围手术期谵妄风险筛查工具对老年髋部骨折患者风险预测效果影响的Meta分析

目的通过Meta分析的方法评价围手术期谵妄风险筛查工具应用于老年髋部骨折患者的分线预测效果。方法计算机检索中国期刊全文数据库(CNKI)、万方数据资源系统、中国生物医学文献数据库(CBM)、Pubmed数据库、Embase、CINAHL、CochraneLibrary等,收集所有关于围手术期谵妄风险筛查工具对老年髋部骨折患者效果影响的诊断试验,使用QUADAS-2对纳入的文献进行质量评价,使用RevMan5.3软件进行统计处理。结果共纳入6篇原始研究,文献的方法学质量整体不高,多数条目评价结果为偏倚风险不清楚。文献的主要涉及6种风险筛查工具,即谵妄预警模型、谵妄预警模型-修改版、老年谵妄风险评估量表、内科谵妄危险因素预警模型、谵妄预测量表、10分制认知筛查量表,前3个工具的灵敏度和特异度分别为SEN=0.72,95%CI(0.62,0.80);SPE=0.64,95%CI(0.58,0.69);SEN=0.75,95%CI(0.53,0.90);SPE=0.92,95%CI(0.84,0.97);SEN=0.68,95%CI(0.59,0.76);SPE=0.73,95%CI(0.66,0.80);内科谵妄危险因素预警模型中将低风险组与中、高度风险组进行比较,灵敏度和特异度分别为:SEN=0.93,95%CI(0.85,0.98);SPE=0.24,95%CI(0.20,0.28);谵妄预测量表中将低、中度风险组与高、极高度风险组进行比较,灵敏度和特异度分别为:SEN=0.89,95%CI(0.75,0.97);SPE=0.75,95%CI(0.59,0.87);10分制认知筛查量表中将正常、中度风险组与高度风险组进行比较,灵敏度和特异度分别为:SEN=0.74,95%CI(0.59,0.86);SPE=0.77,95%CI(0.65,0.87)。其中,谵妄预警模型-修改版、谵妄预测工具、10分制认知筛查量表的SROC面积大小位于前3。结论谵妄预警模型-修改版、谵妄风险预测量表、10分制认知筛查量表风险筛查准确性较高。

Meta-analysis of the effect of perioperative delirium risk screening tools on risk prediction in elderly patients with hip fracture

Objective To systematically evaluate screening tools for peri-operative delirium in elderly patients with hip fracture.Methods CNKI, Wanfang, CBM, Pubmed, Embase, CINAHL and Cochrane Library were retrieved to obtain relevant studies. The retrieval time was from the inception of database to July 16 th, 2018. After screening the literature, the quality of original researches was assessed by means of QUADAS-2. Meta analysis was carried out using RevMan 5.3.Results 6 original studies were included. The overall quality of the literature was not high, and the risk of bias of most QUADAS-2 items were not clear. 6 screening tools( The risk model delirium model, the risk model delirium-revised, delirium elderly at risk, medical risk factor model, a delirium predicting scale, 10-point cognitive screening scale) were involved, the first 3 tools’ sensitivity and specificity were respectively,, and. For medical risk factor model, the low risk group was compared with middle and high risk groups, the sensitivity and specificity was respectively,. For a delirium predicting scale, the moderate risk group was compared with high and extreme high risk group, the sensitivity and specificity was,. For 10-point cognitive screening scale, the normal, the moderate risk group was compared with high risk group, the sensitivity and specificity was respectively,. Among them, the SROC area of the risk model delirium-revised, a delirium predicting scale, and 10-point cognitive screening scale were among the top 3.Conclusion The accuracy of the risk model delirium-revised, delirium predicting scale, 10-point cognitive delirium predicting scale are high, but large sample studies still need to retest the diagnostic accuracy.