@article{ref1,
title="Fallers after stroke: a retrospective study to investigate the combination of postural sway measures and clinical information in faller's identification",
journal="Frontiers in neurology",
year="2023",
author="Jonsdottir, Johanna and Mestanza Mattos, Fabiola Giovanna and Torchio, Alessandro and Corrini, Chiara and Cattaneo, Davide",
volume="14",
number="",
pages="e1157453-e1157453",
abstract="BACKGROUND: Falls can have devastating effects on quality of life. No clear relationships have been identified between clinical and stabilometric postural measures and falling in persons after stroke. <br><br>OBJECTIVE: This cross-sectional study investigates the value of including stabilometric measures of sway with clinical measures of balance in models for identification of faller chronic stroke survivors, and the relations between variables. <br><br>METHODS: Clinical and stabilometric data were collected from a convenience sample of 49 persons with stroke in hospital care. They were categorized as fallers (N = 21) or non-fallers (N = 28) based on the occurrence of falls in the previous 6 months. Logistic regression (model 1) was performed with clinical measures, including the Berg Balance scale (BBS), Barthel Index (BI), and Dynamic Gait Index (DGI). A second model (model 2) was run with stabilometric measures, including mediolateral (SwayML) and anterior-posterior sway (SwayAP), velocity of antero-posterior (VelAP) and medio-lateral sway (VelML), and absolute position of center of pressure (CopX abs). A third stepwise regression model was run including all variables, resulting in a model with SwayML, BBS, and BI (model 3). Finally, correlations between independent variables were analyzed. <br><br>RESULTS: The area under the curve (AUC) for model 1 was 0.68 (95%CI: 0.53-0.83, sensitivity = 95%, specificity = 39%) with prediction accuracy of 63.3%. Model 2 resulted in an AUC of 0.68 (95%CI: 0.53-0.84, sensitivity = 76%, specificity = 57%) with prediction accuracy of 65.3%. The AUC of stepwise model 3 was 0.74 (95%CI: 0.60-0.88, sensitivity = 57%, specificity = 81%) with prediction accuracy of 67.4%. Finally, statistically significant correlations were found between clinical variables (p < 0.05), only velocity parameters were correlated with balance performance (p < 0.05). <br><br>CONCLUSION: A model combining BBS, BI, and SwayML was best at identifying faller status in persons in the chronic phase post stroke. When balance performance is poor, a high SwayML may be part of a strategy protecting from falls.<p /> <p>Language: en</p>",
language="en",
issn="1664-2295",
doi="10.3389/fneur.2023.1157453",
url="http://dx.doi.org/10.3389/fneur.2023.1157453"
}