It’s a year and a half now that the U.S. COVID-19 pandemic has been grinding on.
So, are we done being stupid yet?
Ah, heck no.
A reader alerted me to the latest nonsense going around in the Fox alternative universe. It’s the assertion that COVID-19 vaccines are killing so many people that it’s raising non-COVID, all-(other)-cause mortality numbers.
Wow!
And nobody is talking about it!! Double wow — that means there’s a conspiracy of silence!!!
Or, alternatively, it means that it’s nonsense!!!! But apparently that obvious explanation didn’t seem to cross the minds of the people spreading this!!!!!
So, to summarize, the allegation, from this Fox interview, is that COVID-19 vaccination is:
- Killing people, but only in some countries. Great Britain, Germany, and the U.S. But not in other countries with equal or higher vaccination rates.
- Killing people in some age groups, but not other, more-highly-vaccinated age groups. In the U.S., it’s only killing the 20- (or 30-) to 50 age group. Elsewhere, it’s killing everybody.
- Killing people during some time periods, but not others, and those time periods vary across countries and populations, with a long and variable lag between vaccination and death.
For those of you work in the sciences — or simply have common sense — you’ve probably already spotted what’s wrong with the “logic” of this argument.
It’s called post hoc propter hoc. Dressed, in this case, with a side-order of cherry-picking. The argument is that if (non-COVID) excess mortality went up anywhere, for anyone, in 2021, well, that must be due to vaccination.
Because, of course, nothing else has happened in the past couple of years that might affect the health status of the world’s population. So it has to be vaccination!!!!!!
It almost goes without saying that these allegations of mass vaccine-related deaths, in this Fox interview, are being made by a person with no background in (e.g.) vaccines, mortality data, or health statistics. Instead, this is, of course, yet another allegation about COVID-19 from a right-wing commentator with not the faintest grasp of what he’s talking about.
Yet, by request, I guess I have to look into this in detail.
Let me break this into two parts. The key checkable fact is that there is significant non-COVID-19 excess mortality, in the U.S., in 2021, for the population aged 20 to 50 (or 30 to 50, depending on which sentence you listen to, in this Fox interview). Then, separately, there is the attribution of that excess mortality to a specific cause — COVID-19 vaccination. The interview also contains allegations about two (but only two) European countries, but checking those out correctly requires so much work to understand country-specific deaths data that, for now, I’m sticking to the U.S.A.
So, for my task, this boils down to:
- What’s happened to the U.S. mortality rate for 20 (or 30-) to 50-year olds in 2021?
- Can that plausibly be attributed to COVID-19 vaccination?
Spoiler: I’m just going to start on it today. This post is already too long. I’ll finish it tomorrow.
Some hard-earned advice
There’s a phrase to keep in mind whenever anybody alleges something of this magnitude: Extraordinary claims require extraordinary proof.
At root, this is a claim that mass deaths are occurring as a result of some government-sponsored intervention. (And, secondarily, that there’s a coverup, because nobody is talking about it.)
I learned the hard way, from years of working with Medicare data, any claim about unusual death rates is an extraordinary claim. PBut, as I show below, there’s nothing to back up this claim. It’s the opposite of extraordinary proof.
People in responsible roles in the government take any claims about excess deaths extremely seriously. If you are serious about it, you really need think twice, and check your work, before shooting your mouth off about any claim regarding deaths. In particular, a claim that mass deaths are occurring as a result of some government-sponsored intervention.
But if all you want is attention, the last thing you do is bother to check your facts.
That said, let me now walk through this logically. And check my facts. Because everybody makes mistakes sometimes.
A little background.
These days, most people who talk about U.S. mortality rates have no understanding, at all, about the subject. They don’t know how deaths are reported, or how death certificates are filled out. They don’t understand how cause-of-death is assigned. And they certainly have no understanding of how stable the U.S. mortality rate typically is from year-to-year, the factors that may affect the year-to-year variation in U.S. deaths, or how various arms of the U.S. government publish and use those data.
