Abstract:

Objective This study aims to explore the influencing factors and interaction mechanisms of medication discrepancy in elderly patients with diabetes living in empty-nest conditions，thereby providing a foundation for the development of targeted intervention strategies. Methods By the convenience sampling method，400 elderly empty-nest diabetic patients who were hospitalized in the endocrinology department of a tertiary grade A hospital in Zhuzhou City from February to May，2025 were selected as survey subjects. Baseline data were collected prior to discharge. 2 months post-discharge，patients were surveyed using the General Information Questionnaire，the Self-Efficacy for Appropriate Medication Use Scale，the Summary of Diabetes Self-Care Activity，the Financial Toxicity In Chronic Disease Scale，and the Medication Discrepancy Tool. Potential predictive variables were initially screened through univariate analysis. Significant variables were subsequently analyzed using binary logistic regression，while dependencies and pathways among the variables were explored using a Bayesian network model. Results A total of 400 questionnaires were distributed and 376 valid questionnaires were retrieved. The incidence rate of medication discrepancy among elderly diabetic patients in empty-nest families is 57.71%. The Bayesian network model demonstrated that medication frequency，self-efficacy in rational medication use，diabetes self-management status，and frequency of communication with children are directly correlated with medication discrepancy. In contrast，age，education level，sleep duration，triglyceride glucose index，number of comorbid chronic diseases，and types of medications exhibit indirect relationships with medication discrepancy. Notably，the risk of medication discrepancy is highest（99.7%） among elderly diabetic patients living alone who take medication 3 or more times per day，exhibit poor diabetes self-management，possess low self-efficacy in rational medication use，and do not communicate with their children. Conclusion The incidence of medication discrepancy among elderly diabetic patients living in empty-nest situations is notably high，attributed to the interplay of multiple factors. This issue is influenced by various dimensions，including medication complexity，psychological behavior，and social support. Therefore，clinical medical staff need to adopt systematic strategies to build a multi-level and precise prevention and control system.

Key words: Empty-Nest Elders, Diabetes, Medication Discrepancy, Bayesian Network, Continuing Nursing Care

摘要：

目的 探讨空巢老年糖尿病患者用药偏差的影响因素及其交互网络，为制订精准化干预策略提供依据。方法 采用便利抽样法，选取2025年2—5月在株洲市某三级甲等医院内分泌科住院治疗的400例空巢老年糖尿病患者作为调查对象。出院前收集患者的基线资料，出院2个月使用一般资料调查表、中文版合理服药自我效能量表、中文版糖尿病自我管理行为量表、中文版慢性病经济毒性量表、改良版用药差异评估工具对患者进行调查。通过单因素分析初步筛选潜在预测变量，采用二元Logistic回归进一步筛选变量，并采用贝叶斯网络模型分析各变量间的依赖关系与作用路径。结果 共发放问卷400份，回收有效问卷376份。空巢老年糖尿病患者用药偏差的发生率为57.71%；贝叶斯网络模型显示，服药频率、合理服药自我效能、糖尿病自我管理情况、与子女交流的频率与用药偏差直接相关，年龄、文化程度、睡眠时长、甘油三酯葡萄糖指数、合并慢性病数量、服药种类与用药偏差间接相关；服药频率每天≥3次、糖尿病自我管理情况较差、合理服药自我效能较差、从不与子女交流的空巢老年糖尿病患者发生用药偏差的风险最大（99.7%）。结论 空巢老年糖尿病患者的用药偏差发生率高，是多种因素交互作用的结果，受服药复杂性、心理行为及社会支持等多维度影响，临床医护人员需采取系统性策略，以构建多层次的精准防控体系。

关键词: 空巢老年人, 糖尿病, 用药偏差, 贝叶斯网络, 延续性护理