AI-Guided Outreach System Improves Colorectal Cancer Screening

Colorectal cancer is the second-leading cause of cancer deaths in the United States. Early detection improves survival, yet many eligible adults remain overdue for recommended screening. Health systems need reliable ways to reach high-risk patients before disease progresses. A newly published study shows that AI-guided outreach can raise screening completion and is associated with lower mortality.

The study in Manufacturing & Service Operations Management evaluated a machine learning–guided outreach program implemented at Geisinger Health System. Investigators examined a real-world initiative designed to find patients overdue for colorectal cancer screening and connect them to colonoscopy. The work, titled “Cancer Screening Outreach Guided by Machine Learning: The Benefits of Proactive Care,” assessed whether targeted outreach improves adherence and outcomes.

The program combined predictive analytics with personalized human contact. An algorithm reviewed routinely available information, including complete blood count (CBC) results, age, and sex, to identify patients at elevated risk among those lacking recommended screening. Nurse coordinators then educated flagged patients about colonoscopy and helped schedule appointments.

Patients targeted through this approach were 6% more likely to complete a colonoscopy within three months and 6.9% more likely within six months than similar patients who did not receive outreach. The program was associated with a 6.2% reduction in two-year mortality, a 43% decrease relative to the control group. The analysis covered outcomes from a program operating since 2019.

Authors from the University of Hong Kong, Columbia Business School, Geisinger, and Children’s Hospital of Philadelphia report that artificial intelligence can support proactive care delivery by identifying individuals who stand to benefit from intervention and enabling timely engagement. They note that scaling similar programs requires attention to screening capacity, communication strategies, and disease severity. The findings suggest a practical framework for health systems seeking to improve cancer prevention performance while keeping clinicians central to care.

“Our results showed that a proactive cancer screening outreach program guided by machine learning can significantly improve patient outcomes in addition to achieving higher disease detection rates. The program not only boosts screening participation but also meaningfully reduces mortality,” said Minje Park of the University of Hong Kong.

“This work demonstrates an analytical framework for rigorously evaluating machine learning-aided outreach programs for other cancers and diseases. Establishing unbiased estimates of the impact is critical for capacity planning of screening resources such as colonoscopies,” said Carri Chan of Columbia Business School.

Related Links
Geisinger Health System