STUMP » Articles » How Many COVID Deaths Have There Been in the United States? » 15 January 2023, 14:21

Where Stu & MP spout off about everything.

How Many COVID Deaths Have There Been in the United States?  


15 January 2023, 14:21

About 950,000.


Okay, that was the deaths reported to the CDC through December 17, 2022, and uploaded to the CDC WONDER database as of January 2, 2023.

How difficult is it to count the dead?

Now, you may be thinking that is a lot less than you had heard, and this is in light from this op-ed piece in the Washington Post:

Leana Wen: We are overcounting covid deaths and hospitalizations. That’s a problem.

According to the Centers for Disease Control and Prevention, the United States is experiencing around 400 covid deaths every day. At that rate, there would be nearly 150,000 deaths a year.

But are these Americans dying from covid or with covid?

Understanding this distinction is crucial to putting the continuing toll of the coronavirus into perspective. Determining how likely it is an infection will result in hospitalization or death helps people weigh their own risk. It also enables health officials to assess when vaccine effectiveness wanes and future rounds of boosters are needed.

Two infectious-disease experts I spoke with believe that the number of deaths attributed to covid is far greater than the actual number of people dying from covid. Robin Dretler, an attending physician at Emory Decatur Hospital and the former president of Georgia’s chapter of Infectious Diseases Society of America, estimates that at his hospital, 90 percent of patients diagnosed with covid are actually in the hospital for some other illness.

“Since every hospitalized patient gets tested for covid, many are incidentally positive,” he said. A gunshot victim or someone who had a heart attack, for example, could test positive for the virus, but the infection has no bearing on why they sought medical care.


There is a reason I had to switch to total deaths, all causes, back in April 2020, and it was because of arguments like this.

Anyway, I did a search in Wen’s article for “underlying” as in “underlying cause of death” as opposed to “contributing cause of death”.

It’s not in there. I will come back to this issue in a moment, because let me go to the various COVID death trackers I can find out there.

CDC’s Multiple COVID Trackers

I have mentioned this before, but I do not use media-friendly trackers.

Mainly because I like to grab my own data, but also because it’s focused on “newly reported” deaths, and I HATE THAT CRAP.

But here’s the total from the CDC’s media-friendly COVID tracker.


I grabbed that this morning, Jan 15, 2023.

Here is the not-so-media friendly dashboard that I sometimes use for work, as I know where this data comes from (but this data also has trouble).

Now we start to see the issue:

Don’t see?

Do you see it now?

Yeah, that contributing cause of death.

That’s why my number is about 10% smaller than the “official” number – not merely that it lags about a month from theirs.

All of the ranking tables and mortality analyses I do, I almost always focus on the underlying cause of death. That is, the single cause of death each death certificate must have – the main reason the person died.

There is one and only one underlying cause of death per death certificate. There can be up to 20 contributing causes of death. (I will come back to that link, because I talk about what counts as a COVID death.)

Indeed, I can think of only one set of posts where I looked at contributing causes of death for analysis, and that was alcohol-related deaths. That’s because, obviously, if you’re in a motor vehicle accident caused by alcohol – yes, one needs to dig into contributing causes of death there.

But other than that, it’s the underlying cause of death I care about.

How high is the overcount?

Let’s look at the overall total of deaths that had COVID as underlying cause or contributing cause, to match my original data pull:

So that’s 1,081,858 deaths total, and my original count for underlying cause of death was 947,961.

There is a difference of 133,897 deaths, so that’s actually 14% of the actual COVID deaths (that is, the deaths with COVID as underlying cause of death.)

According to the data pull I’m looking at, 87% of the deaths I’m seeing with COVID somewhere on the death certificate have COVID as the underlying cause of death. That number is not greater than 90%. Hmmm.

Could we get a breakout on that?

Looking at COVID deaths where COVID was a contributing cause

But let us consider the claim here — that COVID is even being overcounted as a contributing cause. Because that’s what Wen is really arguing as well.

I can see a COVID case aggravating other conditions, such as heart disease or diabetes, but homicide? (I will show the counts in a moment)

I can see trying to argue a cause-effect chain of COVID and homicide — say somebody has shot you, the fact that you have COVID may make you frailer or more difficult to treat for the trauma, which contributes to your death. Obviously, the gunshot itself would be the underlying cause of death, but if COVID made it more likely for you to die (lower oxygen levels in the blood, blood clots complicating treatment, etc.), it absolutely should be considered a contributing cause of death.

But let us look at the totals (I will attach the spreadsheet to the substack version of the post for people to look at all the data) – I will just show the top ten sub-chapters from ICD for the underlying cause of death, where they had COVID as a contributing cause of death:

The top three all seem likely – ischemic heart disease (blood clots from COVID could have caused trouble), cancer (difficulty with treatments, COVID making them more frail), and the third is the subchapter that Alzheimer’s falls in. Again, COVID would make someone who is already vulnerable more vulnerable.

I pulled detailed data – it’s in the spreadsheet, but not showing it in this post – but the number four item is essentially non-Alzheimer’s dementias, the next several also make sense… then we hit number 10.

Within that subchapter are a bunch of unrelated accidental causes of death, alas, but the two largest groups contributing to this being number 10 are accidental drug overdoses (only about 600 people, though) and falls (about 1400) – that tends to be old people. Again, what that could be related to is the inability to rescue the people once found, due to their COVID.

I think it’s just fine to have these multiple causes of death on death certificate, when properly detailed. There has been an increasing number of causes of death listed on certificates over time, perhaps due to more detailed standards, but I think because people are older and older at death, and tend to have multiple conditions when they die. Before the pandemic, the average number of causes of death per death certificate was over 3.

If COVID did complicate somebody’s death due to something else, we absolutely want to know that.

But we don’t want to confuse that with a death directly caused by COVID.

Some recommendations for the CDC

So this ends, once again, with me annoyed with the CDC.

Hey, CDC, you want to rebuild credibility and trust?

Here are my recommendations.

1. Split out the actual COVID deaths (COVID as the underlying cause of death) from the ones where COVID is a contributing cause.

2. If you want to report that second number, fine. Report that. Separately.


If you’re going to do that, you may as well go with all deaths from all causes.


I do not care that you think the media need the “newly reported” deaths. No, they do not.


Report the deaths properly, by when they occurred. Explain to the media what that means, and ways they can report on that. The media have issues with quantitative trends, so maybe you can help with that.

And if you don’t know how to explain quantitative trends, CDC, find some people who understand that. That should be part of your rebuild. Less emoting and trying to scare people, more teaching people in a way they can understand. Figure that out.

The CDC knows all the ways “newly reported deaths” distort reality, such as artificial spikes when people are off work for holidays and then report a bunch of deaths when they deal with backlog.

They also know you get absurdities like negative reported deaths when you have to redefine the deaths you’re reporting, as I’m recommending they do here.

Having to back out over 100,000 deaths from your headline number would take a lot of explaining, don’t you think?

Related Posts
The Return of Mortality Monday and RIP, President GHW Bush
Preliminary look at 2022 U.S. Total Mortality - Continuing Excess Mortality
COVID Mortality Idiocies: Irrelevant Comparisons, Bad Context, and Under-estimating Deaths