Exclusion criteria: evaluating the impact of enrollment requirements in SEER-Medicare research

Background Enrollment criteria are routinely utilized in patient selection in SEER-Medicare but little is known about how this may be impacting research outcomes. This study evaluated demographics and survival among pancreatic cancer patients who are included and excluded from SEER-Medicare analyses. Methods Patients >= 66 years old with pancreatic cancer were identified (SEER-Medicare, 2008-2015). Two patient cohorts were compared: included (continuous enrollment in Medicare Parts A and B and no enrollment in Medicare Advantage), and excluded. Mortality was assessed using a Standardized Mortality Ratio. Results Among 49,017 patients with pancreatic cancer, 59.5% were in the included cohort. The excluded cohort was younger (median age 74 vs. 77) with more male (49.9% vs. 47.8%), non-white (33.0% vs. 21.3%) and urban-dwelling patients (91.0% vs. 85.0%). Those excluded had a higher mortality risk (SMR 1.06, 95%CI 1.04-1.07). Conclusions There are significant differences in patient demographics and mortality among those who are and are not routinely included in SEER-Medicare analyses and our study provides a critical opportunity to quantify this potential bias.

Automated summary

This study evaluated the demographic characteristics and survival differences among pancreatic cancer patients who are included and excluded from SEER-Medicare analyses. The study found significant differences in age, gender, race, and mortality risk between the included and excluded cohorts.