Use of expert judgment to bound lung cancer risks

A bounding analytic technique for inferring the contribution of poorly characterized risk factors to a common health endpoint is demonstrated. Lung cancer mortalitywas chosen for the case study because the exposures responsible for the bulk of the deaths are very well-known, and the contribution of other putative causes is a focus of ongoing research and regulatory scrutiny. We elicited expert opinions on the upper and lower bounds on the fractions of the total lung cancer mortality due to individual risk factors. Interactive second-order uncertainty analysis was used to improve the experts' confidence in their bounds. From this information we calculated an upper bound on the residual fraction of deaths due to minor causes not selected by the experts.