目的:探討修訂版長谷川氏癡呆量表(HDS-R)使用於臺灣的老年人口的可行性與臨 床效度。方法:從臺北市萬華區與臺北縣平溪鄉戶籍登記65歲以上老年人口中隨機抽取四分 之三進行篩檢，兩社區中各選取篩檢分數最低之50名個案，另以年齡配對選取比較組50名， 再加上臨床個案，接受HDS-R測試與臨床評估，以求得HDS-R診斷癡呆症之敏感度、特異 度及最佳分界點。結果:191人完成HDS-R測試與臨床評估，認知功能障礙者與正常者各有 45、146人。以共識診斷符合DSM-III-R癡呆症為標準，以17/18分為最佳分界點，HDS-R 之敏感度及特異度93%，89%。不同等級的癡呆症患者間，HDS-R分數有顯著差異， CDR(Clinical Dementia Rating, CDR)與HDS-R分數間呈負相關(Spearman's r=-0.59， p<0.0001)，顯示HDS-R總分也可作為癡呆症嚴重度的指標。各題中，區辨正常與認知功能 障礙者最佳的三題是:語言流暢度，三樣東西之訊息登錄，以及立即回憶剛看過的五樣東西。 結論:修訂版長谷川氏癡呆量表為良好的癡呆症篩檢工具及嚴重度指標，但用於低教育程度的 老年人口，分界點需調低2-3分。

　　　　　Objective: To explore the validity of revised Hasegawa's Dementia Scale (HDS-R) in the elderly population in Taiwan. Methods: Three-fourths of all the elderly residents in an urban and a rural community (N=749) were screened with the Cognitive Abilities Screening Instrument. The fifty subjects with the lowest test score and another fifty age-matched controls were selected from each community. The subjects included community residents as well as patients receiving treatment in various clinical settings. All subjects completed the HDS-R test and a comprehensive psychiatric evaluation. Results: Of the 191 subjects for whom evaluation was completed, 45 cases with cognitive impairment (27 of them fulfilling the DSM-III-R diagnostic criteria for dementia) Were identified by consensus diagnosis. Using a cut-off score of 17/18 to maximize separation between the dementia and normal groups, the sensitivity and specificity of HDS-R for the diagnosis of dementia were 93% and 89%, respectively. The total score of HDS-R decreased significantly as the severity of dementia increased (Spearman's r=-0.59, p<0.0001). Category verbal fluency test, registration of three words and prompt recall of 5 visually recognized objects were the three most effective items to discriminate between normal and the cognitive impaired subjects. Conclusions: The results of this study indicate that the Revised Hasegawa's Dementia Scale is a good screening instrument and severity indicator of dementia in the elderly population of Taiwan. These results also suggest that optimal cut-off score used in the predominately illiterate elderly in Taiwan was 2-3 points lower than that in original Japanese version. (Full Text in Chinese)