Longitudinal Adherence With Stool-Based Testing for Colorectal Cancer Screening in Four United States Health Systems

Longitudinal adherence with annual stool-based screening for colorectal cancer is guideline-recommended. This paper assessed patterns and predictors of repeat stool-based screening over time and associated outcomes.

This retrospective cohort study followed adults aged 50 to 65 years in 4 health systems eligible for repeat screening following a negative index fecal immunochemical test or guaiac fecal occult blood test in 2010 to 2011 for up to 10 years. Repeat stool-based testing patterns were categorized as: consistent (≥75% of rounds completed, inconsistent [<75%], and never repeaters. Multivariable predictors of consistent testing were analyzed, as well as associations between screening consistency and cancer stage at diagnosis.

Among 492,812 individuals, 54% were consistent repeaters, 30% inconsistent repeaters, and 16% never-repeaters. The proportion of never-repeaters declined with more follow-up time, whereas inconsistent repeating increased. Consistent repeating remained fairly constant. Health system was the strongest predictor of consistent screening, with wide variation by site. Consistent repeating was also associated with older age, male sex, being non-Hispanic White, lower comorbidity, body mass index <25 kg/m², public insurance, prior stool-based testing, and a primary care provider visit in the past year. Advanced stage colorectal cancer was more frequent among never-repeaters (19% vs 12% for inconsistent and 12% consistent repeaters). Early-stage cancer was more common among consistent repeaters (56% vs 49% inconsistent and 44% never-repeaters).

Among patients initially screened with an annual stool-based test, about one-half (54%) consistently repeated it, and 16% never repeated during available follow-up. The health system where the patient was cared for was most strongly associated with consistent stool-based testing, and higher adherence was associated with earlier cancer stage at diagnosis.