住院病人护理不良事件发生风险预测模型的构建及验证

Construction and validation of a risk prediction model for nursing adverse events in hospitalized patients

目的:构建住院病人护理不良事件发生风险预测模型,并检验其效能。方法:采用便利抽样法,选取2023年1月—12月江西省某三级甲等医院发生护理不良事件的150例病人和未发生护理不良事件的160例病人作为研究对象。基于医院信息系统收集可能导致住院病人护理不良事件发生的危险因素。采用Logistic回归分析法建立风险预测模型,应用受试者工作特征曲线、Hosmer-Lemeshow拟合优度检验对构建的模型进行评价。结果:Logistic回归分析结果显示,护士职称、“三基三严”入职培训成绩、临床专科培训成绩、一级护理质量督导、病人安全文化、护理人力资源、医院硬件设施设备是护理不良事件发生的影响因素(P<0.05)。根据Logistic回归分析结果构建住院病人护理不良事件发生风险预测模型,模型受试者工作特征曲线下面积为0.850[95%CI(0.808,0.891)],Hosmer-Lemeshow检验结果显示,χ^(2)=9.816,P=0.278,模型最大约登指数为0.584,灵敏度为0.747,特异度为0.838。结论:护士职称“、三基三严”入职培训成绩、临床专科培训成绩、一级护理质量督导、病人安全文化、护理人力资源、医院硬件设施设备对住院病人护理不良事件发生情况有影响,住院病人护理不良事件发生风险预测模型预测效能良好,可为临床护理不良事件防控提供参考。

Objective:To construct a risk prediction model for nursing adverse events in hospitalized patients and test its effectiveness.Methods:A total of 150 patients who experienced nursing adverse events and 160 patients who did not experience nursing adverse events in a tertiary grade A hospital in Jiangxi province were selected as research subjects by convenient sampling method from January to December 2023.Risk factors that may lead to nursing adverse events in hospitalized patients based on Hospital Information System were collected.Logistic regression analysis was used to establish the risk prediction model,and the receiver operator characteristic curve and Hosmer-Lemeshow goodness-of-fit test were used to evaluate the constructed model.Results:Logistic regression analysis showed that nursing professional title,"three basics and three stricts"entry training performance,clinical specialty training performance,first level nursing quality supervision,patient safety culture,nursing human resources,and hospital hardware facilities and equipment were the influencing factors of nursing adverse events(P<0.05).A risk prediction model for nursing adverse events in hospitalized patients was constructed based on the results of Logistic regression analysis,area under the curve of receiver operator characteristic of the model was 0.850(95%CI 0.808-0.891).The Hosmer-Lemeshow test results showed thatχ^(2)=9.816,P=0.278,the maximum Youden index of the model was 0.584,the sensitivity was 0.747,and the specificity was 0.838.Conclusion:Nursing professional title,"three basics and three stricts"entry training performance,clinical specialty training performance,first level nursing quality supervision,patient safety culture,nursing human resources,and hospital hardware facilities and equipment have impact on the occurrence of nursing adverse events in hospitalized patients.The risk prediction model for adverse nursing events in hospitalized patients has good predictive performance and can provide reference for clinical prevention and control of adverse nursing events.