One thing that the COVID-19 pandemic has revealed is the affinity between the antivaccine movement and conspiracy theorists who deny the severity of COVID-19 and the necessity of the public health measures instituted to slow the spread of SARS-CoV-2 (the coronavirus that causes COVID-19). This affinity and alliance between antivaxxers and anti-lockdown protesters and COVID-19 conspiracy theorists who, for example, blame the 5G rollout for COVID-19 came as a surprise to many, but not to those of us who’ve been following the antivaccine movement for a long time. The reason, of course, that the antivaccine movement is based on a conspiracy theory, what I like to refer to as the central conspiracy theory of the antivaccine movement, and groups of conspiracy theorists tend to be attracted to each other, and COVID-19 has been a magnet for medical conspiracy theories. So it comes as no surprise that yet another not-so-dynamic duo of antivaccine propagandists, Amy Becker and Mark Blaxill, have decided to lie with statistics about excess death statistics during the COVID-19 pandemic in order to blame vaccines for killing babies. No, wait. That’s not exactly right. What they’ve done is to note that there has been a decrease in infant deaths during the lockdowns instituted as a response to the pandemic and claim that that indicates that vaccines were killing babies before the pandemic, because, to antivaxxers, whatever happens, be it good or bad, is always an indication that vaccines are bad. Always. It’s their default. These living examples of the Dunning-Kruger effect lay out their “case” (such as it is) in an article posted to Health Choice, Blaxill’s antivaccine group, entitled Why are so many fewer children dying?
The grammar pedant in me can’t help but note the very awkward construction of the title (“so many fewer” just sounds odd). Before I dig in to the ignorance and ridiculousness of the arguments presented, first, let’s look at the authors. We’ve met Mark Blaxill (who’s prior antivaccine group was Safeminds), for instance, on multiple prior occasions, be it when he was promoting horrible antivaccine studies, promoting antivaccine misinformation among the Somali immigrant community in Minnesota who were in the midst of a massive measles outbreak because affluent white antivaxxers had promoted Andrew Wakefield’s MMR pseudoscience among them, or appearing in antivaccine propaganda films. As for Amy Becker, I don’t recall having heard of her before. Unfortunately, because it’s a not-uncommon name I wasn’t able to find out much about her. What I do know is that Health Choice is an antivaccine group whose leadership includes a veritable cornucopia of antivaccine activists, including, for example, Mark Blaxill, Wayne Rohde, Teresa Conrick, Mary Holland, Kim Rossi, John Stone, Jennifer Larson, and Anne Dachel. It’s a veritable who’s who of second tier antivaxxers (first tier antivaxxers including people like Del Bigtree, Andrew Wakefield, and Robert F. Kennedy, Jr.).
This brings us to the paper itself. I will admit one thing, though. I was surprised that Blaxill actually admitted the severity of the pandemic in the introductory paragraph:
Covid-19 is a serious public health issue, but the breathless reporting among the media of positive tests and an ever-rising death toll does little to instruct us about the true nature of the virus and the unprecedented steps taken to prevent its impact. As in many complex and pervasive health phenomena, there are many ways to measure health effects, but in our view the proper measure of impact is not a narrow or intermediate metric, but rather total health outcomes. In the case of a pandemic virus affecting large populations and where the immediate concern is sharp increases in deaths, the best measure of outcomes is not a selective measure of deaths somehow attributed to the disease but instead is deaths from all causes. For perspective, these deaths must be compared to historical death rates from all causes in prior years (Percent of Expected Deaths). As we will show, a balanced view of the broader American Covid19 experience demonstrates both the scale and variability of its negative outcomes in older Americans, especially the elderly, but also some unexpected positives. Surprisingly, U.S. mortality rates have declined among young people during the lockdown, especially among infants. These trends have gone largely unnoticed and remain unexplained.
You can see here that Becker and Blaxill are laying the groundwork to blame vaccines. Indeed, as soon as I saw that last sentence, I knew that the answer to the “mystery” of why mortality has declined among young people would be because they are not getting vaccines because so many pediatrics offices were closed to all but urgent issues and emergencies. Becker and Blaxill then follow with a series of graphs that are a beautiful of example of JAQing off. (You don’t know what that means? Read this.) The first two graphs are relatively straightforward graphs of total deaths per week during the pandemic, with the much higher death toll during the same period in 2020 representing excess deaths over the number expected for a the same time of the year compared to past years. Those excess deaths were almost certainly mostly due to COVID-19. At least Becker and Blaxill admit that COVID-19 has been much, much worse than a typical flu season:
Attributing a Cause of Death (COD) to COVID-19 is not always clear-cut, due to significant overlap among COVID-19, Pneumonia, Influenza, and presumably other primary CODs.
