Fatal Road Crashes Research Articles

Ensemble-based model to investigate factors influencing road crash fatality for imbalanced data

The rapid growth of urbanization and motorization has significantly increased traffic crashes, leading to both loss of life and diminished quality of life for crash survivors and their families. Identifying the factors influencing crash fatality is crucial for reducing such incidents. However, traffic crashes are inherently unpredictable, and crash fatality datasets are often imbalanced. This study provides a comprehensive evaluation of various machine learning (ML) techniques to analyze traffic crash fatality using an imbalanced dataset. It is the first to train eight distinct binary classification models: Classification and Regression Trees (CART), Random Forest (RF), Gradient Boosting Machine (GBM), Extreme Gradient Boost (XGBoost), Logistic Regression (LR), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Naïve Bayes (NB) under three strategies: in isolation, with bagging, and with optimized bagging techniques (Grid Search CV, Random Search CV, and Bayesian Optimization). To handle data imbalance, eight resampling methods were employed, including SMOTE, Random Under-sampling (RUS), Random Over-sampling (ROS), ADASYN, Tomek Links, Near Miss, SMOTETomek, and SMOTEENN. Results show that GBM, combined with Bayesian optimized bagging and RUS, achieved the best performance with a G-mean score of 65.23 and an F1 score of 60.06. This study offers valuable insights into effective ML techniques, data resampling methods, and advanced optimization strategies for imbalanced crash severity datasets.

Tree-based approaches to understanding factors influencing crash severity across roadway classes: A Thailand case study

Existing studies often overlook the nuanced differences between various road classifications and their respective crash dynamics, hindering the development of targeted interventions to mitigate crash severity. To address this gap, this study investigates factors influencing the likelihood of fatality in road crashes across highways, collector roads, and local roads in Thailand using crash data from 2015 to 2021. Highways connect regions with high-speed traffic and large volumes, collector roads link smaller communities with lower traffic density but allow higher speeds, and local roads primarily pass through villages, with narrow pathways, two traffic lanes, and frequent motorcycle use. The study employs machine learning methodologies utilizing tree-based algorithms, including Decision Trees, Random Forest, Gradient Boosting, AdaBoost, Extra Trees, XGBoost, LightGBM, and CatBoost. The XGBoost model delivered superior performance for highways, while Gradient Boosting slightly outperformed XGBoost for local and collector roads. Both models consistently achieved a test accuracy of 0.70, with precision between 0.66 and 0.67, recall ranging from 0.59 to 0.61, and F1-scores from 0.58 to 0.61. The AUC values also consistently ranged from 0.59 to 0.61. SHAP values reveal key factors influencing fatality risk across road types, including speeding, gender disparities, driving under the influence of alcohol, inadequate lighting, and elderly drivers. Specific concerns include reversing on highways, collisions in poorly lit areas on collector roads, and helmet non-use on local roads. The findings support policy recommendations to address speeding, target male and older drivers, prevent reversing incidents, enhance lighting, and promote helmet use. This research deepens our understanding of factors affecting road crash severity and offers valuable insights for improving road safety across various environments.

Investigating and modeling motorcycle crash rates at urban unsignalized intersections in India

Objective The present study focuses on understanding the behavior of motorized 2-wheeler (MTW) riders at urban unsignalized intersections in India. In the Indian context, over 60% of road crash fatalities are attributed to vulnerable road users, with MTWs serving as the predominant contributors, accounting for 44% of total fatalities. Notably, unsignalized intersections have emerged as critical sites for accidents involving vulnerable road users. Methods Postencroachment time is used to assess traffic conflicts of MTW users. Furthermore, the study employs the exceedance property of extreme value theory to calculate crash probabilities. Tobit and grouped random parameters Tobit regression models are developed to model crash probabilities, incorporating variables such as traffic volume, traffic composition, gap acceptance time, intersection characteristics, and intersection conflict area at 4 urban unsignalized intersections in Surat, India. Results MTW riders have the lowest gap acceptance time among vehicles in the traffic stream. Cars and other heavy vehicles readily accept gaps when MTWs are in the conflicting stream at unsignalized intersections, which increases traffic conflicts. MTWs have the highest crash rates in the traffic stream. Among the developed models, the grouped random parameters Tobit regression captures the spatial unobserved heterogeneity of the study sites and outperforms the simple Tobit regression model. The results also indicate that MTW riders are exposed to a higher risk of crashes while turning at unsignalized intersections. The presence of a central traffic island has varied implications; it raises crash rates at 3-legged intersections but lowers them at 4-legged intersections for 2-wheelers. Conclusion The study concludes that MTW crash rates are influenced by traffic and intersectional factors. Increased gap acceptance time correlates with lower crash rates. Countermeasure selections require detailed investigations, because it was observed that the presence of central traffic islands has varied effects on crash rates at 3-legged and 4-legged unsignalized intersections.

