Article Text


357 The two-phase design makes efficient use of expert-time in assessing exposure to occupational carcinogens: a lung cancer case-control study
  1. P Wild
  1. INRS, Vandoeuvre Les Nancy Cedex, France


Objectives To identify dose-response relationship between lung cancer incidence and cumulative exposure to the main occupational carcinogens as assessed by experts making use of an initial algorithmic exposure assessment within the two-phase design.

Methods A population-based case-control study including 246 cases and 531 controls was conducted in North-Eastern France. Detailed occupational and personal risk factors were obtained in face-to-face interviews based on a task-based questionnaire and a series of job-specific questionnaires. In the conceptual framework of a two-phase design, cumulative expert-based exposure scores were obtained in a subset of 215 cases and 269 controls stratified on smoking and a prior algorithmic exposure assessment. The cases and controls included in the expert assessment were chosen in order to over-sample rare exposure categories. The data were analysed using logistic regression models adapted to two-phase data. This analysis makes use of the subset of subjects with expert assessments but also of the initial data set from which the subset was sampled.

Results Expert based cumulative exposure scores were lower than the corresponding algorithmic scores. The correlation between algorithmic and expert-assessed cumulative exposure scores were high for asbestos (r = 0.80) but less so for crystalline silica (CS) (r = 0.60) and polycyclic aromatic hydrocarbons -PAH (r = 0.56). We identified significant dose-response relationships (DRR) for asbestos, CS and Diesel Motor Exhaust (DME) with significant ORs exceeding 2 in the respective highest tertiles of cumulative exposure. The dose-response relationship with PAH was borderline significant but the OR in the highest tertile was 1.96 (95%CI [1.11–3.46]). All DRRs were steeper when using expert-based scores than when using algorithmic scores.

Conclusion Dose-response relationships between lung cancer incidence and cumulative exposure scores could be identified based on expert-assessment in a subset of cases and controls chosen to be informative in the framework of a two-phase design.

Statistics from

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.