Association of socioeconomic status with incidence and outcomes of gastric cancer in a Michigan cohort.

Abstract

25 Background: The purpose of this study is to examine the association between socioeconomic factors and gastric cancer (GC) outcome in a cohort of patients in Michigan. Methods: We reviewed data for patients enrolled in the Michigan Cancer Surveillance Program, with diagnosis between 1990 and 2014 and follow up through 2016. Data extracted were age and stage at diagnosis, gender, race, zip code (as a proxy for family income), receipt of treatments, and survival status at last follow up. The poverty level (PL) was defined as average income by zip code < $24,600. Statistical comparisons were assessed by chi-square tests in contingency tables, by the log-rank test for survival, and Poisson regression for incidence rate comparisons. Results: We identified 10,828 GC patients (22.39% below PL, 64.64% between PL and 200% PL, 12.98% above 200% PL) with complete data for analysis. By race/ethnicity, 68.12% non-Hispanic blacks, 21.05% American Indian/Alaska Natives, 8.59% Asian/Pacific Islanders, 23.83% Hispanics, and 12.63% non-Hispanic whites lived in areas with average income below PL. The socioeconomic status by race among GC patients was significantly different (p < 0.0001). There was significant association of socioeconomic status and receipt of treatments. Patients living in areas below PL tended to have less information available about treatment received: unknown compared to no radiation (OR = 1.45, 95% CI 1.30-1.63), unknown compared to no chemotherapy (OR = 1.39, 95% CI 1.23-1.57), unknown compared to no surgery (OR = 1.49, 95% CI = 1.32-1.69). There was no significant relationship between stage at diagnosis and socioeconomic status (p = 0.153). The Poisson regression results showed that as average income increased in $10,000 increments in the area, incidence of gastric cancer is expected to decrease by 0.0642 (p < 0.0001). As income in the area increases by $10,000, risk of death is expected to decrease (HR 0.977, p = 0.0025). Conclusions: In Michigan, socioeconomic factors are associated with race, incidence of GC, receipt of treatments, and survival, but not stage at diagnosis.