Article Text

PDF

328 Modelling complex mixtures in epidemiologic analysis: Additive versus relative measures for differential effectiveness
  1. G B H Hamra1,
  2. R F M MacLehose2,
  3. D B R Richardson3,
  4. S B Bertke4,
  5. R D D Daniels4
  1. 1University of North Carolina at Chapel Hill School of Public Health, Chapel Hill, United States of America
  2. 2University of Minnesota-Epidemiology, Minneapolis, United States of America
  3. 3University of North Carolina, Chapel Hill, United States of America
  4. 4NIOSH, Cinncinati, United States of America

Abstract

Mixed exposures are often combined into single exposure measures using weighting factors. An example of this arises in radiation epidemiology where doses of distinct forms of ionising radiation (such as alpha, beta, and gamma radiation) are combined based on knowledge of their biological effectiveness relative a reference form of radiation (most often gamma). Similar pooling of mixed exposures may occur with multiple congeners or air pollutants to develop more parsimonious models. The weights used for combining exposures are determined from experimental animal and cellular research, but not observational research. In this work, we show that these weights, which are the ratio of two normally distributed variables, cannot be reliably estimated from observational research. We propose an alterative approach for estimating differences in effectiveness of distinct exposures based on their excess effectiveness compared to a reference exposure. This alternative provides reliable estimates of differences in effectiveness of distinct exposures.

Statistics from Altmetric.com

Request permissions

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.