Does risk stay linear below 10 rad
In plain words
Protection rules assume that halving a small dose halves the risk, with no safe threshold. Direct evidence at such doses is thin and the assumption is contested.
Precise statement
Determine the shape of the dose-response relation for solid-cancer incidence below 10 rad ($1\,\mathrm{rad}=100\,\mathrm{erg}/\mathrm{g}$) of low-linear-energy-transfer radiation: linear without threshold, linear-quadratic, threshold, or non-monotonic. The answer must handle dose-measurement error, confounding and reverse causation at effect sizes of about 5 percent excess relative risk per 10 rad (ERR about 0.5 per 100 rad), falling to about 0.5 percent at 1 rad. An answer is a shape with confidence intervals excluding the alternatives.
What would settle it
Pooled worker and medical cohorts with individual dosimetry of sufficient size and dose accuracy to separate linear from threshold models at the one-percent effect level.
Status in the literature
The INWORKS nuclear-worker analysis (Richardson et al., BMJ 2023) reported solid-cancer mortality ERR of 0.52 per 100 rad, with a larger slope below 10 rad, while threshold and hormesis advocates dispute the dosimetry and confounding control.
Related problems
- More general than Causal cancer risk from diagnostic CT exposure