Abstract: Behavioral and lifestyle interventions (BLIs) form the cornerstone of preventing and treating type 2 diabetes mellitus (T2DM), and are increasingly moving toward customization, dosing, personalization, and precision. Face-to-face interventions are irreplaceable by remote interventions in building doctor-patient trust, fostering harmonious doctor-patient relationships, and enhancing the effectiveness of behavioral interventions. Through diabetes self-management education and support (DSMES), guiding patients in self-management not only improves the efficacy of BLIs but also significantly reduces the workload of healthcare professionals. Behavioral change theory (BCT) and techniques (BCTs) can enhance the efficiency of such interventions and will be widely promoted and applied. Large language models (LLMs) in artificial intelligence (AI) will play a pivotal role in predicting the risk of diabetes onset, providing personalized lifestyle guidance, enabling remote monitoring and self-management, and supporting AI-driven clinical decision-making systems.

Key words: Behavioral and lifestyle interventions, Type 2 diabetes mellitus, Face-to-face interventions, Self-management education and support, Behavioral change theories, Behavioral change techniques, Artificial intelligence

摘要： 行为与生活方式干预（behavioral and lifestyle interventions, BLIs）是 2 型糖尿病预防和治疗的基础，并逐渐走向定制化、剂量化、个性化和精准化。面对面干预在建立医患互信、融洽医患关系和提升行为干预效果方面是远程干预不可替代的。通过糖尿病自我管理教育与支持（diabetes self-management education and support, DSMES），指导患者进行自我管理，既能增加行为与生活方式干预的有效性，也可大幅减轻医护人员的工作负担。行为改变理论（behavioral change theory, BCT）和行为改变技术（behavioral change techniques, BCTs）可提高 BLIs 的效率，建议在 2型糖尿病防治中予以推广和应用。人工智能（AI）大模型（LLMs）在糖尿病发病风险预测、个性化生活方式指导、远程监测与自我管理和临床决策支持系统应用方面将发挥重要作用。

关键词: 行为与生活方式干预, 2 型糖尿病, 面对面干预, 自我管理教育与支持, 行为改变理论, 行为改变技术, 人工智能