Improvement Reframed as Decline: A Critical Review of the Mangano Report on Diablo Canyon and Infant Health

What the data actually show

The Mangano report argues that San Luis Obispo County’s infant mortality moved from about 16% below the California average before Diablo Canyon opened to about 1% above it in 2010–2024, and presents that shift as evidence of harm. It never shows the actual rates behind the comparison.

Those rates fell by more than half. A child born in the county today is more than twice as likely to survive infancy as a child born before 1985. The “worsening” finding comes from substituting a relative comparison for an absolute one: asking not whether the county improved, but whether it improved as fast as the state. A county already performing well before the plant opened has less room to improve, and convergence toward the average is consistent with regression to the mean, demographic shifts, and differential improvements in prenatal care.

The 2019 rate rests on 12 deaths, with a 95% confidence interval of [2.51, 8.50] — a range of nearly 6 per 1,000. The report does not acknowledge that uncertainty.

The dose the report omits

Under 10 CFR 50.36a(a)(2), PG&E measures and reports the radiation dose delivered to the public each year. For 2024 the total was about 0.002 mrem: 1.76 × 10⁻³ mrem from gaseous effluents and 2.17 × 10⁻⁴ mrem from liquid effluents. That is 0.07% of the NRC’s ALARA target and 0.008% of the federal public dose limit. The NRC filing evaluates the dose outside the site boundary as negligible. A single cigarette delivers a dose roughly 900 times higher than a year of Diablo Canyon’s releases.

Where the supporting claims fail

The birth-defects claim rests on 73 cases over nine years with no time series, no independent verification, no geographic breakdown within the county, and no accounting for documented variation in provider reporting. The county rate of 3.31 per 1,000 births is an order of magnitude below rates reported in the European countries with the lowest recorded incidence, which points to reporting methodology rather than a local hazard.

The strontium-90 claim draws entirely on Mangano’s own unreplicated 2003 study, which found California to have the lowest Sr-90 levels of the five states sampled. The 2026 report cites the 31% within-California difference for San Luis Obispo and Santa Barbara counties (127 mBq/g Ca) while omitting that the figure sits below Pennsylvania (154), New York (138) and Florida (130). The same study measured 106 mBq/g Ca near Rancho Seco, closed twelve years before sampling — the lowest reading in the study, and incoherent with the emissions hypothesis.

What rigorous research would require

• Individual-level exposure estimates based on residential proximity, wind patterns, dietary pathways and duration of residence — not county aggregates.
• A matched comparison group of counties with similar demographics, healthcare access and economics that differ mainly in proximity to a plant.
• Time-series data with a genuine pre-startup baseline, to show outcomes changed after operation began rather than were always elevated.
• Independent verification of birth-defect reporting, which is sensitive to provider practice.
• Quantified adjustment for confounders — maternal tobacco use, agricultural chemical exposure, prenatal care access — and a dose-response analysis within the county.

The review notes what the report gets right: Diablo Canyon does release radioactive material, documented in NRC filings and not in dispute; developing infants are more radiosensitive than adults; and updated federal epidemiology near nuclear plants would be valuable, since the National Cancer Institute study most often cited dates from 1990.