Abstract PR005: How racial segregation impacts late stage at cancer diagnosis in Virginia

Abstract

Abstract Background: Racial segregation has often been cited as a “fundamental cause of health disparities” because it creates conditions inimical to health, impacting community investments (structure), access to care and resources. Virginia (VA), which occupies three of the top 15 highest cancer mortality rates in the nation, also has the 23rd highest Black-white health gap (only two spots above Mississippi 25th spot) and the 48th largest Black-white gap in percent of adults with a bachelor's degree. We sought to estimate the relationship between racial segregation and late-stage cancers in VA and the extent to which the relationship is mediated through intervenable factors such as access to care, individual and/or structural barriers. Methods: Our outcome of interest was county-level late stage cancers (defined as the proportion of cancers diagnosed at regional and distant stage from 2013 to 2017), dichotomized at the median (45%). We used county-level data from the Robert Wood Johnson Foundation’s county health profiles across domains of demographics (e.g., % Black, % female), individual (e.g., % smoking, % reporting mental distress), access to care (e.g., % screened for colon cancer - CRC screened, % uninsured, % premature mortality) and structural characteristics (e.g., broadband access, air pollution, violent crime). We built logistic regression (LR) models with backwards selection (removing terms with p ≥ 0.2 and adding those with p <0.1) to identify most salient predictors for each of these domains. We then combined these most salient predictors into one final LR model using the same backwards approach and developed a structural equations model (SEM) to estimate the direct and indirect effect of percent black on late stage at cancer diagnosis. Results: Percent late stage cancer diagnoses and % Black ranged from 35% to 55% and 0 to 76% respectively across the VA’s 133 counties and independent cities. Of the 49 variables considered across domains, % CRC screened, % uninsured, % white, and % female, remained statistically significant at p<0.05. In the final model, the odds of late stage at cancer diagnosis increased ~35% (p=0.05, 95%CI: 1.0-1.8) for every one quintile increase in the % Black at the county level. In SEM models, 32% of the effect of % Black on late stage cancer diagnosis was mediated through % uninsured and an additional 16% through the effect of % uninsured on % CRC screened. There was no significant relationship between %Black and late stage cancer diagnosis (p=0.12).Conclusion: Results from this analysis reveal that in VA, the association between %Black and late stage at cancer diagnosis seems to operate fully through the influence of %Black on access to care variables such as % uninsured and %CRC screening. This suggests that continued investment in access could ameliorate the impact of systemic inequalities such as racial segregation in VA communities. Future efforts to understand the effect of %Black on late stage cancers at more granular geographic levels and across specific cancer types are needed. Citation Format: Savannah Reitzel, Jinlei Zhao, Robert Perera, Robert A. Winn, Katherine Y. Tossas. How racial segregation impacts late stage at cancer diagnosis in Virginia [abstract]. In: Proceedings of the 15th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2022 Sep 16-19; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2022;31(1 Suppl):Abstract nr PR005.