| Objectives Eliminating or reducing job stress is one of the main contents to protect workers’health in the workplace.The steel workers have comparatively high levels of occupational stress.Identifying the influencing factors of job stress and finding the high-risk persons is very helpful for making intervention policies and measures for improving health in steelworkers.A new idea for evaluating job stress will be developed by developing a predictive model.The purposes of this study are to 1 explore the current situation and influencing factors of job stress and find out the high-risk groups of job stress of steel workers in a steel company;2 Clarify the influence of occupational exposure factors(high temperature,noise,dust and carbon monoxide,etc.)on job stress of steel workers;3 The prediction models of job stress and high level job stress of steelworkers were constructed in order to provide convenient and quick tools to identify job stress of steelworkers;4 Build a prediction model of high level stress of steel workers in different types of work,and accurately predict the high level job stress of different types workers.Methods The data are from the baseline data of the health effect cohort study of occupational population in Beijing-Tianjin-Hebei Region.1 A cross-sectional study was carried out.8078 the steel workers who come from a certain steel company in Tangshan city were chosen as the subjects.The information of basic demographic,behavior habits,occupational hazards exposure conditions,current health status and family disease history,were collected by face to face interview.The Job content questionnaire(JCQ)was used to evaluate job stress of steelworkers.The physical examination of subjects were measured at the scene,and the biological samples were collected and tested by professional inspectors.The SAS database was established after data entry.The logistic regression model was used to analyze the influencing factors of job stress and the restricted cubic spline(RCS)function was used to fit the dose response relationship between the cumulative exposure of occupational exposure factors and job stress in steelworkers.2 The 6236 workers were selected as subjects to develop the prediction model of job stress.Lasso regression,random forest and logistic regression was used to screen the predictive variables of job stress.The prediction model of job stress and high job stress was developed by predictive variables which were consist of the factors of job stress and physiological index.The prediction model of different job category steel workers were established.The evaluation of model discrimination was checked by calculating C-statistics.Hosmer-lemeshow test and calibration curve were used to evaluate the model calibration.Nomogram is adopted for the model presentation.Results 1 Among the 8078 subjects,7423 males,accounting for 91.9%;The detection rate of job stress and high level job stress were 49.1%and 26.2%respectively.2 Compared with steel workers aged 30～40,the OR value of job stress risk of workers aged 20～30 was1.469(1.181～1.828).Compared with steel workers with junior or below education level,the OR value of job stress risk of workers with college or above education was 0.767(0.658～0.895).Compared with unmarried workers,the OR value of job stress risk of divorced or widowed workers was 1.459(1.053～2.021).Compared with workers with low family average monthly income,the risk of job stress of workers with family average monthly income more than 2500 yuan/month is 0.738(0.645～0.844).There is low risk of job stress on steel workers with a high level of social support,and high risk in depressed workers.3 The job stress risk of steel workers on short-term contracts or labor dispatch is1.401 times that of steel workers on long-term contracts 1.401(1.170～1.677);The job stress risk of ironworkers is the highest.The risk of job stress in steel workers with CEH more than 712.5(℃·years)was 1.618(1.385～1.890)times that in steel workers with CEH lower than 490(℃·years);The risk of job stress for workers with CED higher than 40(mg/m~3·years)was 1.667(1.449～1.917)times that of workers with CED lower than 19(mg/m~3·years).4 The prediction model of job stress which was developed by logistic regression sensitivity is 61.6%,specificity is 67.6%,the area under the ROC curve(AUC)is 69.1%,right index is 0.293,they were higher than them of other two prediction model which were developed by lasso regression and random forest.5 After adding physiological indexes,the predictive ability of the prediction model was improved.The NRI of lasso regression,random forest and logistic regression model were 3.95%,1.57%and 2.00%,respectively.The IDI were 0.013,0.015 and 0.007,respectively.6 The sensitivity,specificity and AUC of the high job stress prediction model were 38.9%,87.6%and 75.1%.AUC of model of ironworkers,steelworkers,hot-rolled steelworkers,inspection workers and power guarantee workers are 69.2%,71.9%,74.4%,70.5%and 65.7%respectively.Conclusions 1 The steel workers are generally suffering from job stress in the steel company;High education level,high family per capita monthly income and four shift and three operation workers have the lowest job stress risk.Steel workers aged 20-30,divorced or widowed,low social support,poor psychological state,high CEH and CED have the highest risk of job stress.According to the order of hot-rolled steel workers,power guarantee workers,other types of work,steelworkers,repairmen,cold-rolled steel workers and ironworkers,the job stress risk of steel workers increases in turn.2 The prediction model,was developed by logistic regression was the optimal predictive ability than model developed by Lasso regression and random forest.After adding physiological indexes,the predictive ability of each model improved positively.The specificity and prediction ability of high job stress prediction model for steel workers was powerful.The high job stress prediction model of ironmaking,steelmaking and service workers was strong.Figure 43;Table 49;Reference 157... |
PDF Full Text Request