Chulalongkorn Centre of Excellence for Parkinson's Disease & Related Disorders, Department of Medicine, Faculty of Medicine, Chulalongkorn University and King Chulalongkorn Memorial Hospital, Thai Red Cross Society, Bangkok, 10330, Thailand.
Department of Biochemistry and Microbiology, Faculty of Pharmaceutical Sciences, Chulalongkorn University, Bangkok, 10330, Thailand; Vaccine and Therapeutic Protein, the Special Task Force for Activating Research, Faculty of Pharmaceutical Sciences, Chulalongkorn University, Bangkok, 10330, Thailand.
Parkinsonism Relat Disord. 2021 Jan;82:77-83. doi: 10.1016/j.parkreldis.2020.11.014. Epub 2020 Nov 19.
Although risk factors that lead to falling in Parkinson's disease (PD) have been previously studied, the established predictors are mostly non-modifiable. A novel method for fall risk assessment may provide more insight into preventable high-risk activities to reduce future falls.
To explore the prediction of falling in PD patients using a machine learning-based approach.
305 PD patients, with or without a history of falls within the past month, were recruited. Data including clinical demographics, medications, and balance confidence, scaled by the 16-item Activities-Specific Balance Confidence Scale (ABC-16), were entered into the supervised machine learning models using XGBoost to explore the prediction of fallers/recurrent fallers in two separate models.
99 (32%) patients were fallers and 58 (19%) were recurrent fallers. The accuracy of the model to predict falls was 72% (p = 0.001). The most important factors were item 7 (sweeping the floor), item 5 (reaching on tiptoes), and item 12 (walking in a crowded mall) in the ABC-16 scale, followed by disease stage and duration. When recurrent falls were analysed, the models had higher accuracy (81%, p = 0.02). The strongest predictors of recurrent falls were item 12, 5, and 10 (walking across parking lot), followed by disease stage and current age.
Our machine learning-based study demonstrated that predictors of falling combined demographics of PD with environmental factors, including high-risk activities that require cognitive attention and changes in vertical and lateral orientations. This enables physicians to focus on modifiable factors and appropriately implement fall prevention strategies for individual patients.
尽管先前已经研究了导致帕金森病（PD）患者跌倒的风险因素，但已确定的预测因素大多是不可改变的。一种新的跌倒风险评估方法可能会更深入地了解可预防的高风险活动，以减少未来的跌倒。
使用基于机器学习的方法探索 PD 患者跌倒的预测。
招募了 305 名 PD 患者，这些患者在过去一个月内有或无跌倒史。将包括临床人口统计学、药物和平衡信心（由 16 项活动特异性平衡信心量表（ABC-16）评定）在内的数据输入到监督机器学习模型中，使用 XGBoost 探索两个单独模型中跌倒者/复发性跌倒者的预测。
99 名（32%）患者为跌倒者，58 名（19%）为复发性跌倒者。该模型预测跌倒的准确性为 72%（p=0.001）。ABC-16 量表中最重要的因素是第 7 项（扫地）、第 5 项（踮脚尖够物）和第 12 项（在拥挤的购物中心行走），其次是疾病分期和病程。当分析复发性跌倒时，模型的准确性更高（81%，p=0.02）。复发性跌倒的最强预测因素是第 12、5 和 10 项（穿过停车场行走），其次是疾病分期和当前年龄。
我们的基于机器学习的研究表明，跌倒的预测因素将 PD 的人口统计学与环境因素相结合，包括需要认知注意力的高风险活动以及垂直和横向方向的变化。这使医生能够专注于可改变的因素，并为个体患者适当实施跌倒预防策略。
