1What the model predicted
Two groups. Different leading choices.
Researchers used hypothetical survey choices from 1,249 average-risk adults and 400 doctors to predict preferences among four screening tests. The leading choice in each model was:
Patient model
39.4%Next-generation stool DNA
Physician model
95.0%Colonoscopy
Authors’ conclusion: Offering tests that match patients’ preferences may improve screening completion.
These are modeled preferences, not completed screening or doctors’ personal choices. The patient model’s leading share was less than half. Next: what each group compared.
Source: Fendrick et al., 2025. Updated predictions using a 2020–2021 survey. Funded by Exact Sciences. See all four test predictions.
2What each group compared
Only doctors compared precancer detection.
Both groups compared test type, testing frequency, and cancer-test accuracy. Only doctors also compared how well tests detect adenomas—growths that can become cancer.
Information explicitly included in the choice tasks| Comparison | Patients | Doctors |
|---|
| Cancer-test accuracy | Included | Included |
|---|
| Precancer detection | Not included | Included |
|---|
The authors say they left the precancer comparison out of patient tasks to make them simpler.
This describes the information shown in the study. It does not tell us what patients already knew about precancer.
Source: Fendrick et al., 2025, methods and limitations.
3The question left open
Precancer detection mattered to doctors. What about patients?
In the original experiment, precancer detection carried about as much weight as cancer detection in doctors’ choices, across the ranges tested.
The patient questionWould patients choose differently if they compared precancer detection, too?
This study cannot answer that. It did not test how patients’ preferences changed after education.
Our hypothesis: Explaining these differences may influence what patients prefer. The gap between the two groups cannot prove this: they differed in training and experience as well as the information shown.
Sources: Heidenreich 2022 and Fendrick 2025. Original experiment and updated model using the same survey, both funded by Exact Sciences.
4Why we start with education
Explain prevention before choosing a test.
Our takeaway is to include both cancer and precancer detection in the conversation. Finding and removing precancerous polyps can help prevent cancer.
Colonoscopy can find and remove detected polyps during the same exam. That opportunity is why we favor high-quality colonoscopy for screening-eligible, average-risk adults who are medically suited, willing, and have timely access.
“How well do my options find precancer, and is colonoscopy a good starting point for me?”
Discuss preparation, sedation, and risks such as bleeding or a tear in the bowel. Colonoscopy can miss lesions. Other recommended options are available; a positive stool test requires colonoscopy.
For those who can, Colonoscopy First.
Questions to bring to your clinician
This is our educational and advocacy position; the preference study does not establish a best test for everyone. Clinical context: National Cancer Institute and national screening guidelines.