(Or, for that matter, how difficult it is to cast votes in the names of dead people, due in large part to Social Security’s Death Master File. In the U.S., we may not take care of you from cradle to grave, but we sure as hell track you from one to the other, if only to minimize the amount of Social Security fraud that occurs.)
Just for the record, I’m not one of those people. I spent much of my career in the analysis of data from the Medicare program. Roughly one-quarter of all Medicare spending is for individuals in the last year of life. That’s not “the high cost of dying”, that’s largely the high cost of living with severe illness. But given the importance of last-year-of-life spending, I had to learn about death and dying in the U.S. merely to do my job competently. In the last last two decades of my career I had a focus on the cost of Medicare end-of-life care (Reference, Reference).
So, if you care to be among the informed, you can look at a few prior posts here and try to get up to speed.
If you want some basic understanding of what a death certificate looks like, and how those are filed out, go to Post #793, August 2020. That was in response to people who knew nothing about death certificate data distributing disinformation about COVID-19 deaths.
If you want a quick tutorial on how CDC tracks adverse events following vaccines, you can try Post #1208, A Funny Thing About Deaths in the Elderly. That was in response to people who knew nothing about the CDC’s adverse event tracking system, spreading disinformation about deaths and COVID-19 vaccines.
I think I’m seeing a pattern here.
In any case, this post is about U.S. mortality data. And about the people who know nothing about it, who are now spreading disinformation about deaths and COVID-19 vaccines. Again.
Existing research.
I addressed the last round of nonsense regarding COVID-19 vaccination and deaths in “A funny thing about deaths in the elderly (Post #1208, August 10, 2021). That time, the allegation was that the U.S. Vaccine Adverse Events Reporting System (VAERS) proved (proved!!!!) that there were mass deaths from COVID-19 vaccines.
That allegation was, needless to say, incorrect. I walked through that, in that posting.
(And yet, as is the way in the alternative Republican universe, everyone seems to have forgotten that. And so, we’re on to yet a different reason why vaccines are evil. Which we wouldn’t have to touch, if the last one had been right.
You might think that people would eventually tire of being mislead. But they never remember it. So we get a kind of policy Alzheimer’s. Every day is a blank slate, and each new wacky conspiracy is taken afresh, with no bothersome historical context to get in the way of belief.)
In any case, now we have somebody saying that the CDC excess mortality data clearly maybe sort of possible shows that the COVID-19 vaccine is causing mass deaths!!! But there’s a conspiracy of silence about it!!!
And so, I’m once again in the business of tracking through something that I’m almost certain will be yet another bogus analysis, by people who have no clue what they’re looking at.
Why am I so sure this will turn out to be nonsense? “This”, meaning that somehow, unexplained deaths are occurring at a scale large enough to perturb U.S. national mortality data? All linked to COVID-19 vaccination? And yet, nobody has noticed?
To answer that, let me just lay out a just a few of pieces of actual research relevant to the topic. Because facts always exist in a context. And if you don’t know the context of the existing research, you probably aren’t competent to make any judgement about some newly-claimed facts.
0: First, keep this in mind
All of the discussion below is about the side-effects of the vaccine, and only the side-effects. All of it ignores the main effect of vaccination, which is to prevent death from COVID-19 itself. None of the research below provides any assessment of the net benefits of COVID-19 vaccination. It’s an analysis of potential side-effects, only.
1: No excess deaths in the large-scale vaccine clinical trials
All COVID-19 vaccines went through lengthy, multi-phase randomized clinical trials. In those trials, Phase 1 is purely a test of safety. The first thing they do is test the vaccines to make sure they won’t (e.g.) kill you. Then, if they pass Phase 1, Phase 2 tests whether or not the vaccine works. And Phase 3 — if there is a Phase 3 — re-tests that and fine-tunes things like timing and dosage. For example, the analysis of booster shots is formally part of Phase 3 of the vaccine clinical trials.
As a result, there is meticulously-maintained record of adverse reactions and deaths, in a true randomized trial, for a total population that exceeds 50,000 vaccinated persons. That’s for the clinical trials of the three vaccines used in the U.S.
