Department of Physical Therapy, University of Illinois at Chicago, Chicago, IL 60612, United States.
Department of Physical Therapy, University of Illinois at Chicago, Chicago, IL 60612, United States.
J Biomech. 2019 Nov 11;96:109334. doi: 10.1016/j.jbiomech.2019.109334. Epub 2019 Sep 12.
Aging-associated fall-risk assessment is crucial for fall prevention. Thus, this study aimed to develop a prognostic model to predict fall-risk following an unexpected over-ground slip perturbation based on normal gait pattern in healthy older adults. 112 healthy older adults who experienced a novel slip in a safe laboratory environment were included. Their slip trial and natural walking trial immediately prior to it were analyzed. To identify the best fall-risk predictive model, gait related variables including step length, segment angles, center of mass state, and ground reaction force (GRF) were determined and inputted into a stepwise logistic regression. The optimal slip-induced fall prediction model was based on the right thigh angle at slipping foot touchdown (TD), the maximum GRF of the slipping limb after TD, and the momentum change from TD to recovery foot liftoff (LO), with an overall prediction accuracy of 75.9%, predicting 74.5% of falls (sensitivity) and 77.2% of recoveries (specificity). Conversely, a model based on clinical and demographic measures predicted 78.2% of falls and 47.4% of recoveries, resulting in a much lower overall accuracy of 62.5%. The fall-risk model based on normal gait pattern which was developed for slip-induced perturbations in healthy older adults was able to provide a high predictive accuracy. This information could provide insight about the ideal normal gait measures which could be used to contribute towards development of therapeutic strategies related to dynamic balance and fall prevention to enhance preventive interventions in populations with high-risk for slip-induced falls.
与衰老相关的跌倒风险评估对于预防跌倒至关重要。因此，本研究旨在开发一种预后模型，以预测健康老年人在意外地面滑动干扰下的跌倒风险，该模型基于正常的步态模式。纳入了 112 名在安全实验室环境中经历新的滑动的健康老年人。分析了他们在滑动试验之前的自然行走试验。为了确定最佳跌倒风险预测模型，确定了与步态相关的变量，包括步长、节段角度、质心状态和地面反力（GRF），并将其输入逐步逻辑回归。最佳的滑动诱发跌倒预测模型基于滑动脚触地（TD）时的右大腿角度、TD 后滑动肢体的最大 GRF 以及从 TD 到恢复脚离地（LO）的动量变化，总体预测准确率为 75.9%，预测跌倒的准确率为 74.5%（敏感性）和恢复的准确率为 77.2%（特异性）。相反，基于临床和人口统计学措施的模型预测跌倒的准确率为 78.2%，恢复的准确率为 47.4%，整体准确率低至 62.5%。为健康老年人的滑动干扰开发的基于正常步态模式的跌倒风险模型能够提供较高的预测准确率。这些信息可以提供有关理想的正常步态测量的见解，这些测量可用于为与动态平衡和跌倒预防相关的治疗策略的发展做出贡献，以增强对高跌倒风险人群的预防干预。
