Abstract: Background　Due to the increasing aging population，frailty has emerged as a prominent public health issue.Frail older adults have increased risks for falls，disability and death.A good screening tool is needed to early identify frailty among community-dwelling older adults，thereby delivering precise interventions timely.Objective　To compare the consistency and validity of the Fried Frailty Phenotype（FP），FRAIL Scale and Edmonton Frailty Scale（EFS）in screening frailty among community-dwelling older adults.Methods　During July to August 2018，adults aged 75 years or older living in three sub-districts of Xuhui District，Shanghai，were included using the method of convenient sampling.FP，FRAIL Scale and EFS were used to assess the frailty.Activity of Daily Living Scale （ADL） was used as the validity evaluation standard.Correlation analysis，receiver operating characteristic （ROC） curve and Bayes discriminant analysis were used to test the validity and suitability of the three frailty measurements.Results　Of the 2 000 cases included in the study，1 915 completed the measurements，and 1 625 of them（84.9%） who completed effectively were finally enrolled.The prevalence of frailty screened by the FP，FRAIL Scale and EFS was 24.9%（404/1 625），9.2%（149/1 625） and 28.2%（459/1 625），respectively.The Kappa values of FP and FRAIL Scale，FP and EFS，Frail Scale and EFS were 0.371，0.491 and 0.301，respectively （P<0.001）.Congruent validity of FP，FRAIL Scale and EFS with ADL scale were 0.484，0.564 and 0.653 （P<0.001），respectively.The area under the ROC curve （AUC） of FP，FRAIL Scale and EFS for the prediction of activity of dialy living was 0.748，0.736 and 0.787，respectively.The results of Bayes discriminant analysis〔with decline of activity of dialy living as a dependent variable and three frailty screening tools as independent variables〕 showed that the cross validation accuracy of EFS in predicting the decline of activity of dialy living was 75.70%，which was higher than that of FP（63.90%） and FRAIL Scale（67.80%）.Conclusion　FP，FRAIL Scale and EFS play different roles in identifying frailty in community-dwelling older adults.EFS is a good measure for screening the multi-domain frailty；FP may be a better measure for screening physical frailty than the FRAIL Scale.

Key words: Frailty, Activities of daily living, Community, Aged, Edmonton Frailty Scale, Fried Frailty Phenotype

摘要： 背景　人口老龄化和高龄化使得衰弱成为重要的健康相关问题。相较于无衰弱者，衰弱高龄老年人发生跌倒、失能以及死亡的风险显著增加，选择合适的社区筛查工具有助于识别社区高龄衰弱人群并及早给予精准干预。目的　比较Fried衰弱表型（FP）、FRAIL量表和埃德蒙顿衰弱评估量表（EFS）用于社区高龄老年人衰弱筛查的一致性和适用性。方法　2018年7—8月，采用方便抽样法选取上海市徐汇区3个街道75岁及以上老年人为研究对象。采用FP、FRAIL量表和EFS进行衰弱评估，并以日常生活能力量表（ADL）作为效度评价标准，采用相关性分析、受试者工作特征（ROC）曲线和Bayes判别分析检测三种衰弱评估工具的效度及合适性。结果　共发放问卷2 000份，回收1 915份，其中有效问卷1 625份，有效回收率为84.9%。对于1 625例老年人，FP、FRAIL量表、EFS分别筛查出衰弱老年人404例（24.9%）、149例（9.2%）、459例（28.2%）。一致性检验结果显示，FP与FRAIL量表、FP与EFS、FRAIL量表与EFS之间衰弱评估结果的Kappa值分别为0.371、0.491、0.301（P<0.001）。以ADL为校标，FP、FRAIL量表、EFS的校标关联效度分别为0.484、0.564、0.653（P<0.001）。以日常生活能力为结局变量，FP、FRAIL量表和EFS预测日常生活能力的ROC曲线下面积（AUC）分别为0.748、0.736、0.787。以日常生活能力是否下降为因变量，分别以三种衰弱评估工具得分为自变量，进行Bayes判别分析，结果显示，EFS对日常生活能力下降预测的交叉验证准确率为75.70%，高于FP（63.90%）和FRAIL量表（67.80%）。结论　FP、FRAIL量表及EFS对社区高龄老年人衰弱筛查结果存在差异，其中EFS更适合筛查社区高龄老年人的综合衰弱，FP比FRAIL量表更适合筛查社区高龄老年人的身体衰弱。

关键词: 衰弱, 日常生活能力, 社区, 老年人, 埃德蒙顿衰弱评估量表, Fried衰弱表型