Difficulties in research to dig past exposure for linking present illness in epidemiologic point of view

Asbestos related diseases have long latent period. Hence it is very difficult to link past exposure and present illness. There are several bias and misclassification when researchers conduct exposure assessment. Also many problems to detect ill people are present such as survival effect and health worker effect. We found the SMR for lung cancer and malignant mesothelioma among former asbestos textile workers 7.0(2.6~15.3) and 126.2(26.0~368.7) respectively. We investigated present residents around the former asbestos textile factory. The number of source population is 23,461, including people have already died. We searched whole residential record to decide whether present address is representative of past address. When present address is used for surrogate of past environmental asbestos exposure, real past exposure group is 21.0%. In other words, the size of misclassification is 79.0%. We used exposed group who has ever lived in this area when the former asbestos textile factory had been operated, and non-exposed group who has moved this area recently. The exposed group and non-exposed group were matched with cancer registry data. Two large cohorts for the workers and the residents were made. Also job exposure matrix for past occupational exposure and simulation model with CalPuff to estimate past environmental exposure were constructed. The results of those cohorts will be presented. Korea started to compensate environmental asbestos victims from 2011. Composition of malignant mesothelioma in terms of exposure source will be presented.