目的: 通过中介检验探究自我效能在自我管理能力和自我管理行为间的作用机制，及其在不同病程患者间的差异。方法: 以2022年7—9月在山西省、内蒙古自治区的四家医院内分泌科就诊的489例2型糖尿病患者为研究对象，通过一般资料调查表、糖尿病自我管理量表、中文版糖尿病授权简化量表、糖尿病自我效能量表进行调查。采用线性回归模型、Sobel检验、Bootstrap检验，在软件Stata 15.0中进行中介检验，并根据病程是否>5年，划分不同病程组别，进行亚组分析。结果: 2型糖尿病患者自我管理行为得分为6.16±1.41，自我管理能力得分为3.99±0.74，自我效能得分为7.05±1.90。自我效能与自我管理能力(r=0.33)以及自我管理行为(r=0.47)呈正相关(P均 < 0.01)。自我效能的中介效应在自我管理能力对自我管理行为的总效应中占38.28%，且在血糖监测(43.45%)和饮食控制(52.63%)行为中占比很高。病程≤ 5年的患者，自我效能的中介效应约占总效应的40.99%，而对于病程> 5年的患者，中介效应占总效应的39.20%。结论: 自我效能增强了自我管理能力对2型糖尿病患者行为的影响，并且对于病程较短患者来说，这种正向影响更显著。应针对患者的患病特征开展有针对性的健康教育，提升患者自我效能与自我管理能力，激发患者的内在行动力，促进其自我管理行为的养成，并形成疾病管理更加稳定、长效的机制。

Objective: To investigate the mechanism of self-efficacy between self-management ability and self-management behavior and its differences among patients with different disease courses through mediation tests. Methods: In the study, 489 patients with type 2 diabetes who attended the endocrinology departments of four hospitals in Shanxi Province and Inner Mongolia Autonomous Region from July to September 2022 were enrolled as the study population. They were investigated by General Information Questionnaire, Diabetes Self-Management Scale, Chinese version of Diabetes Empowerment Simplified Scale, and Diabetes Self-Efficacy Scale. Mediation analyses were performed using the linear regression model, Sobel test, and Bootstrap test in the software Stata version 15.0 and divided the patients into different disease course groups for subgroup analysis according to whether the disease course was > 5 years. Results: In this study, the score of self-management behavior in the patients with type 2 diabetes was 6.16±1.41, the score of self-management ability was 3.99±0.74, and the score of self-efficacy was 7.05±1.90. The results of the study showed that self-efficacy was positively correlated with self-management ability (r=0.33) as well as self-management behavior (r=0.47) in the patients with type 2 diabetes (P < 0.01). The mediating effect of self-efficacy accounted for 38.28% of the total effect of self-management ability on self-management behaviors and was higher in the behaviors of blood glucose monitoring (43.45%) and diet control (52.63%). The mediating effect of self-efficacy accounted for approximately 40.99% of the total effect for the patients with disease course ≤ 5 years, while for the patients with disease course > 5 years, the mediating effect accounted for 39.20% of the total effect. Conclusion: Self-efficacy enhanced the effect of self-management ability on the behavior of the patients with type 2 diabetes, and this positive effect was more significant for the patients with shorter disease course. Targeted health education should be carried out to enhance patients' self-efficacy and self-management ability according to their disease characteristics, to stimulate their inner action, to promote the development of their self-management behaviors, and to form a more stable and long-term mechanism for disease management.