目的:修訂中文版成癮嚴重度指標(Addiction Severity Index,簡稱ASI),並作初 步信度之探討。方法:將McLellan等人設計之ASI翻譯及修訂成中文版,並且收集46名符合精 神疾病診斷手冊第四版(DSM-IV)中「酒精相關疾患」之診斷者,予以施測此中文版ASI。結 果:46名個案中有27名是由兩位研究者同時會談完成問卷,以計算其一致信度;發現兩位研究 者所評之中文版ASI,其中六大分項及總分之間均達統計學上之顯著相關,另外19名個案由兩 位研究者相隔3至7天分別會談及完成中文版ASI,以計算其再測信度;發現只有身體狀態、法 律狀態及家庭社會關係三個分項之前後兩次評分達統計學上有意義之相關。將全部問卷總 合一起計算,就六大分項分別計算或同時計算,其Chronbach�`值達0.71至0.85之間;六大分項 及總分彼此之間亦達統計學上顯著之相關。結論:修訂之中文版ASI有很好的內在一致性與一 致信度,但再測信度則欠理想。每一個案均需花費40至70分鐘之會談以完成問卷,耗時為此問 卷之最大缺點,限制了其在臨床上的廣泛使用。

　　　　　Objective: To test the reliability of a newly translated Chinese version of the addiction severily index (ASI). Methods: We translated the ASI into a Chinese version. The reliability of the Chinese version ASI was then tested on a group of 46 patients with a diagnosis of alcohol related disorders based on DSM IV criteria. Results: Twenty-seven of these patients were evaluated by two members of the research group simultaneously and the correlations of the rating scores of the two researchers were used to calsulate the concurrent inter- rater reliability. We found that the inter-rater reliability was statistically significant for all the six categories (including medical status, employment/ support, alcohol use, legal status, family-social relationship and psychiatric status)and total score in the ASI. The remaining 19 patients were evaluated by another two researchers separately, with 3 to 7 days (4.2±1.2 days)between the two interviews. The correlation of the two ratings was used to calculate the test-retest reliability. Only the results for the categories of 'medical status', 'legal status' and 'family-social relationship' had a statistically significant test-retest reliability. The internal consistency was found to be good as shown by the high Chronbach �` values in every category and when the six categories were calculated together. There was a statistically significant relation among the scores of the six categories and the total scores. Conclusion: Thi revealed that this newly translated Chinese version-ASI had good internal consistency and concurrent interrater reliability.However, its test-retest reliability was not satisfactory. We spent from 40 to 70 minutes for each interview required by the Chinese version-ASI. Time consumption appears to be the major shortcoming of this questionnair an also and important limitation in clinical use.(Full Text in Chinese)