P0170 - Longitudinal Adherence Demonstrates the Value of Blood-Based Screening Compared to Current Colorectal Cancer Screening Modalities: Results From a Discrete-Event Simulation Model

Shaun P. Forbes, PhD¹, Elifnur Yay Donderici, PhD¹, Nicole Zhang, MPH¹, Behnam Sharif, PhD², Gabriel Tremblay, DBA³, Gregory Schafer, ¹, Victoria Raymond, MS¹, AmirAli Talasaz, PhD¹, Craig Eagle, MD⁴, Amar K. Das, MD, PhD^1
1Guardant Health, Palo Alto, CA; ²Cytel Inc., Toronto, ON, Canada; ³Cytel inc., Waltham, MA; ⁴Guardant Health, Redwood City, CA

Introduction: Poor adherence to colorectal cancer (CRC) screening significantly hinders individual and population health benefits. Barriers for the available screening options lead to approximately 1/3 of people not being up to date. The impact of poor adherence is not adequately considered in leading health economics models, limiting the ability to evaluate population level real-world screening benefits and harms. We developed and validated a discrete-event simulation (DES) model with real-world adherence to evaluate the effectiveness of a blood-based screening test relative to existing screening strategies.

Methods: The CAN-SCREEN (Colorectal cANcer SCReening Economics and adherENce) model evaluated the performance of different CRC screening strategies per 1,000 screened individuals beginning at age 45. Adherence was modeled as: 1) assumed 100% adherence, and 2) longitudinal using a decline model calibrated to real world one-time adherence and population-level cumulative adherence from the National Health Interview Survey (NHIS). Performance for currently available strategies were derived from the literature. Blood-based test performance was per recently reported data (ECLIPSE NCT#04136002; Guardant Health, USA). The 100% adherence model was calibrated with previous CISNET published models (2021). Life-years gained (LYG), CRC deaths averted, and number of colonoscopies were compared to a no screening cohort.

Results: Modeling longitudinal adherence to reflect real-world utilization yields differences in health outcomes and resource utilization between screening modalities (Table 1). Blood based testing is comparable to colonoscopy for LYG (217, 255) and CRC deaths averted (15,15) while leading to fewer colonoscopies over the modeled time period (820, 1996) and is superior to stool-based testing. A blood-based test strategy modeled with real world adherence demonstrated a favorable CRC stage shift, with 83% of CRC diagnoses at stage I – III vs. 76% in colonoscopy and 78-81% in stool-based testing.

Discussion: The CAN-SCREEN model demonstrates that CRC screening strategies integrating test performance with high adherence yield more favorable health outcomes. Existing models assuming 100% adherence are limited in the ability to accurately predict population level health outcomes. Incorporating the blood test, that maximizes adherence due to its ease of use, as an additional strategy could avert additional deaths and reduce the number of colonoscopies performed.



Disclosures:


Shaun P. Forbes, PhD¹, Elifnur Yay Donderici, PhD¹, Nicole Zhang, MPH¹, Behnam Sharif, PhD², Gabriel Tremblay, DBA³, Gregory Schafer, ¹, Victoria Raymond, MS¹, AmirAli Talasaz, PhD¹, Craig Eagle, MD⁴, Amar K. Das, MD, PhD¹. P0170 - Longitudinal Adherence Demonstrates the Value of Blood-Based Screening Compared to Current Colorectal Cancer Screening Modalities: Results From a Discrete-Event Simulation Model, ACG 2023 Annual Scientific Meeting Abstracts. Vancouver, BC, Canada: American College of Gastroenterology.