Abstract:

Objective To explore the potential categories and influencing factors of cognitive function trajectories in elderly hemodialysis patients. Methods A convenience sampling method was used to select elderly patients undergoing hemodialysis in the blood purification center of a tertiary hospital in Beijing，China，from June 2022 to June 2024. A general information questionnaire，the Short Physical Performance Battery，the Patients Health Questionnaire-9，the Impact on Participation and Autonomy Questionnaire，the Hemodialysis Self-Management Instrument，and the Mini-Mental State Examination，were used to collect data at the time of enrollment. Cognitive function was evaluated at the time of enrollment and 6，12，18，and 24 months after enrollment，using latent category growth model to identify trajectory categories，and logistic regression，and decision tree models to analyze the influences on cognitive function trajectories. Results Finally，304 patients were included. A total of 2 trajectories of cognitive function in elderly hemodialysis patients were identified，namely，high-level stable group（63.2%） and low-level declining group（36.8%）. Both logistic regression and decision model results showed that the type of dialyzer，body mass index，physical function，and depression status were the factors influencing potential categories of cognitive function for elderly hemodialysis patients. The areas under the receiver operating characteristic curves for logistic regression and decision tree models were 0.823[95%CI（0.775，0.870）] and 0.816[95%CI（0.765，0.867）]，respectively. Conclusion There is heterogeneity in the cognitive function trajectories of elderly hemodialysis patients，and healthcare professionals can develop targeted interventions according to the influencing factors of the trajectory categories to improve their cognitive function.

Key words: Aged, Renal Dialysis, Cognition, Decision Trees, Latent Class Analysis, Nursing Care

摘要：

目的 探讨老年血液透析患者认知功能变化轨迹的潜在类别及影响因素，为临床护理实践提供参考。 方法 便利选取2022年6月—2024年6月北京市某三级甲等医院血液净化中心的317例老年血液透析患者作为调查对象，采用一般资料调查表、简易体能状况量表、患者健康问卷抑郁症状量表、自主参与测评问卷、血液透析患者自我管理量表和简易智能状态检查量表收集患者入组时的资料，在患者入组时和入组后的第6、12、18、24个月评估其认知功能。采用潜类别增长模型识别轨迹类别，采用Logistic回归及决策树模型分析认知功能轨迹的影响因素。 结果 最终纳入患者304例。识别出2类老年血液透析患者认知功能变化轨迹类别，高水平稳定组（63.2%）和低水平下降组（36.8%）。Logistic回归分析和决策树模型结果均显示，透析器的种类、体重指数、躯体功能、抑郁状态是认知功能潜在类别的影响因素。Logistic回归和决策树模型的受试者操作特征曲线下面积分别为0.823[95%CI（0.775，0.870）]、0.816[95%CI（0.765，0.867）]。 结论 老年血液透析患者认知功能轨迹存在群体异质性，医护人员应根据轨迹类别的影响因素，制订针对性的干预方案以提升其认知功能。

关键词: 老年人, 肾透析, 认知, 决策树, 潜在类别分析, 护理