Conference Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 23 - Statistical Considerations for Epidemiologic Studies of Radiation Risk
Type: Topic Contributed
Date/Time: Sunday, August 7, 2022 : 2:00 PM to 3:50 PM
Sponsor: Section on Statistics in Epidemiology
Abstract #323395
Title: Evaluation of Sources of Bias in Time-Dependent Radiation Dose Response Models for Individual Cohorts and Challenges Associated with Pooling in the Million Worker Study
Author(s): Ashley P Golden* and Sara Howard and Benjamin French and Yeji Ko
Companies: Oak Ridge Associated Universities and Oak Ridge Institute for Science and Education and Vanderbilt University Medical Center and Vanderbilt University
Keywords: Radiation; Epidmeiology; Confounding; Dose-response; Time-varying Covariates; Pooled Analyses
Abstract:

The Million Worker Study (MWS) is a retrospective cohort study assessing the health status of over one million United States workers and veterans who were chronically exposed to low-dose radiation. The study builds upon previously studied cohorts ranging in size from 2,000 to 150,000 workers, each with varying types of exposure and available data on confounders. These cohorts are individually under-powered and subject to bias from measured and unmeasured confounding, particularly in the presence of time-varying covariates. To facilitate combined analyses, each cohort is updated individually with dosimetry methodologies, vital status tracing, and statistical analyses. This presentation outlines the challenges encountered in addressing time-varying covariates, like duration of employment, and in the harmonization of individual cohorts for pooled analyses to achieve sufficient power for detecting dose-response associations in the low-dose range. Examples of the potential impact of these sources of bias are presented through sensitivity analyses and opportunities for methodological approaches are summarized.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2022 program