Receiver Operating Characteristic Analysis of Posture and Gait Parameters to Prevent Frailty Condition and Fall Risk in the Elderly

Abstract

Prevention strategies should be constantly improved to manage falls and frailty in the elderly. Therefore, we aimed at creating a screening and predictive protocol as a replicable model in clinical settings. Bioimpedance analysis was conducted on fifty subjects (mean age 76.9 ± 3.69 years) to obtain body composition; then, posture was analysed with a stabilometric platform. Gait performance was recorded by a 10 m walking test, six-minute walking test, and timed up and go test. After 12 months, subjects were interviewed to check for fall events. Non-parametric analysis was used for comparisons between fallers and non-fallers and between able and frail subjects. ROC curves were obtained to identify the predictive value of falling risk and frailty. Path length (area under the curve, AUC = 0.678), sway area (AUC = 0.727), and sway speed (AUC = 0.778) resulted predictive factors of fall events (p < 0.05). The six-minute walking test predicted frailty condition (AUC = 0.840). Timed up and go test was predictive of both frailty (AUC = 0.702) and fall events (AUC = 0.681). Stabilometry and gait tests should be, therefore, included in a screening protocol for the elderly to prevent fall events and recognize the condition of frailty at an early stage.