The abstract is read at the publisher; the summary is JClub's.
After re-parsing all 299 CDC NNDSS weekly tables (2019–2024), the flood–legionellosis association is null on corrected U.S. surveillance data (IRR 1.015, 95% CI 0.981–1.051; p=0.39).
- Why it matters: Prior national analyses can be distorted by surveillance-data extraction errors and shared seasonality, potentially producing spurious flood–health links.
- What they did: The study re-extracted 15,197 state-weeks and 11,676 legionellosis cases, reconciled 299/299 weeks to printed totals, then fit fixed-effects distributed-lag Poisson/negative-binomial models with state and year-week effects, Fourier seasonality, temperature terms, and permutation/event-study diagnostics.
- The result: The earlier reported lag-0–5 IRR 1.11 (95% CI 1.06–1.17) was driven by inflated 2023 counts from column misreads (weekly national totals roughly 300–400 vs 15–46 after correction), implying the corrected association is not supported.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.