研究目的:本研究目的在探討「飲食態度量表-26題」(EAT-26)在台灣社區使用的最佳切點為何。研究方法:研究過程採兩階段研究設計,先於台南市選取高中男、女校各一所,以「EAT-26」進行第一階段之大規模施測,共得有效問卷2966份,再根據得分以分層隨機抽樣之方式,兩校各抽取120位、共240位學生接受第二階段之診斷性晤談,以尋找此問卷在本土最具鑑別性之切點。研究結果:研究結果發現:與其他切點分數相較,當「EAT-26」總分21分時,「ROC」曲線下面積(AUC)值為0.81,呈現最佳之敏感度、特異度與鑑別效度。研究結論:顯示「EAT-26」在本土使用時,篩選社區飲食異常高危險群之切點分數以21分、稍高於國外之20分為宜。本文最後討論研究之限制、建議與未來研究方向。
Purpose: The purpose of this study was to identify the best cut-off point of ”EAT-26” used in a local community. Methods: One girl and one boy senior high schools in Tainan city were purposively selected and, then, ”EAT-26” was applied to screen 2966 students in both schools. Using stratified random sampling method, 120 students were chosen from each school, a total of 240 students. Researchers interviewed with 240 students to decide whether they meet the diagnostic criteria of any eating disorder, then compared the results between the clinical diagnoses and the ”EAT-26” scores of all study samples to find the best cut-off point in Taiwan. Results: There were two major findings: firstly, using psychiatrist's clinical diagnosis as the golden standard, the areas under ”ROC” curve (ACU) of ”EAT-26” equals to 0.81, which shows a good discriminant validity. Conclusions: The best cut-off point of ”EAT-26” used in this local community is 21. Finally, the suggestions and limitations of the study are also addressed.