Causal inference of effect of motor-vehicle driver's avoidance action on cyclists' injury

doi:10.16265/j.cnki.issn1003-3033.2025.06.0530

Abstract

Abstract:

To verify the impact of drivers' evasive actions on injuries of two-wheeler riders and to identify the conditions under which evasive maneuvers fail or even exacerbate the injuries, based on data from China In-Depth Accident Study (CIDAS), NSGA-II was first used to simulate and optimize the reconstruction of accident scenes. Vehicle speed and two-wheeler speed data at the moment of collision were extracted. A research dataset consisting of independent variables, dependent variables, and 11 covariates was constructed. Simulation-derived variables and investigation-based data fields were integrated to serve as the foundational data for modeling. Secondly, two causal inference methods were adopted, namely the propensity score (PS) weighting - regression analysis combination method considering positive hypotheses and covariate adjustment (inverse probability weighting (IPW) and overlap probability weighting (OW)), to infer the causality between evasive actions and injury severity, and to compare the inter-group balance after processing by IPW, OW and unweighted regression methods. Finally, the causal effect of drivers' evasive actions on the injury severity of two-wheeler riders under different conditions was quantitatively analyzed. The results show that, in general, the evasive actions currently adopted by drivers cannot effectively reduce the injury severity of riders. When the vehicle types are commercial vehicles and the motor vehicle traveling at medium to high speeds, taking evasive driving actions tends to aggravates the injuries of riders. Among these actions, steering maneuvers is more likely to increase the severity of rider injuries.

Key words: hedging action, two-wheeler, riders, injury degree, causal inference

CLC Number:

X910安全人体学

ZHU Tong, LI Wei, ZHAO Yunfei, LI Xiaohu, WANG Peng. Causal inference of effect of motor-vehicle driver's avoidance action on cyclists' injury[J]. China Safety Science Journal, 2025, 35(6): 27-36.