That said, the spike in deaths officially attributed to COVID-19 occurred in tandem with the spike in all-cause deaths, leaving little doubt that Covid19 was the main contributor to the excess of expected deaths between March 22 and May 9.
At least in this 8-week period, the Covid19 pandemic was considerably worse than a typical flu season. To the extent that all-cause deaths fell back to expected levels during May, the excess mortality attributable to the pandemic has passed.
Many states actually saw lower than expected deaths during the period. To be sure, an excess death rate of 5% for the entire U.S. is considerable but also far short of the apocalyptic narrative the pandemic has received. Greater-than-expected death rates were heavily concentrated in the Northeastern corridor. New York City and its surrounding area, including New Jersey, New York State (although possibly not upstate New York), Connecticut, Massachusetts, Maryland and the District of Columbia have so far comprised 6 of top 8 jurisdictions with excess all-cause deaths. New York City was hit especially hard. In a typical spring, New York City could expect 700-800 all-cause deaths per week. From mid-March to mid-May, that number spiked sharply, by ten times that amount, reaching over 7500 deaths in the peak week ending April 11.
Maybe that was true a month ago, but, as we know from recent headlines, as COVID-19 cases are falling in states like New York, New Jersey, Michigan, and other states that were hit early, COVID-19 cases have been climbing in Arizona, Texas, Florida, and other states that have reopened, and these increases can’t be attributed just to increased testing, with some states setting records for new COVID-19 cases:
This is the situation as only two states, New York and Michigan, are on track to get COVID-19 under control.
Besides a detour into figures showing what can only be called “Well, duh!” observations, such as that the greatest toll in excess deaths likely due to COVID-19 has been among people over 65, something that has been known since January or February, Becker and Blaxill try to make this argument:
To the extent that policies have varied across the states, it is not clear that the imposition and/or presence of stringent lock-down policies had much to do with the variation in excess deaths. Less stringent lockdown policies were not associated with higher death rates. In fact, the 5 states that chose not to impose a lockdown are among the roughly 20 jurisdictions with no excess deaths at all.
Several states with the most aggressive lockdowns, including California, Maine, Minnesota and Pennsylvania showed almost no excess deaths effect. Despite huge population centers, California looked nothing like New York City and State. Maine, a mostly rural state, imposed among the more draconian policies with essentially no reason. Minnesota followed a far more aggressive lockdown policy than its neighboring states of Iowa, South Dakota, North Dakota and Wisconsin. Yet it’s Covid19 deaths were among the most concentrated in the country: roughly 80% of Minnesota’s Covid19 deaths occurred among the infirm elderly who were residents of long-term care facilities.
Did aggressive lockdowns stave off the worst-case scenario, preventing vulnerable states from becoming disaster areas like New York City? No controlled experiment will give us that answer. Pennsylvania makes the best case for that argument, with an early excess death pattern that resembled its neighbors in the Northeaster corridor but saw that rate drop precipitously by early April.—. But Pennsylvania is also an unusual geographic unit, with its largest city, Philadelphia, lying on the coast and separated from the western part of the state and its second largest city, Pittsburgh, by the Appalachian Mountains. This anomaly makes it difficult to draw clear conclusions from Pennsylvania’s Covid19 curve.
I have two observations here. First, Becker and Blaxill seem to have left Michigan out, which is odd, given that it’s one of the states on track to control coronavirus, and it imposed one of the more rigid lockdowns as COVID-19 cases were climbing in the second half of March, while being among the slower states to start easing the lockdown and reopenining. (Indeed, hair and nail salons only just opened this week, allowing a certain blogger to get his first haircut since February.). More importantly, Becker and Blaxill are making an argument that it is far too early to make. Again, the pandemic is not done with us yet. In brief, it’s too early to say that lockdowns didn’t decrease excess deaths due to COVID-19 because we don’t know the overall death tolls yet, and the more recent datapoints are incomplete because of that lag in reporting. Indeed, the CDC sources that Beck and Blaxill cite tell you that the numbers of deaths are likely large underestimates because it takes a long time for states and localities to report mortality data to the CDC, as the CDC itself tells anyone using the data:
The provisional counts for coronavirus disease (COVID-19) deaths are based on a current flow of mortality data in the National Vital Statistics System. National provisional counts include deaths occurring within the 50 states and the District of Columbia that have been received and coded as of the date specified. It is important to note that it can take several weeks for death records to be submitted to National Center for Health Statistics (NCHS), processed, coded, and tabulated. Therefore, the data shown on this page may be incomplete, and will likely not include all deaths that occurred during a given time period, especially for the more recent time periods.
