Defining influenza epidemic zones through temporal clustering of global surveillance data

Five influenza epidemic zones were identified from 111 countries (2011–2019 and 2024) using temporal clustering of surveillance detections, revealing distinct seasonal timing patterns.

• Why it matters: Countries need better models of when influenza epidemics occur to time surveillance and public health actions and reduce influenza burden.
• What they did: Weekly influenza-positive test time series were smoothed and imputed with country-specific GAMs, then grouped via ensemble consensus k-means clustering; remaining countries were assigned by population-weighted latitudinal distance.
• The result: Zones A and B peaked November–May (northern-like), D and E peaked June–November (D) and May–November (E) (southern-like), and zone C peaked September–March, with most zone C countries in the tropics.

AI summary of the abstract · may contain errors · not medical advice · read the paper.