Mega-Events: Assessing Road Safety through an Operating Framework. An Application for the Milano–Cortina 2026 Winter Olympic Games

To meet the United Nations and European Union goals of reducing road crash fatalities and injuries, it is also relevant to address the negative externalities due to mega-events on the road network and the local communities, to assess the safety of the road network involved, and to implement appropriate measures for different road environments. Despite their relevance, the literature often overlooks social costs and risks associated with mega-events. This study presents an operating framework for rapidly assessing the safety of the Milano–Cortina 2026—“Via Olimpica” road—which will host a significant proportion of the traffic during the Winter Olympic Games in 2026. The framework proposes a simplified Road Infrastructure Safety Management (RISM) to address the unique challenges posed by the limited time available for screening and implementation by local authorities. The framework integrates four data sources and follows a seven-step procedure. It provides recommendations for improving road safety by identifying critical road sections and blackspots. Road authorities, practitioners, and public administrations may all benefit from the framework, as it makes it easier to prioritise safety improvements within time constraints.

The value of Association Rule Analysis in understanding serious and fatal road traffic crashes - a case study of the N4 toll road between 2015 and 2019

In spite of the fact that the road traffic crash fatality rate in South Africa is significantly higher than the global average, the characteristics and causes of road traffic crashes are still not well understood. Without a clear understanding of the characteristics and causes, intelligence-led countermeasures to reduce crashes cannot be developed or applied. The analysis of local South African road crash data is of particular importance in identifying the underlying problems responsible for the high crash rates across the country. While descriptive analyses can be used to present a snapshot of the crash problem, they are largely inadequate when it comes to throwing light on the underlying patterns and relationships between the contributing factors and causes of the crashes themselves. This paper presents a study of the use of Association Rule Analysis (ARA) on data related to crashes on the N4 between Pretoria and the Mozambican border, between 2015 and 2019, to see what added value ARA may offer to standard descriptive analyses of crash data. Apriori Association Rule algorithm was employed on the dataset to identify what contributory factors were not initially evident in the occurrence of serious and fatal crashes. ARA revealed that pedestrians, negligent driving, overtaking and single-vehicle crashes are common factors likely to be present in a serious crash. However, this study found that descriptive statistical analysis and ARA methods are not mutually exclusive, but rather complementary. Where descriptive statistical analysis focuses on summarising and describing data, ARA focuses on identifying relationships and patterns within the data.

Promoting Emergency Medical Service Infrastructure Equality to Reduce Road Crash Fatalities

Timely emergency medical service (EMS) is critical to mitigate outcomes of severe traffic crashes, especially in rural areas. The urban–rural inequality of the EMS infrastructure has been one of the factors contributing to a higher death rate for patients in rural road crashes. To address the spatial imbalance of EMS infrastructure, this study aimed to provide a methodological framework for evaluating the existing EMS infrastructure by taking South Korea’s EMS infrastructure expansion plans as the case study. Specifically, this study developed a road crash data-driven methodology to promote spatial balance and economical expansion of EMS infrastructure of all types, including EMS stations, hospitals, the helicopter fleet, and landing spots. Geographically weighted binary logit regression and spatial analysis identified strategic locations for prioritizing the EMS infrastructure expansion using crash victim data and road networks to close the gap between urban and rural areas. The analysis of access to existing EMS infrastructure showed that the 16 to 20 min EMS response time, including on-scene time and transport time ranging from 11 to 15 min, are significantly associated with higher crash fatalities. The results also suggested that EMS stations and heliports are inadequate to meet the EMS time thresholds in the central province of Korea. The findings of this research could inform policymakers as they are working toward expanding the EMS infrastructure and creating a more equitable EMS response when it comes to transporting rural road crash patients.

