Artificial intelligence-assisted integrated care to promote colonoscopy uptake (AICC) in China: study protocol for a cluster randomized controlled trial

BACKGROUND: Colorectal cancer (CRC) leads to heavy disease and economic burdens globally. Early screening such as colonoscopy has been demonstrated to reduce both CRC incidence and mortality. However, uptake of colonoscopy among high-risk individuals is low in China, limiting screening efficacy. To improve screening uptake, it is necessary to ensure the engagement of both specialist and general practitioners. And recently, AI techniques show considerable promise in delivering personalized health education. This study aims to evaluate the effectiveness of AI-assisted integrated care, compared with specialty care in improving uptake of colonoscopy among CRC high-risk individuals in China. METHODS: A two-arm, parallel cluster randomized controlled trial will enroll 400 high-risk residents for CRC aged 40-64 years from 40 communities/villages across one urban district and two rural counties in Shandong Province, China. Communities/villages as clusters will be randomized within counties to either a specialty care group or an AI-assisted integrated care group. Participants will be those initially identified as high risk for CRC through a risk assessment questionnaire or fecal immunochemical test. The primary outcomes are colonoscopy uptake at 3 and 6 months post-intervention and the time to colonoscopy. Secondary outcomes encompass CRC screening literacy, beliefs, and intention to undergo colonoscopy. Additionally, we will also conduct a process evaluation and health economic evaluation. DISCUSSION: This study proposes a novel strategy integrating a multidisciplinary team with digital health solutions. We expect that this strategy will provide a feasible approach to improve colonoscopy uptake. The findings could offer practical insights for informing and advancing cancer prevention and control initiatives in China. TRIAL REGISTRATION: ClinicalTrial.gov NCT07261059. Registered on December 2, 2025.