Comparison of first, last, and longest-held jobs as surrogates for all jobs in estimating cumulative exposure in cross-sectional studies of work related asthma

BACKGROUND: Previous occupational studies have used exposure in most recent job as a Surrogate for all jobs or "total work history" exposure. This method may not be valid for diseases in which disability brought on by one job may influence later work history, such as asthma. We investigated different surrogates for total work history for the outcome of asthma symptoms. METHODS: In a cross-sectional study, we quantitatively compared three exposure surrogates (last job, first job, and longest-held job) with the total cumulative work exposure (all jobs) in a cohort of 1778 aluminium industry workers. The chemical exposures we compared were total fluoride dust, inspirable dust, and sulfur dioxide with the respiratory symptoms of wheeze, chest tightness, and rhinitis. RESULTS: When extrapolated over time, all surrogates quantitatively overestimated the gold standard "all jobs" for all three exposures investigated. For the symptom of wheeze, last job was found to be the worst surrogate for all jobs for the three exposure types investigated. Prevalence ratios for fluoride exposure and the symptom of occupational wheeze were last job 1.07 (95% confidence interval, 0.92-1.26), longest job 1.10 (0.94-1.30), first job 1.14 (0.97-1.35), and all jobs 1.27 (1.05-1.53). CONCLUSIONS: Although last job has been found to be a satisfactory surrogate for all jobs in cancer studies, we do not recommend the use of this metric in studies of chronic diseases where development of disability may occur with early exposure. We found that both first job and longest job held were better metrics in studies where the adverse health effect may influence the job history of subject.