Abstract
PURPOSE: We examined whether living in persistent poverty census tracts was associated with disparities in colorectal cancer (CRC) survival and whether the association varied between urban and rural settings. METHODS: Using 2013-2019 state-wide cancer registry and 2013-2021 death records data, CRC patients were classified by tract-level persistent poverty and rural/urban status. Overall and CRC-specific survival were compared using Kaplan-Meier estimation and log-rank tests. Adjusted analyses were conducted using Cox proportional hazard and Fine-Gray competing risk models. FINDINGS: During the study period, 558 (53%) of 1055 CRC patients died in persistent poverty tracts versus 3117 (45%) of 6938 patients in nonpersistent poverty tracts. Of the 3675 deaths, 2269 (61.7%) were from CRC-specific causes. In unadjusted analysis, CRC patients in persistent poverty areas had a higher risk of all-cause (HR, 95%CI: 1.28, 1.17-1.40) and CRC-specific (HR, 95% CI: 1.17, 1.04-1.31) mortality. After covariates adjustment, the relationship between persistent poverty and all-cause mortality (HR, 95% CI: 1.17, 1.06-1.29) and non-CRC-specific mortality (HR, 95% CI: 1.34, 1.15-1.57) remained significant, but CRC-specific mortality did not. In subgroup analyses, persistent poverty was associated with increased overall mortality among urban tracts (HR, 95% CI: 1.22, 1.08-1.38), but not rural tracts. CONCLUSIONS: After covariates adjustment, CRC patients in persistent poverty tracts are more likely to die of all causes and non-CRC causes but not CRC-specific causes than those in nonpersistent poverty areas, suggesting that differences in CRC-specific deaths may be partly attributed to demographics, geography, tumor characteristics, and treatment.