산업재해 예방을 위한 사건보고문화 중재연구

Abstract

The stagnation in the reduction in occupational accident rates has been a worry industrywide. The limited information from accident reports and the lack of incident reporting (IR) are considered contributors to this stagnation. The lack of IR is considered a serious potential risk across the industry since the incident report provides various information about the background situation and sequence of the occurrence. However, previous studies have shown that the under-reporting rates ranged from 29% to 81% and that the reporting was primarily concerned with specific events like serious accidents and property damage. If incidents are not appropriately communicated, the under-reporting will prevent employers, researchers, and policymakers from developing effective preventive strategies. In order to resolve the lack of IR, a right incident-reporting culture (IRC) should be established, where workers report all incidents in a comfortable and psychologically stable condition. The comprehensive understanding of IRC and investigating the effectiveness are crucial for promoting the IRC. However, the few empirical cases and the limited supporting evidence in previous studies reduce the confidence in our knowledge. Therefore, this study aimed to explore the entire process from IRC structure (theme 1) to factors affecting IRC (theme 2), IRC effects (theme 3), and incident data analysis (theme 4). Theme 1 proposed an IRC model and measured the influence of internal psychological factors such as perceptions, attitudes, and backgrounds on the incident-reporting rate (IRR). A questionnaire survey was administered to 614 workers at a chemical company, and the relationship between the factors and IRR was confirmed quantitatively. Theme 2 identified the factors affecting IRC and evaluated the influence degree. The candidate factors were explored through previous studies and plant safety staff. A questionnaire survey was administered to all employees, and 32 affecting factors were identified, along with their influence degree. Theme 3 provided empirical evidence and enhanced the reliability of IRC effects on the accident rates. The study applied a 3-year-intervention at two chemical companies, conducted two preliminary studies, and administered a questionnaire survey to evaluate IRC and willingness to report (WTR). As a result, IR increased by 24.5% and 437.3% in both companies with improved WTR. Following the IR increase, the accident rates decreased, and the causal relationship between IR and accident rates was confirmed with various supporting evidence. Theme 4 presented a text analytics approach for extracting critical information from incident reports. The text mining was used to extract the keywords of the incident causes, and the Dirichlet allocation algorithm was applied to extract latent topics from the incident information. The cause-and-effect diagram was depicted, and the trends of incidents were identified. These results provide a comprehensive understanding of IRC and methodologies to identify the affecting factors, establish IRC, and analyze the incident reports. The study supplements the limitations of previous studies. IRC is suggested to all industries as a universal tool to prevent occupational accidents and will help to resolve the stagnation in the accident rates at the country-level by lowering the company's accident rates.