Multilevel Interventions to Improve Colonoscopy Completion After Abnormal Fecal Immunochemical Testing in a Safety-Net Health System: Preliminary Findings from the IMPACTT Study

Background: Timely diagnostic colonoscopy is a critical step after an abnormal fecal immunochemical test (FIT), yet many patients do not complete follow-up within recommended timeframes, particularly in safety-net settings. Barriers occur across multiple levels and stages of care, suggesting a need for interventions that address both clinic processes and patient-level barriers. Methods: IMProving Adherence to Colonoscopy through Teams and Technology (IMPACTT) evaluated two concurrent interventions across 12 adult primary care clinics in the San Francisco Health Network. A clinic-level Best Practice Bundle (BPB), implemented in a stepped-wedge rollout, included patient tracking, standardized electronic health record documentation, standardized education and communication, and audit and feedback. In parallel, we randomly assigned eligible patients to a patient-level instruction and navigation (PIN) intervention or usual care. PIN provided targeted outreach to patients who were neither referred nor scheduled, as well as multilingual text reminders that allowed patients to confirm, cancel, or reschedule their colonoscopy. The primary outcome was colonoscopy completion within 180 days of an abnormal FIT. This preliminary analysis reports unadjusted comparisons among patients with sufficient follow-up at the data cutoff. Results: Among 2,303 patients with an abnormal FIT, 1,136 were identified before and 1,167 after their clinic implemented the BPB. Colonoscopy completion within 180 days was 44.1% before BPB implementation and 60.1% after implementation (risk difference, 16.0 percentage points; 95% CI, 11.8 to 20.1; p<0.001). Referral to gastroenterology within 42 days increased from 69.8% to 84.2%, and among patients who completed colonoscopy, the median time from FIT to colonoscopy decreased from 131 to 86 days. In the randomized PIN trial, we assigned 1,415 patients to usual care (n=698) or PIN (n=717). Colonoscopy completion within 180 days did not differ between usual care and PIN (58.4% vs 57.7%; risk difference, -0.7 percentage points; 95% CI, -5.9 to 4.6; p=0.9). PIN also did not improve referral within 42 days or scheduling within 56 days. Conclusions: In preliminary unadjusted analyses, implementation of a clinic-level bundle targeting tracking, communication, documentation, and feedback was associated with improved timely follow-up after abnormal FIT, whereas a randomized technology-supported patient instruction and navigation intervention did not improve colonoscopy completion. Prespecified analyses accounting for calendar time, clinic-level clustering, patient characteristics, and concurrent intervention assignment are forthcoming. Trial registration: ClinicalTrials.gov NCT06191185 .