Applying machine learning to identify unrecognized COVID-19 deaths recorded as other causes of death in the United States, 2026, Kiang et al.

[Chandelier]

Applying machine learning to identify unrecognized COVID-19 deaths recorded as other causes of death in the United States

Spoiler: Authors

Kiang, Mathew V.; Li, Zehang Richard; Wrigley-Field, Elizabeth; Raquib, Rafeya V.; Lundberg, Dielle J.; Paglino, Eugenio; Huynh, Benjamin; Bibbins-Domingo, Kirsten; Glymour, M. Maria; Stokes, Andrew C.

Abstract
The actual number of US deaths caused by severe acute respiratory syndrome coronavirus 2 infection has been investigated and debated since the start of the COVID-19 pandemic.

Here, we use machine learning trained on US death certificates from March 2020 to December 2021 to predict 155,536 (95% uncertainty interval: 150,062 to 161,112) unrecognized COVID-19 deaths.

This indicates that 19% more COVID-19 deaths occurred in the US than officially reported.
Predicted unrecognized COVID-19 deaths occurred disproportionately among decedents with less than a high school education; decedents identified as Hispanic, American Indian, Alaska Native, Asian, and/or Black; counties with lower household incomes and worse preexisting health; and counties in the South.

These findings suggest that the US death investigation system undercounted COVID-19 deaths unevenly, hiding the true extent of inequities.

Web | DOI | PMC | PDF | Science Advances

 

[Arnie Pye]

This is suggesting that the US and the UK had almost completely the opposite experiences of recording causes of death during peak Covid.

Here in the UK, people with, for example, terminal cancer who died during pandemic/lockdown were often recorded as having died of Covid rather than their cancer. Also, under normal circumstances influenza kills some people every year - but it almost vanished during Covid according to the death statistics here.

I have always thought that the accuracy of cause of death might have been much lower than usual during Covid, but probably isn't very accurate at any time.

I got the impression at the time that the UK logic involved was :

1) Record as many deaths as possible as being caused by Covid.

2) Then frighten the population with the numbers to get them to stay at home.

I remember reading something about the common "normal" or "ordinary" causes of death in the UK before the pandemic that allegedly almost disappeared when Covid was doing the rounds. And I simply don't believe that.

 

[Yann04]

Also, under normal circumstances influenza kills some people every year - but it almost vanished during Covid according to the death statistics here.

Well Influenza is less contagious than COVID and there were lockdowns and social distancing measures that likely made influenza rare from 2020-2022.

I don’t know much about the UK so perhaps you are right they overcounted. But in most of the world excess deaths were (perhaps still are? I don’t know) very significantly higher than baseline even when correcting for official COVID deaths so there was likely undercounting.

 

[Chandelier]

Also, under normal circumstances influenza kills some people every year - but it almost vanished during Covid according to the death statistics here.

According to these graphs, there’s something going terribly wrong since Covid that might influence the historical assessment on the amount of statistical overcounting that has occurred.

The graphs are from the following posts:
Life expectancy
Sickness days/Retirement/Death in service

 

[Arnie Pye]

I don’t know much about the UK so perhaps you are right they overcounted. But in most of the world excess deaths were (perhaps still are? I don’t know) very significantly higher than baseline even when correcting for official COVID deaths so there was likely undercounting.

I never thought that total deaths from any cause were counted wrongly during Covid (or at any time). It's the actual causes of deaths that I have doubts about. For example, I was with my father in a geriatric hospital and witnessed his death. (It was long before Covid.) I was surprised when I saw his death certificate - it didn't seem to match his actual cause of death at all or what I had witnessed. But since his death was expected and I thought he had died of one (common) thing and the certificate said he died of another (common) thing I didn't think there was any point in querying it.