Weekly mortality and excess/deficit deaths by country and area
Model:
Rendering observations…
Weekly baselines use only STMF-derived weekly values, never the historical monthly supplement. The five-year average is accompanied by the same-week minimum-to-maximum band; Farrington-style uses ±3 weeks; EuroMOMO-style uses W16–25 and W37–44 as baseline weeks. The two lower panels show the excess trend and cumulative excess using the same colors: light red is above baseline; dark red is above prediction limit. With mortality deficits enabled, negative differences are blue and reduce the cumulative mortality difference. Observed and standardized deaths are shown as counts. Crude and age-standardized mortality rates are shown per 100,000 population. Birth-related mortality rates are shown per 1,000 births or deliveries. The age-0 population rate is shown per 100,000 age-0 population and uses a different denominator from birth-related mortality rates.
Countries and areas covered
Deaths, crude rates, and age-standardized rates: 39 countries or areas
England and Wales (GBR), Sweden, United States of America: Uses weekly deaths and mortality rates from HMD STMF.
Terms and aggregation
HMD
Human Mortality Database, which provides harmonized mortality and population data for participating countries. https://www.mortality.org/
STMF
Short-Term Mortality Fluctuations, the weekly mortality dataset provided by HMD. This page uses its age-specific weekly deaths and mortality rates. https://www.mortality.org/Data/STMF
UN WPP
United Nations World Population Prospects, published by the UN Population Division. It supplies annual population data where needed. https://population.un.org/wpp/downloads
Deaths in a week crossing a year boundary are allocated in proportion to the number of days in each period. Population inferred from weekly deaths and rates is accumulated as population-days. Only complete periods from W27 or W36 to the day before the same starting week in the following year are shown. Japanese weekly values are estimates prorated and smoothed from monthly e-Stat data, not directly observed weekly counts.