Association of residence within Hispanic ethnic enclaves and late-stage colorectal cancer diagnosis among Hispanics in four U.S. states
Introduction
Colorectal cancer (CRC) is the second-most common cancer among Hispanic or Latino (hereafter Hispanic) individuals. We investigated the association between residence in Hispanic enclaves (areas with a high concentration of Hispanic individuals, immigrants, and those who speak Spanish and are linguistically isolated or with limited English) and late-stage CRC diagnosis in a population-based study.
Methods
We used data from population-based cancer registries in four states (CA, NJ, NY, TX) to identify Hispanic individuals diagnosed with CRC between 2000 and 2017. We computed Hispanic enclave index using principal components analysis of four census tract-level measures, i.e., percentage of residents who: are Hispanic/Latino, are foreign-born Hispanic/Latino, speak limited English and speak Spanish, and percentage of linguistically isolated households who primarily spoke Spanish. We classified all census tracts into quintiles of ethnic enclave scores (pooled across all four states) and linked these to patient's address at diagnosis. We conducted log binomial regression using robust variance estimators to examine association of enclave residence (in quintiles) with late-stage (regional or distant) compared to early-stage (in-situ or localized) CRC diagnosis. We fit uni- and multivariable models, incorporating age, sex, year of diagnosis, insurance type, state, metropolitan/non-metropolitan census tract residence, and census tract percent poverty as covariates and adjusted for clustering by census tract.
Results
Among 86,766 Hispanic individuals with CRC, 59.6% were diagnosed at late stage. Three-fourths of Hispanic individuals resided in census tracts with higher ethnic enclave scores: 25.8% in Q4 and 47.5% in Q5. The percentage of late-stage CRC diagnoses varied across ethnic enclave quintiles (56.7% in Q1 to 60.2% in Q5; Chi-square across all quintiles p<0.001). Residence in the highest (Q5) compared to lowest (Q1) enclave quintile was associated with increased risk of late-stage diagnosis in the unadjusted model (RR=1.06; 95% CI: 1.03-1.10); however, this association was no longer statistically significant in the fully adjusted model (RR=1.01; 95% CI: 0.98-1.05).
Conclusion
Given that a substantial proportion of Hispanic individuals reside in ethnic enclaves, further research to comprehensively investigate outcomes across the cancer continuum within these communities is warranted. Understanding the complexities surrounding ethnic enclave residence can inform interventions to improve early detection and diagnosis of CRC, ultimately addressing disparities in Hispanic populations' outcomes across the cancer continuum. Future analyses will assess for interaction by sex, neighborhood poverty, and state of residence.