A comparison of frailty indexes based on a comprehensive geriatric assessment for the prediction of adverse outcomes

To compare the ability of different frailty indexes based on a standardized comprehensive geriatric assessment (FI-CGAs) for predicting adverse outcomes.

Prospective cohort study. Geriatric wards of a general hospital.

307 hospitalized patients ≥ 65 years.

The FI-CGA-10D (potential health deficits in ten functional domains), the FI-CGA-10D+CM (aforementioned potential health deficits and co-morbidity burden) and the FI-CGA-MIHD (multiple, individual potential health deficits, including functional deficits, co-morbid diseases, amongst others) were assessed at baseline during the patients` hospital stay. The FI-CGAs were analyzed as categorical (according to a FI-CGA score < and ≥ 0.25) and continuous variables. Patients were followed up over 6 months.

The FI-CGA-10D, FI-CGA-10+CM and the FI-CGA-MIHD predicted 6-month mortality when expressed as categorical (area under the receiver operating characteristic curve (AUC): AUC = 0.611, AUC = 0.637, AUC = 0.616, all p < 0.050, respectively) or continuous variables (AUC = 0.769, AUC = 0.837, AUC = 0.834, all p < 0.001, respectively). AUC comparisons showed that all three FI-CGAs exhibited a comparable ability to predict 6-month mortality when the FI-CGAs were expressed as categorical variables (all p > 0.200) and the FI-CGA-10D+CM and the FI-CGA-MIHD showed a better ability to predict 6-month mortality than the FI-CGA-10D, when the FI-CGAs were expressed as continuous variables (p < 0.001 and p = 0.007, respectively). None of the FI-CGAs predicted any of the other outcomes, i.e., unplanned re-admission to hospital and a fall during follow-up, irrespective of whether the FI-CGAs were expressed as categorical or continuous variables (all p ≥ 0.050).

The more complex FI-CGAs, i.e., the FI-CGA-10D+CM and the FI-CGA-MIHD, revealed better ability to predict 6 month mortality than the more simple FI-CGA, i.e., the FI-CGA-10D.