Fifteen-Year Accident Causation Assessment of the Petroleum Industry in South and Central America Using Statistical Correlations

Abstract In an era of low oil prices, reducing lost time injuries and other consequences of accidents become more attractive than ever. This paper undertook quantitative and qualitative assessment of occupational accidents that occurred in the upstream petroleum industry of South and Central America from 2000 to 2014. Accident reports culled from IOGP database were coded using a 5-tier system in the HFACS model, exported to Excel spreadsheets, then to SPSS and comprehensive correlations run to test the reliability of the coded data and establish inter-relationships between the 19 variables in the HFACS model. Decision-based error was found to have a correlation coefficient of 0.558 with planned inappropriate operations, personal readiness 0.530 and 0.533 with supervisory violations and physical mental condition, respectively, and perceptual error 0.609 with physical environment. These and other results obtained were compared with established benchmarks, explanations made for a few deviations from expected trends and the significance of obtained results discussed. Inferences were thus drawn and recommendations proposed for health and safety management systems that would avert similar accidents in the future. The research validates human error as a prominent causal factor of major accidents in the area under review as the data acquired, processed and interpreted closely corroborate the initial hypothesis, classical accident theories and the findings of a good number of previous works in the literature. Incorporation and sustenance of a robust personnel and process safety culture was therefore recommended for the entire life cycle of oil and gas facilities; from conceptual design to decommissioning, in South and Central America.