A Bayesian network model for analysis of causes and consequences of accidents

Background: Accidents are chains of events that lead to identifiable injuries and illnesses. Among various traumas, traffic accidents have the highest mortality rate. Objectives: The aim of this study was to examine history data for analysis of the causes and consequences of accidents in Kingdom of Saudi Arabia during 2016-2020. Methods: To collect the necessary data, the researchers utilized the Saudi open data portal, named as the National e-Government Portal. The data on consequence, type, seasons, location, and gender were extracted from the database. To analyze the collected data, GeNIe academic software was employed to conduct Bayesian network analysis. Results: In total, 106513 accidents occurred in the Kingdom of Saudi Arabia during 2016-2020. Among them, 40287 and 9492 cases had the consequences of injuries and deaths, respectively. Regarding a sensitivity analysis of accident types, the worst variations in accident consequences were related to accident types of collapse, drowning, and car accidents. Therefore, those could increase the probability of death consequence by 5%, 4%, and 3%, respectively. Regarding a sensitivity analysis of location, for west and east regions, the probability of death consequences decreased by 1%. Moreover, regarding a sensitivity analysis of the season, the probability of deaths decreased by 1% in the autumn season. Regarding a sensitivity analysis of gender, females could decrease the probability of deaths by 4%. Other factors could not make variation in the probability of deaths.  Conclusions: These findings show most important accident types associated with death consequences are collapse, drowning, and car accidents. The locations of the west and east of the kingdom of Saudi Arabia, the season of autumn, and the gender of females could also decrease the death consequences.