(For those of you who don’t understand why double-blind randomized trials are superior to any other form of analysis, it’s simple. The randomization step — randomly selecting who gets the treatment and who gets the placebo — gets rid of all the other factors that might influence the outcomes, such as difference in health behaviors, lifestyle, and so on. By contrast, “observational data” — comparing people who chose to get vaccinated or not — always commingles differences in those populations with the actual impact of the vaccine. The double-blind aspect — neither the subjects nor the doctors know who got vaccine versus placebo — means that nobody’s subjective feelings can influence the results. Again by contrast, in “observational data”, you will find that (e.g.) people who know they got the vaccine are much more likely to attribute vague side-effects to the vaccine.
Whenever you see some goofy-assed “research” results reported in health care, it’s almost a sure bet that they were based on “observational data”, and not on a proper double-blind randomized clinical trial. And if you don’t realize that, the constant stream of dog-bites-man nonsense that is reported in the popular press probably leaves you with the feeling that all research is nonsense. Which is as unfortunate as it is common.)
Let’s take a peek at the actual reported results of the Pfizer vaccine trial, as Pfizer is the most-commonly-provided COVID-19 vaccine in the U.S. These are literally the data that convinced the FDA to approve its use. Here, I refer to the the formal writeup as published in the New England Journal of Medicine.
In this trial, after randomization into treatment and control groups, more than 22,000 individuals were given the Pfizer vaccine, and an equal number got a placebo. Adverse events were tracked pro-actively for two months, while any adverse event (including death) could be reported for up to six months after injection.
Of that 22,000+ population of vaccinated individuals, and other 22,000 who got the placebo, how many died during the study period? (Where “BNT162b2” is the Pfizer vaccine).
“Two BNT162b2 recipients died (one from arteriosclerosis, one from cardiac arrest), as did four placebo recipients (two from unknown causes, one from hemorrhagic stroke, and one from myocardial infarction). No deaths were considered by the investigators to be related to the vaccine or placebo.”
To be crystal clear, they looked hard and long at a population of tens of thousands of individuals who got that vaccine. And they found no evidence of any excess deaths.
If you look around, you can find popular reporting of some longer follow-up period, when more deaths were reported. You still find no excess deaths in the vaccinated population.
2: No excess deaths in analysis of large-scale observation data.
Maybe you just don’t like those extremely-high-quality double-blind controlled clinical trials. Maybe looking at a mere 22,000 people for half a year isn’t enough to satisfy you that the COVID-19 vaccines don’t generate massive additional deaths.
If that’s your bent, then how about an observational data study of 11 million people? Would that be enough to satisfy you? Some careful tracking of health care, death, and vaccination records for 3 percent of the entire U.S. population?
If that’s more to your taste, then read on.
For U.S. deaths data, I of course turn to the CDC’s Morbidity and Mortality Weekly Report (MMWR). As one does. And there, today, featured front-and-center on their website, is this piece of analysis, just published:
Source: Xu S, Huang R, Sy LS, et al. COVID-19 Vaccination and Non–COVID-19 Mortality Risk — Seven Integrated Health Care Organizations, United States, December 14, 2020–July 31, 2021. MMWR Morb Mortal Wkly Rep. ePub: 22 October 2021. DOI: http://dx.doi.org/10.15585/mmwr.mm7043e2.
It’s not hard to spot their main conclusion. I have underlined it in red below.
Source: Same as above.
Admittedly, this is “observational data”. So it lacks the guarantees of double-blind randomized trials. But for me, having done work like this all my life, this one ticked all the boxes that say “carefully done analysis”. It’s done about as well as it can be done short of a controlled clinical trial.
Just a few highlights:
- They tracked 11 million people, in seven large health care organizations. These were split about 60/40 between those who chose to be vaccinated, and those who did not. So there’s no doubt that the sample size is more than adequate.
- Obviously, the vaccine protects against COVID-19 deaths, so they removed all COVID-19 deaths to get an apples-to-apples comparison between the vaccinated and unvaccinated. They didn’t merely toss out deaths that listed COVID-19 as cause of death. They tossed out any death occurring within 30 days of any diagnosis of COVID-19 or any positive COVID-19 test. They excluded all deaths even remotely plausibly related to COVID-19.
