Abstract
Background: Fecal testing (FIT) is an effective screening tool for colorectal cancer. If a FIT is abnormal, a follow-up colonoscopy is necessary to remove polyps or find cancers. Predicting which patients may be at risk for failing to follow-up on an abnormal fecal test could help identify patients in need of early interventions aimed at completing a colonoscopy. Using predictive analytics to determine who is unlikely to follow-up will allow health systems to tailor interventions and outreach.
Understanding the application of predictive analytics across health systems is imperative to providing appropriate care to distinct patient populations. Early interventions could decrease the disease burden if colorectal cancer is found at earlier stages. Methods: We compare two models created in two different settings. The models were created from data from FQHC’s and Logistic and Cox regressions to validate and redevelop a risk prediction model among a retrospective dataset of patients with abnormal FIT results. Results: The initial FQHC risk model included eight variables including race, clinic system, prior no-shows, insurance, prior flu shots, age, indication of anti-coagulation use, and income inequality. However, the model did not validate in a large similar health system in the Pacific Northwest. Risk factors for colonoscopy varied by county, clinics, regions, and patient populations.
Inner and outer setting variables, like referral workflows and access to colonoscopy, contributed to the likelihood of a patient completing their screening. The second model retained most predictors except for anti-coagulation use, and included new predictors like language, homeless status, prior screenings, and a comorbidity index. Conclusions: Application of a risk model helps tailor interventions to help patients complete the continuum of screening for colorectal cancer. However, health systems differ, and precision medicine is best applied when risk is understood in context and interventions are tailored for specific populations predictors.
Citation Format: Amanda F. Petrik, Eric S. Johnson, Matthew Slaughter, Andreea Rawlings, Michael Leo, Jamie Thompson, Ricardo Jimenez, Gloria D. Coronado. The validation and redevelopment of a risk prediction model identifying patients unlikely to complete a colonoscopy following an abnormal FIT test in community clinics [abstract]. In: Proceedings of the AACR Virtual Conference: Thirteenth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2020 Oct 2-4. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(12 Suppl):Abstract nr PO-270.