Historical redlining and survival among children, adolescents, and young adults with cancer diagnosed between 2007-2019 in Seattle/Tacoma

Historic redlining has been associated with inferior survival in adult-onset cancers, however whether there is an association in pediatric, adolescent, and young-adult-onset cancers is unknown. Methods: Cancer cases among individuals living in Seattle and Tacoma between 2007-2019 were identified via Cancer Surveillance System. Case redlining status was determined using Home Owner Loans Corporation (HOLC) data overlayed with 2000 and 2010 census tracts. Kaplan-Meier methods and multivariable Cox proportional hazards models were used to determine 5 and 15-year overall survival and hazard ratio (HR) of death between redlined exposed and unexposed cases. A series of Cox regression models were tested including an unadjusted model, and adjusted models for patient and tumor-characteristics alone, patient and tumor-characteristics with area-level poverty, and patient and tumor-characteristics with area-level poverty and individual insurance status. Finally, a model accounting for interaction between redlining and poverty was included to test for effect modification. Results: Unadjusted overall survival at 5 years was lower (86.0%; 95% CI: 84.0, 87.7) in individuals with cancer exposed to redlining than unexposed individuals (90.5%; 95% CI: 88.9, 91.8). The unadjusted hazard of death for redlined-exposed individuals with cancer was higher than redlined-unexposed (HR: 1.53, 95% CI: 1.25, 1.86). In the fully adjusted model, the hazard of death between individuals exposed to redlining was substantially attenuated (1.17, 95% CI: 0.95, 1.43). There was insufficient evidence of effect modification from area-level poverty in the relationship between redlining and death (p= 0.61). Conclusion: In this retrospective population-based study of children, adolescents and young adults with cancer, residence at the time of diagnosis in an area previously redlined in the 1930s was associated with an increased hazard of death compared to those who were not living in a previously redlined area. Notably, this relationship was attenuated after adjusting for patient and tumor factors, area-level poverty and insurance, suggesting that contemporary socioeconomic status may mitigate the harmful effects of historical redlining.