A causal inference analysis of injury severity in motorciclist crashes

Abstract

The proportion of motorcyclist fatalities has increased in Latin and Asian countries in recent years. The Second Decade of Action for Road Safety 2021-2030 has shown that the systematic view of the Safe System (SS) is the means to achieve zero fatalities. Traditional studies often use a data-driven approach and a single regression, which may not adequately represent this view. The main aim of this dissertation is to develop a causal analysis for motorcyclists from the Safe System perspective using observational data. To consolidate the knowledge of causal inference approaches in road safety studies of the motorcyclists’ severity, Monte Carlo Simulations were elaborated to understand the characteristics of this new approach. Subsequently, an extensive literature review was carried out to formulate a conceptual model based on the Safe System approach for the causal process of motorcyclists' injury. Causal hypotheses were formulated based on the conceptual model and the data collected from motorcyclist crash data on Brazilian highways. The causal model was estimated to evaluate the relationship between alcohol use and the severity of road crashes on federal highways in the state of Ceará, using the SEM methodology. Finally, the results indicated a significant relationship between alcohol use and severity in urban roads. Other findings suggested that weekends and nighttime hours are associated with alcohol use, while heavy vehicles, head-on collisions, and off-peak hours are directly associated with higher severities due to increased crash energy. Furthermore, differences were found in the relationships in rural and urban areas. The use of a causal approach allowed for obtaining more reliable results by controlling confounding variables and incorporating a theoretical framework that includes the Safe System approach. Moreover, it enabled obtaining, in a single model, the interrelationships among the study variables, leading to a better understanding of road safety. Finally, the obtained results can assist decision-makers in developing an action plan capable of achieving a safe system for motorcyclists.