Background
Cancer remains a leading cause of premature mortality worldwide. More than 40% of cases are preventable through modification of behavioral and environmental risk factors; however, current primary prevention strategies face persisitent challenges in personalization, long-term adherence, and sustained behavior change. Although digital health technologies offer substantial promise, most existing tools focus on single risk factors and lack integrated assessment and continuous feedback mechanisms. To address these gaps, we developed the Smart Health Management Digital Platform for Primary Cancer Prevention (SmartHMDP-PCP), a scalable digital system supporting ongoing cancer risk assessment, personalized interventions, and dynamic behavior change in the general population.
 
Objectives
To develop and deploy a scalable digital platform for personalized primary cancer prevention, and to evaluate its feasibility, user engagement, and short-term behavioral impacts across diverse populations in China, generating real-world evidence to inform precision public health implementation and future randomized trials.
 
Methods
Between 2021 and 2025, we conducted mixed-methods evaluations across nine provinces in China to assess deployment feasibility, iteratively optimize platform design and content, and examine short-term behavioral impacts. Approximately 200,000 adults aged ≥18 years were enrolled through community and healthcare networks. Participants completed a baseline cancer risk assessment and up to 17 (women) or 16 (men) exposure modules covering lifestyle, environmental, and behavioral factors. Individual cancer risks were estimated using Rothman-Keller models, and personalized recommendations were delivered through adaptive behavioral counseling and AI-enabled digital assistants. Follow-up assessments evaluated changes in health behaviors over time.
 
Results
SmartHMDP-PCP evolved through four iterative editions, with progressive enhancements in interface design, data architecture, and intervention functionality (Figure 1). The platform expanded from an initial pilot of 7,609 users in 2021 to 148,324 users during 2021–2024, followed by a third edition enrolling an additional 43,542 users by the end of 2025. The forthcoming fourth edition incorporates an AI-powered conversational agent and intelligent knowledge-retrieval system to enhance personalization and sustained engagement.
 
Among 112,678 participants with sufficient data to construct adherence scores based on the 2018 WCRF/AICR recommendations, 23.32% were classified as having healthy behavioral patterns, 60.49% as moderately healthy, and 16.19% as unhealthy. Adherence to healthy behaviors varied substantially across demographic and geographic subgroups, with prevalence ranging from 13.82% to 35.05%. Among 49,207 participants completing follow-up, the proportion adhering to healthy lifestyle patterns increased by 8.91%, accompanied by an 11.81% reduction in the initially unhealthy group. The largest behavioral improvements were observed in alcohol consumption (the proportion of non-drinkers increased by 10.63%) and fast-food intake (the proportion limiting fast-food consumption increased by 8.39%), whereas more modest improvements were seen for decreased red meat consumption (1.60%) and increased fruit and vegetable intake (1.64%).   
 
Conclusion
SmartHMDP-PCP demonstrates strong feasibility, high user engagement, and measurable early behavioral benefits as an integrated, scalable digital solution for personalized primary cancer prevention. These findings provide robust justification for a randomized implementation trial planned for 2026. By integrating multidimensional risk stratification with continuous, individualized behavior support, this platform offers a pragmatic pathway for operationalizing precision public health and accelerating progress toward national and global cancer prevention goals.

Figure 1. Evolution and Implementation Pathway of the Smart Health Management of Digital Platform for Primary Cancer prevention (SmartHMDP-PCP), Including Prototype Design, Functional Optimization, Interface Enhancement, and Public Deployment via WeChat M