Abstract Background Effective screening for colorectal, cervical, and lung cancer requires adherence over time, but little is known about repeat testing in real-world practice. Objective Describe patterns of longitudinal screening adherence and identify patient and system factors associated with repeat testing. Design Retrospective cohort study of colorectal, cervical, or lung cancer screening in 2010–2019. Participants Adults eligible for repeat colorectal (stool-based), cervical, or lung cancer screening following a negative index test in ten regional health systems comprising the US PROSPR consortium. Main Measures Repeat screening based on guideline-recommended intervals. For the colorectal and lung cohorts with opportunities for multiple annual screening rounds, the main outcome was repeat screening categorized as none, inconsistent, or consistent. Results The sample size was: 1,566,346 for colorectal, 216,344 for cervical, and 6,209 for lung cancer screening. For colorectal, cervical, and lung screeners, mean age at index was 58.2, 39.4, and 64.6 years, respectively, and 49%, 55% and 30% were Hispanic and/or non-white. Completion of the next screening round was 62% for colorectal, 56% for cervical, and 56% for lung cancer. For colorectal, over the next two rounds of testing, 53% were consistent, 33% inconsistent, and 14% no repeat screeners. The comparable percentages over 3 + rounds for colorectal were 40% consistent, 50% inconsistent, and 11% no repeat screeners. For lung, over the next two rounds, 47% were consistent, 31% inconsistent, and 22% no repeat screeners. The proportions over 3 + rounds for lung were 44% consistent, 42% inconsistent, and 14% no repeat screening. The health system was the strongest predictor of repeat and consistent testing with three- to ten-fold variation. Conclusions Adherence to longitudinal screening for colorectal, cervical and lung cancer was suboptimal, particularly as the number of testing rounds increased. System-level strategies are needed to increase screening adherence given the strong relationship between health system and outcomes.
Halm et al. (2025) studied this question.