Answers are grounded in this paper and cite their sources.
Methodological modeling reveals severe bias constraints in low-dose radiation epidemiology, highlighting insurmountable limits to detecting cancer risks in public populations.
Whether routine low-dose ionizing radiation from nuclear power plants causes cancer in the public is usually treated as a question of sample size; it is not. A closed-form analysis of study viability resolves any such study into twoboundaries that must both be cleared: a *power* wall (an adequately informative dose contrast) and a *bias* wall (residual same-hazard confounding held below the signal). The bias wall is *invariant to sample size*, so no larger cohortcan move it. With the standard linear excess-relative-rate estimator, its standard-error formula confirmed to reproduce the published INWORKS worker-cohort intervals (France, United Kingdom, United States), and residual per-home indoor radon confounding propagated by Monte Carlo and global sensitivity analysis, the probability of a valid study is about 10⁻⁴ at arealistic ambitious design, and at best a coin flip even at an extreme upper-bound ceiling (the entire U.S. population at the upper-bound fenceline dose), conditional on a residual correlation that cannot be measured without the infeasible study. As acorollary, the same boundary flags when a positive proximity-based association cannot, by itself, be attributed to a radiation effect. The result generalizes to any ultra-low-dose hazard.
No takes yet. Share an insight, caveat, or question.
Adam Stein (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: