Skip to content

Abstract Search

Cancer

Historical redlining and risk of breast cancer in the NIH-AARP Diet and Health Study Aleah Thomas* Aleah Thomas Jonguen Rhee Rena R. Jones Lauren M. Hurwitz Jared A. Fisher Jessica M. Madrigal

Background: The 1930’s federal Home Owners’ Loan Corporation (HOLC) enacted residential racial segregation through ‘redlining.’ This historical practice may contribute to modern residential segregation and social inequities, potentially increasing cancer risk. However, evidence of the association between redlining and cancer is limited. We investigated the association between historical redlining and incident breast cancer in a large prospective cohort.

Methods: We linked georeferenced HOLC grading data (1937-1940) from the University of Richmond to enrollment (1995-1996) addresses for the NIH-AARP Diet and Health Study. A total of 25,944 women (86% of women in the cohort) were included. Multivariable Cox models estimated HRs and 95% CIs for HOLC grade at participant address (best, still desirable, definitely declining, redlined) in relation to incident breast cancer overall (2,341 cases) and by estrogen (ER) and progesterone (PR) receptor status, after adjusting for breast cancer risk factors, census tract level median household income, and ambient air pollution at the residence.

Results: Redlined neighborhoods (vs. other HOLC grades) had lower median household income and had larger proportions of participants from racial and ethnic minority groups and those with lower educational attainment. We found no association between HOLC grade and breast cancer risk overall (HR [redlined vs. best or still desirable]=0.90, 95% CI=0.80, 1.03). We observed suggestive positive associations for hormone receptor-negative tumors [ER-, HR=1.25, 95% CI=0.85, 1.84; PR-, 1.14 (0.85, 1.52)] and inverse associations for hormone receptor-positive tumors [ER+, 0.82 (0.69, 0.97); PR+, 0.78 (0.65, 0.95)].

Conclusion: We did not find evidence that residence in historically redlined neighborhoods is associated with overall breast cancer risk. Possible associations by hormone receptor status warrant further evaluation.