Did we overcount COVID deaths?

Data and Metrics Data Literacy Infectious Diseases

No. In the US, we likely undercounted COVID deaths.  

A common refrain during the pandemic was that people were dying “with” rather than “of” COVID. The implication was that our official statistics were wrong and we were overstating the true danger of the pandemic.

As we wrote about way back in 2020, data nerds like demographers and epidemiologists have tools to overcome the challenge of causes of death not being coded perfectly, especially during a crisis like a pandemic or natural disaster. The trick is to look at deaths from any cause (which we are good at counting), and tally how many deaths we saw above and beyond what we would normally expect. This “excess mortality” gives us a much better picture of the true mortality burden of a crisis like COVID-19.

For most of the pandemic, the US had higher excess mortality than official COVID deaths. A new paper in the Proceedings of the National Academy of Science looked more closely at this data from March 2020 to August 2022. The authors calculated roughly 1.2 million excess deaths due to natural causes over this time (so excluding accidents, homicide, drug overdose), of which 162,886 were not classified as COVID-19.

It’s possible that some excess deaths not counted as COVID were due to “indirect” impacts of the pandemic such as overstrained hospitals or people delaying needed health care. The study looked at the timing of non-COVID excess deaths across regions and counties in the US and found that they mostly moved in sync with spikes in official COVID deaths (happening just before or during COVID waves). This pattern strongly suggests that these deaths were underrecognized COVID deaths.

The study found that undercounted COVID deaths were more common in non-Metropolitan counties of the US, as well as in the South and West:

In this figure, the darker shaded counties had a higher proportion of (seemingly) undercounted COVID deaths compared to official COVID deaths. This geographic variation could be due to limited COVID-19 testing, more deaths outside of hospitals, and/or local differences in death registration by medical examiners vs coroners. Politicalization and stigma against COVID may also have affected the official counts by geography. Read a further summary of the new study here.


We have good evidence that COVID deaths were undercounted rather than overcounted during the pandemic. With over 1.2 million excess deaths in the US through August 2022, we can’t minimize the devastating human toll of the pandemic.

Link to Original Substack Post