Citation

Luo X, Liu Q, Qiu Z. Front. Public Health 2021; 9: 783537.

Copyright

(Copyright © 2021, Frontiers Editorial Office)

DOI

10.3389/fpubh.2021.783537

PMID

35087784

PMCID

PMC8787334

Abstract

This paper firstly proposes a modified human factor classification analysis system (HFACS) framework based on literature analysis and the characteristics of falling accidents in construction. Second, a Bayesian network (BN) topology is constructed based on the dependence between human factors and organizational factors, and the probability distribution of the human-organizational factors in a BN risk assessment model is calculated based on falling accident reports and fuzzy set theory. Finally, the sensitivity of the causal factors is determined. The results show that 1) the most important reason for falling accidents is unsafe on-site supervision. 2) There are significant factors that influence falling accidents at different levels in the proposed model, including operation violations in the unsafe acts layer, factors related to an adverse technological environment for the unsafe acts layer, loopholes in site management in the unsafe on-site supervision layer, lack of safety culture in the adverse organizational influence layer, and lax government regulation in the adverse external environment layer. 3) According to the results of the BN risk assessment model, the most likely causes are loopholes in site management work, lack of safety culture, insufficient safety inspections and acceptance, vulnerable process management and operation violations.

Language: en

Keywords

Bayesian network; falling accidents; fuzzy set theory; human factor analysis and classification system (HFACS); human-organizational factors