护理给药错误事件的危险因素分析及预测模型的构建

目的快速筛查护理给药风险事件,提出全面而有针对性的防范措施,以期指导临床护理工作,确保患者用药安全,提升护理质量,为药物护理质量管理体系提供较为可靠的理论框架和依据。方法采用回顾性病例对照研究设计,选取2016年1月—2018年12月某三甲医院上报的688例护理给药错误事件作为病例组,以病例组例数的2倍数在同时期药物执行数据库中进行随机抽样,抽样1376例未发生给药错误的护理给药事件作为对照组。对给药错误事件发生的相关客观因素及人为因素进行单因素及多因素Logistic回归分析,探讨护理给药错误事件的潜在危险因素,构建护理给药错误事件的风险预测模型。结果单因素分析显示,护士职称、护龄、患者性别、患者文化程度、患者年龄、给药班次、交接班时段、工作时长、给药途径、药物执行的科室是护理给药错误事件的影响因素(P<0.05)。通过多因素Logistic回归分析将护龄、科室、给药途径、患者年龄、是否处于交接班时段纳入最终的护理给药错误事件风险预测模型(P<0.05)。绘制并计算ROC下面积(AUC)=0.765,AUC>0.7,显示该模型有较好的临床预测能力。结论护理给药错误预测模型的构建可为临床药物护理提供理论依据,对于保障医院药物护理管理体系更具针对性及实用性,能够保证护理安全。

Risk factors analysis and prediction model of nursing medication errors

Objective To quickly screen the risk events of nursing medication,propose comprehensive and targeted preventive measures to guide clinical nursing work,ensure the safety of patient medication,improve the quality of care,and provide a more reliable theoretical framework and basis for the quality management system of drug care.Methods In this retrospective case-control study,totally 688 nursing medication errors reported from a ClassⅢGrade A hospital from January 2016 to December 2018 were selected as the case group,and 1376 cases without nursing medication errors were randomly selected as the control group from the drug execution database at the same time.Univariate and multivariate Logistic regression analysis was performed on the objective and human factors related to the occurrence of medication errors to explore the potential risk factors of nursing medication errors and construct a risk prediction model for nursing medication errors.Results Univariate analysis showed that nurses'job title and working years,patients'gender,education level and age,dosing shift,time of shift,working hours,route of administration,and department of drug administration were the influencing factors of error events in nursing medication(P<0.05).Multivariate Logistic regression analysis was used to incorporate nursing years,department,route of administration,patient's age,whether or not in the shift period into the final risk prediction model for nursing medication errors(P<0.05).The receiver operating characteristic curve(ROC)was drawn and the area under the ROC(AUC)was calculated.The AUC was 0.765,which was>0.7,showing that the model had sound predictive power in clinical practice.Conclusions The construction of the prediction model for nursing medication errors can provide a theoretical basis for clinical drug care,which is more targeted and practical for the hospital drug care management system and can ensure the safety of nursing.