The map presents an almost geological appearance. Throughout the eastern United States, counties shimmer in deep purple wherever nuclear plants are concentrated: along the Great Lakes corridor, the mid-Atlantic, and the Connecticut River Valley. Journey west past the Mississippi, and the shading fades to a light yellow. There are fewer plants and fewer individuals drawn into their gravitational influence.
This map is also, based on a new Harvard study released today in Nature Communications, a rough estimate of increased cancer risk.
Researchers from Harvard T.H. Chan School of Public Health dedicated years to creating what they believe is the most thorough national overview of cancer mortality in relation to nuclear power plants in the U.S. Their findings are unsettling, especially considering our current situation: counties located nearer to functioning nuclear plants display higher cancer death rates than those farther away—and this link persists even after accounting for poverty, smoking, obesity, race, healthcare access, and most other conceivable factors.
The statistics the team uncovered are remarkable. Over a span of 19 years (2000 to 2018), they attribute approximately 115,000 cancer fatalities—around 6,400 annually—to proximity to nuclear facilities. This number comes with broad confidence intervals, and the researchers stress that correlation does not imply causation. However, it is significant. For reference, a prominent 2023 Science article estimated that emissions from coal power plants result in about 20,900 all-cause deaths each year. The nuclear cancer statistic, as uncertain as it may be, equates to roughly 20% of that coal mortality toll—and this figure includes only cancer-related deaths.
The research team, led by doctoral candidate Yazan Alwadi and senior author Petros Koutrakis, utilized a different methodological tactic compared to prior studies in this field. Previous U.S. research generally concentrated on an individual plant and its neighboring community, which limits statistical power and enforces binary classifications: you’re either within a certain radius or not. Alwadi and collaborators instead implemented what they refer to as continuous proximity—a cumulative tally of inverse distances to every operational nuclear plant within 200 km of each county’s center, summed across all plants and averaged over 10-year intervals to accommodate the long latency periods typical of radiation-induced cancers.
This is important. A county situated between two plants might rank high on this measure even if it’s not particularly close to either one. The method realistically represents how cumulative environmental exposures function.
The results exhibited consistency across various sensitivity tests. When the team modified the distance threshold—from 200 km down to 100 km, in 10 km increments—the associations remained intact. When they experimented with different averaging periods (2, 5, 10, 15, 20 years), the outcomes stayed stable. These results are not the fragile findings that vanish upon scrutiny of the assumptions.
Older adults represent the population most at risk, and the age-gender patterns are notable. The peak associations appear a decade apart by gender: the highest relative cancer mortality risk for females is found in the 55–64 age demographic, while for males, it resides in the 65–74 range. This gender-differentiated timing aligns with known disparities in radiation sensitivity and cancer biology, although the study was not designed to dissect that aspect. At equivalent closest distances, the relative risk for the highest-exposure categories reaches approximately 1.20—a 20% increase compared to the most distant counties.
“Our research indicates that residing near a nuclear power plant may pose a measurable cancer risk—one that diminishes with distance,” states Koutrakis, the Akira Yamaguchi Professor of Environmental Health and Human Habitation at Harvard. “We advise that further studies be conducted to explore the matter of nuclear power plants and health effects, especially at a time when nuclear energy is being championed as a clean solution to climate change.”
That concluding statement is where the finding takes on a politically sensitive tone. Nuclear energy is undergoing a revival in climate conversations—new constructions, reactor restarts, government incentives—specifically because it produces electricity without carbon emissions. The possibility that proximity to plants could entail a cancer risk fits uncomfortably within that narrative.
To clarify the study’s capabilities and limitations: it does not quantify actual radiation exposure. It assumes, possibly naively, that all plants contribute uniformly to exposure regardless of their design, age, or operational practices. The county-level data cannot trace where individuals actually resided or relocated. And because it is ecological—examining populations, not individuals—it cannot dismiss the influence of unmeasured confounders.
The research team is transparent about these issues. The study specifically notes that its exposure measure reflects geographic closeness instead of actual radiation experienced by people, and it does not address particular cancer types, mechanisms, or the childhood cancer aspect, which requires entirely different methods.
The global literature surrounding nuclear power plants and cancer is indeed convoluted. Some national studies—in the U.S., UK, and Canada—have reported no significant findings. In contrast, others, from France, Spain, and South Korea