ESCALA DE FRAGILIDADE DE EDMONTON: ESTUDO DE ACURÁCIA DA DETECÇÃO DO IDOSO FRÁGIL

Abstract

There are several instruments for detecting frailty syndrome (FS) in older adults, including the Edmonton Frailty Scale (EFS), which is based on the multidimensional concept of frailty, evaluating EFCHS domains, such as cognition, health status, social support, among others. Few studies have evaluated the psychometric properties of EFS. In Brazil, some studies used the EFS, but there is a divergence in the cutoff point to identify the frail elderly. Objective: To investigate the accuracy of EFS for the diagnosis of frailty in Brazilian community-dwelling elderly. Materials and Methods: Accuracy study with two frailty detection instruments: EFS and the Physical Frailty Index (EFCHS) (reference standard). The instruments were applied to elderly people attended at a health center in Brazil. Sensitivity, specificity, predictive values and area under the ROC curve for EFS were determined. Results: The prevalence of frailty among the 700 elderly participants in the study was 34.6% frail and 34.43% pre-frail and 44.7% frail and 24.1% pre-frail according to the EFCHS and EFS, respectively. The agreement analysis of the two scales showed a moderate association (kappa coefficient of 0.66). Sensitivity, specificity, positive predictive value and negative predictive value of EFS were, respectively: 95.04%; 81.88%; 73.5% and 96.9%. The area under the ROC curve was 0.938 for a cutoff equal to or greater than 6. Conclusion: The results show an excellent accuracy of the EFS in the diagnosis of FS, with high sensitivity and excellent specificity; suggesting a new cutoff point (equal to or greater than 6) for detecting FS in Brazilian community-dwelling elderly