A pair of Harvard-led studies released in recent months found that people living closer to operational nuclear power plants in the United States experienced higher rates of cancer mortality and cancer incidence between 2000 and 2018. The research, which spans every U.S. county and drills down to ZIP code-level data in Massachusetts, estimates that roughly 115,000 excess cancer deaths nationwide and approximately 20,600 additional cancer cases in one state alone may be linked to proximity to reactors. The findings land at a moment when federal policy is leaning toward nuclear expansion as a climate solution, raising hard questions about who bears the health costs of that strategy.
A Nationwide Pattern Across All U.S. Counties
The larger of the two studies, published in a peer‑reviewed journal, examined cancer mortality data from all U.S. counties over an 18-year window ending in 2018. Led by researchers at the Harvard T.H. Chan School of Public Health, the analysis measured each county’s proximity to operational nuclear power plants using a summed inverse-distance metric within a defined radius. That approach captures not just whether a plant sits nearby but how many plants are within range and how close each one is. Counties with higher proximity scores showed statistically elevated cancer death rates even after the researchers controlled for income, smoking prevalence, healthcare access, and environmental exposures.
The study is described in a summary for journalists as the first 21st-century nationwide analysis to cover every county in the country. Senior author Petros Koutrakis, a professor of environmental sciences at Harvard, framed the results as a signal that health impacts should be factored into energy planning. The approximately 115,000 attributable cancer deaths estimated over the study period represent a substantial toll, though the researchers were careful to describe a statistical association rather than a confirmed causal mechanism. That distinction matters: the study does not claim reactors directly cause cancer, but the correlation persisted across a battery of adjustments designed to strip out confounding variables.
Massachusetts ZIP Codes Tell a Sharper Story
A companion study narrowed the lens to Massachusetts, where researchers evaluated cancer incidence at the ZIP code level from 2000 to 2018. Published in an environmental health journal, this analysis used an inverse-distance-weighted proximity metric to nuclear plants within 120 km and adjusted for fine particulate matter (PM2.5) pollution alongside multiple sociodemographic covariates. The finer geographic resolution allowed the team to detect patterns that county-level data might blur. ZIP codes closer to plants like the now-shuttered Pilgrim Nuclear Power Station showed significantly higher cancer incidence, with the researchers estimating roughly 20,600 attributable cases statewide and an attributable fraction of about 3.3%.
One of the more striking details from the Massachusetts analysis is the distance gradient. Cancer risk dropped off beyond approximately 30 km from a plant, suggesting a geographic threshold where the association weakens. The study also broke results down by sex and age group, finding that cancer risk near nuclear plants generally increased with age. That age pattern aligns with what epidemiologists would expect if chronic low-level exposure were a contributing factor, since older residents would have accumulated more years of proximity. The Massachusetts Cancer Registry, a state-run surveillance system that tracks diagnoses across consecutive five-year periods from 2000 to 2020, provides an independent data source that could support further local verification of these patterns around communities near Pilgrim and the former Yankee Rowe site.
What the Studies Cannot Prove
Both papers are observational, and their authors are explicit about the limits of that design. A correlation between proximity and cancer rates does not establish that radiation from normal plant operations is the cause. Nuclear plants operate under strict emission limits set by the U.S. Nuclear Regulatory Commission, which maintains an official roster of operating power reactor units and monitors compliance. Routine releases of radioactive material from properly functioning reactors are generally considered to be far below thresholds associated with health effects in existing regulatory science. Critics of proximity studies have long argued that confounders such as industrial activity, chemical pollution from co-located facilities, or population self-selection could explain elevated cancer rates near power plants.
The Harvard researchers addressed several of those objections by adjusting for socioeconomic, behavioral, environmental, and healthcare factors. Still, no observational study can fully eliminate unmeasured confounders. The national analysis, for instance, operates at the county level, a geographic unit large enough that exposure misclassification is a real concern. The Massachusetts study’s ZIP code resolution is tighter, but ZIP codes are postal constructs, not epidemiological catchment areas. Both studies also end their observation window in 2018, meaning they do not capture any changes following recent plant closures or shifts in reactor operations. The Pilgrim plant, for example, shut down in 2019, and no post-closure cancer data from that period is included in either analysis.
Older Reactors and the Question of Legacy Exposure
One thread worth pulling from the data involves reactor age. The U.S. nuclear fleet skews old, many of the plants operating during the 2000 to 2018 study window had been running for decades, and older facilities may have different emission profiles than newer designs. The federal energy statistics on nuclear generation and capacity by state show that a relatively small number of large reactors account for most nuclear electricity, often at sites that began operating in the 1970s and 1980s. A geocoded dataset of operable plants derived from federal reporting forms provides coordinates and plant attributes that could support further distance-based analysis, including tests of whether older reactors are more strongly associated with nearby cancer rates than newer units.
Age matters because reactor designs, fuel management practices, and regulatory standards have evolved over time. Earlier generations of plants were built under different assumptions about acceptable releases and long-term waste handling, and they have accumulated decades of routine emissions, maintenance activities, and unplanned incidents that do not necessarily rise to the level of reportable accidents. The Harvard studies did not explicitly stratify their analyses by reactor age or design type, but their exposure metrics implicitly capture legacy exposure in communities that have lived near the same plant for many years. Future research could merge reactor vintage data with cancer registries to test whether the strength of the association between proximity and cancer varies with how long a reactor has been in service or whether it has undergone major uprates or refurbishments.
Weighing Nuclear Risks Against Other Energy Choices
The renewed focus on health near nuclear plants comes as policymakers weigh nuclear power’s role in decarbonizing the grid. Proponents argue that nuclear generation delivers large quantities of low-carbon electricity and can displace fossil fuel plants whose emissions of particulate matter, nitrogen oxides, and other pollutants are well-documented drivers of cardiopulmonary disease. A broad review of air pollution health impacts underscores how strongly fine particles and related pollutants are linked to mortality, highlighting the public health gains that can come from phasing down coal and oil. In that framing, nuclear power’s benefits are measured not only in avoided carbon dioxide emissions but also in reduced deaths from dirty air.
The Harvard proximity studies complicate that calculus, suggesting that communities near reactors may face their own, more localized cancer burdens. Unlike the diffuse harms of fossil fuel pollution, any nuclear-related risk is geographically concentrated around plant sites, often in coastal or riverine towns with long histories of industrial activity. For residents in those areas, the question is not abstract: they must weigh the climate benefits that nuclear plants provide to the wider grid against potential health costs borne close to home. The new findings do not offer a definitive answer to whether nuclear power is “safe enough,” but they do indicate that health risk assessments should be updated with contemporary data and that plant siting, relicensing, and decommissioning decisions should account for emerging evidence on cancer patterns in neighboring communities.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
