"A closer look at U.S. deaths due to COVID-19" retracted Johns Hopkins journal article

"The CDC classified all deaths that are related to COVID-19 simply as COVID-19 deaths. Even patients dying from other underlying diseases but are infected with COVID-19 count as COVID-19 deaths. This is likely the main explanation as to why COVID-19 deaths drastically increased while deaths by all other diseases experienced a significant decrease.

“All of this points to no evidence that COVID-19 created any excess deaths. Total death numbers are not above normal death numbers. We found no evidence to the contrary,” Briand concluded.

In an interview with The News-Letter, Briand addressed the question of whether COVID-19 deaths can be called misleading since the infection might have exacerbated and even led to deaths by other underlying diseases.

“If [the COVID-19 death toll] was not misleading at all, what we should have observed is an increased number of heart attacks and increased COVID-19 numbers. But a decreased number of heart attacks and all the other death causes doesn’t give us a choice but to point to some misclassification,” Briand replied.

In other words, the effect of COVID-19 on deaths in the U.S. is considered problematic only when it increases the total number of deaths or the true death burden by a significant amount in addition to the expected deaths by other causes. Since the crude number of total deaths by all causes before and after COVID-19 has stayed the same, one can hardly say, in Briand’s view, that COVID-19 deaths are concerning."

----------------

The editorial board of this Johns Hopkins student journal retracted the article after they were confronted by faculty wokists and Covidians.  pl

https://web.archive.org/web/20201126163323/https://www.jhunewsletter.com/article/2020/11/a-closer-look-at-u-s-deaths-due-to-covid-19

Comments

[Serge]

If the data in the following link is correct the crude number of total weekly deaths in US has increased ~10%, with a peak of over 20% during the spring:
https://ourworldindata.org/grapher/excess-mortality-raw-death-count?tab=chart&stackMode=absolute&region=World

Maybe Briand has access to other figures.

[Caroline Collins]

There is a plausible explanation for the static rate of dying during a pandemic. Let’s say that normally, 100,000 people die in our imaginary country. A virulent disease takes hold and kills 10,000 in one year.

The fear of the disease causes behavioral changes. People stay home. Doctors cancel appointments. Prescriptions aren’t written or refilled. Only emergencies go to hospitals. Harfly anyone drives. Many are out of work.

Under those conditions, you eliminate many kidney toxicity deaths caused by prescription drugs, many lethal falls caused by hypotension caused by blood pressure and other drugs, all the hospital-acquired infection deaths, all the traffic deaths, and many exhaustion-related deaths. Probably only 50,000 die,

In the final counting of deaths in 2020, it is possible that the number of deaths will be lower than usual.

[Deap]

Which is more fatal - covid or attacks by wokists? Thanks from bringing this critical Johns Hopkins data analysis study back into focus. Why are wokists so invested in only keeping "covid" alive, well..... and deadly? Cui bono. Personal or political?

Camus, in The Plague, did deal with the psychological dependencies such devastations create in those who have a tenuous hold on life in the first place - the plague gave them meaning -- back in the 1950s' - another time of global transition post WWII.

[TravelFoodFitnessGuide]

or you can also read from doctors at https://www.factcheck.org/2020/12/flawed-analysis-leads-to-false-claim-of-no-excess-deaths-in-2020/

"Quick Take

An economics professor’s flawed interpretation of U.S. mortality data has prompted a viral, false claim that COVID-19 hasn’t led to more deaths than normal this year. In fact, multiple analyses have found there to be a higher-than-normal number of deaths during the pandemic — as much as 20%, according to some studies.

Full Story

Viral headlines and social media posts are propelling the erroneous claim that there have been “no excess deaths” in the U.S. this year, suggesting that concern over COVID-19 is overblown.

That’s wrong.

As we’ve previously explained, excess deaths are deaths above the number expected in a given time period.

The metric in the context of COVID-19 can help assess the impact of the pandemic, including by capturing direct and indirect effects of the virus.

