进食评估问卷调查工具-10对吞咽障碍筛检价值的Meta分析

摘要

Objective To evaluate the diagnostic value of the Eating Assessment Tool (EAT-10) in screening dysphagia by using 2 and 3 as cut-off value with systematic review. Methods A retrieval on the diagnostic value of EAT-10 in dysphagia was done all through PubMed, Embase, the Cochrane Library, CBM, CNKI, WANFANG and VIP databases (from January 2008 to November 2016)."Quality As-sessment of Diagnostic Accuracy Studies" (QUADAS) was used to appraise included papers. Meta-Disc1.4 software was used to analyze heterogeneity and the combined effect of the data. Results Three papers were included and 394 subjects were included in the review. By using 2 and 3 as the cut-off value of EAT-10, the value of the weighted sensitivity was 0.84 [95%CI (0.79, 0.88)] and 0.77 [95%CI(0.71, 0.82)], the specificity was 0.64[95%CI (0.55, 0.72)] and 0.74 [95%CI(0.66, 0.81)],and the area under the receiver-operating-characteris-tic-curve (AUC-ROC) was 0.8621 and 0.8343, which illustrated the diagnostic accuracy of EAT-10 on dysphagia at the medium level. Conclusion The diagnostic ability of EAT-10 is good, but the sensitivity and specificity are different by using 2 and 3 as the cut-off value. The subjects should consider the specific clinical situation to determine the best cut-off value.%目的 采用Meta分析方法探讨评价进食评估问卷调查工具-10(the eating assessment tool,EAT-10)诊断分界值为2和3分时,其对吞咽障碍的筛检价值.方法 计算机检索2008年1月至2016年11月PubMed、Embase、Cochrane图书馆、中国生物医学文献数据库(CBM)、中国期刊全文数据库(CNKI)、万方数据库、维普数据库中使用EAT-10筛查吞咽障碍的诊断/筛检试验.采用"诊断精确性研究质量评价"(quality assessment of diagnostic accuracy studies,QUADAS)工具评价文献质量,Meta-DiSc1.4软件进行异质性检验和合并效应量的分析.结果 共纳入3篇文献,合计394例研究对象.EAT-10诊断分界值为2分和3分时,其合并敏感度分别为0.84[95%CI(0.79,0.88)]和0.77[95%CI(0.71,0.82)],合并特异度分别为0.64[95%CI(0.55,0.72)]和0.74[95%CI(0.66,0.81)],EAT-10诊断分界值为2分时其漏诊率低于3分时,但误诊率要高于3分时;EAT-10诊断分界值为2分与3分时,其Q*指数分别为0.7928和0.7666,EAT-10诊断分界值为2分与3分时,其受试者诊断特征(receiver operating characteristic carve,SROC)曲线下面积(the area under carve,AUC)分别为0.8621和0.8343,表明分界值为2分时其诊断准确性相对高些,而且不管在2分时还是3分时EAT-10对吞咽障碍的诊断准确性均处于中等水平.结论 EAT-10能有效筛检出吞咽障碍患者,但分界值为2分和3分时其筛检效能有所差异,应综合考虑具体的临床情况确定最佳分界值.