Provisional counts are weighted to account for potential underreporting in the most recent weeks. However, data for the most recent week(s) are still likely to be incomplete. Only about 60% of deaths are reported within 10 days of the date of death, and there is considerable variation by jurisdiction and age. The completeness of provisional data varies by cause of death and by age group. However, the weights applied do not account for this variability. Therefore, the predicted numbers of deaths may be too low for some age groups and causes of death. For example, provisional data on deaths among younger age groups is typically less complete than among older age groups. Predicted counts may therefore be too low among the younger age groups. More detail about the methods, weighting, data, and limitations can be found in the Technical Notes.
In other words, again, it’s almost certainly too early to draw any firm conclusions like what Beck and Blaxill are trying to draw. Becker and Blaxill acknowledge the second cautionary note by the CDC in an endnote and state that they “deliberately excluded the most recent four weeks of available data.” However, they also say that the charts were for the week ending May 16, pulled on June 6, which is only three weeks after the end of the last week examined. Given that the CDC notes that the potential for underestimation of the number of deaths per week due to incomplete data is worse for exactly the age groups that Becker and Blaxill are trying to use to blame vaccines for SIDS, I can’t help but question their methodology even more. Also, without a statistical analysis, we’re just looking at a trend that might or might not exist. Everything Becker and Blaxill did was based on just “eyeballing” the charts; it wasn’t based on anything resembling a rigorous analysis.
All of this, including the sloppy analysis, the JAQing off, the rampant speculation, however, is just a buildup to the main point that Becker and Blaxill want to make, the “climax” of their incompetent analyses of COVID-19 death statistics. In doing so, they provide the rope with which their argument can be easily hanged:
But the pandemic experience has brought on a surprising effect on this expected death rate among children. Starting in early March, expected deaths began a sharp decline, from an expected level of around 700 deaths per week to well under 500 by mid-April and throughout May. As untimely deaths spiked among the elderly in Manhattan nursing homes and in similar settings all over the country, something mysterious was saving the lives of children. As springtime in America came along with massive disruptions in family life amid near universal lockdowns, roughly 30% fewer children died.
Was this a protective effect of school closures? Were teenagers getting themselves into risky situations at a lower rate? No. There was very little effect among school age children or adolescents.
Virtually the entire change came from infants. Somehow, the changing pattern of American life during the lockdowns has been saving the lives of hundreds of infants, over 200 per week.
The first thing that one must note here is that we’re dealing with relatively small numbers, namely because far fewer children than adults die every year. Now, you might wonder why Becker and Blaxill claim that virtually all the decline came from infants. For that, all you need to do is to look at this graph that they reproduce:
So, from between the week of February 1 and May 16, the number of deaths of infants under 1 year old fell from roughly just under 400 a week to around 250 a week, a greater than 35% decline. But what about Becker and Blaxill’s claims that there was “very little” effect among school age children or adolescents? As an absolute number, that’s sort of true, but as a percentage? Not so much. Again, look at the graph. Between the weeks of February 1 and May 16, deaths of children 5-14 years old fell from over 100/week to roughly 75 a week, a fall of close to 25%, not much less than that among infants. During the same time period, among children aged 1-4 years, the number of deaths fell from around 70-90 a week to 50-60 a week, a similar decline. Of course, one thing that has to be noted about these numbers is that they are so small that it’s had to say if this trend is even real, particularly given the lag in reporting that means that later statistics are almost certainly undercounts, because of the several week delay in processing some death certificates. Basically, Becker and Blaxill probably jumped the gun, because they are relying on statistics that are incomplete, the more recent the statistics, the more incomplete. In any event, even if the later statistics hold up, the case Becker and Blaxill are making is not convincing.
