Abstract
PURPOSE: the phenotype defined by Fried et al. is one of the main operationalizations of frailty. Santos-Eggimann et al. pioneered the adaptation of the phenotype criteria to the Survey of Health, Ageing and Retirement in Europe (SHARE, http://www.share-project.org/). Using the adapted criteria, Romero-Ortuno et al. created and validated the SHARE Frailty Instrument for Primary Care (SHARE-FI). In SHARE-FI, the cut-offs for the phenotypic categories (i.e. non-frail, pre-frail and frail) are automatically derived from latent variable modelling, while Fried et al. (and also Santos-Eggimann et al.) use a rule based on the number of criteria present (Ncriteria): ≥ 3: frail; 1 or 2: pre-frail; 0: non-frail. The aim of the present study was to compare the mortality prediction of these two different approaches (latent variable modelling in SHARE-FI vs. Ncriteria in Santos-Eggimann et al.). SUBJECTS AND METHODS: the subjects were 15,420 women and 12,742 men from the first wave of SHARE. A correspondence analysis was used to assess the degree of agreement between phenotypic classifications. The ability of the continuous measures (Ncriteria and SHARE-FI score) to predict mortality (mean follow-up of 2.4 years) was compared using receiver operating characteristic plots and areas under the curve (AUC). RESULTS: in both women and men, there was a high degree of correspondence between phenotypic categories. The two continuous measures performed similarly as mortality predictors (women: SHARE-FI-AUC = 0.77; Ncriteria-AUC = 0.75. Men: SHARE-FI-AUC = 0.76; Ncriteria-AUC = 0.72). CONCLUSION: the two approaches to the SHARE operationalized frailty phenotype performed equally well to predict mortality.