Methods of Exposure Assessment for Community-Based Studies: Aspects Inherent to the Validation of Questionnaires

Abstract Possible sources of inaccuracy in the data collecting stage and various methods to evaluate the quality of data are reviewed. Interviews of surrogates are often used in epidemiological studies, even if they are likely to introduce a higher level of inaccuracy than when subjects themselves are questioned. The preliminary results of a study on the accuracy of work history information obtained from the relatives of the subjects in a case-control study of lung cancer are presented. A mail questionnaire requested the same information from the surrogates as was asked of the original cases. Accuracy was measured for major occupational categories by evaluating sensitivity and specificity and calculating k statistic agreement. Some major occupational categories, which are well represented and traditional in the study area, showed a high specificity and sensitivity and a high k agreement. Various methods of exposure identification are reviewed. Job exposure matrices generated a priori or a posteriori were ...