Colon cancer screening: New data point to ways to boost uptake and follow up

Tens of millions of middle-aged and older Americans haven’t gotten their recommended checks for early signs of potential colon cancer, either through the “gold standard” of colonoscopy or a non-invasive test.

Now, a pair of studies examines two aspects of the screening process: the preferences of patients and physicians for all currently available screening options, and the impact of a 2023 policy change that cut costs for those who get an abnormal result on a home-based stool test and then need a colonoscopy.

The preference study shows that 75% of adults eligible for screening would prefer a non-colonoscopy option based on a sample of their stool or blood as their initial test.

Respondents received information about the nature, accuracy and frequency of all currently available options.

However, only 5% of physicians chose non-colonoscopy screening as their preferred option for their patients after receiving similar information.

The other study shows that follow-up colonoscopies after abnormal home stool tests rose 41% after the policy took effect, even though the total number of colonoscopies didn’t rise.

Pre-cancerous polyps can be seen and removed during a colonoscopy, making the colorectal cancer screening process a key tool to prevent cancer.

Both studies, published in the journal Current Medical Research and Opinion, were led by A. Mark Fendrick M.D., a University of Michigan Medical School professor with a decades-long interest in the prevention and early detection of colorectal cancer.

The research was funded by Exact Sciences, the company that makes the Cologuard and Cologuard Plus home-based stool testing options and employs several of the studies’ authors.

The studies did not focus only on that company’s product; other stool-based tests and a newer blood-based screening test were also included.

“Given that one-third of insured adults in this country who qualify for no-cost colorectal cancer screening under the Affordable Care Act are not getting screened, we need to better understand the reasons why eligible individuals are not receiving this potentially life-saving preventive service,” said Fendrick, the lead author of one paper and senior author of the other, and a paid advisor to Exact Sciences.

“We hope these findings will help inform the process of encouraging screening and follow up tests, ultimately leading to a reduction in the number of detected cases of this preventable cancer.”

Fendrick, who leads the U-M Center for Value Based Insurance Design, has been a longtime advocate for reducing patients’ out-of-pocket costs for care that delivers a high level of health benefit for an individual.

This includes costs related to screening for multiple types of cancer, including follow up tests needed to confirm or rule out a cancer diagnosis after an initial screening test.

Patient and physician preferences for colorectal cancer screening

To see which type of colon cancer screening tests would be preferred by individuals in the target group for screening, and by physicians, Fendrick and his colleagues conducted a predicted choice probability study.

They did an online survey of 1,249 adults between the ages of 45 and 75 who have no individual or family history of colorectal cancer, and 400 physicians divided equally between primary care doctors and gastroenterologists.

They asked both groups to choose a preferred screening option for themselves or their patients, from among colonoscopy, several types of stool-based testing, and blood-based testing.

Each respondent received information about the nature of each test, how often it needs to be repeated for ongoing screening, how often it gives a true positive result when a person does have cancer (sensitivity), and how often it gives a true negative result when a person doesn’t have cancer (specificity).

The physicians also received information about each test’s ability to detect non-cancerous growths in the colon, called adenomas or polyps. Insurance coverage or cost to the patient were not addressed.

In all, 39% of the screening-age individuals chose multi-target stool DNA tests (mt-sDNA, the product made by Exact Sciences), 25% chose colonoscopy; 21% chose the blood test, and 15% chose a stool test called a fecal immunochemical test or FIT, which looks for microscopic signs of blood that could be related to cancer.

Among people who had had a colonoscopy in the past, colonoscopy and mt-sDNA were preferred by nearly equal percentages (34% and 32%, respectively).

Among those who had never had any screening test for colorectal cancer, or had had a test other than colonoscopy, mt-sDNA preference was far higher than all other options.

Among physicians, 95% chose colonoscopy and just over 4% chose mt-sDNA for their patients; less than 1% chose either of the other two tests.

There was no difference between primary care and gastroenterology practitioners.

“Understanding patient preferences is critical to encouraging screening, and in this survey we show that consumers are willing to adopt the shift to noninvasive testing even if it means more frequency and less accuracy than the ‘gold standard’ of colonoscopy,” Fendrick said.

“But it is extremely important that we clearly convey at the time of screening to those who choose non-invasive tests that a colonoscopy must be performed after an abnormal result, which up to 10% will receive depending on the modality chosen.”

That could involve having patients formally ‘commit’ to follow-up colonoscopy if needed when they choose a home-based test.

It could also include help with navigating logistical issues that accompany colonoscopy, such as scheduling, bowel preparation, and the need for a driver to accompany the patient.

Fendrick is working with others at U-M Health to increase follow-up testing among patients at U-M primary care clinics.

Pre- and post-cost reduction study

The idea of seeing colorectal cancer screening as a process, rather than a single test, formed the basis for two federal policies that took effect in January 2023.

The policies required private insurance companies and Medicare to make follow-up colonoscopies available without cost to patients who had an abnormal result on a stool-based screening test, including co-pays, co-insurance and deductibles.

In their new paper, Fendrick and colleagues examined national insurance claims data for 10.8 million colonoscopies performed in 2022 and the first 11 months of 2023.

They were able to see which ones involved patients who had had an abnormal result in the last six months on a non-invasive stool test that looks for blood or DNA.

The number of these follow-up colonoscopies saw a relative increase of 41% from 2022 to 2023, even though the total number of colonoscopies performed each month didn’t change appreciably.

The absolute increase was 1.5%, because the percentage of all colonoscopies that were coded as follow-ups to abnormal stool tests went from 3.6% to just over 5% of all colonoscopies.

Given the relatively fixed supply of colonoscopy appointments in the United States, this suggests that increased noninvasive testing could lead to a shift in use of colonoscopies, rather than an increase to procedures performed.

This is desirable, says Fendrick, because patients are already facing delays in colonoscopy scheduling thanks to a guideline change in 2021 that made 20 million Americans between the ages of 45 and 49 eligible for no-cost screening, as well as ongoing catch-up for patients who delayed screening during the height of the COVID-19 pandemic.

At the same time, clinicians and health systems are interested in maximizing the efficient use of colonoscopy teams and facilities but not overwhelming them, says Fendrick.

Papers cited:

“Patient and physician preferences among colorectal cancer screening tests: updated predictions from a discrete choice experiment,” Current Medical Research and Opinion. DOI:10.1080/03007995.2025.2576596

“Completing the colorectal cancer screening process: impact of eliminating cost-sharing for follow-up colonoscopy,” Current Medical Research and Opinion. DOI: 10.1080/03007995.2025.2577763