Occupational life course and lung cancer risk in men. Findings from a socio-epidemiological analysis of job-changing histories in a case-control study

Psychosocial factors in general, and the social and psychological conditions of work and occupation in particular, have attracted little attention to the epidemiological investigation into the occupational-related causes of lung cancer. There is some evidence available concerning the impact of loss events--including job losses--on cancer development. During our research presented in this article, we examined job changes and job transitions in the occupational biography of men with regard to their circumstances in terms of (in-)voluntariness both as individual events and the job-changing histories of individuals. We expected the job-changing histories of lung cancer cases to be more involuntary than those of population controls, and vice versa. Our sample contains 391 male case-control pairs. Cases are defined as newly-diagnosed men of German nationality with a histologically- or cytologically-confirmed diagnosis of primary lung cancer. Population controls were drawn randomly from the municipality records of residents and individuals matched to cases 1:1 by age and region. Occupational histories were reconstructed, and information about other classical and suspected risk factors was collected during standardized interviews by trained interviewers. We categorized the reported reasons for job changes, and classified them with respect to voluntariness on a seven-stepped rating scale from -3 'very involuntary' to +3 'very voluntary', defining relations within the empirically established categories. The job-changing histories of individuals are described by using the average score of voluntariness. We measured the degree of voluntariness for the occupational history: (1) by the rounded average score; and (2) by the quintiles of the average score. On the basis of individuals matched odds ratios (OR) estimates are derived for the score of voluntariness unadjusted and adjusted for smoking, asbestos exposure and socio-economic group by conditional logistic regression. The investigation into job-changing histories using the average score of voluntariness lead to similar results for both measuring methods for the total study group. When the rounded score was used, job-changing histories which were assessed as 'very involuntary' and 'involuntary' (score -3 and -2) showed an elevated odds ratio of adjusted 1.41 (95%-CI: 0.57; 3.48), and when they were assessed as 'moderately involuntary' (score -1) the adjusted odds ratio was 1.59 (95%-CI: 1.02; 2.48), as compared with job-changing histories assessed as neutral (score 0). When the degree of voluntariness was measured by quintiles, the first quintile (score -3- 0.6-3: OR 0.63) were insignificant. The analysis was also conducted for three broad age groups, and for biographies reduced by the five years prior to the interview or the diagnosis of lung cancer respectively. Methodological limitations of the case-control design and methodical implications--missing values and recall-bias in particular--as well as possible explanations for these findings are discussed.