Great site.Eugene Wrayburn wrote:
The I can't get the numbers to exactly agree to the ONS England and Wales numbers (which is slightly annoying), but they're basically the same - within ~1k.
What this also gives is the option of using the "adjusted five year average", ie "National five-year average data for each week and place, scaled down by a factor (XXX) to fit the observed total number of deaths nationally over the first 10 weeks of the year. This can be thought of as an 'expected' trend, given what was observed in the first 10 weeks of the year, and assuming the trend over time follows the five-year average. This baseline will produce larger excess deaths than the first method using unadjusted five-year average deaths."
This is useful because, the first 10 weeks of the year (ie those before any official Covid-19 deaths) had lower all-cause deaths than the 5-year average, suggesting that had there been no Covid-19, the deaths for the year would have been lower than that average. Adjusting for this adds about 12,000 deaths to the total, so around 71k deaths to 29 May (and that's only Eng & Wal). Obviously this method can be debated but it does suggest that the ~60k excess deaths is an underestimate of the Covid-19 impact.
