Modelling developmental trajectories of novice drivers with high and low crash liability

道路交通事故是对公共健康的最大威胁之一，特别是对青年人而言，给社会带来巨大的社会和经济代价。新手司机在伤亡统计中的比例过高，在他们驾驶生涯的前六个月，风险特别高。有证据表明，驾驶技能的不足和采取危险驾驶方式的决定（例如，超速行驶和在危险的地方超车）会导致新手的脆弱性。在所有驾驶员中，平均碰撞率在随后的18-24个月内迅速下降，但直到驾驶员达到约30岁才达到平稳状态。虽然新手驾驶员道路交通伤害的发生率有据可查，但对于为什么在驾驶的头三个月内撞车事故责任特别高，然后随着时间的推移稳步下降，仍然了解得相对较少。我们认为，不同类型的新手驾驶员将遵循不同的碰撞风险和行为发展轨迹，而不是所有驾驶员都遵循相同的发展路径。例如，我们对拟议数据集的初步分析表明，在通过驾驶考试时，碰撞涉及的风险态度不同。在风险态度量表中得分最高的三分之一的参与者显示，最初的碰撞参与率很高，但在前6个月后急剧下降。风险态度较低的参与者的初始撞车风险较低，他们的风险没有显示出类似的下降模式。我们现有的资源意味着我们的初步分析只能使用基本的技术。整个项目将采用先进的统计建模技术，以确定新手驾驶员的亚组，这些亚组遵循碰撞参与和驾驶员行为的不同发展轨迹。然后，我们将探讨在何种程度上的轨迹是预测的可修改的风险因素，如驾驶行为和经验，在学习驾驶。这些结果将指导教育、评估和管理政策（例如，分级发牌），以鼓励驾驶者进入较低风险的轨道，而不是较高的风险。

Road traffic crashes are one of the greatest threats to public health, particularly for young people, with a substantial social and economic cost to society. Novice drivers are overrepresented in casualty statistics, with risk being particularly high in the first six months of their driving careers. There is evidence that both deficiencies in driving skills and decisions to adopt risky driving styles (e.g., speeding and overtaking in dangerous places) contribute to novice vulnerability. Averaged over all drivers, the rate of collisions declines rapidly over the subsequent 18-24 months but does not plateau until drivers reach approximately 30 years of age. While the prevalence of road traffic injuries for novice drivers is well documented, there is still relatively little understanding of why crash liability is particularly high in the first three months of driving before decreasing steadily over time. Rather than all drivers following the same developmental pathways, we believe different types of novice driver will follow different trajectories of crash risk and behavioural development. For example, our preliminary analyses of the proposed dataset indicate that crash involvement differs on the basis of risky attitudes at the time of passing the driving test. Participants scoring in the top third of a risky attitudes scale showed an initially high rate of crash involvement that declined steeply after the first 6 months. The participants with less risky attitudes were at lower initial risk of crash and their risk did not show a similar pattern of decline. Our available resources meant our preliminary analysis could only use basic techniques. The full project will employ advanced statistical modelling techniques to identify sub-groups of novice drivers following different developmental trajectories of crash involvement and driver behaviour. We will then explore the extent to which trajectory is predicted by modifiable risk factors such as driving behaviour and experiences in learning to drive. These results will guide policy on education, assessment and management (e.g., graduated licensing) of learner and novice drivers to encourage drivers onto the lower risk trajectories rather than the higher risk.