A validation of information on occupation--data from a nested case-control study

Abstract

Objectives: To evaluate the validity of information about employment obtained from self-respondents and proxies in a nested case-control study.

Methods: We interviewed 230 self-respondents and 652 proxies covering 237 cancer cases and 645 controls. Questions about employment as a bus driver, other jobs, and residence were asked. Golden standard was employment data obtained from personnel records.

Results: All categories of respondents tended to over-report the length of employment, and respondents aged above 70 years had significantly decreased odds ratio for disagreement between interview and company files compared with respondents below the age of 60 years. There were no significant differences in whether the respondent was a case or control as there were no differences in whether the respondent was a self-respondent or a proxy. The validity of the reported employment time was statistically independent of the interval between death and proxy contact. Compared with the self-responders, the proxies reported significantly fewer jobs and residential areas.

Conclusions: In general, data obtained from cases and controls not knowing that the study was about cancer and without the interviewer's knowing who were cases and who were controls appeared good. The quality of information obtained from proxies appeared good as well. There seems to be no problem in interviewing very old people about a specific job when they are mentally prepared for it. The time interval between death and proxy contact was without importance.