Use of Uncertainty Analysis to Quantify the Impact of Measurement Error on an Estimate of Effect for Occupational Exposure to Tcdd-Contaminated Chemicals and Ischemic Heart Disease

Abstract

Introduction Over the last few years novel advances in epidemiologic methods have refocused attention on causal relationships between exposure and disease, renouncing simple associations that often lack quantitative adjustment for study imperfections. Probabilistic uncertainty analysis is a mathematical technique for assessing how systematic error may affect study results. 1-7 Commonly employed in the fields of engineering and policy and risk analysis, 8-9 this method has not yet become established in epidemiology and has never been applied to dioxin research. Here, the impact of exposure and disease misclassification on the odds ratio (OR) for ischemic heart disease (IHD) mortality in a cohort of trichlorophenol (TCP) workers exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) is quantified using probabilistic uncertainty analysis to illustrate how this innovative method can be utilized in epidemiological dioxin studies.