Predictors of Endoscopic Colorectal Cancer Screening over Time in 11 States

Background: Colorectal cancer (CRC) ranks third among all cancer sites in incidence, and second in cancer-related mortality. CRC screening is essential for early detection of colorectal cancer and both reduced morbidity and increased survival among cancer victims, but screening rates are suboptimal.

Objective(s): To study predictors of CRC screening behavior in a cohort of Medicare-enrolled persons over a period during which coverage for CRC screening increased (2000-2005). The study population includes 272,077 men and women who are age 65+ in 2000 and remain alive through 2005, located in 11 SEER Registry states that span all census regions and exhibit heterogeneous socio-demographics.

Methods: We use multilevel logistic regression of a binary probit choice model over two time intervals (2000-2002, 2003-2005) and assess empirically whether predicted effects are different across states and time. Gender, previous cancer diagnosis, disability as the original reason for entitlement, and dual eligibility are key person-specific factors of interest. Social segregation, commuter intensity, availability of CRC screening facilities, availability of oncologists and gastroenterologists, poverty, and income disparity are key environmental variables.

Results : The geographic variation in utilization and its associated factors provide insights regarding the heterogeneity of supply-demand systems and realized utilization of CRC services across the landscape. Men are not universally more likely than women to utilize CRC screening; African Americans are not universally less likely to be screened than whites; and disparities between minorities and whites appear to decrease over time. Policy interventions to improve screening rates should consider local community characteristics.