Analysis of the severity of occupational injuries in the mining industry using a Bayesian network / 한국역학회지

ABSTRACT

OBJECTIVES: Occupational injuries are known to be the main adverse outcome of occupational accidents. The purpose of the current study was to identify control strategies to reduce the severity of occupational injuries in the mining industry using Bayesian network (BN) analysis. METHODS: The BN structure was created using a focus group technique. Data on 425 mining accidents was collected, and the required information was extracted. The expectation-maximization algorithm was used to estimate the conditional probability tables. Belief updating was used to determine which factors had the greatest effect on severity of accidents. RESULTS: Based on sensitivity analyses of the BN, training, type of accident, and activity type of workers were the most important factors influencing the severity of accidents. Of individual factors, workers’ experience had the strongest influence on the severity of accidents. CONCLUSIONS: Among the examined factors, safety training was the most important factor influencing the severity of accidents. Organizations may be able to reduce the severity of occupational injuries by holding safety training courses prepared based on the activity type of workers.