Association Between Socioeconomic and Insurance Status and Delayed Diagnosis of Gastrointestinal Cancers.

INTRODUCTION: Association between socioeconomic status (SES) and stage at diagnosis in gastrointestinal (GI) cancers is poorly described. Relationship between low SES and stage at diagnosis as well as the mediating role of insurance status (IS) was examined. METHODS: The Surveillance, Epidemiology, and End Results database was queried for esophageal, gastric, liver, biliary, pancreatic, colon, and rectal cancers diagnosed in 2012-2016. Relationship between census-tract SES index quintiles and late diagnosis (distant disease at diagnosis) was examined. Uni and multivariable logistic regressions were performed. Mediation analyses were conducted to determine the degree to which IS (private/Medicare versus Medicaid/uninsured) mediates the relationship between SES and late diagnosis of cancer. RESULTS: Analysis included 236,713 adult patients from 18 Surveillance, Epidemiology, and End Results areas. In univariable analysis, lowest SES quintile was significantly associated with late diagnosis for all cancers except gastric and biliary cancers. In multivariable analysis controlling for age, gender, marital status and race, this association remained significant for liver (odds ratio (OR) 1.41 [95% confidence interval (CI) 1.25-1.58]), pancreatic (OR 1.13 [95% CI 1.06-1.21]), and rectal (OR 1.31 [95% CI 1.20-1.42]) cancers. Further controlling for IS showed the largest effect size reduction for rectal cancer (OR 1.18 [95% CI 1.09-1.29]), with IS mediating 36.5% (P < 0.0001) of SES effect. CONCLUSIONS: Low SES is an independent risk factor for late diagnosis in liver, pancreas, and rectal cancers. Insurance is not a critical mediator of difference by SES for most GI cancers, with the exception of rectal cancer. Further research is needed to understand factors beyond IS that can account for SES differences in late diagnosis for GI cancers. Insurance related differences for rectal cancer deserves further attention.