The project focuses on the Wismut cohort of 58,974 male uranium miners exposed to hazards like radon, uranium dust, and silica. Exposure was reconstructed using job exposure matrices, with early years (1946–1955) based on expert estimates. Many studies simplify the analysis by assuming fixed exposures and ignoring measurement errors.
To address this, a Bayesian hierarchical approach was developed, combining disease, measurement, and exposure models, and analyzed using a Metropolis-Hastings algorithm. Simulations show this method outperforms standard approaches and provides more realistic risk estimates.
Contact at StaBLab: Johannes Piller, Marie Scherzer
Project partner: Federal Office for Radiation Protection, Department of Radiation Epidemiology and Risk Assessment, Project number: 3624S12223