Application of Seemingly Unrelated Regression (SUR) in Determination of Risk Factors of Fatigue and General Health among the Employees of Petrochemical Companies

Background: In this cross-sectional study, 501 employees of petrochemical companies were selected by simple sampling method. Methods: Data were collected using Swedish Occupational Fatigue Inventory (SOFI-20), Occupational Fatigue/Exhaustion Recovery (OFER-15), and General Health Questionnaire (GHQ- 28). To identify the factors associated with fatigue and general health, we used ordinary least squares regression (OLS) and SUR and the results were compared. The analysis showed that satisfaction, mental disorder and sleepiness were the important factors associated with fatigue among these workers. However, the SUR estimator provided higher precision of the estimates than the OLS estimator as the parameters obtained by SUR are characterized by lower standard errors. As the models are intended to predict the fatigue risk factors, we particularly focused on the SUR method because it assesses the precision of the model in predicting fatigue determination. SUR estimators performed consistently better than the OLS estimators since SUR takes the correlation between error terms into account. Results: The findings showed that the study population were young and almost had a low job tenure. The correlation test showed that there was a significant relationship between fatigue and general health with job satisfaction (P=0.05), sleep disorder (P=0.01) and mental disorder (P=0.001). Finally, the analysis showed that fatigue as the result of work was affected by some organizational and individual risk factors, among which “general health status” in general fatigue and “job satisfaction and mental disorders” in mental, physical, shift work, chronic and acute fatigue had the most effect. Conclusion: The prevalence of fatigue among the study population was assessed high. Thus, elimination and reduction of casual risk factors are necessary to reduce the prevalence of fatigue at work environmental.