An Al-Driven Framework for Personalized and Diversified Teaching Strategies in Medical Education: A Cross-Sectional Study

Background As times evolve, it becomes increasingly evident that the outdated, one-size-fits-all approach to medical education is failing to meet students' diverse learning needs. In teaching, educators often rely on their own experience and intuition to determine their instructional methods. However, such approaches may not suit every student, resulting in a significant gap between theory and practice. To bridge this gap, we initiated research to establish a new AI-driven framework. This framework enables in-depth analysis of student learning data, automatically matching and recommending personalized teaching plans best suited to each learner. It provides an evidence-based and data-driven solution to the challenges facing medical education. Methods Based on cross-sectional survey data from 442 medical students in the East China region, we created a two-phase framework. First, we created a binary logistic regression model. This model analyzes the key factors that affect students' ability to adapt to online learning, and this profile serves as the basis for model analysis. Secondly, we developed an intelligent system. The system transforms the results of the model into useful ideas through diagnosing and interpreting student profiles, matching strategies, and recommending paths. Results The model has statistical significance. Its ROC curve area is 0.751.The following key variables serve as important predictors of students' adaptability in online learning: it is worth mentioning that a smooth network (OR = 2.006) and the use of specific software such as 'Smart Tree' (OR = 3.177,) are also helpful. On the contrary, using simple applications like 'QQ Group' (OR = 0.572) makes it more difficult for students to adapt to online learning, and issues related to learning and mental health also make it harder for students to adapt. Conclusion The core achievement of our research is the verification of a brand-new, data-driven framework. Specifically, this framework adopts an artificial intelligence model, which can generate a unique learning profile for each student based on the key factors we have verified through experiments. Based on this profile, the system will be able to automatically provide the most suitable personalized teaching recommendations for the student. We firmly believe this approach can help us achieve a truly student-centered learning model, significantly enhancing educational efficiency.