It can account for unreported COVID-19 deaths and for other factors — such as people dying from other causes, say avoiding medical attention, as well as drops in other deaths due to pandemic-related restrictions.

The Centers for Disease Control and Prevention in October published a report that found that from late January through Oct. 3, there were an estimated 299,028 excess deaths.

Two-thirds were estimated to be directly attributed to COVID-19.

The COVID-19 death toll has surpassed 275,000, as of Dec. 3, according to Johns Hopkins University data. Cases are rising around the country, and projections anticipate that the death count will climb past 300,000 by the end of the year.

The viral claim about “no excess deaths” — which has been amplified by unreliable websites and disseminated on Facebook and Twitter — stems from an online presentation by an economics lecturer at Johns Hopkins.

But it’s flawed for multiple reasons, experts say.

In a Nov. 13 webinar, Genevieve Briand, the assistant director for the university’s Applied Economics master’s program, looked at select CDC data, such as weekly reported deaths, to reach the faulty conclusion that there is “no evidence that COVID-19 create[d] any excess deaths.”

“Total death number are not above normal death numbers,” she claims. “We found no evidence to the contrary.”

Her claims were then relayed in a story in the Johns Hopkins News-Letter,
a student-run publication,
which said the analysis showed COVID-19 “has relatively no effect on deaths in the United States.”

The publication later retracted the report. An archived version of the story, however, has been shared on Facebook 10,000 times according to CrowdTangle analytics data.

The student publication’s retraction also became part of the false narrative online, triggering claims that the university had published a “study” then retracted it.

“Johns Hopkins Study Mysteriously Disappears after it Revealed, In Spite of COVID, No More Deaths in 2020 Than In Prior Years,” read the false headline on the Gateway Pundit, a far-right website known for spreading misinformation.

Briand, however, confirmed in an email to us that she had not published her analysis anywhere apart from the webinar.

But Dr. Steven Woolf, a professor in the Department of Family Medicine and Population Health at the Virginia Commonwealth University, told us that “multiple studies have shown that total deaths in the US are about 20% higher than expected.”

Woolf and colleagues relayed as much in an article published in the Journal of the American Medical Association in October.

“Although total US death counts are remarkably consistent from year to year, US deaths increased by 20% during March-July 2020,” the authors wrote.

Studies from others, including a paper by two University of Oxford economists, have reached similar conclusions.

Woolf said in an email that the lecture video instead presented a “very unsophisticated analysis, in which the speaker uses crude death counts for one year (2018) as the basis for comparison and performs simple subtraction from 2020 values for her assertions.

Any reputable analysis of excess deaths is based on statistical modeling that computes seasonal averages over a period of many years, and this is done to adjust for random variation from year to year.”

“For example, our studies use Poisson regression modeling. Statistical modeling is necessary to know whether differences in death counts are statistically significant,” he said.

“The speaker admits to merely eyeballing the trend lines for 2020. Researchers prefer to use statistical modeling to know whether a supposed change in counts is random variation or a statistically significant trend.”

The CDC’s National Center for Health Statistics excess death analysis (which uses such modeling) has found deaths above normal rates week after week since the end of March — as readers can see in the center’s chart below.

Briand’s video doesn’t consider that analysis or any other previous studies on the issue.

Instead, she looks at some public CDC data and, in part, homes in on a spike in deaths in spring 2020, observing that it does not correspond with an increase in all deaths — such as heart disease, chronic lower respiratory disease, influenza and pneumonia

— as was the case during a spike in deaths in early January 2018.

“Our death number increased in 2020 and we have less heart attack than in 2018,” she said. “Where have all the heart attacks gone?”

But she’s misleadingly comparing different points in time for the two years.

Bob Anderson, chief of NCHS’ Mortality Statistics Branch, said in an email statement that the video’s claims are born out of “a fundamental lack of understanding of the seasonal nature of mortality.”

While deaths typically increase in the winter, peaking around the New Year, “[d]eaths then decline in the spring through the summer and then increase again in the fall through the winter,” Anderson said. “But, what we are seeing in 2020 with the COVID-19 pandemic is very different.

