Abstract: Previous studies on accident consequences of two-wheeled vehicle accidents mainly focused on discussing the overall injuries of cyclists, while few studies focused on the injuries of specific body parts of cyclists. In order to reveal the correlation and cause of injuries of various body parts, the injury severity of seven main parts such as the head, chest, upper and lower limbs of the two wheel (TW) driver were addressed as the dependent variables, and 44 main independent variables were selected to represent the pre-crash behavior, crash position and other information, which was based on 2799 two-wheeled vehicle accidents sourced from the China in-depth accident study (CIDAS) data. A multivariable model with random parameters was used to better explain the relationship between the severity of each injury site and the heterogeneity that was not observed in the data. The results show that the gender and age of TW driver, TW vehicle characteristics, TW driving behaviors prior to crash and TW collision positions will have a significant impact on the injuries of different parts of TW driver. The head, chest and lower limbs are proven to be three most vulnerable parts during TW crashes, which deserve further study. The middle-aged TW driver is a random parameter obeying normal distribution in the head injury model, indicating that the middle-aged TW driver will have heterogeneous effects on the severity of head injury of TW driver. It is extremely necessary to strengthen the safety design of cars and two-wheeled vehicles to protect the chest and abdomen of two wheeled cyclists. Wearing helmets can not only protect head from injuries, but also play an important role in protecting from upper limb injuries.

Key words: traffic engineering; two-wheel cyclists; body region; CIDAS data; random parameter model

摘要： 现有针对二轮车事故后果的研究主要侧重于讨论骑行者总体伤害，少有针对骑行者具体身体部位的受伤情况相关研究。为揭示各身体部位受伤之间的关联性及致因，基于中国深度事故数据（CIDAS）中2 799起二轮车事故，以二轮车骑行者头部、胸部、上下肢等7个主要部位伤害严重程度为因变量，选取44个主要自变量以表征碰撞前行为、碰撞位置等信息。采用带随机参数的多变量模型进行研究，以更好解释各受伤部位严重程度间的关系及数据中未能观测到的异质性。结果表明：二轮车骑行者性别及年龄、车辆属性、碰撞前二轮车驾驶行为及二轮车被撞位置均会对二轮车骑行者不同部位的伤害造成显著影响；头部、胸部及下肢在二轮车事故中是最脆弱的3个部位，值得进一步研究；中年二轮车骑行者在头部伤害模型服从正态分布的随机参数，即中年二轮车骑行者会对二轮车骑行者头部伤害严重程度存在异质性影响；加强汽车及二轮车的安全设计从而保护二轮车骑行者胸部和腹部极有必要；佩戴头盔不仅能够保护头部伤害，对于保护上肢伤害也有重要作用。

关键词: 交通工程；二轮车骑行者；身体部位；CIDAS数据；随机参数模型