Effect of colorectal cancer screening on 1-year mortality and medical costs: Real-world evidence from the largest commercial medical claims database in the US.

Abstract

10579 Background: Colorectal cancer (CRC) is the third leading cause of cancer-related deaths in the United States with increasing trends in both mortality and incidence in individuals under age 50. Whether CRC screening reduces mortality recently becomes controversial in some observational studies due to a lack of data and a possibility of screenings confounded with social determinants of health (SDoH) and patient characteristics that were not properly addressed. Methods: We used the Optum’s de-identified Clinformatics Data Mart Database that provides information of healthcare plan, major SDoH, the CRC diagnosis, treatments, medical service and prescription charges, and survival data since 2007 for over 75 million patients to re-examine benefits of CRC screenings. Our CRC study cohort (N=99,704) included persons first diagnosed with CRC between 2010 – 2020, with 3-year pre-diagnosis continuing insurance coverage to identify pre-diagnosis screening, and 1-year follow-up to identify death and associated costs. Primary outcomes are 1-year mortality, the post-diagnosis medical and prescription cost increments, and health plan change after CRC diagnosis. Augmented inverse probability weighting (AIPW) models were employed to estimate the benefits of CRC screening, adjusting the probabilities of pre-diagnosis screenings due to insurance cover types (commercial/employer, Medicare advantage, and dual Medicare-Medicaid or low-income subsidy (LIS) covers), and healthcare plans (pre-paid, only-in-network, pay-less in-network, and fee-for-service plans), ages, sex, race, residential states, and calendar year. Results: In the study cohort, 37.8% have pre-diagnosis screening records, 18% died within 1 year. Women, persons under age 45, and persons with dual covers were more present in the non-screening group than the screening one. CRC screening was associated with a significant reduction of one-year mortality after adjusting for age, sex, race, health plans, and other factors (death rates in the non-screening group vs the screening one were 23% vs 12%, adjusted OR = 0.57, 95%CI = 0.55 - 0.59). Persons with no screening records experienced higher one-year medical cost increases than those who had pre-diagnosis screenings (17.8% vs 14.8%). No significant difference in health plan changes and/or disruption among the two groups was obvious for persons having either dual- or commercial-cover beneficiaries, yet the screenings slightly induced more these plan changes in the Medicare advantage beneficiaries (OR = 1.06, 95%CI = 1.01 - 1.11). Conclusions: The real-world evidence from the US largest medical claims database suggested that, after adjusting for the screening probability due to healthcare plans and other confounders, CRC screening dramatically reduced one-year death risk, and slightly eased the 1-year financial burdens after CRC diagnosis.