- They fully accounted for the fact that you have to be alive in order to get vaccinated. (Which sounds dumb, but it’s a necessary step, and it was good to see that they did that with an industry-standard “pseudo-event date” method. That is, they took the distribution of actual vaccination dates, and imposed that set of dates on the un-vaccinated comparison group. This ensured that individuals in both samples were known to be alive as of that vaccination or equivalent pseudo-vaccination date.)
- They accounted for differences in preventive health behavior in general by drawing the population without COVID-19 vaccines from those who had gotten a flu shot in the last couple of years. So the entire study population consists of people who are willing to be vaccinated. It’s just that some didn’t get the COVID-19 vaccine.
- They accounted for differences in the age/sex/race mix of the vaccinated and un-vaccinated populations.
- And, in the end, the COVID-19-vaccinated population had a much lower mortality rate.
To be clear, this sort of observational data research is not the gold standard. The most accurate thing you can say is that if vaccines did cause mass deaths, they were too small to be observed against the background of pre-existing health status differences between those choosing to be COVID-19 vaccinated and not.
And if you’re interested in a higher-mortality population, this study of nursing home residents found no short-term increase in mortality following vaccination, compared to an un-vaccinated control group. So if the vaccine kills old people, it most assuredly doesn’t kill them very quickly.
3: But wait, there’s more …
Hilariously enough, in looking at research, I found a serious study showing an increase in non-COVID deaths in the younger (age 25-44) population, for the U.S., for a few months. Just exactly as that Fox interview alleges.
What’s the catch? That was a study of 2020 — the year before the vaccines were introduced. This is from a Research Letter in the Journal of the American Medical Association Network, so I’m not sure how tightly that was reviewed. That said, the JAMA imprimatur means it was probably credibly well done. FWIW, their takeaway is that COVID-19 was probably under-diagnosed, at that time, in that generally healthy young population.
Clearly, the only possible explanation is that those vaccines are so damned dangerous, they were killing off young people in the year before they were administered. (That’s sarcasm, in case that wasn’t crystal clear.)
But seriously, this, by itself, means that you can’t make much of it, even if you can identify a 2021 period of excess mortality for that population. For all I know, apparent excess mortality in that population is a regular thing, owing perhaps to above-average instability in the measured death rate, because the death rate for that population is miniscule.
Sometimes, in research, you just can’t make up stuff that’s half as good as the real thing. In this case, it sure looks like there was excess non-COVID mortality in the young, occurring at the time of the initial U.S. COVID-19 outbreak … well before vaccines were even on the horizon.
4: To summarize the existing research
Serious researchers have looked for any evidence of excess non-COVID-19 mortality related to COVID-19 vaccination. The conclusion so far is that such excess mortality is:
- Not in the randomized clinical trials of the vaccines.
- Not in an 11-million-person study of private health care plans.
- Not in a study of hundreds of nursing homes.
But, you do seem to see something, for a few months, at the height of the first COVID-19 wave, in the younger population, in the year before the vaccines were in use. The researchers who spotted that attributed it to under-diagnosis of COVID-19 in that otherwise healthy population. Given how hard it was to get tested at that time, I think that’s at least plausible.
Now let’s take an initial look at the U.S. mortality data.
US Mortality data: First, use your eyes.
Let’s do a little reality check first. I can easily get U.S. mortality data by cause, graphed, from this scholarly source. So let’s start there.
The story we are investigating is that, somehow, there’s an increase in 2021 non-COVID deaths that is linked to COVID-19 vaccination. So let’s start with the vaccinations, from CDC, displayed over the same time period as the subsequent mortality data. And contrast that, by eye, with U.S. non-COVID deaths over the same period.
The top graph is newly vaccinated individuals, from the U.S CDC COVID data tracker. The second is U.S. deaths for all causes other than COVID-19, from the scholarly source above. Matching time periods for both graphs.
