Abstract
Background: In Louisiana, colorectal cancer (CRC) incidence and mortality exceed national rates. Census tract, sex, and racial disparities across the state are well documented. This study examined whether there were subpopulation differences in associations between CRC screening, area deprivation index (ADI), and patient characteristics. Methods: This retrospective observational study included patients aged 50 to 75 years who received care within Ochsner Health in Louisiana between July 1, 2012, and December 31, 2020. Logistic regression models were used to generate adjusted odds ratios (95% CI). Results: A total of 75,344 patients met eligibility criteria for inclusion in the data analysis (60% female, 36% Black, 56% with spouse/partner, 42% Medicare/Medicaid,17% living in high deprivation areas, 41% with 2+ chronic conditions, 56% never smoked, 51% obese). Living in areas with less deprivation (state decile 1-3 vs 8-10: 1.19 [1.14-1.24]), number of comorbidities (3+ conditions: 1.15 [1.12-1.17]), and prior outpatient visits (1.63 [1.58-1.67]) increased odds of CRC screening. Male sex (0.82 [0.79-0.84]), age group 55 to 59 years (0.97 [0.95-0.99]), and Medicaid insurance (0.89 [0.86-0.92]) decreased odds of screening. ADI was collinear with sex, race, marital status, body mass index, and smoking status. In subgroup analyses, between-group differences in strength of associations of CRC screening with ADI and patient characteristics varied most prominently by race. Conclusion: There may be an unmeasured social context explaining persistent racial differences among factors associated with CRC screening. A combination of census tract and individual-level social determinants may guide population health management for at-risk subpopulations.