Development of a multi-payer cohort for comprehensive statewide small area estimation of cancer screening in North Carolina

Abstract

Background: Reliable sub-state cancer screening estimates are needed to identify geographic disparities and guide interventions, but survey data often lack sufficient precision for small-area analyses. Multi-payer claims data offer a promising alternative, though constructing screening-eligible cohorts is complicated by transitions across payers over time. Methods: We used the University of North Carolina Cancer Information and Population Health Resource to develop a longitudinal multi-payer cohort of insured North Carolina residents from 2003 to 2020. Deterministic linkage based on Social Security Number (SSN), date of birth, and sex was used to identify unique individuals across private insurance, Medicare, and Medicaid. As a use case, we constructed annual colorectal cancer (CRC) screening–eligible cohorts for 2012–2020 using two definitions of 10-year enrollment. Results: The parent cohort included 12,010,374 unique individuals with 74,779,932 person-years of follow-up. Median continuous enrollment was 60 months (interquartile range, 19–122), and 24.7% linked to multiple payers over time. In 2020, the annual CRC screening cohort covered an estimated 56% of insured North Carolina residents aged 45–75 years. A flexible enrollment definition increased cohort size by approximately 10% with minimal demographic differences. Statewide up-to-date CRC screening prevalence was approximately 80% and stable across the study period, with geographic variation observed across small areas. Conclusions: This resource provides a scalable framework for constructing screening-eligible populations from multi-payer claims and generating small-area cancer screening estimates. Impact: This infrastructure has particular value for identifying screening disparities in underserved populations that are poorly captured by traditional survey methods.