Columbia Business School AI Cuts Colorectal Cancer Deaths 43% & Boosts Screening

A new artificial intelligence (AI) model developed by researchers at Columbia Business School has been linked to a 43% reduction in colorectal cancer mortality and a more than 200% increase in screening rates, according to a study released Wednesday.

The AI model was implemented in 2019 at Geisinger Health System in Pennsylvania. It identifies high-risk patients aged 51 to 75 who are overdue for colorectal cancer screenings. The system analyzes patient data, including blood test results, age, and sex, to calculate a cancer risk score. Patients with scores exceeding 0.150 are flagged as high-risk and contacted by care coordinators who facilitate colonoscopy appointments.

The study, titled “Cancer Screening Outreach Guided by Machine Learning: The Benefits of Proactive Care,” found that patients identified as high-risk by the AI were 214% more likely to undergo a colonoscopy within three months, and 117% more likely within six months, compared to a control group. Over two years, the AI-guided approach resulted in a 6.2 percentage point decrease in colorectal cancer mortality, representing an overall reduction of 43%.

Colorectal cancer claims the lives of over 52,000 Americans annually, and nearly half of adults eligible for screening do not receive recommended checkups. The AI model aims to address this gap by proactively identifying and engaging those most at risk.

The research has been accepted for publication in the journal Manufacturing & Service Operations Management. The research team was led by Columbia Business School professor Carri W. Chan.

The system analyzed over 62,000 risk assessments during the study period, processing approximately 450 assessments each week.