Sunday, May 16, 2021

U.S. Excess Mortality was Not Higher in 2017

Yesterday over at Millenial Star, Geoff B. asserted that, according to a recent study, the U.S. had higher excess deaths in 2017 than in 2020. He then went on to bemoan government efforts to stop the spread of COVID and asserted that all of the worry was just media and political hype, since 2020 was less deadly than a normal year. Why, he asks, did we fall for the propaganda?

The claim that the U.S. had higher excess death in 2017 seemed unbelievable to me, but at first glance that's what the paper seemed to say. Looking closer, however, it appeared that Geoff--or whoever brought it to his attention--misunderstood the point of the study. Comments at MS are moderated and unfortunately it appears that, rather than argue the merits, my comment was rejected for unknown reasons. Luckily I saved a copy, which I have reproduced below in its entirety. Fewer people will see my explanation here, but I still think it's worth posting.

I don’t think the authors are saying what you think they are saying.

What you think they are saying: U.S. excess deaths in 2017 were greater than deaths from COVID in 2020, therefore 2020 was no big deal in comparison to 2017.

What I think they are saying: Using Europe as a mortality reference, since 2000 the U.S. has a trend of increasing excess deaths, to the point that in 2017 the U.S. lost an excess of people (vs Europe) comparable to COVID in 2020. In other words, mortality in the U.S. sucks compared to Europe. The fact that much of that excess mortality occurs in younger people, resulting in higher years of life lost (YLL) compared to the 2020 pandemic simply highlights the tragic nature of the status quo.

I offer three items in support of my view:
1. The method section states: “We estimate the number of US deaths that would not have occurred at age x, year t if the United States had the set of age-specific death rates of the European standard…”
2. A university press release that quotes the authors.
3. This CDC dashboard (scroll to the bottom) where the effect of COVID compared to other years (including 2017) is visually obvious. If you play with the options, you can find that the total U.S. predicted excess deaths (in comparison with itself, not Europe) for 2020 is 570,621 – 696,637.

So to to re-state, what we consider to be a normal number of deaths is actually an excess of deaths when compared to Europe, and the magnitude of that difference is comparable to 2020 U.S. COVID deaths.

Below is a picture from the CDC dashboard I mentioned. It shows U.S. weekly deaths beginning in 2017, with COVID deaths in blue and non-COVID deaths in green. The orange line indicates the threshold for excess deaths. Bear in mind that the highest point of the peak on the right marks the beginning of 2021, so the nunmbers for 2020 don't tell the whole COVID story. Moreover, based on provisional data, there were 3.36 million total deaths in 2020 compared to 2.81 million in 2017, with COVID as the third leading cause of death.

9 comments:

Tygan,  5/16/2021 02:23:00 PM  

Thanks for the clarification!

Quantumleap42,  5/16/2021 06:36:00 PM  

Hmmm, my comment never went through either. Mysterious.

Last Lemming,  5/16/2021 06:42:00 PM  

Thank you for posting this. I don't even bother trying to comment at M* any more and I was about to suggest that you take this on when I saw the post.

You have have explained what the study was trying to accomplish, and I found that helpful because it was not obvious to me. What was obvious to me was that for the purposes M* was assuming, the cited study was simply not useful. They were comparing excess deaths in 2017 with COVID-attributed deaths in 2020. That is apples and oranges. If you compare excess deaths (as the study defines) them in both years, 2020 loses badly.

But what if M*'s interpretation had been correct? Well that would mean that all of the mitigation efforts made in 2020 worked--the exact opposite conclusion M* wants one to reach. But alas, they did not work THAT well.

Jared*,  5/16/2021 07:27:00 PM  

Quantumleap42:

Mysterious indeed. The charitable side of me wonders if there was a technical glitch, but a subsequent comment lauding him for having the courage to post on something NO ONE is talking about made it through.

Last Lemming:

I had the same thought initially--that perhaps all of the mitigation efforts actually reduced total deaths. Unfortunately, no.

I think the study could have been worded more clearly. Once I realized what they were doing, it all made sense.

Anonymous,  5/17/2021 05:17:00 AM  

Another thank you for your explanation. I didn’t bother to click through to M* from the headline at Mormon Archipelago because I knew the post, whatever it was, would be untrue. When I saw you posting on the same topic, I immediately clicked through because YOU always know what you’re talking about.

Jared*,  5/17/2021 08:32:00 PM  

[Blush] Thanks for the vote of confidence, Anonymous. I'm as fallible as anyone else, but I try to keep a healthy respect for the limits of my knowledge.

gomez,  5/20/2021 07:20:00 AM  

Hi Jared, I've only just seen your blog but just wanted to highlight that I've had the exact same experience as you trying to comment on this particular blog post. I've tried posting a couple of times, tried emailing the admin a couple of times but no response.

My comment was inline with the M* comment policy, still, I can understand not wanting to post my comment (or yours) as it makes Geoff look a little silly for clearly misreading the PNAS study. But at least take the post down or modify it to correctly reflect the data. But instead, he allows other comments through reinforcing his incorrect reading of the study. Not only that, he has taken one of those comments as the basis for a new post in which he bemoans unscientific (irony alert) promoters of lockdowns and says one of the ways to combat that is to engage on blogs/social media. Ha.

Ideology is obviously much more important than factual accuracy.

Hunter,  5/20/2021 04:19:00 PM  

Re: M*: it's ironic, all the talk of cancel culture by the political right, that the admins at M* are so afraid of dissenting opinions or countervailing facts. It shows a fragility to their arguments.

Anyhow, to this post's point, thanks for taking some time to offer your perspective here.

Jared*,  5/20/2021 07:50:00 PM  

gomez, thanks for the report. The funny thing is even if he admitted he was wrong, he could still maintain his policy position since it is rooted in values more than empiricism. Earlier this year I and others took issue with something Geoff posted and, to his credit, he deleted it from his post since he considered it a distraction from his main point. On the other hand, he cleansed the comments of any complaints, so now it looks like nothing was ever amiss. I considered it a small win nonetheless.

Hunter, the opinions expressed at M* can be controversial, so I can certainly understand them not wanting to have interminable arguments over everything they post. But if they want M* to be more than just another propaganda mill, they'll need to be a little more responsible.

  © Blogger templates The Professional Template by Ourblogtemplates.com 2008

Back to TOP