In the spring of 2020, during a period where we normally should be seeing declining mortality, deaths continued to increase and were at unusually high levels through the spring.”

The comparison of the two time periods used in the video is “not appropriate,” he said.

Anderson added that in 2018, “deaths followed the normal pattern (although at a higher level than normal).

In 2020, deaths did not follow the normal pattern … they should have been declining in the spring, but instead increased substantially.

In addition, we have an additional peak in deaths in late July/early August that seems to have been completely ignored in this analysis.”

“With regard to heart disease, it is true that heart disease deaths did not increase in April 2020 to the same extent that they did in January 2018,” he said.

“But, in January, heart disease deaths are normally at a peak.

"In April, heart disease deaths are normally on the decline."

"Even so, the number of heart disease deaths has been largely above normal throughout the pandemic. The same is true for other causes of death, especially diabetes, hypertension, and dementia.”

"So far this year, through the week ending Nov. 21, there have been 598,659 reported heart disease deaths in the U.S.
—
according to incomplete, provisional death count data published by the CDC on Dec. 3.

In all of 2018, there were 655,381 deaths from heart disease. "

[Walrus]

Col. Lang, I refer you to the CDC excess deaths data.

https://www.cdc.gov/nchs/nvss/vsrr/covid19/excess_deaths.htm

I also happen to live in the state of Victoria where, at great cost, we won the latest battle against covid (but not yet the war). We have had no new cases for thirty days.

Another effect of the covid precautions is to reduce the transmission of other diseases like flu.

By the way, our economy is bouncing back very strongly too.

My point is that the bug is real, its a menace because it is capable of overwhelming your health system, not because it is particularly lethal. Masks and behaviour changes can stop the virus.

[voislav]

The same data pulled from the CDC website shows the 2020 deaths from all causes in the US at 2.27 million by 11/21/2020 vs. 2.04 million in all of 2019. So her conclusion that there is no increase in deaths is wrong, probably because she started pulling out data subsets (such as by cause of death or age) that may or may not be reported reliably.

The expected increase in the number of deaths from all causes due to population increase and ageing is about 1%, and this is far in excess of that.

The best way to analyse this is to look at the largest dataset possible, which is all deaths, and see if that agrees with conclusions drawn from a subset. If not, which is the case here, then data selection introduced a bias. So the professor writing the paper messed up her methodology and had her paper rightfully pulled. Given this is a student journal, there is likely no peer review that would have caught the error.

This is nothing unusual, I'd say that about 20% of papers I get for review have serious methodology issues and require major revisions or all pulled before publication altogether. Normally this would not be news, but this year is what it is.

One thing to note is that the numbers for deaths from all causes on the CDC website seem to be lower than reported in other sources. I've seen the numbers of 2.17 and 2.39 million for 2019 and 2020, so I am not sure where the difference is coming from without doing a deeper dive into the counting methodology. All sources show a 220-240,000 difference between 2019 and 2020.

Source for the CDC numbers:
https://data.cdc.gov/d/muzy-jte6/visualization

[Yeah, Right]

The John Hopkins student journal article is mainly an interview with Briand following her presentation of that data at a John Hopkins webinar.

So one step removed from the presentation and, therefore, it is very difficult to judge the quality of that data.

But if you want watch the webinar itself, then you can do so here: https://www.youtube.com/watch?v=3TKJN61aflI

Note that it says "Comments are turned off."

Can't imagine why....

[Yeah, Right]

I am puzzled by this claim by Briand: "If [the COVID-19 death toll] was not misleading at all, what we should have observed is an increased number of heart attacks and increased COVID-19 numbers."

Pardon?

Every oldie who dies from Covid-19 is one less oldie who can't ever make it to the "heart attack" column of the Ledger Of Death.

Briand: "But a decreased number of heart attacks and all the other death causes doesn’t give us a choice but to point to some misclassification,"

See above. I think there are many other choices that could explain that observation.

