Adults who returned a mailed stool test had about 43% lower risk of dying from colorectal cancer (rate ratio 0.57), while adults merely invited to screen had about 26% lower risk (rate ratio 0.74), according to a study of 376,511 people published August 20, 2026 in JAMA Network Open. The study tracked Sweden’s Stockholm-Gotland screening program for up to 14 years.
Key takeaways
- Being invited to screen was tied to about 26% lower colorectal cancer death (rate ratio 0.74, very likely between 6% and 42% lower).
- Returning the kit was tied to about 43% lower death (rate ratio 0.57, very likely between 20% and 60% lower).
- Both rest on a new statistical correction. The raw comparison showed about 8% lower death and was not statistically significant.
What the study found
Stockholm and Gotland began routine screening in 2008. Birth cohorts of people aged 60 to 69 were randomly assigned to a study group invited to screen in 2008 to 2012 (203,692 people) or a control group invited later, in 2013 to 2015, or never (172,819 people). Follow-up ran through December 31, 2021.
There were 856 colorectal cancer deaths in the study group and 812 in the control group, 1,668 in all: about 4.2 deaths per 1,000 invited people versus about 4.7 per 1,000 controls. That gap is small, and it shows. Adjusting only for age and follow-up time, the rate ratio was 0.92, roughly 8% lower risk, with a range running from 17% lower to 1% higher (95% CI 0.83 to 1.01). That range crosses zero benefit, so on its own it does not establish an effect.
The authors then corrected for two things that dilute any real program. First, 455 of the study group’s deaths came from cancers diagnosed more than two years after that person’s last invitation, when they were no longer being screened. Second, 66% of the control group was invited at least once anyway, so the comparison group was not unscreened. After those corrections the rate ratio was 0.74, a 26% reduction (95% CI 0.58 to 0.94). Correcting further for invited people who never returned a kit produced the 43% figure (95% CI 0.40 to 0.80). An earlier evaluation of the same program, without these corrections, found a 14% reduction.
Dr. Kumar’s take
The 43% figure moving around the internet is not what a health system gets for mailing kits. It is what a person gets for opening the envelope. The number a health department can plan around is 26%, and even that assumes an invitation system as organized as Sweden’s.
The distance between those numbers is behavioral, not biological. Johannes Blom, the study’s corresponding author, put it plainly: “Although screening is offered free of charge and the test is simple to carry out, around a third of people do not submit a sample.” The test cannot find a tumor in a kit sitting on a kitchen counter.
The drop-off does not stop at the mailbox. The authors note that 13% of people with a positive stool test never had the follow-up colonoscopy, left unadjusted here, and that in their earlier work those people had four times the colorectal cancer risk. Screening is a chain: kit returned, test read, colonoscopy completed. Every broken link erases the benefit. Researchers are also testing whether gut bacteria can flag colorectal cancer in the same sample.
The 43% is also a modeled estimate. The corrections are reasonable and the reasoning is transparent, but as Karolinska’s announcement says, “the results are based on statistical adjustments for various types of bias, meaning that some uncertainties remain.” The adjustments are the finding.
How does this stack up against colonoscopy, or the new blood tests? The study cannot say. It compared invited people with less-invited people, not one test against another. What it does show is the variable the marketing skips: real-world benefit is accuracy multiplied by how many people finish the process. A less sensitive test everyone completes can beat a better test people avoid.
One detail matters for American readers. This program used the old guaiac card until October 2015, then switched to FIT, the test most US programs mail today. Adherence rose by more than 10%, and the authors say their estimate “could be an underestimation regarding FIT-screening.”
What it means for you
If a screening kit arrives, the useful action is to return it, not to research it. Colorectal cancer caught early has a good prognosis, which is the reason these programs exist. If the test comes back positive, the colonoscopy is not optional. The Swedish data suggest the people who stop there carry the most risk.
When weighing options with your clinician, favor the one you will actually complete on schedule over the one that sounds most thorough in the exam room. The same logic runs through other screening debates, including whether a blood test can prevent deaths from prostate cancer.