Investigating the severity of single-vehicle truck crashes under different crash types using mixed logit models

Introduction: Almost 90% of fatal road crashes occur in developing countries. Among these countries, Iran has a noticeable fatal crash rate of 21.47 deaths per 100,000 persons. Improving the safety of trucks is of particular importance in Iran where road freight is used to transport almost 90% of the commodities. Researchers have suggested dichotomizing crashes into single- and multi-vehicle categories and found that when this is performed vast differences can be identified between the mechanisms behind these categories of crashes, particularly when investigating truck crashes. Method: This study investigated single-vehicle truck crashes in Khorasan Razavi province in Iran from 2013 to 2021. Likelihood ratio tests were employed to show that separate models are statistically valid for different crash types. Subsequently, three mixed logit crash-type models were developed to investigate 5,703 single-vehicle truck crashes. Results: Four significant variables were exclusive to collisions with an object (brake failure, ABS, primary roads, and rainy or snowy weather), five significant variables were associated with run-off-road crashes (driving a loaded truck, speed limit (>60 km/h), paved shoulders, driving uphill, and inability to control the truck), and three significant variables were associated with overturn crashes (overloaded truck, curved roads, and changing direction suddenly). In all crash types, both fastening the seatbelt and speeding were found to be significant factors. Conclusion: The research highlights the need to analyze single-vehicle truck crashes using distinct crash type models and highlights the unique contributing factors of three common single-vehicle crash types. Practical applications: The study presents recommendations for policy to address key crash risks for trucks in Iran, including education and training to improve driver experience, compliance with seat belt usage, enforcement of speeding, and vehicle technologies to monitor drivers.

Fatalities in value chains—an attempt to classify road traffic crashes in accordance with the United Nations General Assembly resolution 74/299

Large corporations are today expected or obliged to report on accidental deaths and serious injuries to employed or contracted employed as a part of reporting on sustainability and workplace safety. Data about road crashes are part of such events and are therefore, but not separately, collected and presented. In Europe, 40% to 60% of all work-related accidents resulting in death has been reported to be road traffic accidents. In 2020, the Stockholm Declaration urged all corporations to report on their safety footprint including their entire value chain. The aims of the present study were to use a new definition of safety footprint and to quantify those killed as employed and at work, and those killed in a crash where the other part was at work, as so called third parties, to transports for duty with employed drivers. The Swedish Transport Administration (STA) in-depth database of fatal crashes was used, that covers all fatalities classified as road traffic related and consists of information from the police, medical journals, autopsy reports, accident analyses performed by STA, and witness statements. All fatalities excluding suicides or those caused by sickness occurring during year 2019 were investigated (n = 214). 11% (23/214) of the fatalities occurred when the killed person was at work and 16 while commuting. 37% of the fatal accidents occurred when the killed road user or the other part was at work. In total, almost half of the fatalities in the road transport system were related to work in some way when including both the fatally injured and their collision partners. A larger proportion of non-privately owned and procured vehicles was found for the vehicles of the collision partners compared to the vehicles of the fatally injured. In approximately one third of the fatal accidents a procurement of a transport service was involved. The Swedish Work Environment Authority (SWEA) identified 10 of the 23 fatalities at work investigated and none of these accidents was found to be investigated by the police as a crime related to the work environment. In conclusion, almost half of the fatalities in the road transport system in 2019 were related to work in some way, either the fatally injured or their collision partners were at work or while commuting. When including the third-party casualties, the problem becomes much bigger and more complex. In Sweden fatalities related to work are underreported, as the SWEA does not receive basic data. Efforts are needed to improve reporting of work-related road fatalities. It was found that the police did not investigate road traffic fatalities as death at workplace. It is crucial that the police start to follow the intention of regulations linked to workplace safety. If not, the possibility to collect relevant data for organizations to report on their safety footprint is limited. It is complicated to collect, classify and analyse value chain fatal crash information, mainly due to that the police do not investigate fatal road crashes as possibly work-related events. It is recommended that organizations manage their own data collection if they wish to report on their safety footprint data.

