Objective: To study the reliability and validity of the Chinese Eating Assessment Tool (EAT-10) in evaluation of acute stroke patients with dysphagia.
Methods: The inpatients of stroke were assessed with Chinese EAT-10. As a golden standard for evaluation of dysphagia, videofluoroscopic swallow study (VFSS) test was used to judge the reliability and validity of EAT-10.
Results: A total of 130 qualified questionnaires were collected. The Cronbach's alpha coefficient for Chinese EAT-10 scale was 0.845. The total score of each item was related. The lowest or highest correlation coefficient for the item 2 or 3 was 0.271 or 0.772. The retest reliability was greater than 0.7, which met the requirements. According to the investigator consistency reliability test, the value collected from the investigator in the item 2 kept constant. The consistent correlation coefficient of the remaining nine items was more than 0.7. The consistency between each item and the mean score was high. The EAT-10 with the cut-off point at 1 was an optimal cut-off point. With the cut-off value of 1 (EAT-10 score ≥ 1), the sensitivity and specificity for EAT-10 was 77.9% and 66.1%, respectively. The positive predictive value (PPV) and negative value (NPV) was 71.6% and 73.2%, respectively, with 2.30 LHR+ and 0.33 LHR- for dysphagia.
Conclusion: The Chinese EAT-10 has a good reliability and validity in evaluation of the acute stroke patients with dysphagia. The sensitivity and negative value are the best with the cut-off value of 1 (EAT-10 score ≥ 1). It offers a good way to discriminate dysphagia, impaired efficacy, penetrations, and aspirations in acute stroke patients.
目的：探讨进食评估问卷调查工具-10(eating assessment tool-10，EAT-10)中文版在急性期脑卒中后吞咽障碍评估中的信度和效度。方法：选择急性期脑卒中后住院患者180例，采用EAT-10中文版量表进行筛查评估，并用吞咽障碍的诊断金标准——视频透视吞咽功能检查(videofluoroscopic swallow study，VFSS)为效标进行检验。结果：180名入选患者中有130名问卷合格并顺利行VFS检查，EAT-10中文版总量表Cronbach’s α=0.845，各条目与总分均存在相关，相关系数最低的为条目2(r=0.271)，相关系数最高的为条目3(r=0.772)，重测信度均为0.7以上，重测信度符合要求。经调查员一致性信度检验，条目2有一位调查员的结果是恒定值，量表其余9个条目的一致相关系数均>0.7，各亚项和总分均值间一致性较高。分别对EAT-10分界值3，2，1进行效度检验，发现分界值1为最理想分界值，灵敏度及阴性预测值最高，判断吞咽障碍的灵敏度为77.9%，特异度为66.1%，阳性预测值71.6%，阴性预测值73.2%，阳性似然比2.30，阴性似然比0.33。结论：EAT-10中文版仅适用于已有饮水和进食经历的患者，EAT-10中文版对评估急性期脑卒中患者有良好的信度和效度，当分界值为1，EAT-10总分≥1时灵敏度和阴性预测值最佳，能够较好地预测急性期脑卒中患者吞咽障碍、吞咽能力受损、渗透和误吸。.