About three in 10 patients with one or two nonadvanced adenomas were advised to return for surveillance colonoscopy sooner than recommended in the years after a 2020 guideline update, according to a Kaiser Permanente Northern California study published in Clinical Gastroenterology and Hepatology.
Adherence to the recommended seven- to 10-year interval rose from 68% in 2021 to 72% in 2024. By comparison, more than 90% of patients with three or four nonadvanced adenomas were advised to return within the recommended three to five years each year since the update.
The findings suggest that “the potential for over-surveillance is substantially more likely” after removal of one or two nonadvanced adenomas than after removal of three or four, wrote Jeffrey M. Hendel, MD, first author and chief internal medicine resident at Kaiser Permanente Medical Center in San Francisco; his research advisor and senior author, Jeffrey K. Lee, MD, MPH, a gastroenterologist at Kaiser Permanente Medical Center in San Francisco; and colleagues.
In 2020, the U.S. Multi-Society Task Force on Colorectal Cancer extended the recommended surveillance interval from five to 10 years to seven to 10 years for patients with one or two nonadvanced adenomas. For patients with three or four nonadvanced adenomas, it extended the interval from three years to three to five years.
To assess how often endoscopists followed the recommendations, the researchers conducted a cross-sectional study of 41,606 Kaiser Permanente Northern California patients who had one to four nonadvanced adenomas removed during colonoscopy from 2019 through 2024. The analysis included 33,787 patients with one or two nonadvanced adenomas and 7,819 with three or four. The colonoscopies were performed by 281 endoscopists.
Patients were aged 45-75 years, had adequate bowel preparation, and underwent a complete examination to the cecum. All were at average risk for colorectal cancer (CRC). Nonadvanced adenomas were tubular adenomas smaller than 10 mm with no high-grade dysplasia or villous features.
The researchers excluded patients with a history of colorectal polyps or CRC, inflammatory bowel disease, hereditary CRC syndromes, or a family history of CRC. They also excluded patients with advanced adenomas, serrated lesions, proximal hyperplastic polyps, or at least five nonadvanced adenomas.
Surveillance recommendations were obtained from manual review of post-colonoscopy flowsheets. The primary outcome was the proportion of patients each year assigned an interval within the guideline-recommended range. The researchers used multivariable logistic regression to identify patient and endoscopist factors linked to nonadherent recommendations from 2021 through 2024. No electronic decision-support tools or other interventions were introduced during the study, the authors noted.
In 2019, before the guidance changed, 88% of patients with one or two nonadvanced adenomas were advised to return within the recommended five to 10 years. Adherence to the new seven- to 10-year interval was 62% in 2020 and increased from 68% in 2021 to 72% in 2024. Each additional year from 2021 through 2024 was associated with 6% lower odds of a nonadherent recommendation.
During that period, 28%-32% of these patients were advised to return in less than seven years. None were advised to wait more than 10 years.
Endoscopists rarely chose intervals in the middle of the recommended range. From 2021 through 2024, 46%-47% of patients with one or two nonadvanced adenomas were advised to return in 10 years and 20%-25% in seven years. No more than 1.5% received an eight-year recommendation, and none received a nine-year recommendation.
Adherence was higher for patients with three or four nonadvanced adenomas. In 2019, 81% were advised to return in the then-recommended three years. After the guidance changed, 90%-91% received a recommendation within the new three- to five-year range each year from 2020 through 2024.
From 2021 through 2024, 51%-52% of these patients were advised to return in three years and 38%-40% in five years. No more than 1.4% received a four-year interval. About 2.5% were advised to return sooner than three years, while about 7% were told to return after more than five years.
Among patients with one or two nonadvanced adenomas, several factors were linked to shorter-than-recommended intervals, including older age, male sex, a body mass index of 30 or higher, a history of smoking, and higher scores on the Charlson Comorbidity Index. Patients with a score of 3 or higher had 30% higher odds of receiving a shorter-than-recommended interval than those with a score of 0.
Endoscopists with 15-24 years in practice had 20% higher odds of assigning a shorter-than-recommended interval than those with fewer than 15 years; the association was not statistically significant for those with 25 or more years. In contrast, shorter-than-recommended intervals were less common among patients undergoing diagnostic rather than screening colonoscopy, Hispanic patients, and non-Hispanic Asian patients. Among patients with three or four nonadvanced adenomas, none of the factors studied was linked to nonadherence.
The authors said endoscopists’ tendency to assign the shortest or longest interval in each range may help inform future guidance. Guideline developers could consider recommending specific intervals, such as seven or 10 years instead of a seven- to 10-year range, the authors wrote.
The authors noted that the findings may not apply outside integrated health systems, particularly fee-for-service practices, where financial incentives might lead to shorter intervals. The exclusion of patients with a family history of CRC or with serrated lesions — factors previously linked to shorter surveillance intervals — may also limit generalizability. Because the study was observational, some patients or recommendations may have been misclassified, and unmeasured factors may have affected the results.
The researchers assessed recommendations based only on adenoma findings, and endoscopists may have assigned shorter intervals “intentionally and appropriately” based on perceived elevated risk, Dr. Hendel and Dr. Lee wrote. Pandemic-related resource constraints also may have influenced the assignment of longer intervals in 2020 and 2021.
The authors reported no conflicts of interest.
Expert insight
Dr. Hendel and Dr. Lee spoke with GI & Hepatology News about the study.
Why are endoscopists less likely to follow surveillance recommendations after removal of one or two nonadvanced adenomas?
Dr. Hendel and Dr. Lee: The exact reasons remain unclear, but several factors have been suggested or studied including lack of knowledge of current guidelines, lack of confidence in the evidence base supporting these recommendations, suboptimal bowel preparation, practice habits, medicolegal concerns, patient-related factors, and potentially financial incentives.
When, if ever, is a surveillance interval shorter than seven years appropriate for these patients?
Dr. Hendel and Dr. Lee: Usually, a shorter interval is justified when there is concern about the adequacy of the colonoscopy examination itself (e.g., suboptimal bowel preparation) or when patient-specific factors place the patient outside the “average-risk” population the 7-10 year recommendation was designed for. For example, one could justify a shorter interval for individuals with a personal history of inflammatory bowel disease or hereditary colorectal cancer syndromes or a strong family history of colorectal cancer.
What practical steps could improve adherence and reduce unnecessary surveillance colonoscopies?
Dr. Hendel and Dr. Lee: Several system-level and educational interventions could help close this adherence gap. One of the most promising approaches is embedding a clinical decision-support tool directly within the electronic health record that automatically calculates and prompts endoscopists with the guideline-recommended surveillance interval based on the baseline colonoscopy findings entered, such as the number, size, and histology of polyps, bowel prep quality, and completeness of the exam. This would reduce reliance on memory or outdated habits at the point of care. Complementing this, automated, guideline-based result letters could be generated to state the recommended surveillance interval directly to the patient and primary care provider, reinforcing the correct interval and creating an additional layer of accountability beyond the endoscopist's own documentation.
Building on this infrastructure, the decision-support tool could also include a "hard stop" requiring the endoscopist to document a specific reason whenever they recommend an interval that deviates from the guideline default. This would prompt a deliberate pause to consider whether a shorter interval is truly justified by factors such as incomplete resection, poor bowel preparation, or relevant patient history, rather than defaulting to habit, while still preserving clinical flexibility when a shorter interval is genuinely warranted. Finally, formally incorporating guideline-concordant surveillance recommendations as a tracked quality indicator, as some professional societies already recommend, could help align institutional and financial incentives with appropriate rather than overly frequent surveillance.