Identification and Analysis of fatal Road Crash Black spot Clusters in an Urban Setting in South Coastal India

ABSTRACT Introduction: Deaths from road crashes form the leading cause of mortality in India. Streamlining road crash data systems are essential for building robust prevention strategies. This study explores objectivisation of fatal road crash data by spatiotemporal analysis (geographical information system [GIS] technology) in an urban setting in South India. Aim: To identify clusters of fatal road crash black spot clusters in an urban setting and to analyze crash-related variables in clusters. Settings and Design: Secondary data analysis of fatal road crashes in Puducherry. Methods: Fatal road crash data from 2016 to 2018 were collected from South Traffic Police records. Spatiotemporal analysis was done using GIS to map high-density locations (black spots); these were further grouped into clusters. Crash-related variables in each cluster were studied to identify profiles of crash victims, alleged offenders, and risk factors. Results: Raw data accessed in descriptive format were converted to analyzable objective format using a self-developed data extraction template. A total of 154 fatal road crashes occurred in Puducherry South during the study period. Total 11 black spots and 3 clusters were mapped. One particular stretch of National Highway witnessed maximum (59%) black spots. Clusters differed from each other for variables such as age (of both victims and causing persons), time of the crash, and causing vehicle type. Intercluster similarities were observed in the preponderance of males (82.3%), youth (mean age: 28.9 years), vulnerable road users (92.6%), rainy season (43.4%), and weekends (46.7%) witnessing most fatal crashes. Conclusion: Standardized, objective format for data capturing, and seamless mechanism for data processing are essential. The crash location is a key index variable for data systems, with the feasibility to superimpose other data layers.

Differences in road safety ‘logics’ between high income countries (HICs) and low income countries (LMICs)

Existing models and frameworks in relation to road safety that had been developed in High Income Countries (HICs) may not be adequate for Low and Middle Income Countries (LMICs). Associations between various indicators were examined, providing further evidence of different ‘logics’ in relation to road safety in HICs and LMICs. Analyses were undertaken by comparing indicators from different sources, such as the WHO and ESRA. From these analyses it has appeared that relationships which hold in HICs do not necessarily hold in LMICs.Examples are given in relation to factors affecting the road crash fatality rate, risky behaviour in traffic, and support for road safety policy measures. Given the importance of the road safety challenge in LMICs, more research is needed to develop models and frameworks that are more appropriate and fit better the road safety context in LMICs.

Let there be light: Daylight saving time and road traffic collisions

AbstractThis paper examines the impact of daylight saving time (DST) on fatal road crashes in Australia. I exploit within year transitions to and from DST in a regression discontinuity in time framework. To examine the long run effect of the policy, I use trials of DST implemented in various states, and a DST extension due to the Sydney Olympics. Neither the transition to or out of DST, nor the long run effects of DST have an impact on fatalities on the road. However, there is evidence of reallocation of accidents over the day due to ambient light.

Comparing Female and Male Pedestrians’ Relative Vulnerability Based on Crash Locations

While pedestrian fatalities represent 23% of all road crash fatalities in Iran, fatalities seem more common amongst female pedestrians than male pedestrians. This study investigated female pedestrian fatalities and contributing factors such as age, education, and visibility. Using Iran’s 2014 death registry data, a relative vulnerability ratio was estimated for different groups stratified by crash locations. Factors that increase the risk of fatal injuries for female pedestrians included: being under 15 years old or 41-64 years old; lacking formal education; being single; not receiving trauma care in time due to resting at home or long transfer time to a hospital; not wearing contrasting clothing in urban areas; and being involved in a crash on rural roads or at dusk/dawn on intercity roads.

Factors propelling fatalities during road crashes: A detailed investigation and modelling of historical crash data with field studies

One of the major concerns in developing countries like India is to maintain traffic safety under mixed and heterogenous scenario. Although zero accidents is the need of the hour, the first step to attain it is ensuring zero deaths and no serious long-term disabling injuries in road crashes. To reduce the road crash fatalities, explicit and detailed studies have been conducted by utilising historical road crash data of two emerging smart cities of India - Bhubaneswar and Visakhapatnam. Traffic flow data and characteristics of road infrastructure has also been collected by performing field studies at accident prone locations. Various factors including vehicular characteristics, road user characteristics, and road infrastructure have been analyzed using various non-parametric tests to identify the contributing factors resulting in fatalities. It is observed that out of 14 variables used for study, 8 factors were significantly related to fatal crashes. These included categories of victim and accused, 85th percentile speed, presence of road markings, availability of sight distance, etc. The significant factors were subjected to binary logistic regression to determine the odd’s ratio of significant factors. The logistic regression predicted 79% of deaths correctly. Crash fatality prediction models are developed using both Classification and Regression Tree (CART) classification tree with 83% accuracy. Although CART classification led to higher accuracy, binary logistic regression is more robust as it considered more significant factors as compared to CART. Subsequently, a severity index has been proposed based on proportions of actual fatal crashes and usage of K-means clustering technique. The proposed indices shall be really helpful in traffic safety management, specifically in reduction of fatalities during road crashes.

