Identifying Fallers Using Clinical Balance Measures in Community-Dwelling Adults with Lower Extremity Amputation: A Cross-Sectional Study

The purpose of this study was to administer clinical balance measures in community-dwelling adults with unilateral lower extremity (LE) amputations and determine optimal cut-off scores for identifying fallers. A cross-sectional study was conducted on a convenience sample of 40 independently ambulatory participants with unilateral LE amputations and a prosthetic device (20 fallers and 20 nonfallers) who were tested during a single session using: the Amputee Mobility Predictor with Prosthesis (AMP PRO), Functional Reach (FR), Single Limb Stance (SLS) and the Timed-Up-and-GO (TUG). There was a statistically significant difference on the AMP PRO (4.75 points.; 95% CI 1.10–8.40) for fallers (36.95 out of 47, SD 7.43) compared with nonfallers (41.70 out of 47, SD 3.13) and on the TUG (4.17 s., 95% CI 0.65–7.69) for fallers (14.84 s, SD 7.33) versus nonfallers (10.67 s, SD 2.64). The AMP PRO had a moderate Area Under the Curve (AUC) predictive value of 0.70 with an optimal cut-off score of 39.5 out of 47 total points (sensitivity 80%, specificity 60%) for identification of fallers. For the TUG, there was a moderate AUC value of 0.68 (sensitivity 70%, specificity 60%) when using a cut-off score of 10.03 s. There was a moderate to high correlation between the AMP PRO and TUG both between (r?=??0.77) and within (r?=??0.68) groups indicating that they measure similar constructs. Limitations included a heterogeneous sample and self-report of fall history. Both the AMP PRO and the TUG provide acceptable sensitivity and specificity for identifying fallers among community-dwelling adults with LE amputation.