Figures/Tables 7

Table 1

Fig.1

Fig.2

Fig.3

Fig.4

Table 2

Table 3

References 25

[1]	国家统计局. 中国统计年鉴[M]. 北京: 中国统计出版社, 2022:771.
[2]	HUANG Yuan, ZHOU Qing, KOELPER C, et al. Are riders of electric two-wheelers safer than bicyclists in collisions with motor vehicles?[J]. Accident Analysis & Prevention, 2020, 134: DOI: 10.1016/j.aap.2019.105336.
[3]	LUBBE N, JEPPSSON H, STERNLUND S, et al. Injury risk curves to guide safe speed limits on Swedish roads using german crash data supplemented with estimated non-injury crashes[J]. Accident Analysis & Prevention, 2024, 202: DOI: 10.1016/j.aap.2024.107586.
[4]	殷豪, 林淼, 王鹏, 等. 基于XGBoost的二轮车碾压事故致因研究[J]. 安全与环境工程, 2023, 30(5): 19-27,45.
	YIN Hao, LIN Miao, WANG Peng, et al. Study on the cause of two-wheeler run-over accidents based on XGBoost[J]. Safety and Environmental Engineering, 2023, 30(5): 19-27,45.
[5]	王仲旭, 陈泽平, 鲁誉, 等. 两轮车事故影响因素分析及损伤机理研究[J]. 汽车技术, 2022(10): 56-62.
	WANG Zhongxu, CHEN Zeping, LU Yu, et al. Research on factors affecting two-wheeler accidents and injury mechanisms[J]. Automobile Technology, 2022(10): 56-62.
[6]	周华, 张巧玉, 母尧尧, 等. 十字路口乘用车与二轮车典型危险场景聚类及推演[J]. 中国安全科学学报, 2020, 30(4): 100-107. doi: 10.16265/j.cnki.issn1003-3033.2020.04.016
	ZHOU Hua, ZHANG Qiaoyu, MU Yaoyao, et al. Clustering and deduction of typical dangerous scenarios between passenger vehicles and two-wheelers at crossroads[J]. China Safety Science Journal, 2020, 30(4): 100-107. doi: 10.16265/j.cnki.issn1003-3033.2020.04.016
[7]	周华, 刘峰, 李平飞, 等. 电动二轮车侧倾角对骑车人头部损伤影响研究[J]. 中国安全科学学报, 2019, 29(5): 79-84. doi: 10.16265/j.cnki.issn1003-3033.2019.05.014
	ZHOU Hua, LIU Feng, LI Pingfei, et al. Impact of electric bicycles' roll angle on cyclist's head injury[J]. China Safety Science Journal, 2019, 29(5): 79-84. doi: 10.16265/j.cnki.issn1003-3033.2019.05.014
[8]	REN Qiaoqiao, XU Min. Heterogeneity in crash patterns of autonomous vehicles: the latent class analysis coupled with multinomial logit model[J]. Accident Analysis & Prevention, 2025, 209: DOI: 10.1016/j.aap.2024.107827.
[9]	MILTON J C, SHANKAR V N, MANNERING F L. Highway accident severities and the mixed Logit model: an exploratory empirical analysis[J]. Accident Analysis & Prevention, 2008, 40(1): 260-266.
[10]	WANG Chenzhu, ABDEL-ATY M, HAN Lei. Effects of speed difference on injury severity of freeway rear-end crashes: insights from correlated joint random parameters bivariate probit models and temporal instability[J]. Analytic Methods in Accident Research, 2024, 42: DOI: 10.1016/j.amar.2024.100320.
[11]	ZHANG Pengxiang, MAHMUD A, VIKASH V, et al. Comparison of safety effect estimates from propensity scores-potential outcomes framework and empirical Bayes before-after method: case study of adaptive traffic signal control[J]. Journal of Safety Research, 2024, 91: 258-270. doi: 10.1016/j.jsr.2024.09.011 pmid: 39998528
[12]	董傲然. 面向骑行者保护的电动两轮车事故重建优化与损伤仿真研究[D]. 西安: 长安大学, 2023.
	DONG Aoran. Research on accident reconstruction optimization and damage simulation of electric two-wheeled for cyclist protection[D]. Xi'an: Chang'an University, 2023.
[13]	李明, 唐姣, 宋欣筑, 等. 基于响应面法的抑尘剂组分配比优选模型[J]. 中国安全科学学报, 2023, 33(1): 88-94. doi: 10.16265/j.cnki.issn1003-3033.2023.01.0175
	LI Ming, TANG Jiao, SONG Xinzhu, et al. Optimization model of dust suppressant component ratio based on response surface methodology[J]. China Safety Science Journal, 2023, 33(1): 88-94. doi: 10.16265/j.cnki.issn1003-3033.2023.01.0175
[14]	CRAMER H. Mathematical methods of statistics[M]. Princeton: Princeton University Press, 1999:286.
[15]	ROSENBAUM P R, RUBIN D B. The central role of the propensity score in observational studies for causal effects[J]. Biometrika, 1983, 70(1): 41-55.
[16]	LI Yan, LI Liang. Propensity score analysis methods with balancing constraints: a monte carlo study[J]. Statistical Methods in Medical Research, 2021, 30(4): 1119-1142. doi: 10.1177/0962280220983512 pmid: 33525962
[17]	涂博祥, 秦婴逸, 吴骋, 等. 倾向性评分加权方法介绍及R软件实现[J]. 中国循证医学杂志, 2022, 22(3): 365-372.
	TU Boxiang, QIN Yingyi, WU Cheng, et al. Introduction of propensity score weighting methods and implementation by R software[J]. Chinese Journal of Evidence-Based Medicine, 2022, 22(3): 365-372.
[18]	韩贺东. 基于稳定权重目标最大似然估计模型的研究及其在真实世界研究中的应用[D]. 上海: 中国人民解放军海军军医大学, 2020.
	HAN Hedong. Modeling of stabilized weight based targeted maximum likelihood estimation and its application in real-world studies[D]. Shanghai: Naval Medical University, 2020.
[19]	MOLINA J, SUED M, VALDORA M. Models for the propensity score that contemplate the positivity assumption and their application to missing data and causality[J]. Stats in Medicine, 2017, 37(24): 3503-3518.
[20]	高绒绒, 高倩, 王彤. 倾向性评分逆概率加权法中极端权重的处理方法[J]. 中国卫生统计, 2023, 40(6): 956-960.
	GAO Rongrong, GAO Qian, WANG Tong. Treatment of extreme weight in inverse probability weighted method of propensity score[J]. China Health Statistics, 2023, 40(6): 956-960.
[21]	LI Fan, MORGAN K L, ZASLAVSKY A M. Balancing covariates via propensity score weighting[J]. Journal of the American Statistical Association, 2018, 113(521): 390-400.
[22]	LI Fan, THOMAS L E. Addressing extreme propensity scores via the overlap weights[J]. American Journal of Epidemiology, 2019, 188(1): 250-257. doi: 10.1093/aje/kwy201 pmid: 30189042
[23]	于淼, 张生奎, 王利华, 等. 肝酶对钢铁工人代谢健康型肥胖与高尿酸血症关联影响[J]. 中国安全科学学报, 2024, 34(2): 231-238. doi: 10.16265/j.cnki.issn1003-3033.2024.02.1184
	YU Miao, ZHANG Shengkui, WANG Lihua, et al. Effect of liver enzymes on association between metabolically healthy obesity and hyperuricemia in steelworkers[J]. China Safety Science Journal, 2024, 34(2): 231-238. doi: 10.16265/j.cnki.issn1003-3033.2024.02.1184
[24]	朱彤, 王驰恒, 李青, 等. 基于集成学习的驾驶人避险安全效果影响因素研究[J]. 安全与环境学报, 2022, 22(6): 3243-3251.
	ZHU Tong, WANG Chiheng, LI Qing, et al. Study on factors influencing the avoidance effect of drivers based on integrated learning[J]. Journal of Safety and Environment, 2022, 22(6): 3243-3251.
[25]	SUNBIN Y, KUMAGAI J, MORITA T, et al. Who to sacrifice? modeling the driver's dilemma[J]. Transportation Research Part A: Policy and Practice, 2023, 178: DOI:10.1016/j.tra.2023.103872.

