Neighborhood-level deprivation mediates racial and ethnic disparities in HCC diagnosis in Texas

Background:Texas has the highest HCC rates in the United States, and the greatest burden is among Hispanics. Racial and ethnic disparities in HCC incidence have multiple underpinning factors. We conducted a mediation analysis to examine the role of neighborhood disadvantage (Area Deprivation Index) as a potential mediator of the association between neighborhood race and ethnicity distribution and neighborhood HCC case counts in Texas.Methods:The primary outcome measure was counts of new HCC diagnoses per census tract based on Texas Department of State Health Services Texas Cancer Registry data. The primary exposure of interest was the race and ethnicity-based Index of Concentration at the Extremes (non-Hispanic Black ICE or Hispanic ICE). We assessed Area Deprivation Index as a potential mediator of the association between Black/Hispanic ICE and HCC case counts. We adjusted the analyses for selected census tract characteristics.Results:We analyzed 4934 census tracts containing 13,632 new HCC diagnoses reported to Texas Cancer Registry between 2016 and 2020. Racial minority (Black/Hispanic ICE)-concentrated neighborhoods had a higher socioeconomic disadvantage. The results of the mediation analyses showed that compared to non-Hispanic White-concentrated census tracts, non-Hispanic Black-concentrated census tracts and Hispanic-concentrated census tracts had higher case counts of HCC (total effects: adjusted case count ratio: 1.03 [95% CI, 1.02-1.04] and adjusted case count ratio: 1.09 [95% CI, 1.08-1.10], respectively). Approximately 48% and 15% of the neighborhood-level disparity in HCC case counts were attributable to neighborhood socioeconomic disadvantage in Black and Hispanic minoritized neighborhoods, respectively.Conclusions:Neighborhood HCC case counts varied by neighborhood race and ethnicity distribution. The variations were partly explained by neighborhood deprivation, with a stronger effect among Black-concentrated census tracts.