U.S. counties located closer to operational nuclear power plants record higher cancer death rates than counties farther away, according to a new nationwide study led by Harvard T.H. Chan School of Public Health.
Researchers analyzed cancer mortality data from 2000 to 2018 across every U.S. county. They compared those figures with the distance to operating nuclear power plants. Additionally, they examined plants in Canada located near the U.S. border. The team published the findings in Nature Communications. However, the authors said the results do not prove that nuclear plants cause cancer.
They conducted what they described as the first 21st-century national review of this question. Previous U.S. research often focused on a single plant and nearby communities. Consequently, earlier studies offered a limited picture of potential risk. The researchers used a method called continuous proximity. This approach measured how close counties were to all nearby plants. Furthermore, the model estimated the combined impact of multiple facilities rather than one site.
They gathered plant location and operating data from the U.S. Energy Information Administration. Meanwhile, they obtained county-level cancer mortality data from the Centers for Disease Control and Prevention. The team adjusted for factors that can influence cancer risk. These included income, education, and racial composition. In addition, they controlled for smoking rates, body mass index, temperature, humidity, and distance to hospitals.
Even after those adjustments, counties nearer to plants showed higher cancer mortality rates. The link appeared strongest among older adults. Additionally, the association weakened as distance from a plant increased.
Read more: Europe’s first NuScale SMR advances with Romanian approval and AI push
Read more: NextEra explores nuclear growth as AI drives record U.S. electricity demand
Potential public health impact is measurable
The researchers estimated that about 115,000 cancer deaths over the study period related to plant proximity. That figure equals roughly 6,400 deaths per year. Consequently, the authors described the potential public health impact as measurable. Senior author Petros Koutrakis said the findings suggest distance from a plant may matter. He indicated that risk appears to decline the farther people live from a facility. However, he cautioned that the study cannot establish a direct cause-and-effect relationship.
He called for more detailed research into nuclear power and health. Furthermore, he noted that governments increasingly promote nuclear energy as a climate solution. The team said the results align with a prior Massachusetts study. That earlier work found higher cancer incidence closer to nuclear plants. Conversely, studies in other countries have produced mixed findings.
The authors acknowledged several limitations. They did not measure radiation exposure directly. Instead, they assumed each plant exerted an equal effect in the statistical model. Additionally, the analysis relied on county-level data rather than individual medical records. The researchers said future studies could use more precise exposure measurements and finer geographic detail.
.
Nicholas Geary
February 25, 2026 at 10:34 pm
Interesting that they could control for income, education, racial composition, smoking rates, body mass index, healthcare access, and distance to hospitals without having any individual medical records. That’s not even enough to show what proportion of decedents in a county were residents, or how long they lived there. Let’s see if anyone else can replicate these results.