I think an argument could be made that a pandemic should also lead to an increase in other causes of death (health care being overwhelmed, societal breakdown, etc.). Sounds reasonable.

But an argument could also be made for the opposite proposition: a pandemic can lead to a decrease in other causes of death because it snatches people away before they get to die of Something Else.

I don't know the answer.

Briand believes she does, so good luck to her.

[Barbara Ann]

That would sure help explain this:

"Update as of December 3, 2020: The model used to generate influenza in-season preliminary burden estimates uses current season flu hospitalization data. Reported flu hospitalizations are too low at this time to generate an estimate."

https://www.cdc.gov/flu/about/burden/preliminary-in-season-estimates.htm

[Polish Janitor]

In the early days of Covid-19, whenever reports like this emerged that were skeptical of the official 'Covid mass extinction' narrative they were condemned as conspiracy theories, but what about this? Is Johns Hopkins a conspiracy theory spreader now?

I have come believe that the whole Covid phenomenon had a lot to do with 'leveling the field' for 2020 elections in favor of Dems, since they could not under any regular scenario have been able to defeat Trump, so their effort in manufacturing 'Biden' as the guy who defeated Trump was nothing more than adding insult to injury. It is like, "see not only we can cheat on a massive scale, but we will also select a delusional and unpopular do-nothing hack as the President of the United States, just because we can and you just gonna have to accept it!"

[JohninMK]

Thank you for putting that link up. I meant to read it when it first came out but they deleted it very quickly before I could.

Clearly one of the most dangerous bits of research to the 'message' that had to be suppressed.

[fotokemist]

It was reported last week that the failure rate of minority students in Math is up 600% and in English 500%. Do these Covidians realize what the have done and continue doing? A couple of weeks ago a similar mortality report based on German data was published.

Meanwhile, our constitutional rights are being crushed to force us to engage in ineffective actions (porous masks, "social distancing") that benefit from neither theoretical nor experimental support, while effective actions (HCQ + Zn) are blocked. (If you have any doubt, look up the COVID-19 death rate in Uganda where HCQ is widely used to prevent malaria.)

It is becoming more difficult to remain optimistic about our country.

[Eric Newhill]

Yeah, Right,
I keep telling people that the correct metric is years of expected life lost; not a raw death count.

If it was done my way - the way that insurance actuaries look at these things - then a clearer picture would emerge. It takes care of the issues and questions you note.

Sigh - no one ever listens.

Also, even if one wants to do it wrong and compare death counts between years, then that should be done on an age adjusted deaths per 100,000 rate. The rate should be weighted for age because it's not just that there are more people in the country than previous years. The people are also older (the boomers aging). There is a much bigger difference in the increase of probability of death when someone moves from 75 years old to 76 than there is moving from 32 to 33.

If one insists on not looking at years of expected life lost, the age adjusted rate per 100K should be compared to several different years I the past. Deaths per 100K fluctuate quite a bit over the years. IMO, 2020 is going to have a rate fairly similar to 2010 if things keep trending as they are.

The CDC has some important and, frankly, strange caveats and notes at the bottom of its excess deaths data/graph. I think everyone should read them.

The CDC itself notes that 1/3 of their excess deaths in 2020 are not due to covid. It's right there on their page.

[Eric Newhill]

Barbara Ann,
One would think that if masks, social distancing and stay at home worked, there wouldn't be a flu season this year.

[Deap]

How many angels can dance on the head of a pin? Health care reduced to unsupported metrics is not health care; it is an indoor parlor game. Usually for profit.

WSJ today screams Sweden's covid experiment is a failure with red bars shooting off the graph - along with the US until you read the actual numbers that justify this high impact graphic and headline - Sweden hit by surge of cases - Sweden now reporting 5 deaths per million, compared to Denmark's 1 deaths per million.

And of course the leading bad guy is Trump's US with a staggering 6 deaths per million. According to this front page chart.

[FWH]

Glad that somebody has tried to separate covid deaths and deaths from other MEDICAL causes that we all know have been mixed. Explaining total excess deaths was probably beyond the scope of what was being done.

