Objective Case-control studies use job- and industry-specific modules to systematically obtain details about the subjects’ work tasks. This reduces respondent burden over asking the same questions for every job but relies on accurate module assignment to ensure relevant questions are asked.

Methods In a multi-center hospital-based case-control study of lymphoid and myeloid malignancies in Asia, lifetime occupational histories were collected for every job held by a subject. It comprised open-ended job title, work tasks, employer and its product/services, and yes/no screening questions pertaining to paints/stains, solvents/glues/degreasing agents, and engineered woods. These responses were used in a keyword-based algorithm to automatically assign each job to one of 23 modules, including ‘generic exposure’ and ‘work location’ modules. We evaluated the proportion of jobs assigned the work location module that were redirected to more specific modules and the prevalence of ‘yes’ responses to solvent-exposed tasks in the generic exposure module as indirect measures of the sensitivity of capturing relevant work task information.

Results The work location (29%) and generic exposure (21%) modules were the most frequently assigned to the 31,398 jobs. Of those jobs assigned the work location module, 8% were redirected to the generic module based on location of ‘construction’ and 12% were redirected to other modules. In total, 7,359 (23%) jobs were assigned the generic module: 38% based on screening questions with no relevant keywords identified; 50% based on ambiguous keywords; and 11% based on work location module. For these three pathways, the prevalence of ‘yes’ responses to any of the module’s task and bystander questions were 77%, 44%, and 23%, respectively.

Conclusions The screening and the work location questions were successful in identifying jobs involved in tasks of interest that would have been missed if we had solely used the algorithm keyword search for jobs, tasks, and industries of interest.