目的：長期照護機構中的住民常因疾病或衰老而伴隨有平衡功能的問題。本研究的目的在比較平衡電腦適性測驗 (balance computerized adaptive testing, Balance CAT)及柏格氏平衡量表 (Berg balance score, BBS)之再測信度、最小可偵測變化值 (minimal detectable change, MDC)及施測效率。方法：於某護理之家中招募具備基本認知功能的住民參與本研究，進行前後二次評估，間隔約3周。

結果：共78位住民完成評估。Balance CAT的前後測平均值各為3.1分與3.0分，差異平均值為-0.04分 (p ＞ .05)。組內相關係數及MDC (%)各為0.95與1.4分 (14%)。BBS前後測平均值各為10.5分與11.4分，差異平均值為0.9分 (p ＞ .05)。組內相關係數及MDC (%)各為0.98與5.6分 (10%)。使用Balance CAT與BBS之平均時間分別為42.7秒與102.8秒；Balance CAT平均施測項目為3.5項。結論：Balance CAT與BBS應用於長期照護機構中之住民均具有良好之再測一致性與MDC。施測效率以Balance CAT較佳。本研究僅於單一護理之家收案，且參與個案須具備基本認知功能，故研究結果難以概化到所有長期照護機構之住民。

Objective: The residents of long-term care (LTC) facilities usually have balance problems due to illness or infirmity. The purpose of this study was to compare the test-retest reliability, minimal detectable change (MDC), and efficiency of the Balance Computerized Adaptive Testing (Balance CAT) and the Berg Balance Scale (BBS) when administered to LTC residents. Method: Residents of one LTC facility with sufficient cognitive function were recruited to participate in this study and complete the assessments twice at an interval of about three weeks. Results: A total of 78 residents completed two sessions of assessments. The mean scores of pre- and post-assessments of the Balance CAT were 3.1 and 3.0, respectively, and the mean difference was -0.04 points (p ＞ .05). The intraclass correlation coefficient (ICC) and MDC (%) of the Balance CAT were 0.95 and 1.4 points (14.0%), respectively. The mean scores of the BBS were 3.07 and 3.02, respectively, and the mean difference was 0.9 points (p ＞ .05). The ICC and MDC (%) of the BBS were 0.98 and 5.6 points (10.0%), respectively. The mean administration times of the Balance CAT and the BBS were 42.7 seconds and 102.8 seconds, respectively. The mean number of items for one administration of the Balance CAT was 3.5. Conclusion: The Balance CAT and the BBS both have nearly excellent test-retest reliability and MDC. The Balance CAT has better efficiency than the BBS when administered to LTC residents. The limitations of this study include that the participants came from a single nursing home and needed to have basic cognitive function to complete the tests. The results of this study cannot be generalized to all residents of LTC facilities.