Reduction of workplace accident rates using mathematical statistical models

Abstract:

The objective of this research was to select the best regression model that would allow identifying the variables of the Occupational Health and Safety Systems, which have a greater impact on the occurrence of accidents, to project programs of improvements in order to reduce accident occurrence rates. The research was conducted with a sample of 24 small and medium-sized Ecuadorian companies. The causal factors to be investigated were Occupational Health Management, Occupational Risk Prevention Management, Management of Natural Hazards/Anthropic Risks and Document Management. The mathematical models subjected to analysis to determine the relationship between the causal factors and the number of accidents were the Poisson, Negative Binomial and Logistics Regression models. STATGRAPHICS Statistical Software was used to determine the model with the best goodness of fit. Statistical inference was made by comparing Poisson, negative binomial and logistic regression models, the latter being the one that presented the best fit. The application of the designed intervention plans made it possible the observation of improvements in the performance of these systems, which was evidenced by a significant reduction in accident rates.

Año de publicación:

2021

Keywords:

• Multivariate statistics
• Performance improvement
• Regression Models
• Safety and Health Audits
• Workplace accidents
• Logistics models

Fuente:

scopus