变量		占比/%
因变量	损伤严重程度:无伤或者轻伤、重伤或致命伤	72.5、27.5
自变量	避险驾驶动作:无避险动作、刹车和转向、仅刹车、仅转向	64.3、16.22、16.29、3.19
协变量	机动车驾驶人性别:男、女	86.8、13.2
	骑行者性别:男、女	72.1、27.8
	机动车驾驶人年龄:≤30岁、31~40岁、41~50岁、51~60岁、>60岁	33.7、33.5、23.0、8.1、1.6
	骑行者年龄:≤30岁、31~40岁、41~50岁、51~60岁、>60岁	20.2、15.5、27.6、21.6、15.1
	机动车驾驶人驾龄:0~1年、2~5年、6~10年、>10年	7.9、28.5、26.2、37.3
	骑行者佩戴头盔:是、否	18.9、81.1
	机动车车辆类型:商用车、乘用车	83.0、17.0
	两轮车车辆类型:电动自行车、自行车、摩托车	48.6、8.3、43.1
	机动车车速:≤10、11~20、21~30、31~40、41~50、51~60、>60 km/h	10.5、20.7、23.3、17.2、11.8、7.8、8.6
	两轮车车速:≤10、11~20、21~30、31~40、>40 km/h	10.8、25.9、35.3、17.1、10.9
	事故发生地点:市区、非市区	51.1、48.9
	路面干燥程度:干燥、潮湿	88.0、12.0
	道路类型:平直路段、弯曲路段、交叉口、其他(停车场等)	23.1、4.7、68.9、3.3
	天气条件:晴天、阴天、雨雪大雾天气	75.1、21.7、3.2
	路面材料:沥青、水泥、砂石	68.5、30.5、1.0
	交通密度:自由流、稳定流、不稳定流	73.6、24.3、2.1

分组	加权方法	避险动作	OR	差异占比/%	P值	ASMD 平均值	ASMD 最大值
总体	未加权	刹车和转向^*	2.355	—	< 0.01^*	0.11	0.46
		仅刹车	0.979	—	0.862
		仅转向^*	1.999	—	<0.01^*
总体	IPW	刹车和转向	1.053	-55.29	0.070	0.01	0.03
		仅刹车	0.993	1.43	0.758
		仅转向^*	1.134	-43.27	0.027^*
总体	OW	刹车和转向	1.046	-55.58	0.150	0	0
		仅刹车	0.998	1.94	0.940
		仅转向^*	1.128	-43.57	0.020^*

分组		避险动作	OR	差异占比/%	P值
机动车车辆类型	乘用车	刹车和转向	1.026	-55.795	0.479
		仅刹车	0.980	-1.110	0.586
		仅转向	1.061	-43.892	0.350
	商用车	刹车和转向^*	2.124	-9.809	< 0.01^*
		仅刹车	1.341	44.504	0.077
		仅转向^*	2.054	-79.166	<0.01^*
道路类型	交叉口	刹车和转向	1.135	-54.709	0.070
		仅刹车	1.118	-22.199	0.124
		仅转向	1.177	-27.301	0.068
	非交叉口	刹车和转向	1.019	-55.463	0.625
		仅刹车	0.948	18.057	0.171
		仅转向^*	1.140	-48.323	0.028^*
机动车车速/ (km·h^-1)	0~30	刹车和转向	1.028	-18.927	0.626
		仅刹车	0.975	24.046	0.581
		仅转向	1.100	-34.988	0.250
	31~60	刹车和转向^*	1.220	-38.693	< 0.01^*
		仅刹车	1.090	34.071	0.110
		仅转向^*	1.206	-28.129	< 0.01^*
	>60	刹车和转向	1.054	4.460	0.661
		仅刹车	1.247	2.381	0.095
		仅转向	1.123	11.741	0.448