The scale of the study is what sets it apart. Between January 1, 2017, and December 31, 2019, nearly 4.75 million adults aged 45 to 85 years were mailed a multitarget stool DNA, or mt-sDNA, test kit. After applying rigorous eligibility criteria designed to isolate average-risk individuals, the final analytic cohort comprised 2,574,759 adults who together contributed more than 14.9 million person-years of follow-up through the end of 2024. Of these, 1,818,918 people, or 70.6 percent, returned their kits, while 755,841, or 29.4 percent, did not. During follow-up, 39,401 incident colorectal cancer cases were identified using claims-based diagnosis codes, providing an enormous evidence base for the comparison.

The headline finding is striking. Colorectal cancer incidence was 248.95 per 100,000 person-years among kit returners, compared with 301.89 per 100,000 person-years among non-returners. In the primary adjusted analysis using Poisson regression, returning the kit was associated with a 55 percent lower colorectal cancer incidence rate, with an incidence rate ratio of 0.45 and a confidence interval of 0.45 to 0.46, a result that held at a p-value below 0.0001. Supportive time-to-event analyses told the same story: kit return was associated with improved colorectal cancer-free survival, with a hazard ratio of 0.76, and Kaplan-Meier curves showed a persistent separation between the two groups throughout follow-up.

What makes these numbers credible rather than merely impressive is the battery of statistical safeguards the researchers deployed. Observational screening studies are notoriously vulnerable to confounding and to a subtle trap known as immortal time bias, which arises when the time before an exposure occurs is incorrectly attributed to the exposed group. To counter this, the team modeled kit return as a time-varying exposure in their Poisson models, meaning that each person contributed follow-up time to the unexposed category until the moment they returned their kit, and only then shifted into the exposed category. This design ensures that the analysis does not credit non-returners with protected time they never actually used for screening.

The researchers then attacked measured confounding from two additional directions. First, they used propensity score matching to construct a balanced cohort of 1,511,602 individuals, pairing returners with non-returners who shared similar age, sex, race and ethnicity, geographic region, and shipment year. In this matched cohort, kit return was associated with a 42 percent lower colorectal cancer incidence rate. Second, they applied inverse probability of treatment weighting, which reweights the entire population rather than discarding unmatched individuals, generating a weighted pseudo-population of more than 5.1 million. Under this framework, the association appeared even stronger, with a 72 percent lower incidence rate. Standardized mean differences confirmed that covariate balance improved dramatically after both procedures, and weight diagnostics showed no instability.

The biology behind the association is straightforward once the mechanics of stool-based screening are understood. The mt-sDNA test combines assays for aberrant DNA markers shed from colorectal neoplasia with fecal hemoglobin detection, and it has demonstrated higher sensitivity for colorectal cancer and better detection of advanced precancerous lesions than the fecal immunochemical test, though at the cost of lower specificity and more false positives. Because colorectal cancers typically develop over years through the adenoma-carcinoma sequence, screening works by catching precursor lesions while they can still be removed during follow-up colonoscopy, before malignant transformation occurs. A person who returns the kit enters that pipeline; a person who does not, by definition, cannot benefit from it.

The authors are careful to frame the finding as associative rather than causal, and their caveats deserve attention. Kit return alone cannot reduce cancer incidence; any benefit depends on completing follow-up colonoscopy after a positive result and on subsequent clinical management. The study did not directly evaluate colonoscopy completion, so the observed association likely reflects the combined effect of the entire screening pathway rather than the test itself. There is also the possibility of healthy adherer bias, the well-documented phenomenon in which people who complete one preventive health behavior tend to engage in others, from medication adherence to routine checkups, in ways that administrative claims data cannot fully capture. Residual confounding by healthcare engagement therefore remains a plausible partial explanation.

Additional limitations shape how the results should be read. Outcomes were identified from claims-based diagnosis codes rather than adjudicated clinical records, leaving room for misclassification, and tumor stage at diagnosis was not reliably available, so the study could not determine whether returners who developed cancer were diagnosed earlier. Subsequent alternative screening completed during follow-up was not captured, and the cohort was restricted to insured individuals in the continental United States, limiting generalizability to uninsured or underinsured populations. Notably, the study was funded by Abbott, the manufacturer of the mt-sDNA test and the employer of several authors, a disclosure that readers should weigh alongside the study’s methodological strengths.

Why does this matter now? Colorectal cancer is projected to account for 158,850 new cases and 55,230 deaths among US adults in 2026, making it the third most diagnosed cancer and the second leading cause of cancer death in men and women combined. Yet in 2023, only about 63.5 percent of eligible adults were up to date with screening, well short of the national goal of 80 percent. Mailed outreach programs have proven effective at boosting participation, but this study exposes the weak link: a substantial share of people who receive a kit never send it back. The 29.4 percent non-return rate in this cohort represents millions of missed opportunities for early detection and prevention.

The practical implications are already taking shape. Patient navigation, reminder systems, and structured engagement strategies designed to push kits from mailbox to laboratory could translate into measurable population-level reductions in colorectal cancer incidence. The consistency of the association across unmatched, matched, and weighted analyses, and across both incidence-rate and survival models, suggests the signal is robust to the methodological choices that often sink observational screening research. For the roughly one in three Americans who set aside a mailed screening kit, the message from this study is blunt and personal: the test sitting unopened on the counter is not a screening test. Only the one that comes back, sealed and shipped, can do its job of finding cancer before it finds you.

Subject of Research: Association between return of mailed multitarget stool DNA colorectal cancer screening kits and subsequent colorectal cancer incidence

Article Title: Return of Mailed Multitarget Stool DNA Test Kits and Subsequent Colorectal Cancer Incidence in the United States: A Retrospective Cohort Study

Article References: Goyat, R., Karlitz, J. J., Wang, W., Ozbay, A. B., Montminy, E., & Greene, M. (2026). Return of Mailed Multitarget Stool DNA Test Kits and Subsequent Colorectal Cancer Incidence in the United States: A Retrospective Cohort Study. Advances in Therapy. https://doi.org/10.1007/s12325-026-03818-7

Image Credits: AI Generated

DOI: 10.1007/s12325-026-03818-7

Keywords: colorectal cancer, cancer screening, stool DNA test, mt-sDNA, screening adherence, mailed outreach, epidemiology, cohort study, cancer prevention, colonoscopy, health disparities, screening completion