Background There are more than ten kinds of occupational health risk assessment models in the world. Each model has its own advantages and limitations due to its different technical principles. Different models may generate inconsistent results of risk assessment towards a single hazard. At present, there are few reports on the differences among various occupational health risk assessment models.

Objective The purpose of this study is to understand the differences among six commonly used occupational health risk assessment models, and to provide references for the formulation of relevant standards and for the methodological study and application of occupational health risk assessments.

Methods The printing and laminating positions of seven small printing companies in Zhejiang Province were selected as study subjects in 2019. Because they were exposed to the same risk factors, the inherent risks of both positions were estimated by comparing occupational hazard concentration ratio (the ratio of hazard concentration to occupational exposure limit), and then were used to determine the accuracy of results calculated by each risk assessment model. Six common occupational health risk assessment models (i.e., U.S. EPA model, Singapore model, UK COSHH model, Australian model, Romanian model, and ICMM model) were used to assess the identified occupational health risk factors of the selected printing and laminating positions. A quantitative comparison of the risk assessment models was conducted using risk ratio (RR) to verify the accuracy and analyze the parallelism and correlation of the assessment results.

Results The average xylene and ethyl acetate concentration ratio of the laminating positions (0.33±0.40) in the selected seven printing companies was greater than the ratio of the printing positions (0.08±0.07) (P < 0.05), thus it was the basis for judging the accuracy of model evaluation results that higher inherent risk of xylene and ethyl acetate in laminating positions than that in printing positions. According to U.S. EPA model and Singapore model, the mean RR of xylene and ethyl acetate in laminating positions was higher than that in printing positions (P < 0.05), which was consistent with the inherent risk of the two positions. The assessment results of other models failed to distinguish the risk difference between the two positions. The parallelism analysis results showed that RR_{U.S. EPA}=0.65±0.34, RR_{UK COSHH}=0.50±0.15, RR_{Singapore model}=0.36±0.15, RR_{Australian model}=0.34±0.15, RR_ICMM=0.34±0.08, and RR_Romanian=0.20±0.10 (P < 0.05). The correlation analysis results showed that the concentration ratio of each occupational hazard in the laminating and printing positions of the seven companies was correlated with the RR of each model (P < 0.01), and the correlation coefficient was the largest for RR_U.S.EPA model and RR_{Singapore model} (r=0.761 and 0.792, P < 0.01), and the smallest for RR_{UK COSHH} model (r=0.330, P < 0.01). The correlation analysis results also showed that there was a correlation between RR of each model except that between RR_U.S.EPA model and RR_{UK COSHH} model; the correlation coefficient (r) between RR_U.S.EPA model and RR_{Singapore model} was 0.601, the values between RR_U.S.EPA model and other RRs were < 0.5 (P < 0.01), and the values between RR_{Singapore model} and other RRs were greater than 0.7.

Conclusion Both U.S. EPA model and Singapore model have better assessment accuracy and strong correlation with the field survey based occupational hazard concentration ratio. U.S. EPA model and UK COSHH model are relatively independent, while Singapore model has strong correlations with other models.