When Are Risk Analyses on Job Titles Informative?

Abstract

In the January issue of the Annals of Occupational Hygiene, we published a paper by Garabrant et al. (2016a), ‘Mesothelioma among Motor Vehicle Mechanics: An updated review and Meta-analyses’. The paper described an update from a 2004 metaanalyses performed largely by the same authors. The updated meta-analyses included 10 case–control studies, 1 cohort study, and 5 proportional mortality ratio/ standardized mortality odds ratio (OR) studies. The results of the meta-analyses indicated that the summary risk estimates were generally <1 and that no statistically significant increases in mesothelioma were observed. The authors therefore concluded that ‘this meta-analysis of the epidemiologic studies provides evidence that motor vehicle mechanics, including workers who were engaged in brake repair, are not at an increased risk of mesothelioma’ (Garabrant et al., 2016a). This conclusion is similar to the conclusion of the 2004 paper that concluded, based on seven studies of which six studies were also included in the current meta-analyses, that ‘employment as a motor vehicle mechanic does not increase the risk of developing mesothelioma’ (Goodman et al., 2004). Although there were some methodological differences between the two meta-analyses in how the quality of the studies was scored the two ranking systems showed overall good correspondence and resulted in a similar classification in quality categories. Analyses by quality of the studies did not show an association between study quality and the log OR and of the 10 studies only 2 studies had an OR > 1. All together, the conclusion of the papers seemed therefore scientifically justified. However, there is a question in how one should interpret this result. As Teschke (2016) pointed out in her letter to the editor, the issue of interest is whether chrysotile asbestos exposure among mechanics causes mesothelioma. Mechanics perform a variety of tasks and not all mechanics will be involved in break repair. The ones that do in turn may only do so for part of the time. Using broad categories of job titles (e.g. garage workers, auto repair and related services, and auto engineers) are therefore likely poor proxies of chrysotile asbestos exposure. As such although the conclusion of the paper is correct in that no elevated risks are found among motor vehicle mechanics as a whole this cannot be equated to mean that there is no risk of mesothelioma among mechanics that were involved in break repair and that had potential chrysotile exposure. In their answer to the letter to the editor, Garabrant et al. (2016b) indicated that the results may indicate that motor vehicle mechanics and brake workers are perhaps not at increased risk of mesothelioma as they may not have sufficient exposure to observe an increased risk. Again, this statement is correct at the job-title level but may not hold for specific subgroups and/or individuals that may have been higher exposed. As such, the question can and perhaps should be raised in how far job-title-based analyses are informative and when. A similar debate arose during the International Agency of Research of Cancer (IARC) on the evaluation of the carcinogenic risk among firefighters. Firefighters are exposed to numerous known carcinogens among which benzene, diesel exhaust, formaldehyde, and polycyclic aromatic hydrocarbons (IARC, 2010). However, epidemiological studies based on having a job as a firefighter have rendered heterogeneous results with some indications of elevated risks of testicular and prostate cancer and non-Hodgkin lymphoma. Based Ann. Occup. Hyg.