Phase-wise injury integrated severity modeling of road accidents: a two-stage hybrid multi-criteria decision-making model

The road accident severity analysis of any geographic area is usually based on the total number of accidents, while the significance of minor and grievous injuries is often ignored. Furthermore, most research works on road accident severity applies statistical models and machine-learning techniques with complex interdependencies among the input variables, while very limited multi-criteria decision-making (MCDM) models have been developed to analyze the injury severity of road accidents. This study evaluates the road accident severity of 30 Indian cities having a million-plus population with the introduction of a two-stage MCDM model: the analytical hierarchy process (AHP) and the multi-objective optimization on the basis of ratio analysis (MULTIMOORA). The model uses the annual road accident data provided by the Indian Ministry of Road Transport and Highways. The proposed model identifies the weights of every injury type based on experts' inputs and concludes in the injury integrated severity (IIS) values-based robust rankings of the selected Indian cities. Moreover, the phase-wise injury integrated severity (PIIS) analysis divides all 30 Indian cities into three phases according to the IIS values. The model estimates that Delhi, Chennai, Jaipur, and Bengaluru are the cities with the highest IIS values and got placed within Severity phase I. Moreover, the phase-wise clustered result is uniquely represented with the Pareto distribution to interpret the cumulative severity. The results demonstrate that the proposed model provides a decision-making tool for road safety stakeholders and has the potential to support the severity-based prioritization of road safety intervention programs.