Finding causation in occupational fatalities: A latent class analysis

Latent class analysis of occupational accidents patterns among Iranian industry workers

Occupational accidents (OA) are among the main causes of disabilities and death in developing and developed countries. The aims of this study were to identify the subgroups of OA and assess the independent role of demographic characteristics on the membership of participants in each latent class. This cross-sectional study was performed on 290 workers between 2011 and 2017. Data gathering was done using the reports of accidents recorded in filed lawsuits. Descriptive statistical analysis was done using SPSS 16 and LCA was done using PROC LCA in SAS9.2. For latent classes were identified; namely "critical due to distractions and lack of supervision" (40.1%), "critical due to lack of safety knowledge" (27.9%), "critical due to fatigue and lack of supervision" (13.1%), and "catastrophic" (18.8%). After adjusting for other studied covariates, being illiterate significantly increased the odds of membership in "critical due to fatigue and lack of supervision" (OR = 4.05) and "catastrophic" (OR = 18.99) classes compared to "critical due to distractions and lack of supervision" class. Results of this study showed that the majority of workers fell under the latent class of critical due to distractions and lack of supervision. In addition, it should be noted that although a relatively small percentage of the workers are in the catastrophic class, the probability of occurring death is quite high in this class. Focusing on the education of workers and enhancing manager's supervision and employing educated workers could help in reducing severe and catastrophic OA.

Comparison of machine learning methods in predicting binary and multi-class occupational accident severity

Although Machine Learning (ML) is widely used to examine hidden patterns in complex databases and learn from them to predict future events in many fields, utilization of it for predicting the outcome of occupational accidents is relatively sparse. This study utilized diversified ML algorithms; Multinomial Logistic Regression (MLR), Support Vector Machines (SVM), Single C5.0 Tree (C5), Stochastic Gradient Boosting (SGB), and Neural Network (NN) in classifying the severity of occupational accidents in binary (Fatal/NonFatal) and multi-class (Fatal/Major/Minor) outcomes. Comparison of the performance of models showed Balanced Accuracy to be the best for SVM and SGB methods in 2-Class and SGB in 3-Class. Algorithms performed better at predicting fatal accidents compared to major and minor accidents. Results obtained revealed that, ML unveils factors contributing to severity to better address the corrective actions. Furthermore, taking action related to even some of the most significant factors in complex accidents database with many attributes can prevent majority of severe accidents. Interpretation of most significant factors identified for accident prediction suggest the following corrective measures: taking fall prevention actions, prioritizing workplace inspections based on the number of employees, and supplementing safety actions according to worker’s age and experience.