Associations of area level deprivation with survival outcomes among patients with newly diagnosed colorectal cancer considered across the lifespan

Background: Colorectal cancer (CRC) is the second overall leading cause of cancer-related deaths in the United States (US). The Area Deprivation Index (ADI), a neighborhood socioeconomic disadvantage measure, reflects multiple CRC survival-related exposures including access to resources, healthcare, quality of life, and other opportunities for economic mobility. These exposures may vary by birth cohort, making it important to examine ADI across age of onset. The purpose of this study is to investigate the associations of ADI with overall (OS) and disease-free survival (DFS) by age of onset (early: <50 years (EO), average: 50-64 years (AO), older: ≥65 years (LO)) in the ColoCare Study. Methods: We included US data from the ColoCare Study, a prospective multicenter cohort study of newly diagnosed stage I-IV CRC patients. National- and state-level ADI were calculated using residential addresses captured at study entry. We compared ADI across ages of onset and estimated hazard ratios (HRs) and 95% confidence intervals (95% CI) for the associations of ADI with OS and DFS using multivariable Cox proportional hazard models, overall and stratified by age of onset, race, and ethnicity. Results: There were 2,477 participants from the ColoCare study with ADI data, with a median follow-up time of 3.6 years (SD=2.5) starting at the date of diagnosis and ending at the date of death, last follow-up, or censored at 5 years. On average, individuals in the LO group had higher national ADI (mean=4.08 [2.81]) compared to the AO group (mean=3.93 [2.85]; P<0.001) and early onset group (mean=3.49 [2.72]; P<0.001]). Survival associations of the ADI based on national percentiles tended to be stronger than those based on the state. Considered per 10-percentile increase in national-level ADI, there was an 8% (95% CI=1.01, 1.15; P=0.02) and 10% (95% CI= 1.01, 1.21; P=0.03) higher overall mortality risk among LO and EO groups, respectively. After adjusting national-level ADI for race, there was a 6% and 3% higher overall mortality risk among the LO and AO groups, respectively. Among average onset individuals, the association of the ADI measures with OS was closer to the null and not statistically significant. Associations of ADI with DFS were weaker and not statistically significant among the EO and LO groups, but the ADI was positively associated with DFS among the AO groups (HR DFS =1.09; 1.01-1.16; P=0.02). Though not statistically significant, the associations of ADI with OS and DFS were strongest among Non-Hispanic Black individuals (HR OS = 1.12; 95% CI=0.92, 1.36; P=0.27; and HR DFS =1.14; 0.93-1.40; P=0.20). Conclusion: We observed that older CRC survivors tended to live in more deprived areas; however, across the lifespan, area level determinants of health were associated with poorer CRC survival outcomes.