測量工具之反應性研究回顧

A Review on the Responsiveness Studies of Measurement Tools

王湘慧(Hsiang-Hui Wang) ； 廖華芳(Hua-Fang Liao)

《物理治療》 29卷4期 (2004/08) Pp. 263-270

摘要

自第三次醫療革命，臨床開始著重成效評估，因此所使用的測量工具便要能評鑑出病人治療前後的改變，以顯現出治療成效，即測量工具必須具有反應性。本文綜合相關文獻，列出反應性之定義、研究方法及測量之統計方法。所謂良好反應性是指此測量工具能偵測到一個確切已有的發生改變量，例如：受試者經一段時間後的臨床改變或最小臨床有意義的改變。反應性的研究方法可依據「層次」、「個體內的改變或個體間的差異」及「改變值的種類」三個向度分為30種，而反應性測量的統計方法也隨不同的「改變值的種類」而異，包括：可信賴之改變指數、配對t檢定、效應值、ROC曲線（ Receiver Operating Characteristic curves ）等等。但以不同反應性測量的統計值比較不同工具之反應性，似乎無法得到一致的結果，所以不應只以單一反應性測量的統計值來代表一測量工具之反應性。（物理治療 2004 ; 29 ( 4 ) : 263 － 270 )

關鍵字

反應性 ； 測量工具

並列摘要

During the third stage of medical revolution, the focus of treatment has been directed toward enhancing the quality and effectiveness of health care. Therefore the measurement tools physical therapists use need to show changes after treatment, the responsiveness. Only measurement tools with a large effect of responsiveness are capable of distinguishing the change from the monotony. We review the related articles, and describe the definitions, study methods, and indices of responsiveness in this article. The large effect of responsiveness is that a measure is able to detect the change or minimally clinical important change over a particular prespecified time frame. The studies of responsiveness are classified into 30 categories based on the dimensions of “setting” , “which”, and “type of change/difference”. The statistic methods of responsiveness, which are based on the classification of “type of change/ difference” indices, include reliability change index, paired t-test, effect size statistics, receiver operating characteristic curves, etc. The different indices of responsiveness may provide different rank orderings of scales in terms of responsiveness. We then suggest using a number of indices of responsiveness to represent the responsiveness of one measure. (FJPT 2004;29(4):263-270)

並列關鍵字

Responsiveness ； Measurement