Adjustment for Suspected Misclassified Smoking Data in an Historical Cohort Study of Workers Exposed to Acrylonitrile

摘要: Objectives: To examine the association between exposure to acrylonitrile (AN) and lung cancer mortality after properly addressing misclassification possible positive confounding of smoking history. Methods: Subjects were 992 white males who employed for three or more months 1960 1996 at an AN chemical plant in Lima, OH. There 15 identified cases deaths. Smoking histories obtained 90.3% cohort 54.2% as having “ever smoked”. Though there few “unknown” histories, variable was determined be misclassified RR ever smoked related only 1.08 (95% CI=0.26, 6.18). We addressed potential by presence suspected data determining if a reasonable adjustment available would change risk levels original Lima relationship using Monte Carlo simulation bias adjustment. Conclusions: After running simulation, we found that mean given differing decreased adjusting simulated data. However, results from must interpreted with caution analysis limited number cases. In this cohort, concluded positively confounded mortality. Public Health Relevance: Properly history studies is critical validity study results. As seen study, habits impact certain health outcomes. Researchers attempt address whenever possible.