Profile of low and middle-income countries with increases versus decreases in road crash fatality population rates and necessity of motorcycle safety

Introduction: Road crash fatalities have increased significantly in Low- and Middle- Income Countries (LMICs) between 2006 and 2016. This study presents how road safety characteristics have changed in LMICs by comparing data over time and relationships between the road crash fatality increase and a wide range of data from LMICs. Parametric and nonparametric methods are used to test significance. Method: There were 35 countries in the Latin America and Caribbean region, the Sub-Saharan Africa region, the East Asia and Pacific region, and the South Asia region, where the population rate of road crash fatalities consistently increased as per country reports, World Health Organization and Global Burden of Disease estimates. In these countries, the proportion of fatalities involving motorcycles (including powered two or three-wheelers) substantially increased (44%) over the same time (statistically significant). In these countries, the helmet-wearing rate was only 46% for all passengers. These patterns were not observed in LMICs with decreasing population fatality rates. Results: Motorcycle helmet usage rates strongly correlate with decreasing fatalities per 10,000 motorcycles in Low-Income Countries (LICs) and LMICs. Effective interventions (including increasing helmet usage) are urgently needed for motorcycle crash trauma in LMICs, especially where the economy and motorization rapidly grow. National strategies for motorcycle safety, conforming to the Safe System principles, are recommended. Conclusions: For evidence-based policy formulation, there is a need to continue strengthening data collection, sharing, and use.

An analysis of influential factors associated with rural crashes in a developing country: A case study of Iran

Road traffic deaths continue to rise, reaching 1.35 million in recent years. Road traffic injuries are the eighth leading cause of death for people of all ages. Note that there is a wide difference in the crash rate between developed and developing countries and that developed countries report much lower crash rates than developing and underdeveloped countries. World Health Organization reports that over 80% of fatal road crashes occur in developing countries, while developed countries account for about 7% of the total. The rate of road crashes in developing countries is higher than the global average, despite some measures reducing deaths over the last decade. Numerous studies have been carried out on the safety of urban roads. However, comprehensive research evaluating influential factors associated with rural crashes in developing countries is still neglected. Therefore, it is crucial to understand how factors influence the severi-ty of rural road crashes. In the present study, rural roads in Mazandaran province were considered a case study. The Crash data collected from the Iranian Legal Medicine Organization covers 2018 to 2021, including 2047 rural crash-es. Dependent variables were classified as damage crashes and injury-fatal crashes. Besides, independent variables such as driver specifications, crash specifications, environment specifications, traffic specifications, and geometrical road specifications were considered parameters. The logit model data indicate that factors associated with driver and crash specifications influence rural crashes. The type of crashes is the most critical factor influencing the severity of crashes, on which the fatal rate depends. The findings suggested that implementing solutions that minimize the effect of the factors associated with injury and death on rural roads can reduce the severity of crashes on rural roads that share the same safety issues as the case study. Further studies can also be conducted on the safety and mechanics of the vehicle by focusing the research on the types of vehicles and the sources of the damage.

Estimation of Willingness to Pay and Value of Statistical Life for Road Crash Fatality Reduction for Motorcyclists: A Case Study of Patna, India

Estimation of Willingness to Pay and Value of Statistical Life for Road Crash Fatality Reduction for Motorcyclists: A Case Study of Patna, India

Alcohol Price Floors and Externalities: The Case of Fatal Road Crashes

AbstractIn May 2018, Scotland introduced a minimum unit price on alcohol. We examine the impact of this policy on traffic fatalities and drunk driving accidents. Using administrative data on the universe of vehicle collisions in Britain and a range of quasi‐experimental modeling approaches, we do not find that the policy had an effect on road crash deaths and drunk driving collisions. The results are robust to several sensitivity exercises. There is no evidence of effect heterogeneity by income and other predictors of alcohol consumption or cross‐border effects. A brief discussion of the policy implications of our findings is provided.

