Nuclear power plant proximity study sparks controversy

“I do not think this study advances the field of radiation epidemiology in any way,” said Bahadori. “It is common for articles like this to garner substantial media attention. People see headlines and make assumptions on causality often without understanding the real scientific value of the work. Results from ecological studies must be scrutinized because of the limitations associated with study design.”

The premise: The study looked at data from U.S. counties collected from 2000 to 2018 and calculated the sum of inverse distances from all operational nuclear power plants within 200 kilometers of the county. It then took this number as a proxy for the radiation exposure to residents of that county and compared that to the county’s cancer mortality rates, while controlling for covariates including annual county rates of educational attainment, median household income, poverty level, racial composition, population density, current smoking prevalence, and mean BMI.

Study limitations: The authors acknowledged the following limitations:

Their data do not include any direct measurements of radiation exposure, solely depending on geographic proximity as a proxy.

They analyze all cancer types combined and do not analyze childhood cancers.

The analysis is conducted at the county level, which lacks the ability to capture individual-level outcomes.

They use a formula for calculating the attributable fraction that assumes a causal relationship between exposure and outcome.

They do not incorporate residential histories or mobility.

They assume equal radiation exposure from all nuclear power plants at a given distance.

Bahadori additionally noted that the study does not appear to account for the well-established lag time between radiation exposure and cancer mortality.

Emily Caffrey, assistant professor of health physics at the University of Alabama–Birmingham, also added that the study does not account for medical exposure to radiation.

Distance as a measure of exposure: The fundamental assumption of the study, that proximity to a nuclear power plant can be used as a proxy for radiation exposure, is under question.

Caffrey said, “Distance to a plant is not a direct proxy for individual radiation exposure; emissions under normal operation are typically very low. . . . Physical proximity to a plant does not directly represent radiation dose; doses from normal plant operation are orders of magnitude lower than many natural and medical sources.”

When asked about this, study author Petros Koutrakis, a professor of environmental health at Harvard University, said, “It is self-evident that the closer you are to a source the higher the exposure,” citing that other studies in environmental epidemiology also use distance as a surrogate.

Jim Smith, professor of environmental science at the University of Portsmouth, said in a Science Media Center UK response to this study, “There is no evidence whatsoever that radiation doses are significant to people living near nuclear sites. Nor is there any evidence that there is a significant change in dose over the large distances (tens of kilometers) considered in this study.”

The results: The study found a correlation between county cancer mortality rates and distance to nuclear power plants, citing the strongest association in females aged 55–64 and in males aged 65–74.

Bahadori noted that “the researchers did not perform any back-of-the-envelope calculations using estimates of effect size from other studies to support or refute their results” and that other experts have said that the trend and size of the correlation are questionable.

In a Science Media Center Spain response to this study, cancer epidemiologist Amy Berrington de Gonzalez said, “The pattern of risks with age are not what we would expect with radiation exposure, which is usually higher risks with younger age at exposure not the reverse. Secondly, the size of the risk is much higher than we would expect based on the typical radiation exposure levels around nuclear power plants.”

Between the study’s severe limitations, absence of evidence to support key assumptions, and results that don’t align with what would be expected were a causal relationship between nuclear power plant proximity and cancer mortality rates to be found, some consider the study to be flawed to the point of being disqualifying.

American Nuclear Society Executive Director/CEO Craig Piercy said, “As a society of scientific and technical professionals, we support open inquiry even if we don't like the results. But this study is so severely flawed, it is unworthy of any serious consideration.”

Lacking the ability to establish causality, Caffrey said the study “stokes unnecessary fear” without meaningfully contributing to our understanding of low-dose radiation risk.

Broader context: Despite decades of effort, nailing down conclusive results on the impact of low-dose radiation exposure has been elusive, but the public perception of its impact, heavily shaped by ALARA (as low as reasonably achievable) and the linear no-threshold model, leads to concerns that a "report highlights" document released by the National Academies of Science, Engineering, and Medicine said have influence on both national and individual levels, including patient acceptance of medical diagnostic procedures and policy decisions related to nuclear power and waste management.

“For meaningful policy or causal claims, more targeted research incorporating individual exposure, mechanistic data, cancer type specificity, and improved confounder control is required,” said Caffrey.

Koutrakis said that despite its limitations, the study adds value by considering national-level data over 18 years.

Bahadori pointed to the Million Person Study, which is currently tracking 880,000 U.S. radiation workers or atomic veterans, and the International Workers Study, which includes 309,932 workers in France, the United Kingdom, and the United States, as high-quality, individual-level analyses that are actively contributing to our understanding of radiation health effects.