机器人辅助步态训练:需要更多随机对照试验吗?或许未必!
本文为观点文章,基于Kuo等人利用Lokomat机器人训练数据与机器学习算法预测神经康复患者功能性步行分类的研究,讨论固定式机器人辅助步态训练的研究优先方向。作者指出随机对照试验未显示该类设备优于可比干预,且有临床实践指南甚至推荐不使用,但设备仍广泛使用,需通过设备数据收集分析、更好患者选择和重视动机因素来优化应用。
要点
- 固定式步态机器人随机对照试验未显示优于可比干预的临床有效性。
- 已有临床实践指南甚至推荐不使用这类固定式步态机器人设备。
- 基于设备训练数据的机器学习预测可辅助优化机器人辅助步态训练应用。
核心结论
- 固定式机器人辅助步态训练随机对照试验未显示优于可比干预的临床有效性,需通过数据收集分析、患者选择和动机因素优化应用。
研究设计与样本
- 本文为观点文章,围绕Kuo等人使用Lokomat机器人训练数据与机器学习算法预测神经康复患者功能性步行分类的研究展开论述。
主要发现
- Kuo等人将患者机器人辅助步态训练干预初期的数据输入不同机器学习算法,可在患者仍参与干预时预测功能性步行分类。
临床怎么用
- 基于设备数据收集与分析可优化固定式步态机器人的应用,并需更好的患者选择和重视参与者动机因素。
局限性
- 本文为观点文章,未报告原始研究数据或系统评价结果。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
I was encouraged by the recent article by Kuo et al. entitled "Prediction of robotic neurorehabilitation functional ambulatory outcome in patients with neurological disorders" to write an opinion piece on the possible further development of stationary robot-assisted gait training research. Randomized clinical trials investigating stationary gait robots have not shown the superiority of these devices over comparable interventions regarding clinical effectiveness, and there are clinical practice guidelines that even recommend against their use. Nevertheless, these devices are still widely used, and our field needs to find ways to apply these devices more effectively. The authors of the article mentioned above feed different machine learning algorithms with patients' data from the beginning of a robot-assisted gait training intervention using the robot Lokomat. The output of these algorithms allows predictions of the clinical outcome (i.e., functional ambulation categories) while the patients are still participating in the intervention. Such an analysis based on the collection of the device's data could optimize the application of these devices. The article provides an example of how our field of research could make progress as we advance, and in this opinion piece, I would like to present my view on the prioritization of upcoming research on robot-assisted gait training. Furthermore, I briefly speculate on some drawbacks of randomized clinical trials in the field of robot-assisted gait training and how the quality and thus the effectiveness of robot-assisted gait training could potentially be improved based on the collection and analysis of clinical training data, a better patient selection and by giving greater weight to the motivational aspects for the participants.
引用
- Labruyère R. Robot-assisted gait training: more randomized controlled trials are needed! Or maybe not? J Neuroeng Rehabil. 2022. PMID: 35676742. doi:10.1186/s12984-022-01037-z
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。