COMMENTARY
Methodological Challenges in the Statistical Analysis of Epidemiology Studies: use of Average Exposure Metrics in Historical Cohort Designs
Thomas B. Starr1, *, Gary M. Marsh2
Article Information
Identifiers and Pagination:
Year: 2016Volume: 3
First Page: 238
Last Page: 242
Publisher ID: MEDJ-3-238
DOI: 10.2174/1874220301603010238
Article History:
Received Date: 22/03/2016Revision Received Date: 31/08/2016
Acceptance Date: 17/9/2016
Electronic publication date: 31/10/2016
Collection year: 2016
open-access license: This is an open access article licensed under the terms of the Creative Commons Attribution-Non-Commercial 4.0 International Public License (CC BY-NC 4.0) (https://creativecommons.org/licenses/by-nc/4.0/legalcode), which permits unrestricted, non-commercial use, distribution and reproduction in any medium, provided the work is properly cited.
Abstract
An important methodological challenge in the analysis of historical occupational cohort data is choosing the most appropriate metric for the average exposure of the workers under study. We describe and illustrate the many issues associated with this challenge using a recent re-analysis by Kopylev [1] of lung cancer mortality in the National Cancer Institute (NCI) acrylonitrile cohort study. Kopylev proposed the routine use of both Average Exposure and Average Intensity when analyzing epidemiological cohort data. However, due to the methodological issues that arise with these metrics, we have concerns about the validity of his finding of a significant positive association between workers’ acrylonitrile exposure and increased lung cancer mortality in a subset of the NCI cohort. These include 1) the opportunity for substantial selection bias to have impacted the results; 2) the failure to account properly for latency; 3) the absence of a convincing biological rationale or other a priori justification for Kopylev’s preferred exposure metrics; 4) the absence of meaningful differences in Average Exposure- and Average Intensity- based risk estimates; 5) the lack of a logical basis for using either of these exposure metrics and 6) the conclusion that smoking was not a significant positive confounder, which is at odds with all other such findings for this cohort.