Occupational case-control studies: I. Collecting information on work histories and work-related exposures.

The strength and credibility of evidence from occupational case-control studies largely depend on the validity and precision with which the work history is reported and the exposure is assessed. We discuss the two steps which ultimately lead to an exposure decision. The first step involves the exchange between the respondent and an interviewer. The latter is usually naïve to occupations and workplace exposures and, as such, is limited to asking generic and open-ended questions about the workplace. Often, this type of information is too nonspecific to assess exposure. In the second step, an expert reviews the information reported on each occupation and decides on exposure status without contacting either the interviewer or respondent. Exposure assessment is not, therefore, integrated with data collection and, in fact, is usually not initiated until after all the interviews are completed. As such, the exposure expert does not have an opportunity to resolve questions before making the exposure decision. To improve the quality and specificity of data collected, we have developed over 40 sets of close-ended questions (branch questions) which are specific to defined occupations. These branch questions, incorporated into a computer-assisted telephone interview, are asked if selected occupations or their synonyms are reported. Second, to link the data collection process with the assessment process, we have developed a procedure called SCORE (Subject Corrected Occupational Report) which provides the industrial hygienist with a cost efficient method to ask questions directly of respondents. Shortly after each interview is completed, a computerized version of the work history is reviewed by the industrial hygienist who develops questions when more information is needed. Subsequently, respondents are mailed a form listing their reported work history along with the questions. After two mailings, 73% of participants in a pilot study returned the SCORE form.