None of that stops them from inferring a cause. Those of you who’ve followed the antivaccine movement likely guessed that this was where Becker and Blaxill were going: Straight to what they always go to, vaccines, in this case, the lack of vaccines caused by delayed well-baby visits due to the lockdowns in response to the COVID-19 pandemic:
We have no specific data on the trend in SIDS deaths during the pandemic. We have, however, heard anecdotal reports from emergency room (ER) doctors suggesting some have observed a decline in SIDS. One doctor who says he might see 3 cases of SIDS in a typical week has seen zero cases since the pandemic and associated lockdowns began.
This leap to a conclusion is unsurprising, given the antivaccine myth that vaccines cause sudden infant death syndrome (SIDS). They most definitely do not, even though antivaccine cranks love to promote this myth, the Vaccine Court once made a dubious ruling suggesting a link, and grieving mothers who attribute their baby’s SIDS to vaccines make for sympathetic and compelling figures, which is why antivaxxers try to convince mothers who’ve lost babies to SIDS that it was vaccines that caused it. In fact, although it’s very clear that vaccines do not increase the risk of SIDS, it’s actually likely that they decrease the risk of SIDS.
Ignoring all this science, Becker and Blaxill boldly engage in JAQing off over whether vaccines might be responsible for this decline in mortality:
One very clear change that has received publicity is that public health officials are bemoaning the sharp decline in infant vaccinations as parents are not taking their infants into pediatric offices for their regular well-baby checks. In the May 15 issue of the CDC Morbidity and Mortality Weekly Report (MMWR), a group of authors from the CDC and Kaiser Permanente reported a sharp decline in provider orders for vaccines as well as a decline in pediatric vaccine doses administered. (8) These declines began in early march, around the time infant deaths began declining.
Again, when you’re a hammer, everything is a nail, and when you’re an antivaxxer, everything must have something to do with vaccines, particularly if you think you see a correlation between a negative outcome with vaccines or a correlation of a positive outcome with a lack of vaccines.
You know what? If Becker and Blaxill want to include anecdotes, I’ll include one too. It’s from a pediatric intensive care doctor, who responded to a thread on Twitter in which an antivaxxer named Jeff claiming a correlation between decreased infant mortality and decreases in vaccine uptake:
See? This is a potential other explanation for the decline in the number of infant deaths (if, in fact, it turns out to be real, which, given the small numbers involved, it might not be). Another potential explanation is the huge decline in miles driven in automobiles, which likely resulted in a decline in deaths due to auto collisions. Given the small numbers of deaths of children under 14 per week at baseline, it wouldn’t take that much of a decline to partially or completely explain the declines Becker and Blaxill are touting. Of course, antivaxxers like Becker and Blaxill don’t even consider it. Instead, they conclude with a flourish of unfounded speculation:
Covid19 is unique among recent pandemics in that the mortality toll is measurable, real and convincing. It is also nearly certain to be transitory, but that won’t stop the propaganda juggernaut from rolling forward. However, as the saying goes, “the best laid plans of mice and men often go awry.” What no one would have predicted in advance of Covid19 is that the extreme lockdown response has produced a natural experiment that actually calls into question the very actions—widespread, mandated vaccines for all–that the infectious disease and public health community have been pushing for years. We should mourn the deaths of the elderly Manhattan nursing home residents but also take heed of the hundreds of avoided infant deaths. Only with that kind of balance will we draw the proper lessons from the pandemic and the lockdowns that have followed in its wake.
Note the offensive comparison. Sure, many tens of thousands of elderly people have died horrible deaths because of COVID-19, but won’t you think of the children, specifically the several hundred who, for whatever reason, didn’t die because of the lockdown? It’s how antivaxxers think. Again, it’s all about the vaccines, first, foremost, and always. Think of it this way. Even if Becker and Blaxill have a point, they are saying, when you boil down their comparison, that it was worth it to have tens of thousands of elderly people die if it saved at least a ten- to twenty-fold smaller number of children—and “proved” that vaccines cause SIDS, to boot! If you think I’m being uncharitable and don’t think that’s what they meant to infer, then answer me this: Why did they also spend so much verbiage (and two charts) looking at average life expectancy per age cohort and quality-adjusted life-years saved or lost by age group?
The bottom line is that Becker and Blaxill made broad, sweeping generalizations in the form of JAQing off just by “eyeballing” raw figures for which the later numbers are likely to be significant underestimates, all in the service of antivaccine pseudoscience in the form of the myth that vaccines cause SIDS. Antivaxxers have been predicting that SIDS would decline during the lockdown because of the decline in vaccination, and this is their latest pathetic attempt to show it.