Commentary: standardized coding of occupational data in epidemiological studies.

The evaluation of occupational exposures in epidemiological studies is complex because of the multiple potential exposures in the workplace, the varying determinants of exposure between people, the many jobs people hold in a lifetime, and the different reasons for taking or leaving a job. Mannetje and Kromhout 1 show that beyond these well-recognized difficulties there are several more basic issues that are not adequately dealt with in epidemiological studies, particularly the occupational classifications used, and the coding of this information. Their recommendations on the use of standard classifications (e.g. International Labour Organization [ILO]/International Standard Classification of Occupations [ISCO]), the improvement of coding, and the use of additional databanks, will help provide more reliable and comparable results in epidemiological studies regarding occupation. Pooling of studies evaluating occupational exposures typically requires extensive and very time-consuming re-coding of information on occupation and industry that are partially or even totally incompatible. 2 Furthermore, those of us involved in multicentre studies know that this is a particularly serious problem in those studies, since recording of information and coding varies between centres. Similar concerns in a wider context led to initiatives for the development of core questionnaires to be used in epidemiological studies. 3 In earlier years, a basic classification by industry or major occupations sufficed to identify occupational risks. 4,5 Despite the serious limitations of exposure assessment methods that use only information on occupation and industry, these methods have helped identify specific risks and should continue being used. Surprisingly, in several situations they may even be the best proxy we have for the evaluation of combined and complex exposures. More powerful methods have been developed both for industry-based and population-based studies, including methods for the collection of more detailed information initially, and also elaborate ways for the evaluation of this information. Collection of occupational information can be done, for example, through computer-assisted interviews, repeated interviews with selected subjects, and use of modular questionnaires. 6,7 Evaluation of the data includes assessment by experts, 7,8 and the use of job exposure matrices based on extensive population-based measurements. 9 The identification of exposures still remains a complex issue, but in studies focusing on occupation the available methods have dramatically improved exposure assessment. One of the main problems many of us encounter refers to the methods applied in studies that are not principally focusing on occupational exposures and in which the interview time dedicated to the evaluation of these …