[FWH]

Anecdote: I once had a glamorized image of an Australian. I saw him as strong and independent; tough as a drought hardened roo or a salt water croc. I saw him as beholding to no man, including the English masters who once sponsored him. This image is gone now, washed away with the flood of news propaganda and the effect that I know that it has on people. I realize now that Australians are pretty much just like me.

[Mark K Logan]

The hospitals do list COVID as cause of death a bit more than it could be argued they should, but in their defense they do that typically. An example: Someone with type 2 diabetes is seriously injured by a bear and the hospital was unable to manage the person's blood sugar level through the emergency and extensive treatments...the cause of death would still be listed as "Bear attack".

[TV]

Science?
is this another field where "woke" people cheat their way to advanced degrees and use "science" to push their agendas?
Many "scientists" depend on grants - usually from the Fed or left wing foundations.
How much of their "science" is compromised by the money?
Remember the "scientists" (including some at the CDC) who signed a statement that burning, looting and rioting took precedence over COVID-19 social distancing and mask-wearing?
Would you trust any of them to tell you the temperature at which water boils?

[Deap]

Allegedly "flu season" is the result of kids going back to school in the Fall and brining bugs home. Closing schools would therefore impact what has long been called a "normal flu season".

[BillWade]

We are truly living in a sort of hell now. I wonder how many deaths are due to obesity rather than covid, flu, and colds. Yet, nothing is being done to help the obese. I think there should be a national mandate that all obese people be confined to their homes, not be allowed to work, and be given a manadatory caloric intake that will eventually get their weight down to normal. For these people; Christmas, New Years, Mardi Gras, Easter, 4th of July must be cancelled. We must save the lives of the fat, no exceptions whatsover. We'll start out slow, make this work for 15 days and then let them go to McBurgers, then lock them down for a few months - sorta slow frog boil them. We should pass the peak at some point, no?

[Zar]

Colonel,

As it happens, I spent a few hours studying that article, since I also thought it was interesting that it was taken down.

Long story short, I agree with Walrus. Two reasons.

1. I couldn't reproduce the author's results with the CDC data. I could've just been using slightly different categories, but I should've seen similar patterns.

2. The article compares rises in deaths from COVID to drops in deaths from other causes from week to week in order to draw conclusions about excess deaths. But any such matches (which I couldn't spot myself) might be coincidence. Let's say:

You receive 5 apples and 5 oranges in week 1.
You receive 3 apples and 7 oranges in week 2.
You receive 2 apples and 8 oranges in week 3.

Even if other people count the fruit differently, or the pattern only holds for three weeks, you can write an article about that. But unless you prove a relationship, your article might get taken down.

The better method to determine excess deaths is to just compare deaths from year to year. And the CDC already provides data on that. Even including the more interesting rumors I've heard, hospitals and the CDC aren't faking deaths altogether.

[Deap]

Trump murdered 250,000 people vs. 250,000 were going to die this year anyway. Which one makes the better campaign slogan?

[j. casey]

methinks the Johns Hopkins student struck a nerve. I suspect the covidians will start to wind down the op once Gates' "Emergency Use Authorization" is awarded in the US by the end of the month. Because the music is winding down now, and I expect that the total deaths in the US for 2020 will not be especially higher than an average of all deaths over the last ten years. Been thinking about Duvall in Apocalypse Now, squatting on the beach, saying "Some day this war is gonna' end," wistfully. I think that's where the covidians are now, realizing, wistfully, that soon the op will have its mRNA-experimental-vaccine course and we won;t have this fun and exciting drama and Little Red Guard work to do anymore.

[Zar]

After reading the above comments, I think I shouldn't have said "coincidence." Too strong. Some missing heart attacks deaths (for example) can probably be explained by people dying of COVID before they had a chance to have a heart attack, or by hospitals fudging the cause of death. Not clear how many, but fewer than Briand thinks. And I shouldn't have implied that the newsletter article was written by Briand.