., "redlining"), the US, state, and local governments forcibly segregated communities by race and denied these marginalized communities access to financing and other forms of investment, a residential pattern that largely persists through present day.[65–67] The effect of these segregationist policies can be measured in terms of disparate exposures (e.g., increased concentration of alcohol retailers),[68,69] poorer access to healthy food, and ultimately worse cancer and other health outcomes.[71–73] Indeed, we observed higher counts of alcohol and fast-food retailers, and correspondingly higher rates of AUD and obesity, in census tracts that were categorized into the most disadvantaged ICE-Race quintile and had the highest odds of TNBC.
This study was limited by its single-site, cross-sectional design. Findings may not be generalizable to populations from other geographic areas. Nevertheless, given the notably elevated rates of TNBC in Delaware, the results of this study can help to inform local cancer control and prevention efforts while providing a methodological proof of concept that can be replicated for other geographic areas. Without patient residential histories, it is unclear to what degree prior neighborhood exposures may have been associated with TNBC risk. Furthermore, while we did have access to patient-level measures of race and insurance status, we did not have patient-level measures of other relevant exposures (e.g., alcohol use) and were not able to determine ethnicity (Hispanic vs. non-Hispanic) at the patient level. Further, we limited our investigation of neighborhood to include ICE measures, given there is no standard set of measures used to measure neighborhood deprivation. It is possible other socioeconomic indices (e.g., Yost index, ADI) could also provide additional insights into the impact of neighborhood on TNBC. Future research should be conducted on cohorts from a range of geographic areas, with more detailed patient- and area-level measures of exposure, to further characterize the multilevel relationships between race, SES, and TNBC.
Breast Cancer Res. 2022;24(37) © 2022 BioMed Central, Ltd.
Copyright to this article is held by the author(s), licensee BioMed Central Ltd. This is an Open Access article: verbatim copying and redistribution of this article are permitted in all media for any purpose, provided this notice is preserved along with the article's original citation.