If you see some cause-and-effect relationship between those two curves, you have a much more vivid imagination that I have. Among other things:
- There was a small peak in non-COVID-19 deaths in April 2020, before vaccines were even available.
- The peak rate of vaccination was April 2021. At that time, non-COVID-19 deaths were gently declining.
- In fact, U.S. non-COVID-19 deaths peaked in the winter of 2020, and were declining throughout the entire period during which the population was getting intensively vaccinated.
And yet, there was more month-to-month variation in the death rate shown above than is typical or normal for the U.S. In a typical year, that U.S. mortality data is pretty much a flat line, other than for a little seasonal rise during the cold weather. So there certainly was some unusual variation in non-COVID-19 deaths in 2020 and 2021. But it clearly bore no relationship to COVID-19 vaccinations. At least, not in any way that is obvious to the eye, or related in terms of timing.
I wonder what could possibly explain those odd rises in U.S. non-COVID-19 deaths in April 2020, in the winter of 2020/2021, and then again in August 2021?
Here’s the same graph of U.S. deaths, but this time including COVID-19 deaths as the top segment.
Notice any correlation? At this point, the explanation for the variation in non-COVID-19 deaths is pretty clear. They coincide with the peaks in COVID deaths. There’s clearly something about COVID-19 deaths that is spilling over into other causes of death. The big spikes in COVID-19 deaths have been creating secondary peaks in non-COVID deaths.
All this graph establishes is the fact of that, not the reason why. But it’s easy enough to point to a few plausible mechanisms.
First, those could simply be miscoded COVID-19 deaths, done deliberately or otherwise. In some parts of the country, or some social circles, there’s probably quite a stigma associated with dying from a non-existent disease that was made up by liberals. And so, in the same way that other stigmatized diseases have been under-reported on death certificates (such as dementia), COVID-19 may have been deliberately under-reported, at the family’s behest.
Second, those could be direct medical spillovers from COVID-19 onto other diseases of the frail. Of all persons who die with COVID-19 mentioned on the death certificate, six percent are said to have died from other causes, with COVID-19 as a secondary diagnosis. (I documented that back in one of my earlier analyses of COVID-19 deaths, cited above). For some frail individuals, COVID-19 may have hastened an otherwise-impending death from other causes, and thus caused a coincident peak in non-COVID deaths, broadly speaking.
Third is the potential for crowd-out of health care access. In some areas, COVID-19 patients have crowded the hospitals and other facilities to the extent that persons with non-COVID diagnoses had difficulty obtaining care. I think it’s at the edge of plausibility that some of the increase in non-COVID deaths might have been caused by that. But I doubt that it’s much. I would expect it to be quite rare that an individual ill enough to be nearing death would have been denied live-saving inpatient care under almost any circumstances.
Summary to this point.
Everything in the serious medical or epidemiological literature says that there’s no excess non-COVID-19 mortality associated with getting vaccinated. The clinical trials show nothing. The observational data shows lower mortality among the vaccinated (almost certainly an artifact of better health status or health habits).
By eye, the U.S. mortality data show no linkage whatsoever to the rate of COVID-19 vaccinations.
Instead, it’s crystal clear that something about the peaks in COVID-19 deaths did, in fact, spill over onto non-COVID deaths. Whether that’s mere mis-reporting of cause of death, or has some actual medical basis (COVID-19 hastened death in some frail people reported as having died from other causes, or COVID-19 caused additional deaths via hospital crowding), there is no way to tell from a simple graph.
Tomorrow, I’ll try to find whatever-it-was that prompted somebody to say that there were “excess non-COVID-19 deaths” in the 20- (or 30-) to 50 year old population.
Spoiler alert: Near as I can tell, CDC only defines excess deaths for total deaths, not deaths by cause. And CDC is the entity that defines any sort of official U.S. excess deaths measure, using a reasonably complex algorithm to do so. What I’m trying to say is that, as far as I can tell so far, there is no such thing as excess non-COVID-19 deaths, defined by the U.S. CDC. So where that particular bit of mis-information came from, I cannot yet say. I’ll try to track down what I can.