Self-regulation and tailgating behaviours among young drivers

Young drivers aged 17 to 25 years are overrepresented in road crash fatalities globally. In Queensland, Australia, tailgating is an illegal behaviour that has been attributed to at least 8,000 rear-end crashes per year. Notably, it has been found that young drivers tailgate more frequently than older drivers. This study utilised self-determination theory to understand self-regulation behind tailgating among young drivers. This theory has yet to be applied to understand tailgating but has been successful in understanding other risky driving behaviours, suggesting self-determination theory may have predictive utility in explaining tailgating. Participants (n = 609) aged 17 to 25 years (Mage = 20.61 years) completed an anonymous online survey measuring engagement in tailgating, external and internal regulation, trait self-regulation, and aggressive driving. Results indicated that young drivers less influenced by legal and non-legal sanctions are likely to engage in tailgating more frequently (lower external regulation). Young drivers influenced by safe driving values were less likely to engage in tailgating (greater internal regulation). Poorer trait self-regulation and higher measures of aggressive driving were predictors of more frequent tailgating in young drivers. The results support the development of informative campaigns centred on the self-determination theory with the aim of educating young drivers on the implications of tailgating and safe following distances.

Development of crash prediction models by assessing the role of perpetrators and victims: a comparison of ANN & logistic model using historical crash data

Road traffic injuries cost countries 3% of their annual GDP. In developing countries like India, every year around 150,000 people die on roads. The type of vehicles involved in a crash contribute majorly to the outcome of casualty (injury/death). Barring few studies, literature are less regarding the role of vehicle as perpetrator and victim on road crash fatalities. Historical crash data has been used in the present study to examine the role of vehicles (both as perpetrator & victim). The study reveals that victim’s effect is more as compared to perpetrator/accused for determining the outcome of crash. Heavy vehicles as perpetrator, and self-hitting vehicles along with pedestrians as victims have higher fatality rates. Binary logistic regression and artificial neural network (ANN) has been utilized for developing prediction models. Binary logistic model predicted around 75% of outcomes correctly with default cut-off value (0.5). However, based on reported crash data, where 19% of total crashes lead to deaths, 0.19 has been proposed as cut-off value which increases the accuracy of the predictions. Accuracy of ANN technique directly depends on the number of crashes reported for a definite pair of perpetrator and victim and the type of validation technique used (Holdback/K-Fold) along with the type of hidden layer chosen for the study based on different types of sigmoid activation function. ROC curves in ANN suggest that the analysis can predict 75% of the outcomes which can be increased by deleting the pairs of vehicles which are present/have occurred in very less number. A comparison has been made between the two techniques based on their advantages and limitations. The developed models can be used as safety indicators based on composition of traffic flow on urban roads.

A registry-based investigation of road traffic fatality risk factors using police data: A case study of Hyderabad, India

This study presents a comprehensive methodological approach for the identification and segmentation of the key risk factors associated with fatal road crashes. Hyderabad, an Indian metropolis with significant annual fatal crashes, is selected as the case study city. Data containing the date, time, and location of the crash, number of injuries and fatalities, accused and victim vehicle details are collected from Hyderabad Traffic Police, and a comprehensive registry-based crash database is developed. Based on the database, a Cross-sectional study is conducted, and risk ratio (RR) is used as a measure to test the association between the risk factors and fatal outcomes. Logistic regression, log-binomial regression, and robust Poisson regression models were also used to understand the association of fatal crash outcomes with different attributes. RR and associated confidence intervals are further used to classify the factors into three groups: significant, insignificant, and non-risk factors. Subsequently, the Apriori algorithm is used to determine the interrelationship/association between the risk factors leading to a fatal crash outcome. Using the Apriori algorithm, a set of association rules involving three factors leading to a significant number of fatal crashes are identified. Finally, the derived results are combined to segment the factors associated with fatal crashes into six specific segments: Very high, High, Moderate, Low, Very low, and Extremely Low-risk factors. Such identification and segmentation of potential risk factors associated with fatal crashes would help the planning agencies to formulate mitigation measures for a low and medium-income country (LMIC) like India.