Read on Twitter
There's been a lot of discussion about "excess mortality", which I've mentioned previously. A short thread about some limitations. https://twitter.com/peripatetical/status/1253115315677888514
This method has been traditionally used for influenza, by looking to see how many more deaths occur than is expected based on a baseline and markers of influenza activity https://pubmed.ncbi.nlm.nih.gov/19374272/
When looking at excess mortality potentially attributed to COVID-19, it is possible that we may be either underestimating or overestimating COVID mortality. https://www.ft.com/coronavirus-latest
Underestimation may occur because some causes of death may be less common at this time, particularly road trauma. Additionally, there may be delays in death occurring, or being registered or reported.
Overestimation may occur because not all deaths that occur at this time are a direct result of COVID-19. So many things are different now than at this time last year.
For example, some people may not be coming to hospital with heart attacks or strokes and may die due to delayed treatment. This might still be regarded as an indirect consequence of COVID, even though they didn't have infection.
In Australia, studies have looked at excess mortality for influenza, but regular data are only published for NSW. The last available report was from March and there isn't any signal of excess mortality
https://www.health.nsw.gov.au/Infectious/Influenza/Publications/2020/march-influenza-report.pdf
https://www.health.nsw.gov.au/Infectious/Influenza/Publications/2020/march-influenza-report.pdf
A phenomenon known as mortality displacement can occur, where a period of higher mortality may be followed by a period of lower mortality. This may occur because deaths in the frail elderly are only brought forward by a few months.
Studies in influenza haven't conclusively demonstrated mortality displacement (also awfully termed "harvesting") but this may be different for COVID. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6462785/
Thus, reported deaths with confirmed COVID might be regarded as a minimum estimate, and estimates of excess mortality are probably closer to a maximum estimate of COVID mortality.
Excess mortality probably includes acute deaths indirectly caused by COVID due to disrupted health services, but still may not capture them all (eg disrupted preventive health or primary care that may lead to delayed cancer diagnoses)
