Modelling participation in road accidents of drivers with disabilities who use hand controls

Almost 200 million persons with disabilities face specific difficulties in everyday life. Private vehicles provide persons with disabilities with a high level of flexibility, a high level of time efficiency, and a better quality of life. It is sometimes necessary to make vehicle modifications to enable persons with disabilities to drive. One of the most frequent modifications is hand controls. Although drivers with disabilities who use hand controls face the same risk of road accidents as non-disabled drivers, predictors of road accidents for drivers with disabilities who use hand controls have not been the subject of earlier research. The predictors show which factors influence the occurrence of road accidents of drivers with disabilities who use hand controls. This paper aims to develop a model that describes the participation in road accidents of drivers with disabilities who use hand controls and recognises contributing predictors. A multidisciplinary team of experts identified twenty-three predictors that impact road accidents of drivers with disabilities who use hand controls. Bayesian logistic regression models have identified speeding, alcohol consumption, mobile phone usage, and especially fatigue as risky behaviours. This paper proposes several important measures that would improve the safety of drivers with disabilities using hand controls.近2亿残疾人在日常生活中面临各类特定障碍。私家车可为残疾人带来极高的出行灵活性、时间利用效率与更优的生活质量。为使残疾人能够驾驶车辆，有时需对车辆进行改装，其中最常见的改装项目之一为手动操控装置（hand controls）。尽管使用手动操控装置的残疾驾驶人与健全驾驶人面临的道路交通事故风险并无二致，但此前尚无针对该类驾驶人的道路事故预测因子的相关研究。此类预测因子可揭示哪些因素会影响使用手动操控装置的残疾驾驶人的道路事故发生情况。本研究旨在构建一种模型，既可描述该类驾驶人的道路事故卷入情况，又可识别相关贡献性预测因子。由多学科专家组成的团队共识别出23个影响该类驾驶人道路事故发生的预测因子。通过贝叶斯逻辑回归模型（Bayesian logistic regression models）分析，研究团队识别出超速、饮酒、使用手机，尤其是疲劳驾驶为高危驾驶行为。本研究最后提出了多项可提升使用手动操控装置的残疾驾驶人出行安全的重要改进措施。