When COVID-19 Deaths Go Up, Other Deaths Go Down (IMPORTANT CORRECTION)
UPDATE: This article has been significantly revised since its original publication. Specifically, the CDC numbers used in the original article were incorrect. Below is first the original article and at the bottom the new numbers with an explanation why the original numbers were incorrect. The new numbers do not support the thesis of the article nearly as well as the old numbers.
The result is that the overall numbers of weekly deaths are roughly the same in 2020 (this suggests a roughly 5% increase but see this criticism) as they were in the 4 years prior. (Do note that the 2020 numbers for the most recent weeks may still be revised upwards)
As to why this is so, there are likely multiple reasons, but part of the explanation may have to do with the way COVID-19 deaths are counted, as described by the Illinois Director of Public Health, Ngozi O. Ezike, in this (frankly shocking) video: If someone dies and at the time of death was also known to have tested positive for COVID-19 they are automatically counted as a COVID-19 death. In other words, deaths are listed as COVID-19 deaths even when the cause of death was clearly something else.
If infection rates in a population are at 10% for example, that will move 10% from the Other Deaths category into the COVID-19 Deaths category. Depending on whether people who die are more or less likely to test positive for COVID-19 than the general population, that 10% could be higher or lower.
Per week, starting in Week 9 when the US recorded its first COVID-19 death:
- Blue is COVID-19 deaths
- Green is Other Deaths
- Red is Influenza Deaths
- Yellow is Pneumonia Deaths
- Orange is Total Deaths
The CDC data used in the original article are misleading. The article assumed ‘COVID-19’ deaths and ‘Pneumonia deaths’ were mutually exclusive, i.e. that ‘COVID-19 deaths’ has only those deaths that list COVID-19 as primary cause of death while ‘Pneumonia deaths’ has only those deaths that list pneumonia as primary cause of death. But this is false: A COVID-19 death can be counted in both the COVID-19 category and the Pneumonia category. So the proper way to count ‘Other Deaths’ is not to take ‘Total Deaths’ and subtract ‘COVID-19 deaths’ and ‘Pneumonia deaths’ and ‘Influenza deaths’ from it, but to take ‘Total Deaths’ and subtract the category ‘Pneumonia Deaths or Influenza Deaths or COVID-19 Deaths’ from it. The new results are below.
To be sure, it is still quite possible that a substantial part of the increases in total deaths are the result of increases not in deaths from COVID-19 but deaths with COVID-19, and hence that the problem I described in the original article — how deaths are counted as COVID-19 deaths even though the actual cause of death was different — is still inflating the number of COVID-19 deaths. It’s just that the category of COVID-19 deaths does not seem to be displacing ‘Other Deaths’ to nearly the same extent as originally thought.